Previous day | Jump to hour: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | Next day

Seconds: Show Hide | Joins: Show Hide | View raw
Font: Serif Sans-Serif Monospace | Size: Small Medium Large

Click in the nick column to highlight everything a person has said.
The Logo icon identifies that the person is a core developer (has commit access).

#rockbox log for 2006-04-24

00:00:20amiconnI though all codecs deliver S3.28, and that's converted to 16bit only after applying all the dsp stuff
00:00:37lostlogicamiconn: crossfade is post DSP
00:00:43mareviqumm... any chances of 4GB iPod Mini 1G partition table being aviable in the near future from the page?
00:00:46amiconnah ok
00:01:40lostlogicMikachu: petur: compiler would bitch if the (short *) happened first, wouldn't it?
00:01:43amiconnMikachu: You think ((short *)(fade_out_chunk->addr)) would be better?
00:02:05 Quit TCK (Read error: 110 (Connection timed out))
00:02:58peturinstead of learning precedence, I started using too many ()()() :(
00:03:32amiconnpetur: Yeah, using too many parentheses just in case doesn't help readability
00:03:39peturwe never got c/c++ at school so I learned it while I needed it - skipping some parts...
00:03:51amiconnIt wil still help to sort out such problems when in doubt
00:04:01Mikachui remember doing opposite parenthesis to cast the pointer in our java class
00:04:39amiconnAnd yes, the compiler should have complained if the (short*) cast happens first, but I don't really trust gcc anymore...
00:04:44Mikachuand i would think most people know those rules, i'm just not so good at all parts of C, like petur :)
00:05:18*lostlogic applies temporary variable POWER to the problem to see if it's an order of bullsheet problem.
00:05:19amiconnWe had a similar problem in talk.c back then...
00:06:19*amiconn still didn't find a way to make the lcd init reliable and efficient at the same time
00:06:25amiconn*remote lcd init
00:07:30petur - at the bottom
00:09:08amiconn rev. 1.27 in case anyone's interested
00:09:24 Quit midkay (Read error: 104 (Connection reset by peer))
00:09:25peturtypecasting is a higher level than -> or am I not thinking right again?
00:09:34 Join midkay_ [0] (n=midkay@
00:10:07lostlogicpetur: hmmmmmmmmmmmm, not sure.
00:11:28mareviqso, another question: will someone hack into rockbox support for mac partitioning? (NOT hfs, these are two different things), as I can get it to run if the ipod is formatted in a mac-like fashion with a fat32 partition there... it's just that rockbox won't see that partition
00:11:35Mikachuthat table doesn't look quite right to me
00:11:40lostlogicI'm testing with a temp var, it actually looks cleaner so I might just keep it
00:11:51amiconnpetur: On that page, lower level means earlier evaluation, so no
00:12:12lostlogiclower is last
00:12:14Mikachubut maybe it's jsut unclear
00:12:25Mikachuah right, that's * as in dereference
00:14:14Mikachustill, * is above + and * happens before +, and (type) is above ->
00:14:15amiconnThe table goes from highest to lowest priority. -> is level 2, (cast) is level 4
00:14:26amiconnSo -> happens first
00:14:44Mikachuwhat does he mean with level 5 then?
00:15:05lostlogicusing a temp variable mostly fixed it.
00:15:06amiconnHmm, that's strange...
00:15:11lostlogicdamn, GCC is one _dumb_ mofo.
00:15:36preglowis that so........
00:16:08Mikachu(i didn't see the -> in level 2 before)
00:17:29amiconnHmm, level 5 seems to be something c++ specific
00:17:40amiconnDoesn't appear in my old c book
00:18:02mareviquhm.... I take that as "no"...
00:18:11lostlogicmareviq: correct.
00:18:15preglowoooh, a nano skin that's actually quite decent looking
00:18:24 Quit Massa ("IceChat - Its what Cool People use")
00:18:34 Join twisted` [0] (
00:18:35lostlogicmareviq: that I'm pretty sure would be an accepted patch, but nobody's looking at it afaik
00:18:42twisted`I really need to add this chan to autojoin
00:18:51twisted`sharpe: I was able to create an account for you :)
00:19:24preglowwe need to find some decent wpses for release...
00:19:32*twisted` slaps lostlogic around with some corrupted justice
00:19:44 Quit mareviq (Client Quit)
00:20:00Mikachumareviq: why not just repartition it?
00:20:00 Quit Mikachu (Success)
00:20:00peturis Linus going to add the margin patch or not?
00:20:14preglowno idea
00:20:23twisted`damn he's been idle for 14hrs
00:20:24twisted`bad sharpe
00:20:25preglowif he is, he should do it soon
00:22:11 Quit ender` (" You don't have to think too hard when you talk to teachers. -- J. D. Salinger")
00:22:35twisted`sharpe: anyway, when ur there, lemme know, so I can send you the account information
00:22:50 Join ashridah [0] (
00:23:43preglowcan anyone tell me why the hell my nano refuses to play anything now? :>
00:23:55 Join Mikachu [0] (
00:25:24peturbe nice to it?
00:25:53 Quit Moos (Read error: 104 (Connection reset by peer))
00:27:41 Join Moos [0] (
00:28:09preglowcvs build doesn't work either
00:28:22preglowany other nano people here?
00:28:37 Quit petur ("Zzzz")
00:28:40Mikachui guess me
00:29:13*Mikachu cvs update
00:29:17preglowdoes it work?
00:30:06Mikachuplayback.c: In function ‘load_next_track’:
00:30:06Mikachuplayback.c:1979: warning: suggest explicit braces to avoid ambiguous ‘else’
00:30:16preglowthe disk icon just keeps displaying
00:30:18preglowthen nothing happens
00:30:39Mikachulostlogic: been coding python lately? :)
00:30:56*preglow crosses himself
00:31:21lostlogic:-P 8fixes*
00:31:51preglownice one
00:31:52lostlogicthat's what I get for throwing in quick fixes for stuff I notice as I'm working on something else.
00:32:10Mikachui added braces before testing, and it works here
00:32:59lostlogicMikachu: do songs at the end of playlists no longer cut themselves off?
00:33:18Mikachui haven't experienced that bug
00:33:22lostlogicah, ok
00:34:17preglowMikachu: still doesn't work here
00:35:21Mikachuthen i don't have anything better than worksforme
00:37:58Mikachupreglow: i moved the volume change from main wps to the quick screen, i think it works well
00:38:20Mikachusomeone thought it was a bad idea but i don't remember why
00:39:56*Galois votes "not working"
00:40:16Mikachuthis must be the first time something works for me but not someone else :)
00:40:20Galoiswhen I try to play a song the playback just stops cold
00:40:26 Join Ninjaw [0] (
00:40:31Mikachui'm using resume of an old playlist
00:40:36Galoisbuttons etc. still work
00:40:38Ninjawhi boys
00:40:46Galoisoh, wait, I just got everything to freeze
00:40:52Ninjawi'm listening my freshly rocked ipod nano
00:40:53lostlogichmm... I take it, it's broken.
00:41:05Mikachuokay, stopped and inserted a dir, and it still works
00:41:09preglowGalois: what player?
00:41:15Galoisnano 4GB. CVS build.
00:41:19preglowahh, good
00:41:28preglowwhat happens here is just that the disk icon shows for a long time
00:41:31preglowthen nothing happens
00:41:35Galoisnow the only thing that works is the backlight
00:41:59amiconnWhy does slow plug-in of the remote work fine now (which it doesn't with cvs), but quick plug-in does not??
00:42:01lostlogicno breaking for me ;(
00:42:41preglowstop just hung the entire thing for me
00:42:50preglowonly backlight works
00:43:03*Galois resets the ipod. Still can't get it to play files.
00:43:05lostlogicusnig voice? crossfade? anything?
00:43:07amiconnHmm, perhaps fire a series of several inits. We can only blind-init anyway...
00:43:15GaloisI know I have voice and crossfade off
00:43:16preglowgimme a sec
00:43:16Mikachui don't use voice, crossfade or crossfeed
00:43:38preglowi had voice enabled
00:43:49preglowdoesn't matter anyway
00:44:11lostlogicpreglow: did this just start today?
00:44:19Ninjawi dont see ho wthe fuck rockbox is changing the usb access ?
00:44:26Ninjawlooks like the same to me
00:44:30preglowlostlogic: yeah, but i've been coding coldfire for the last couple of days, so i've had ample time to not notice
00:44:53lostlogicpreglow: hmm, but then the question is what's breaking you and galois but not me or mikachu?
00:45:10 Quit mikearthur (Read error: 104 (Connection reset by peer))
00:45:17 Join mikearthur [0] (
00:45:19Ninjawanyone knows if it will allows macos9 acces ?
00:45:23preglowlostlogic: i'm on default settings now
00:45:24lostlogicNinjaw: 1) be more polite 2) we use apple's disk mode.
00:45:28MikachuNinjaw: what is your actual question?
00:45:51lostlogicpreglow: selecting a playlist or a file?
00:45:52Ninjawi dunno apple disk mode
00:45:59Ninjawmy question is
00:46:02Galoisdoes the file format matter? I use ogg here
00:46:08preglowlostlogic: file, i have no playlists
00:46:13Mikachuogg here too
00:46:25preglowi just fixed it
00:46:30preglowby removing the voice file
00:46:38preglowlemme check if putting it back breaks it again
00:46:43lostlogicso it's something to do with the voice buffer setup.
00:46:46Galoishey let me try that
00:46:53amiconnWhat is the problem?
00:46:57Ninjawdo u think if i try the rocked ipod nano on a macos9 i will finally access the files ?
00:46:58*lostlogic about to commit more breakage while we're still recovering from his last breakage :)
00:47:03amiconnCurrent cvs playing fine here
00:47:14 Quit PaulJ (".")
00:47:21preglowconfirmed, it's the voice file
00:47:27amiconnmp3, not crossfade, voice enabled
00:47:43lostlogicso having a voice file on nano causes it to not work.
00:47:46preglowit's buffering and everything, correct track count, but no pcm acion
00:48:00Ninjawwhat about macos9
00:48:05lostlogicsome kind of race with the voice thread vs. the codec thread
00:48:10amiconnOh wait, I don't have the 2 latest commits in
00:48:14Galoisyeah I had a voice file too
00:48:14preglowlostlogic: allocated mp3 buffer size looks correct
00:48:38lostlogicthe last commits should only impact track changes
00:48:41preglowGalois: and removing it fixed it, yes?
00:48:41lostlogicnot startup
00:48:58lostlogicNinjaw: we have no idea what you are asking still.
00:48:58Ninjawwhat is fisk mode screen on ipod ?
00:49:12preglowNinjaw: what it sounds like. it just lets you access the ipod as a disk
00:50:00Galoismy last build was about a day ago. It was working fine with the voice file there.
00:50:06Ninjawhow the ipod says "i'm in disk mode"
00:50:13Ninjawwhat's teh screen ?
00:50:31lostlogicGalois: was that before or after I did the big voice rework commit
00:50:49Ninjawmy whole point is: will it finally work on macos9 (damn)
00:50:56amiconnHmm, current cvs seems to have an USB problem...
00:51:08preglowNinjaw: we don't know, we can't control disk mode and none of us uses macos9
00:51:23Ninjawyeah ok
00:51:27preglowdisk mode is apple code
00:51:28Galoislostlogic: before
00:51:31preglowwe can't control it
00:51:36Ninjawyeah yeah
00:51:42Ninjawnow could u describe it
00:51:49Ninjawis there any screen
00:51:54amiconnStill playing fine with latest cvs + voice. both when resuming and when starting a track from the browser
00:51:57preglowit's more than self-evident
00:52:08Ninjawwhat ya mean
00:52:09preglowperhaps it's this specific voice file...
00:52:11preglowit's rather old
00:52:18amiconnThis is on h340
00:52:20lostlogicamiconn: somehow it seems to be nano specific!? because you have mini2g and h300 and I have 5g...
00:52:38lostlogicpreglow: theoretically could have to do with the size of the voice fiel
00:52:43lostlogicpreglow: what is the size, in bytes?
00:52:44Ninjawi hate macos9
00:53:11Galois-rw-rw-r−− 1 djao djao 1408600 Sep 28 2005 english.voice
00:53:28Ninjawdoes this unfamous disk mode is supposed to work as a "usb hard drive"
00:53:29preglowwant me to upload it?
00:53:34preglowNinjaw: yes
00:53:37kapheineIf anyone is really bored, the other day I made a robotfindskitten plugin. I only put keybindings in for the iPod but you could easily add new ones. I put it up here:
00:53:39Galoisit's just MS Mary
00:53:43Ninjawis only itunes activating this disk mode ?
00:53:44preglowyeah, ms mary
00:53:49amiconnlostlogic: So we can blame preglow's ata hack...
00:54:00preglowamiconn: doesn't work with my new code either
00:54:16preglowthen again, that's not fully tested
00:54:25Mikachukapheine: and the .h?
00:54:26Ninjawwait does play+select is activating disk mode ,
00:54:53preglowNinjaw: we're not an ipod support channel. you might have more luck in getting answers in a more suitable channel
00:55:23 Quit _Lucretia_ (Connection timed out)
00:55:33amiconnJust tried on mini2g. Working fine there as well, voice + mp3 + no crossfade...
00:55:33Ninjawis rockbox forcing in anyway a stupid disk mode that i couldnt have before (on macos9 lol)
00:55:36kapheineMikachu: Sorry, forgot about that.
00:55:51preglowamiconn: and why would my ata hack suddenly start affecting voice files?
00:55:57preglowit's been working for a long time now
00:56:04amiconnForgot the ;)
00:57:05lostlogicthe lack of disk spinup time is causing a race somehow. I'll be trying to figure it out.
00:57:19NinjawOH MY
00:57:24amiconnlostlogic: At least it must be a quick race
00:57:27preglowin which case my ata hack certainly is still doing its job
00:57:39 Join _Lucretia_ [0] (
00:57:44amiconn..since it doesn't affect mini2g with its significantly shorter spinup time
00:57:48Ninjawwhat the ?
00:57:59lostlogicany crossfade users? Please test crossfade on latest CVS.
00:58:07Ninjawlost i tried
00:58:16Ninjawhavent heard any crossfade sorry
00:58:36Ninjawi also just discovered the disk mode
00:58:37Mikachukapheine: haha, nice :)
00:58:53Mikachukapheine: i haven't heard of this before though
00:58:57Ninjawsorry for u but it appears my RockBox wasnt activating it ?
00:59:25 Join KlrSpz [0] (
00:59:31preglowbah, the musepack seeking patch doesn't work on ipods
00:59:46KlrSpzthere's a few things i've found today that aren't working right
00:59:51KlrSpzliek crossfading :(
01:00:01Ninjawi cant exit the disk mode
01:00:28 Quit kaediil (Remote closed the connection)
01:00:31lostlogicKlrSpz: when the latest CVS build finishes, please retest crossfade, I just reworked it.
01:00:37kapheineMikachu: robotfindskitten? It has been ported to many platforms :P
01:00:41KlrSpzlostlogic, not a problem
01:00:51Ninjawlostlogic not me ?
01:00:57Mikachukapheine: do i click randomly on the letters?
01:00:58kapheineMikachu: I only partially implemented the color support. If you are bored and want something to do, go for it :P
01:01:15kapheineMikachu: No you control the robot (#) and walk into the objects, trying to find the kitten
01:01:30Mikachuah right i thought the frontpage had a javascript
01:01:43Ninjawso what CVS got working crossfade ?
01:02:06kapheineMikachu: Oh you meant the webpage :P
01:02:09lostlogicpreglow: btw, latest CVS should fix your bad response tmie on track changes ;)
01:02:11Ninjawand .... diskmode
01:02:20Mikachukapheine: yeah
01:02:21preglowlostlogic: great and good
01:02:26Mikachukapheine: should the icons just have random colors?
01:02:58NinjawIs that normal Rockbox does not boot on disk mode ?
01:03:00lostlogicpreglow: I'm really bad at stoping refactoring −− I went into pcmbuf.c to look for why CFM_FLUSH was being retarded and came out with crossfade refactored to all hell :-\
01:03:49kapheineMikachu: Yeah. I already assign the colors, I just don't have the code to actually change to the correct color when displaying the different characters.
01:04:04Mikachulcd_set_foreground or so i think
01:04:36kapheineIf I'm just changing the color for a single character displayed by snprintf, so I still use that, or is there another method? I haven't looked into it yet.
01:04:39KlrSpz rb->lcd_set_background(LCD_RGBPACK(0,0,255)); /* Pure blue */
01:04:49 Join chi [0] (n=chi@tor/session/external/x-5a1bae9db4c5d4aa)
01:04:51KlrSpzerr, only with foreground
01:05:35chihow large should the first partition on a 2gig nano be?
01:05:51kapheineHmm okay, welp currently I just store colors as a number between 1-6 (because this is how the program I ported did it), so that could be changed or a lookup table can be made to map 1-6 to some RGB values
01:05:56preglow80 megs or something
01:06:09chior boot partition i mean
01:06:25KlrSpzkapheine, enumeration?
01:06:32chipreglow, was that 80 megs for me?
01:06:37crashdkapheine: cant you just #define colour_blue LCD_RGBPACK(0,0,255)
01:06:47 Quit ashridah ("Leaving")
01:06:55kapheineKlrSpz: sure, or that. I just have lookup table fetish
01:07:04KlrSpzlike a hash table?
01:07:08crashdlookup tables are a waste of time in this instance, you've only got the 6 values ;)
01:07:08kapheinecrashd: Sure but the point is to easily map numbers to colors
01:07:14crashdyaeh, sure
01:07:30KlrSpzit'd prolly be more effective to just use defines
01:07:32kapheineKlrSpz: Simplier than that :P as long as LCD_RBGPACK makes a numeric value, we can just have a simple array
01:07:47kapheineint colors[] = { LCD_RGBPACK(...), LCD_RGBPACK(...), ... };
01:07:49Mikachukapheine: use set_foreground and then putsxy to put a char at x,y
01:08:09KlrSpzdoes it require any cross compiling?
01:08:17KlrSpzyou're making me want to write some classes n shit
01:08:33kapheineWhat? I'm just talking about an array of integers, that seems pretty simple to me :P
01:09:20KlrSpzmy only prob with that is if you don't know what the index's represent you have to look and figure it out. it's not very descriptive
01:09:39Mikachukapheine: in draw() you can just add lcd_set_foreground before the lcd_putsxy and it should work fine
01:09:55kapheineOh well in this case we don't care if 0 means blue or red or whatever. Because the only point is to give objects a random color
01:10:12 Quit Ninjaw ("Ninjaw P.B. baby")
01:10:19KlrSpzahh ok
01:10:31Mikachu rb->lcd_set_foreground(obj.color);
01:10:31Mikachu rb->lcd_putsxy(obj.x*font_w, obj.y*font_h, buffer);
01:10:51KlrSpzwhoops, wrong window
01:11:02kapheineLooks good to me, just gotta make sure obj.color is an LCD_RGBPACK value
01:11:57KlrSpzugh, i knew i should have put on sun screen.. i'm burning like mad
01:11:58preglowpatch works in sim, so it's not an endian issue
01:12:09preglowwhat's left of ipod problems except endian problems?
01:12:16Mikachuyeah that would help
01:12:16Mikachuan int between 1 and 6 is just different shades of black :)
01:12:16Mikachuapparently different shades of green
01:12:16DBUGEnqueued KICK Mikachu
01:12:16Mikachuwow, matrixy :)
01:12:16 Quit Mikachu (Success)
01:12:46KlrSpzwhat are ya'll making?
01:13:21kapheineKlrSpz: us?
01:13:30kapheineWe were talking about my robotfindskitten plugin
01:15:30 Join Mikachu [0] (
01:15:49Mikachukapheine: okay what i got now works, but doesn't guarantee objects aren't completely black
01:16:56kapheineMikachu: If it's easier, you can just change the randcolor function to return some nicer LD_RGBPACK values
01:17:05Mikachustatic inline int randcolor(void) { return rb->rand() % LCD_RGBPACK(255,255,255); }
01:17:13Mikachu robot.color = LCD_RGBPACK(255,255,255);
01:18:23kapheineI think it'd be better to have some predefined colors, having the colors totally random will (as you said) make some completely black items, but could also have colors that really aren't that different, or just hard to see :P
01:19:43preglowamiconn: what do you think about a transparent selector bar mode? :)
01:21:13KlrSpzwhy does the rb class have it's own random number generator?
01:21:23preglowi'm just playing around with some themes with backdrops now, and it'd be cool if the backdrop was visible through the cursor bar
01:21:31 Quit matsl ("Leaving")
01:21:37preglowright now it just overwrites it with the bg colour
01:21:55KlrSpzi'd like that, if you could change the hilight foreground
01:21:57 Quit Acksaw ("I'm off, see ya later!")
01:22:04amiconnTransparency would really eat power
01:22:13 Join Mikaelh [0] (
01:22:24crashdisnt there a patch that lets you do that preglow ?
01:22:28amiconnColour gfx isn't exactly fast right now...
01:22:30crashdim sure ive seen some wps' that allow you to do that
01:22:31preglowcrashd: no idea
01:22:33KlrSpzwould you really apply any transparency in that case? or just not hilight the background?
01:22:42 Quit Mikaelh (Client Quit)
01:22:46preglowKlrSpz: i think the bar is nice
01:23:04KlrSpzoh so you want some alpha stuff huh?
01:23:19amiconnThe selector bar uses the foreground colour for the background pixels and lets the background shine through the letters?
01:23:47preglownot the letters, no, just the backgroudn
01:23:54KlrSpzcan you draw a demo?
01:23:58KlrSpzto be precise?
01:24:01preglowby no means
01:24:12preglowthis isn't exactly the most important idea i've had in aeons
01:24:17CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
01:24:17*amiconn doesn't understand
01:24:28*KlrSpz doesn't either, but agrees with the importance of the idea
01:24:31preglowforget it, it's not vital to my continued existance
01:24:40KlrSpzwhat about a nice EQ?
01:24:47preglowwhat, another one?
01:24:57KlrSpzwell it seems a little odd with the default one
01:25:02KlrSpzor it did to me
01:25:04KlrSpzmaybe i was too lit
01:25:16preglowwell, you need to tell what is odd with it
01:25:21lostlogicpreglow: would it be correct that telling vorbis to seek as the very first action within a song would be a Bad Thing (TM)?
01:25:35preglowlostlogic: shruggage
01:25:36***Saving seen data "./dancer.seen"
01:25:51KlrSpzi think i'd like to see an option at least for going to verticle bars instead of horiz
01:26:03preglowKlrSpz: that'll be fixed in the future
01:26:15lostlogicpreglow: hmm, I think it is cuz I just got a data abort in vorbis_synthesis after trying to make it do that in order to deal with the "seeking in previous track when the decoder is on next track" bug
01:26:44preglowi would have thought resuming meant doing that anyway
01:27:01KlrSpzi'd also like to be able to define out-file paths, and maybe masks for the filenames (not really needed, but would be neat)
01:27:11Mikachupreglow: what about just drawing the letters in another color and not doing a bar at all?
01:27:28KlrSpzMikachu, that's what i was saying earlier
01:27:33lostlogicpreglow: resuming uses a different mechanism "offset" not "seek" not sure what the difference is, but it seems to make one.
01:28:36KlrSpzare there docs on the site about setting up the compile environ?
01:28:50MikachuKlrSpz: sorry, didn't read everything
01:29:10KlrSpzMikachu, yeah i figured, but i agree that i'd like that
01:29:15preglowMikachu: might work, but i like the extra visual cue the bar gives
01:29:35Mikachuwhat about a rectangle drawn around the text?
01:30:05preglowyou mean just like now? :)
01:31:01KlrSpzpreglow, not filled
01:31:36KlrSpzpolymorphism would make that easy
01:31:59preglowit's not exactly hard without it either
01:32:10preglowwe use function pointers now anyway, which is exactly the same
01:32:12Mikachuit already supports another mode
01:32:55Mikachuomg i accidentally discovered a new feature in gvim
01:33:04Mikachuif you doubleclick a parenthesis, it will select all text until the matching one
01:33:04KlrSpzwell you'd be able to pass it an object of which you'd like to draw
01:33:09KlrSpzinstead of calling a diff function
01:33:29KlrSpzthen the object you pass would have it's own DRAW method
01:33:32preglowMikachu: !
01:35:40Mikachupreglow: you use gvim?
01:35:48Mikachuor are you laughing at me for using it?
01:36:04preglowMikachu: i'm using gvim
01:36:18amiconnKlrSpz: Rockbox is plain C, not C++, so there are not objects
01:36:28KlrSpzwhy is that fortunate?
01:36:36KlrSpzsounds unfortunate to me
01:36:48preglowit helps us to avoid dynamic memory allocation
01:36:56KlrSpzi see
01:37:01preglowcoding c++ with no dynamic memory isn't very fun
01:37:12KlrSpzyeah, kinda defeats the purpose
01:37:15amiconnI sure as hell wouldn't touch rockbox if it would be written in C++
01:37:28lostlogichmm... correctly detecting "Playlist automatically ending because it's the end of the playlist" in order to save a NULL resume position is nontrivial.
01:37:29preglowKlrSpz: plus, a lot of the old timers here have an irrational hate when it comes to c++ ;)
01:37:35KlrSpzi wasn't under the axiom of minimal memory usage
01:37:50KlrSpzlo siento
01:37:52Mikachui think a lot of the arguments from linux not using c++ also apply here, and i say that without remembering any of them
01:37:59kapheineirrational? :P
01:38:10Mikachui don't like c++ because i don't know c++ :)
01:38:10preglowKlrSpz: nah, not minimal, we just want to save every last possible byte of memory for the mp3 file buffer
01:38:25preglowKlrSpz: if we had dynamic memory, we'd have to reserve a ton for the allocation pool
01:38:27KlrSpzsounds minimal to me, and i understand it
01:38:46lostlogicbreak time.
01:38:48preglowc++ is good enough
01:38:57KlrSpzi wasn't really considering the device's specs
01:38:58preglowtoo full of kludges, but ok
01:39:10KlrSpznow that i'm in that perspective, it's clear
01:39:18preglowand it tries to be c compatible, which is bad
01:39:22*amiconn tries to avoid oo in general
01:39:24preglowmakes it heaps more complex than it should be
01:39:28amiconnIt gives me headaches
01:39:37preglowwell, it's another way to think
01:39:41preglowoo in general is nice for bigger thing
01:39:42KlrSpzi love OOP
01:40:00preglowit pretty much forces you to make a good interface
01:40:03kapheineI'm with amiconn on that one :P
01:40:04preglowwhich is good
01:40:16KlrSpzthat's mostly what i do.. thinking in non oop would be hard for me at first
01:40:40preglowi mostly use c++ for work
01:40:45KlrSpzso yeah, actually... let's talk compiling environment.. is there any docs on it?
01:40:50preglowwell, yeah
01:40:52preglowwindows or linux?
01:41:01preglowlinux is the best
01:41:08preglowthen you just need to set up cross compilers
01:41:09KlrSpzif linux, i don't have xfree compiled or anything
01:41:11preglowand you're good to go
01:41:15preglowyou don't need xfree
01:41:23KlrSpzi was reading about a simulator that sounded neat
01:41:25preglowall you need is the cross compilers, cvs and an editor of your choice
01:41:35preglowfor the sim you'll need sdl and something sdl supports
01:41:43preglowsdl might support svgalib still :>
01:41:44KlrSpzk, so how do i get the cross compilers and cvs going?
01:41:58Tobulostlogic: shouldn't it resume at the start of the playlist?*
01:42:28KlrSpzpreglow, i really don't want to bother anyone, so if there are docs i can read them.. otherwise if you'd kindly help me get it going i'd also appreciate that
01:42:32Mikachuyou can probably use aalib with sdl
01:42:43lostlogicTobu: no, it should clear the resume position if playback ends
01:42:50preglowKlrSpz: the last link should be ok
01:43:37 Join TCK [0] (
01:43:46preglowstill no one with spdif here? :>
01:43:46Galoisyou can do sort-of pseudo OO coding in C, like the GNOME project does
01:44:00preglowGalois: lots of people do that, and when taken to extremes,it's the worst thing i know
01:44:16preglowbut when done with moderation, it's very neat
01:44:27preglowfor spdif people:
01:44:34preglowgive me feedback here, i'll read the logs
01:44:50Tobulostlogic: if you mean the position to resume at reboot, I disagree
01:45:31lostlogicTobu: I didn't design it, I'm juts trying to make it functino as designed.
01:46:12amiconnpreglow: Suddenly I am thinking of libflac, can you tell me why? ;)
01:46:52 Quit Farpenoodle ("Large sacks of cheese")
01:48:26preglowamiconn: nah, can't imagine why
01:48:35preglownow enough about that, i plan to sleep soon
01:48:49preglowthinking about libflac does not make for a good night's sleep
01:49:52blindxSo can I play video on my iPod using rockbox? (I don't have a video iPod, but I know that was a feature with iPodLinux)
01:49:52preglowamiconn: :)
01:50:00amiconnlostlogic: Something is wrong with playlist ends...
01:50:17lostlogicamiconn: :( specifically?
01:50:24amiconnMy album should have ended just now, but instead the playback code started repeating the last track
01:50:57 Quit Kohlriba ("Fast alle Menschen sind Regenwürmer")
01:51:16lostlogicamiconn: dangit
01:51:23preglowthat mail pretty much sums up what i feel about libflac, but in more polite terms
01:51:59 Join whatboutbob [0] (
01:52:16preglowwhatboutbob: in the nick of time
01:52:20preglowwhatboutbob: wanna test spdif?
01:52:38whatboutbobof course. :)
01:52:48preglowwhatboutbob: can you compile yourself now, or should i make you a build?
01:53:12whatboutbobafter an epic battle yesterday i can now compile again...
01:53:26lostlogicamiconn: bagh, that code looks right, and it worked when I tested it a while ago. :(
01:53:30preglowthen just use that patch
01:53:32whatboutbobdid you see there's another user in the forum that's reporting the same issue?
01:53:45preglowwhatboutbob: beware, i just discovered you need to have optical out option ON
01:54:06 Join Membrillo [0] (
01:54:08 Part Zzompp ("ERC Version 5.1.2 (CVS) $Revision: 1.815 $ (IRC client for Emacs)")
01:54:49preglowfinished build is here:
01:54:54preglowit will clobber your settings
01:55:22Membrilloanyone know why this page isn't updated anymore?
01:56:33Mikachupreglow: haha, that's some mail (flac)
01:56:34 Quit midkay_ ("Leaving")
01:57:08 Quit filoktetes (Connection reset by peer)
01:57:36preglowMikachu: better than they deserve
01:58:00 Join mtnbkr [0] (n=mtnbkr@tor/session/external/x-16df2ca415bfa12a)
01:58:09 Part Poka64 ("Parting with a reason")
01:59:38whatboutbobpreglow: ah...was that was stuffing us up?
02:00:03preglowwhatboutbob: no, there were further bugs too, which i THINK i've solved
02:00:10preglowlike i said, works here
02:00:22whatboutbobhehe...just like the last one. :)
02:00:34preglowthat sometimes worked here, heh
02:00:35whatboutbobgrr...ok...vmware's hanging on me. *sigh*
02:00:37preglowi can't make this one crash
02:00:41Membrillono one knows aout thay page? ^^
02:00:42whatboutbobany chance i could get a precompilation?
02:00:50preglowwhatboutbob: there's a link a page up
02:01:01whatboutbobbah. blind. thanks.
02:01:18preglowMembrillo: i've never seen it...
02:01:44 Join chi- [0] (n=chi@tor/session/external/x-1fc89c5b32068463)
02:01:45MikachuMembrillo: just look at the frontpage
02:02:09KlrSpzpreglow, is gcc 3.4.5-r1 good enough? it says 3.4.4 in the wiki, jus tnot sure
02:02:34Membrillonah, it was made especially for me. I use it in my rockbox updater app. i was just wondering why it doesnt stay updated
02:02:37preglowyeah, should be just fine
02:04:15Membrilloummm, weird, i just updated my build and im now getting codec errors whenever i try to play .mp3s
02:05:17 Quit tvelocity ("Ex-Chat")
02:05:30whatboutbobpreglow: no dice. :( I enabled optical output and i'm still getting stuck @ 8kb
02:05:45KlrSpzpreglow, so should i go latest on everything else except the gcc core?
02:06:22Membrilloah, fixed. just had to regresh settings
02:06:24lostlogicKlrSpz: you just need the latest ibnutils and gcc 3.4.6
02:06:31lostlogicKlrSpz: if you are building for coldfire
02:06:33 Quit obo (Read error: 110 (Connection timed out))
02:06:46lostlogicthen use 4.0.3
02:07:06KlrSpzi can use gcc-config to toggle between them right?
02:07:28lostlogicKlrSpz: uhh... this is a cross compiler, it's not something you'll use for building anything for your system.
02:07:39lostlogicare you using gentoo's crossdev or something? I don't recommend it.
02:07:45KlrSpzso you want me to compile it somewhere else?
02:07:48preglowwhatboutbob: i don't get it, i switch on my player, go to the recording screen, switch from mic to digital, and when i'm back in the recording screen, i can immediately hear the spdif input. when i record, everything goes as it should
02:08:14whatboutbobwait...i reset .cfg and it works!
02:08:16lostlogicKlrSpz: you're building a cross compiler, not a host compiler, even if you install a 4.0.3 cross compiler in /usr it won't overwrite or be used in place of your host cc
02:08:22*preglow thanks god
02:08:29whatboutbob...or at least i'm getting levels...
02:08:41KlrSpzlostlogic, yeah unless the package system removes it?
02:08:46preglowwell, if it crashes the first time, i'm not too hopeful
02:08:56preglowwhatboutbob: nothing in earphones?
02:09:02lostlogicKlrSpz: are you using crossdev?
02:09:13KlrSpzdon't think so, no
02:09:20KlrSpzi'm on gentoo
02:09:36lostlogicgatherd that already.
02:09:38KlrSpzhaven't done much manual compiling on this, mostly c# stuff (which is cake)
02:09:53lostlogicyou can't juts emerge a cross compiler.
02:09:56blindxSo can I play video on my iPod using rockbox? (I don't have a video iPod, but I know that was a feature with iPodLinux)
02:10:09lostlogicyou can either use crossdev, or you can compile it by hand and install it in /usr/local or your home dir.
02:10:16preglowblindx: no video
02:10:20KlrSpzhmm ok
02:10:24blindxshucks. in the future?
02:10:27preglowblindx: sure
02:10:27 Quit stoffel (Remote closed the connection)
02:10:27KlrSpzwhich do you suggest?
02:10:32preglowblindx: but it's not a big deal for us
02:10:45blindxit's not a big deal, per se. but it would be pretty neat.
02:10:47lostlogicKlrSpz: dunno, home machine I use it in home, server I have it in /usr/local
02:11:10KlrSpzi was thinking of making a user, and putting it in it's home
02:11:28Mikachuthat would be a bit silly
02:11:29preglowwhatboutbob: status update?!?!
02:11:32lostlogicKlrSpz: why would you do that? Then your own user would have a hard time using it!?
02:11:33KlrSpzwould it?
02:11:45lostlogicKlrSpz: either put it in th euser you will develop on's home dir, or in /usr/local.
02:11:45KlrSpzlostlogic, why is that?
02:11:50lostlogicJust trust me, seroiusly.
02:12:03 Quit chi (Read error: 110 (Connection timed out))
02:12:08Mikachui have mine in /usr/local/ipod-chain :)
02:12:36KlrSpzso making a user, and putting it in that user's home would be a bad idea? i mean i would log in as that user to do everything of course
02:12:44amiconnUgh, I just noticed the remote scroll settings aren't settable...
02:12:45lostlogicOn my server I just have mine in /usr/local/{lib,bin,share,whatever} because it means they're in everyone's path so rbclient and my user can use them easily
02:12:52whatboutbobstart/stop fie thus far.
02:12:59lostlogicKlrSpz: just seems islly to have an extra user for this...
02:13:09MikachuKlrSpz: i don't understand why you would want another user for it, it's not like someone would put exploits in rockbox cvs
02:13:12lostlogicKlrSpz: this is a separate C compiler, it will not interract in any way wiht things you are compiling for your host system.
02:13:15whatboutbobtransferring files to review now.
02:13:24amiconn(but my improved remote lcd init is rock-solid now)
02:13:24KlrSpzi'm not afraid of exploits, just trying to keep projects separated
02:13:33MikachuKlrSpz: mkdir
02:13:50KlrSpzjust get's ugly
02:14:14lostlogicKlrSpz: sup'to you, it's certainly not what I'd do.
02:14:35*whatboutbob stands and applauds preglow.
02:14:49KlrSpzlostlogic, i'm more afraid of it getting lost in all the crap i have on there
02:15:04lostlogicKlrSpz: then just install it in /usr/local where it's safe from your /home stuff
02:15:17preglowwhatboutbob: i just hope it can stand further testing
02:15:26lostlogicthen it'll be in your path, and you can compile rockbox without having to worry about it
02:15:30preglowwhatboutbob: anywho, thanks for testing
02:15:32preglowi gotta sleep now
02:15:36lostlogicg'night preglow
02:15:43Mikachupreglow: good night
02:15:47whatboutbobnight preglow. well done.
02:16:26whatboutbobpreglow: do you mind if i post the link for the precompiled version on the forum for further testing?
02:18:15 Quit KlrSpz ()
02:19:08Membrillois the h300 having pwrmgt issues lately? my h300 keeps shutting off randomly even if plugged into AC
02:19:37blindxcan rockbox play .wmas?
02:20:03 Quit TCK (Read error: 110 (Connection timed out))
02:20:05Membrilloif they do, it wont be for quite a while
02:20:15Membrilloif it does i should say
02:20:28Mikachunot quite following that
02:21:22Membrilloif rockbox does eventually play .wmas, it won't be for quite a while. how is that?
02:21:31Mikachuah, okay
02:23:04Membrilloso who should I talk to about my conundrum with that CVS page?
02:24:11Mooswhich one?
02:24:38Membrilloi think its just got something to do with the URL system of changing
02:24:50Membrillotheres a 'dist' in a lot of the URLs now
02:25:09 Quit muesli__ (Read error: 110 (Connection timed out))
02:25:38Moosthis one? there is in the home page
02:26:24Membrillono, thats updated daily, last5 updated with each commit. it was exactly the same as the one on the frontpage. I just need it for the rockbox updater app im finishing
02:26:52MoosBagder: you can fixe the front to "hourly" now :)
02:27:20Moosbut you can ask Bagder for this, probably sleeping at this hour
02:27:29MikachuMembrillo: shouldn't be too hard to parse the frontpage either
02:28:45Membrilloits a lot more difficult. I've just built a simple web browser into my app and i just pointed to that page. No parsing needed... getting that page to update again would be very useful
02:30:00Moostime to sleep here, night or whatever all
02:30:09 Quit Moos ("Glory to Rockbox !!!")
02:31:10 Join dj-fu [0] (
02:31:31 Quit macdonalder ("bye bye ;D")
02:34:57 Quit actionshrimp (Read error: 104 (Connection reset by peer))
02:34:57 Join |Unknown| [0] (
02:36:06amiconnlostlogic: The wps multi-update bug is also still there...
02:36:22lostlogicamiconn: you are the bearer of bad news.
02:36:55lostlogicamiconn: manual track change? On or off of disk buffer?
02:37:23dj-fuis there a cvs changelog anywhere?
02:37:24amiconnJust resumed, and the wps updates every 2..3 seconds roughly 10 times...
02:38:00|Unknown|can someone fix this theme?
02:38:12KyomiWhat in the everliving hell is wrong with my h320?
02:38:15|Unknown|idk why but the main background in the wps file doesnt show up
02:38:29|Unknown|and the main background file desnt show up
02:38:34lostlogicamiconn: oh, is it a problem, or just that it gets updated lots? I did hta ton purpose so that the next track info would populate earlier
02:38:55KyomiI'm trying to copy a 1.81GB rar file to my h320 and 1/4th the way through it says it can't find it and the usb disconnects/reconnects
02:39:03KyomiIt's like it's not allowing me to copy it over
02:39:09KyomiAny other kind of file works fine
02:39:18lostlogicKyomi: are you using rockbox disk mode or iriver's disk mode?
02:39:21lostlogictry the other
02:39:25amiconnlostlogic: Updating all info multiple times is a problem, because that restarts scrolling lines
02:39:30amiconnLooks really odd...
02:39:39Kyomilostlogic: it happens with both
02:39:52lostlogicamiconn: ooohhh... forgot about that. bah, will hafta come up with a better sol'n
02:39:58KyomiThe only thing I haven't tried is bootloader usb mode
02:40:18lostlogicKyomi: then it's probably hardware. You could also try setting your computer to a lower speed USB mode, or using a different computer.
02:40:19MikachuKyomi: maybe the problem is with the file on your computer
02:40:23lostlogicor that
02:40:28dj-fuIt's probably a windows issue
02:40:31*dj-fu nods
02:40:43*|Unknown| wants to know if someone fix this theme?
02:40:51Kyomilostlogic: You know... copying a 1.81GB file through internet is really unpleasent
02:41:00amiconnOkay, just tried again. Also happens after normal trackchange...
02:41:19MikachuKyomi: nah, takes about 8 minutes ;)
02:41:40KyomiI had this problem before with a 70somethingMB video file
02:41:41lostlogicamiconn: each time a track is read from disk it's updated currently, like I said, it was deliberate, because i forgot that updating more than once is a bad thing.
02:41:47KyomiIt was the only one that wouldn't copy
02:41:57KyomiBut I just zipped it and it worked fine
02:41:58Mikachulostlogic: on automatic trackchange, nearing the end of the from-song, should the info change to the next file a few seconds/moments before it switches songs?
02:42:04Mikachu(swcodec on nano still)
02:42:05|Unknown|Mikachu: It would take me over a month to download a 1.81GB file );
02:42:26lostlogicMikachu: depends if you are crossfading or not
02:42:31Mikachunot doing so
02:42:41 Quit Tobu ("Ex-Chat")
02:42:44lostlogicMikachu: then no, it shouldn't change tracks until the next track is heard.
02:43:20KyomiAlright... lets try extracting it directly onto the H320 :P
02:43:22Mikachuokay, probably something i did then
02:44:16KyomiThats odd
02:44:20lostlogicMikachu: could be a bug too...
02:44:22KyomiIt extracts it in bursts
02:44:33 Join midgey34 [0] (
02:44:41KyomiIt looks like it's going to crash but then it zooms through it really fast then stops again
02:44:47Mikachulostlogic: i'll try with the regular buffers
02:45:26KyomiI swear my h320 hates me
02:46:32KyomiOn that note
02:47:36KyomiDoes anyone know a good place to buy a more powerful battery for an h320?
02:48:11lostlogicamiconn: try CVS I just committed, should do what you want it to.
02:48:22 Quit twisted` (Remote closed the connection)
02:48:24lostlogicKyomi: ebay, just get an ipod 1/2 gen battery, and don't forget to switch the leads
02:48:31KyomiThis thing freaking hates me
02:48:35|Unknown|can someone convert this to 5g theme?
02:49:10Mikachu|Unknown|: maybe the lack of answers the first 34 times you asked means no?
02:49:46|Unknown|i wonder if i ask 34 more times
02:49:49|Unknown|ill get a response
02:50:10Mikachuhrm, looks like i have to make clean too
02:50:15Kyomilostlogic: I've tried bootloader usb, rockbox usb, original firmware usb and they all reset themselves at the same place
02:50:21Membrilloas a general rule. No response = not possible or people are too busy
02:51:25lostlogicKyomi: then it's either hardware, or it's your computer.
02:51:30KyomiThe file is 1.81GB and it says I have 1.82GB
02:51:34Kyomifree on the h320
02:51:48Kyomilostlogic: Strange thing is I can copy over anything else without a problem
02:52:00MikachuKyomi: can you copy the file somewhere else than your h320?
02:52:19KyomiI dunno
02:52:20Kyomilemme see
02:52:54lostlogicKyomi: and it's obviously ont a rockbox problem, so you might try asking somewhere like misticriver instead of #rockbox.
02:53:30KyomiI had this problem before with a video file transfering through rockbox.. but I just zipped it and it worked fine
02:53:47Mikachuokay, buffer size back to vanilla
02:53:58midgey34|Unknown|: someone else already posted a version similar to that for the 5G on the WPSGallery, but you'll need a number of patches applied and you'll need to compile on your own
02:54:20 Join Tobu [0] (
02:54:31amiconnlostlogic: Ugh! Now I get the playback hang...
02:55:02amiconn..but somehow my cvs up told me it updated more files than just playback.c ??
02:55:16Mikachulostlogic: hm yeah it doesn't happen now
02:55:22Mikachulostlogic: you're off the hook :)
02:56:02lostlogicw. t. f. *cvs ups and updaets his player*
02:56:14KyomiMikachu: I can copy that file to another hard drive fine
02:56:25KyomiMaybe it's the usb port?
02:56:48*amiconn hates maybe-bug
02:56:55*amiconn has an idea...
02:57:09 Quit Membrillo ()
02:57:45amiconnHmm, shit
02:57:59amiconnNot the cause I thought..
02:58:16amiconnIt worked after rebooting with voice enabled
02:58:19KyomiThe "shit" bug is contagious... it's already claimed lostlogic and amiconn
02:58:39Mikachudon't let the shit bugs bite
02:58:50lostlogicdung beatles? I thought the hang happens when a voice file is present but voice is disabled
02:58:58amiconn...but that's not the case
02:59:05KyomiWhen I try to move it to my player 1/4th the way though it says cannot find the file specified
02:59:53lostlogicI almost always operate with voice disabled and a voice file present
03:00:17amiconnlostlogic: Now the info reset only happens once ~2sec after resuming
03:00:32lostlogicright, in order to update the next track info
03:00:44amiconnWell, if it's a race, you may experience the hang sooner or later
03:00:45lostlogictehre is not currently a separate update for next track
03:00:58lostlogicindeed, I'm on the look out for the race.
03:01:12amiconnI got it once now
03:01:35amiconnWPS was shown and the disk icon was lit, but time didn't count and no music
03:01:41lostlogicI'm sure if I look at startup sequences that are effected by spin up times it'll become point
03:01:49amiconnI was still able to stop & get out of the wps that way
03:02:15amiconnRetrying several times always had the same effect.
03:02:22amiconnA reboot fixed it
03:02:45lostlogicso it's perhaps a bootup race between threads with the voice_init call or something...
03:03:23 Quit TeaSea ("Leaving")
03:03:35amiconnJust 2 lines of text on the remote look funny...
03:03:42amiconn(24px font)
03:06:28 Join Paul_The_Nerd [0] (
03:07:03midgey34lostlogic: I just cvs uped to the most current commit and all track changes seem to be broken
03:07:05 Quit |Unknown| ()
03:07:14dj-fuis there a cvs changelog anywhere
03:07:16dj-fu? :D
03:07:22midgey34manual and normal both freeze the player (h320)
03:07:40midgey34using MP3 files
03:08:40lostlogicmidgey34: can you tell me anything else specific about your player that might impact it? I'm skipping around during and not during fill, on and off buffer without trouble.
03:08:54lostlogicmidgey34: do you use crossfade?
03:09:08midgey34scratch that, track changes work fine with crossfade disabled
03:09:15lostlogicso crossfade is broken :(
03:09:20midgey34seems so
03:09:32lostlogicwhat are your crossfade settings?
03:09:45Mikachuhm, setting the seconds += line to 10 works, but 15 doesn't
03:09:58Mikachuconclusion, i will use 10
03:10:09lostlogicMikachu: weird.
03:10:29Mikachumy test isn't entirely scientific since i didn't use the same track for testing
03:10:33lostlogicMikachu: have you seen a positive battery life benefit from using that large buffer?
03:11:03Mikachui haven't done any battery life tests..
03:11:24lostlogicmidgey34: ...?
03:12:15midgey34yep, setting crossfade to shuffle and to track change only both freeze the player but if its set to always, it doesn't freeze sometimes, but there's no fading
03:12:22Mikachuokay, tried the same track now with the same result
03:12:31lostlogicmidgey34: and your settings?
03:12:39midgey34for crossfade?
03:12:43Mikachuthis is not with the normal plugin buffer size btw
03:12:43midgey34one sec, let me reboot
03:13:06 Quit Tobu (Remote closed the connection)
03:14:21midgey34Fade-in delay: 0 s, Fade-In duration: 1 s, Fade-Out Delay: 1 s, Fade_out Duration: 1 s, Fade-Out Mode: Crossfade
03:15:10 Join zelfric [0] (n=tharg@tor/session/external/x-aeb72f0631fc8bfe)
03:15:17Paul_The_Nerdlostlogic: This might be of interest to you:
03:16:58KyomiAlright.... this is really creepy
03:17:04KyomiThat file I said wouldn't copy.....
03:17:17KyomiI just zipped the .rar and it seems to copy fine now
03:17:32KyomiDoes it just not like solid archived .rars?
03:19:53 Join Daishi [0] (
03:23:35lostlogicmidgey34: I can't reproduce any kind of freeze with those settings, shuffle, track skip, or always, can you give me any further information?
03:23:59amiconnMono bitmap drawing on 16bit LCDs is definitely in need of optimisation :/
03:25:40***Saving seen data "./dancer.seen"
03:28:55midgey34lostlogic: what else would be useful to know?
03:30:34lostlogicmidgey34: ok, so settings are the time settings mentioned above, with track skip only crossfade enabled, and you are playing music of what format, and are skipping when the target track is on or off buffer?
03:31:13mtnbkrHi people... Anyone know if a loan of a 2nd Gen 10GB ipod to a dev be helpful in getting rockbox support for it?
03:33:56midgey34lostlogic: crossfade is actually set to shuffle and I'm using MP3 (192kbps or VBR encoded with lame) I just created a playlist and let a song play for 1.04 and hit >> and the player froze
03:34:36midgey34HDD is not spinning
03:34:46lostlogicmidgey34: ty *tries that*
03:37:20lostlogicbagh, worked flawlessly! /me joins amiconn hating maybebugs
03:37:44midgey34hmm, ill clear my other settings and try again
03:39:18lostlogicsigh, I should really fix the bug that freezes playback if I hit play and next in rapid succession... been in my TODO for a really long time.
03:40:02 Join JdGordon [0] (
03:41:06midgey34hmm, works with settings cleared and crossfade seems to be working
03:41:20lostlogicthe world hates you.
03:41:36lostlogicbut seriously if you do figure out exactly what setting caused the problem, let me know :)
03:41:40midgey34seems so
03:41:43midgey34I will
03:43:40lostlogicbagh, I should do laundry. I hate doing laundry.
03:45:13 Join Arrogant_ [0] (
03:52:08 Quit Arrogant_ ("Leaving")
03:54:32 Quit mikearthur (Connection timed out)
03:55:06ShonkyWinner! I have cracked the encryption on the Gigabeat :) Totally primitive and really only just obscuring things.
03:56:40lostlogicwho else was working on the gigabeat port? are you working with him?
03:57:16Paul_The_Nerdlostlogic: Okay, pretty easy to reproduce bug has been found.
03:57:25lostlogicdo tell
03:57:35lostlogic(so I can go on ignoring bugs :-P)
03:57:37Paul_The_Nerdlostlogic: Play a song. While that song's playing, leave the WPS, browse to another song, and click it.
03:57:47lostlogicon the wiki already
03:57:53Paul_The_Nerdlostlogic: Gotcha
03:58:04lostlogicI have some inkling of how to kill it, but haven't actually killed it yet
03:58:06ShonkyMarkun. We are talking and he gave me a few bits of info, plus he started disassembling the bootloader which is where the decryption is done. Without that I wouldn't haven't gotten this far so quickly.
04:00:03lostlogicPaul_The_Nerd: The wiki page is definitely a good resource both fo ryou and bug reporters, as it gigves some idea what's been fixed recently as well as what issues I already know about, plus some friendly people (hardeep) have picked up some bugs from the wiki page nad fixed them so I don't have to.
04:00:48 Quit Daishi ("Client exiting...")
04:02:07 Quit Arrogant (Connection timed out)
04:03:13 Join TCK [0] (
04:05:16Paul_The_Nerdlostlogic: Yeah. I know about the wiki page, I just didn't check it with this one because the guy said it'd been happening since the 20th, but it's never happened to me until I just CVS upped, and I swear it didn't even happen with the build I made this morning (though it almost must've I suppose.)
04:08:00midgey34lostlogic: it appears the problem does have to do with crossfade. setting the player to shuffled with crossfade on shuffle freezes at track changes. setting the player to shuffle with crossfade off works fine. setting the player to no shuffle with crossfade on shuffle works fine. setting the player to shuffle off and crossfade always also freezes the player
04:09:29midgey34with a cleared player with shuffle and crossfade enabled on shuffle, it freezes
04:09:41 Join bagawk_ [0] (
04:11:04lostlogicmidgey34: very strange, because I almost always use shuffle | repeat-shuffle, and that's the mode I was testing crossfade on :(
04:12:49midgey34this time though, the hdd was spinning
04:13:41lostlogic:( still having no luck reproducing
04:13:47lostlogicbbiab, laundry
04:21:51 Quit bagawk (Read error: 110 (Connection timed out))
04:27:14midgey34lostlogic: here's exactly what I do to get a freeze
04:27:34midgey34I turn on the player and hold record to clear settings. Then I go into the menu and set Crossfade to Shuffle, fade-in delay: 0s, fade-in duration: 1s, fade-out delay: 1s, fade-out duration: 1s, fade-out mode: crossfade. Then I set shuffle to on. Go back to the browser, navigate to "/Music/Artic Monkeys/Whatever People Say I Am, That's What I'm Not" hold navi and choose playlist > insert. Then I wait until the virtual led dissappears
04:29:20midgey34that's about as specific as I can be
04:36:25lostlogicmidgey34: after all that, you press next?
04:36:48lostlogicmidgey34: I wonder if it has to do with creating the playlist with navi vs. selecting the first file to play fromt he dir.
04:37:45midgey34yes I hit next (looks like >>| on the player)
04:40:16dj-fuquick question, what does Crossfeed actually do?
04:40:18midgey34and doing the same thing but selecting the file instead of inserting them by holding navi freezes as well
04:40:38lostlogicmidgey34: you are on H3x0, right?
04:40:44midgey34dj-fu: mixes some of the left channel to play in the right speaker and vice versa
04:40:48midgey34lostlogic: yes
04:41:02dj-funow, why would you want to do that?
04:41:04lostlogicmidgey34: this is officially either platform specific, or at least performance specific, but I'll see what I can find :-\
04:41:13Paul_The_Nerddj-fu: The easiest solution is to google the term "crossfeed"
04:41:40dj-fuit sounds better, ic
04:41:47Paul_The_Nerddj-fu: It mixes a very small amount and introduces a slight delay to the mixed portion to better simulate the sound you'd hear from speakers as opposed to earphones.
04:41:48midgey34dj-fu: some tracks have a very heavy difference in sound in each channel
04:42:13midgey34lostlogic: are you on h1x0?
04:42:33lostlogicmidgey34: ipod 5g
04:42:37dj-fuI'll reboot to windows and try the crossfeed plugin for foobar, as I can't seem to get nonskipping playback with crossfeed enabled on my nano
04:43:06Paul_The_Nerddj-fu: It's only really useful if you're using ear/headphones, so make sure to plug some in instead of using speakers
04:43:09midgey34could have something to do with gcc being dumb?
04:43:23dj-fuyeah, the speakers on my lappy are shocking - I always use headphones
04:43:25*Paul_The_Nerd goes to build an h120 build.
04:43:38Paul_The_NerdI'll see if I can reproduce
04:43:39lostlogicmidgey34: I'm guessing it's something to do with timing that varies between the platforms due to performance differences.
04:43:43lostlogicPaul_The_Nerd: thanks
04:44:07midgey34hmm what do I type in bash to see what gcc I'm running
04:44:18Paul_The_Nerdm68k-elf-gcc −−version
04:44:19lostlogicgcc -dumpversion
04:44:27lostlogicmm, yeah, what he said
04:44:29dj-fuis someone working on using the second core on the iPod's? from what I've been told, when this is solved I'll be able to have crossfeeded-equalisered music, without skipping?
04:45:03Paul_The_Nerddj-fu: It's not a priority at all since the iPod port is still more for devs than users.
04:45:04midgey34alright I'm on 3.4.5 compiling until cygwin
04:45:33dj-fuah, I understand it's still very new, but it's really the only showstopper I've seen
04:45:49dj-fui'm even content with using it, while skipping due to the fact I can just drag'n'drop music :P
04:46:04Paul_The_Nerddj-fu: I don't get any skipping on my Nano, but then I don't use the equalizer.
04:46:32dj-fuyeah, with eq & crossfeed disabled, using zezayer wps, I still get occasional skips
04:46:43Paul_The_NerdSo don't use crossfeed
04:46:58midgey34dj-fu: what files and bitrates?
04:47:13dj-funo Paul_The_Nerd, eq and crossfeed are both disabled
04:47:21dj-fumidgey34, mp3, mostly cbr, some are vbr
04:47:25dj-fuall between 128-320k
04:47:38dj-fuis there a better supported format?
04:47:39lostlogicdj-fu: does that wps have a peakmeter on it?
04:47:46dj-fulostlogic, no, it does not
04:47:47lostlogicogg is faster on ipod.
04:47:54dj-fuI see
04:47:59*dj-fu ponders converting 10gb of music to ogg
04:48:04 Part ProgramZeta
04:48:08Paul_The_Nerddj-fu: Odd, on my Nano, I don't get any skipping on MP3 music...
04:48:15lostlogiconr I on my 5g
04:48:23lostlogicunless I turn on crossfeed or eq
04:48:43dj-fuyeah, i was told to disable crossfeed & eq but still get skips occasionally, so I dunno
04:49:08dj-fuit's random though, I was listenining to a 100mb dnb set yesterday, no skips for the entire thing
04:49:36lostlogicdepsnds on the content of the music −− I get occasional skips in very difficult sections of live albums
04:50:24midgey34hmm to add to the "bugs" I found one with shutdown
04:51:18midgey34if you shut down while in the wps and the wps has a backdrop, everything is cleared but the backdrop remains and the splash says "shutting down"
04:51:40lostlogicah, not mine :)
04:52:09midgey34hmm, I'm running a version of HiPod 1.4 that works with cvs (no patches)
04:52:26midgey34I modified it to get rid of the margins etc
04:53:16Paul_The_Nerdlostlogic: Okay, I tried it with an album of FLACs, it did not happen. With an album of MP3s and it did freeze
04:53:53lostlogicPaul_The_Nerd: *sigh* I hate bugs I can't reproduce.
04:56:15 Quit lostnihilist (Read error: 104 (Connection reset by peer))
04:56:17Paul_The_Nerdlostlogic: Yeah, I can confirm the same thing *doesn't* happen on my Nano (though y'know, I don't know how much that says, being the Nano and all it seems to sometimes be a unique case for bugs)
04:56:54lostlogicdoes'nt happen on nano, nor 5g, but does on 140 and 340
04:57:14lostlogicwonder if it has to do with FIQ vs. DMA
04:57:49midgey34I wonder if the x5 has problems as well; I assume it would.
04:57:49 Join lostnihilist [0] (
04:57:59Paul_The_NerdIt seems odd that it happens when Crossfade is enabled, but in a situation where the music doesn't actually crossfade.
04:58:02lostlogicprobably yeah, but would beinteresting
04:58:06 Quit dj-fu ("Leaving")
04:58:18lostlogicPaul_The_Nerd: yeah... could be a helpful clue...
04:59:46midgey34hmm, h300-sim crashes too...
04:59:53midgey34doing the same thing I did above
05:01:03midgey34here's what stderr.txt reads at the end
05:01:05midgey34Warning: skipping 626 bytes of garbage
05:01:05midgey34Header: fffbb004, Ver 0, lay 3, bitr 192, freq 44100, chmode 0, mode_ext 0, emph 0, bytes: 626 time: 1280/49
05:01:05midgey34Space between ID3V2 tag and first audio frame: 0x272 bytes
05:01:05DBUGEnqueued KICK midgey34
05:01:05midgey34First frame is at 57b
05:01:05midgey34backlight: on
05:01:14 Join Rob2222_ [0] (
05:01:40midgey34but I think it crashes before any debug info is written
05:01:48lostlogiccan one of you see if it also freezes if the mode is mix instead of crossfade?
05:02:40midgey34well, it didn't crash on the sim
05:02:46midgey34ill test on target now
05:03:16Paul_The_Nerdlostlogic: Works on Target.
05:03:33lostlogicok, so manual skip crossfade on iriver breaks
05:03:43lostlogiccan either of you test on an ipod sim?
05:03:52Paul_The_Nerdlostlogic: By "works" I mean, "the problem does not occur"
05:03:55lostlogicto see if it's a code difference between them, or just performance
05:04:56midgey34hmm mixing works but I don't hear any actual mixing
05:05:39 Quit Kyomi ()
05:06:06lostlogicmidgey34: your particular settings are interesting, and I don't think any mixing -should- be heard in your case.
05:06:26midgey34well I just set fade-in duration and fade-out duration to 10s
05:06:36lostlogicand no mixing? :(
05:06:38midgey34and I still don't hear a mix
05:07:41midgey34hmm, that tim it seems like it mixed
05:07:56midgey34but the other two times were abrupt
05:08:08lostlogicbuffer level may not have been sufficient.
05:08:24Paul_The_Nerdlostlogic: I don't know if it's a related bug, but *sometimes* I get the WPS after doing the "Insert" operation, it shows the metadata for the first song it's going to play, but never progresses past 0:00 and doesn't let me leave. Backlight and volume are still responsive.
05:08:45lostlogicprobably unrelated, but also a bug
05:09:23midgey34hmm yes, mixing seems to work fine here
05:09:27midgey34ignore what I said earlier
05:09:53lostlogicok, so it's most likely in the crossfade_process_buffer, on platforms that perform better than ipod.
05:09:59lostlogicnow we're getting somewhere :)
05:10:24 Quit TCK (Read error: 110 (Connection timed out))
05:14:06 Quit zelfric ("sudden death")
05:14:42lostlogicno ipod sim tests?
05:15:28midgey34ill build it right now
05:16:43 Quit Rob2222 (Read error: 113 (No route to host))
05:18:30Paul_The_NerdI can't test the sims right now. I have no X, and no SDL.
05:18:50lostlogicPaul_The_Nerd: terrible.
05:18:58lostlogicI have a theory.
05:19:14Paul_The_NerdI moved to the old vmware and haven't bothered to manually install X or download the new image.
05:19:22 Join webguest08 [0] (
05:19:51midgey34go cygwin go!
05:20:16Paul_The_NerdI was on cygwin in the past, but it annoyed me for various reasons.
05:20:22Paul_The_NerdPlus, a VMWare image is vastly more portable for me.
05:21:18midgey34I agree, but I haven't had time to set an image up
05:21:39midgey34while he wait, I have a screendump of the other shutdown bug I mentioned
05:22:09lostlogicmidgey34: can you try changing the text of the pcmbuf_descs function from pcmbuf_size / PCMBUF_TARGET_CHUNK to pcmbuf_size / 24576
05:22:17lostlogicsee if that changes your bug at all.
05:22:38lostlogic(in pcmbuf.c
05:23:31Paul_The_Nerdmidgey34: That's just because it uses whatever backdrop image is currently in use. Since you're in WPS, it uses that one.
05:23:48whatboutbobhey guys: i've been getting a few 'check hdd connection' msgs on my h120 of late. I *think* it stems back to a couple of weeks ago when I had to pull the usb cable out without 'disconnecting' it in windows.
05:24:02whatboutbobi've tried formatting the drive to no avail. any other suggestions?
05:24:32Paul_The_Nerdwhatboutbob: That *usually* means either the drive isn't seated well, or it's going out.
05:25:09whatboutbob'going out' = dying a disgraceful death?
05:25:28midgey34lostlogic: the sim doesn't crash but the behavior is rather strange
05:25:36midgey34could be my computer though
05:25:42***Saving seen data "./dancer.seen"
05:26:03Paul_The_Nerdwhatboutbob: Possibly
05:26:19midgey34I'm getting skips in the audio and the peak meters drop to zero then it starts back up
05:26:24kapheineIs there a reason Rockbox doesn't pause playing when the headphones are removed on the iPod nano, or is it just not implemented yet? I can't find anything on the mailing list or wiki about it
05:26:38midgey34not implemented
05:26:55whatboutbobpaul_the_nerd: I'll open her up and have a look anyways. I think i've got a spare drive kicking around. I may swap them out to see if it makes any difference.
05:26:57midgey34there was a patch, but I think it wasn't the correct way of handling it
05:27:07lostlogicmidgey34: hmph, can you test on target?
05:27:13kapheinemidgey34: Do you happen to know where the patch is?
05:27:55Paul_The_Nerdkapheine: Just search the patch tracker
05:28:16midgey34lostlogic: those comments were in regards to the 5g sim with no changes in code
05:28:27midgey34ill make the changes now
05:28:38lostlogicmidgey34: oh, gotcha... this gets stranger and stranger.
05:28:44kapheinePaul_The_Nerd: Yeah I did, nothing seems to be coming up
05:29:04lostlogicso no bug on 5g sim, but yes oddness
05:29:07kapheineOh, I was just looking for the wrong thing
05:29:08lostlogicI hate computers.
05:29:12lostlogicI _REALLY_ hate computers.
05:30:00Paul_The_Nerdkapheine: I think this is it
05:30:18midgey34plau_the_nerd: that's it
05:30:27Paul_The_NerdIt's not well described.
05:30:35Paul_The_NerdAuto-pause implies to me that no user intervention is required.
05:30:49whatboutbobpaul_the_nerd: dumb question: is there a better option than 'format' to wipe a drive?
05:31:13Paul_The_Nerdwhatboutbob: Not really
05:31:42kapheinePaul_The_Nerd: Yeah, I was searchign for something about headphones, which is why I missed it. I'm reading through the discussion about the patch now, thanks.
05:33:50 Quit Paul_The_Nerd (Read error: 104 (Connection reset by peer))
05:34:33midgey34lostlogic: still freezes on target
05:34:39 Join Paul_The_Nerd [0] (
05:34:41lostlogicmidgey34: well so much for that theory.
05:36:36midgey34in regard to the other "bug," that is the intended behavior?
05:36:46lostlogicnot intended, but no big thing.
05:37:12midgey34I agree that its not a big deal, but I thought it should be pointed out
05:37:22lostlogicsure, file a bug so it's not forgotten, if you wish :)
05:37:56 Join nave7693 [0] (
05:39:00Paul_The_Nerdmidgey34: For it to show your menu backdrop, it'd have to swap out that portion of the buffer though. I mean it's a rather tiny operation, but it seems like changing the behaviour would just add a very slight bit more delay and computation too shutting down. (Probably not a humanly measurable delay, but still...)
05:39:52 Part nave7693
05:41:41CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
05:41:41*lostlogic found a bug which may be the cause of midgey's problem, not just a theory, but a real bug
05:42:09 Join RoC_MM [0] (
05:42:39Paul_The_NerdOooh, real bugs!
05:43:56lostlogiccommitted, see if that changes it.
05:44:40 Join nownot [0] (
05:45:01nownoti just put rockbox on a 5g ipod, how do i put photos on there
05:45:58midgey34Paul_The_Nerd: I don't think any backdrop should be drawn when shutting down the player, I think the entire screen should be cleared. Right now it shows my WPS backdrop
05:45:59lostlogic... you copy them from a USB host device (ie a computer)
05:46:23nownotlike directory wise or what ever
05:46:30lostlogichowever you like.
05:46:31nownoti put them in the photos folder but they dont see them
05:46:40whatboutbobpreglow: i was getting codec failure for any playback using the rockbox.iriver file you uploaded for optical recording testing. I reverted and all is well.
05:46:43 Join piroko [0] (
05:47:02lostlogicdid you browse to the photos folder in the rockbox browser?
05:47:23lostlogicwhat file format are the images?
05:47:43Paul_The_Nerdwhatboutbob: Well, you *just* downloaded rockbox.iriver which means its version didn't match your codec versions. That's not really a valid bug unless you had codecs built with that version.
05:47:54whatboutbobpreglow: also, i left my iriver unattended recording digital...upon my return the unit had shut down and a 0byte file was on the drive that i can't seem to salvage anything from. Not sure if this could be my hdd playing up tho.
05:48:10lostlogicnownot: I dunno, should work.
05:48:19RoC_MMthose 0 byte files are tricksters
05:48:21Paul_The_Nerdnownot: Do you see the Photos folder inside of Rockbox?
05:48:27whatboutbobpaul_the_nerd: thanks. thought that might be the case but wasn't sure so thought i'd mention it.
05:48:47Paul_The_Nerdwhatboutbob: It's possible your unit idle turned off before the file ever got flushed to disk the first time while recording...
05:50:07nownotyes well there is a shortcut named Photos and a folder named Photos in root on the ipod
05:50:33Paul_The_Nerdnownot: And you copied the files into the Photos folder by normal drag and drop in disk mode?
05:50:50pirokothe only thing keeping me away from rockbox right now is the freeze-after-71-minutes bug. I'm depressed...
05:51:03Paul_The_Nerdpiroko: What freeze after 71 minutes bug?
05:51:09lostlogicon 4g
05:51:22whatboutbobPaul: My idle off is set to 10minutes, so that makes that unlikely...right?
05:51:36Paul_The_Nerdwhatboutbob: Depends. How long total did you leave the unit?
05:51:50whatboutbobtoo long...couple of hours.
05:51:51Paul_The_Nerdnownot: And you don't see the photos when you browse into that folder while running Rockbox?
05:52:10whatboutbobit was recording for at least 5 mins before i left.
05:52:30Paul_The_Nerdwhatboutbob: I'm thinking that it's possible that recording is being counted as "idle" time (maybe a bug). Have you tried recording more than 10 minutes without touching the player, and watching it for an idle offf?
05:52:42nownotok created a new folder names pic on root, threw 3 files in there jpg. all about 60KB and it doesnt see them
05:52:55Paul_The_Nerdnownot: And what "File view" mode are you in?
05:52:58nownotwhen i browse in that folder on my ipod
05:53:19whatboutbobPaul: yup. first thing i did thereafter. Got p to 12 minutes before i shut down manually. File was saved correctly (as one would expect).
05:53:28nownotin which option
05:53:30Paul_The_Nerdwhatboutbob: Hm... Odd.
05:53:41lostlogicnownot: hold th emenu botton
05:53:47lostlogiclook at the bottom of the screen
05:53:51lostlogicwhat does it say
05:54:03nownotthere we go
05:54:08nownoti only had music showing
05:54:15Paul_The_NerdWell that would explain it. :-P
05:55:17lostlogicmidgey34: Paul_The_Nerd: any word on if my commit stopped dumb crossfade freeze?
05:55:23whatboutbobPaul: As I mentioned, it *may* be my hdd. I'll have a look-see now...I want to check for battery swell anyways.
05:55:43pirokolostlogic: I can test crossfade too. I use it all the time
05:55:49lostlogicpiroko: you have an ipod
05:55:54lostlogicit's not borken on ipod afaik :0
05:57:05pirokooh. oops
05:57:18lostlogicpiroko: our poor ipods are too slow to trigger the bug or something
05:57:58nownothow do i exit picture mode
05:58:22lostlogicmenu and select or hold menu or hold play or some damn thing... it's on the wiki somewhere :)
06:00:36 Join amiconn_ [0] (n=jens@rockbox/developer/amiconn)
06:01:14nownotok another dumb question does rockbox support video playback for 5g ipods?
06:01:17*Paul_The_Nerd needs to get a USB hub.
06:01:21Paul_The_Nerdnownot: No.
06:01:26 Join Manoj [0] (
06:01:33Manojhi folks
06:01:50nownotso for a 5g why is this supuroerere
06:01:55nownotyou know the word
06:02:08nownotyep thats it
06:02:10Paul_The_Nerdnownot: Well, it's superior on the assumption that you bought an MP3 player primarily to listen to music.
06:02:27ManojI just installed rockboc from 20060419 in an ipod video −− and when the machine is unplugged, rockbox seems to be there (no music on this machine yet).
06:02:28nownoti see
06:03:00Paul_The_Nerdnownot: Now, you can always just boot retail to watch your movies anyway. But most people would agree the audio capabilities of Rockbox are *well* beyond what any retail firmware offers now.
06:03:14ManojHowever, as soon as I plug it in, an apple logo appears, the display greys out, and it just says do not disconnect
06:03:27ManojI can't get any other response −− until I disconnect
06:03:32Paul_The_NerdManoj: This is normal. It uses Apple's disk mode.
06:03:33nownotok, how do i choose what i want to boot, it seems to auto boot into rockbox
06:03:40lostlogicManoj: ... that's what it'll do if you plug in the USB to go to disk mode :-p
06:03:48lostlogicnownot: hold menu during bottup to go to apple OS
06:04:26ManojOh. What does disk mode mean? That I should just mount the ipod, and copy music over?
06:04:35lostlogic... seems reasonable...
06:05:01midgey34lostlogic: I still get freezes
06:05:07lostlogicmidgey34: hates you.
06:05:08midgey34but I must be going
06:05:15lostlogicmidgey34: kk, see ya later
06:05:19midgey34good luck solving the problem
06:05:23ManojI guess that makes sense, since I can point any music player on my machine to access the ogss on the mounted ipod partition. Thanks, guys
06:05:25lostlogichehe, thanks
06:05:34 Quit midgey34 ("Download Gaim:")
06:05:41lostlogicManoj: I'd keep a copy both on ipod and on computer
06:05:53lostlogicManoj: little hard drives you carry with you aren't a good place to keep an only copy of anything.
06:06:04Manojlostlogic: well, I have it on my home machine, but not on my laptop
06:06:28Manojwhen I travel, the ipod is it, as far as sotred music goes
06:06:40Manojstored, even
06:06:55lostlogicmakes sense.
06:06:59lostlogicthat's what I do as well.
06:07:36lostlogicIf I were you, I'd also copy an updated rockbox build onto the player (latest CVS).
06:07:53lostlogicthere are constant bug fixes, so regular updates are adviseable, and use CVS builds, not dailies, IMO
06:09:10ManojI plan on, as soon as my elf-gcc finishes. I fell in love with the Green5g theme, and the patches are not applying cleanly to today;s CVS
06:09:38lostlogicmost gccs are elf, I assume you mean arm-elf-gcc :-P
06:10:04Manojonce I set up arch branches tracking CVS, and build parrallel branches tracking the theme, I'll go to more upto date builds.
06:10:14Manojyeah, arm-elf. Been a long day :)
06:10:24lostlogiccrazy patch using bastages
06:11:05Paul_The_NerdI'm a crazy patch using bastage.
06:11:41lostlogic*sigh* what is midgey's stupid bug. this is very frustrating.
06:11:47Paul_The_NerdThough mine are "USB goes into Retail instead of Disk" "Starfield shows FPS (predating the FPS plugin)" and "SID codec"
06:12:19lostlogicPaul_The_Nerd: yeah, yours are less core, so less likely to suddenly stop applying
06:12:41Paul_The_NerdThey've been broken here and there in the past.
06:12:55 Quit whatboutbob ("CGI:IRC (Ping timeout)")
06:13:14Paul_The_NerdBut never anything major
06:13:25lostlogicafter 3.0, I'm going to hafta put both pcmbuf and filebuf metadata on-buffer, and then work witih someone to redo album art
06:14:01Paul_The_NerdAnd then have millions upon millions of fans when you become the dev to commit album art.
06:14:19Paul_The_NerdOh, I know your tricks
06:14:39lostlogichehe, if I can just get voice working omre smoothely, I'll have some very dedicated blind fans :)
06:15:27Paul_The_NerdYeah, I hadn't thought of how valuable Rockbox would be to the blind community until I saw a bunch of people asking about it when voice stopped being usable.
06:16:14lostlogicyah −− I think it's onw the best it's been on swcodec, but still needs work and that "nano doesn't boot with a voice file" bug definitely needs fixed too.
06:18:02Paul_The_NerdNano/Voice related bug?
06:18:39 Quit amiconn (Read error: 110 (Connection timed out))
06:18:39 Nick amiconn_ is now known as amiconn (n=jens@rockbox/developer/amiconn)
06:18:41Paul_The_NerdNano boots fine with a voice file
06:18:45Paul_The_NerdDoesn't seem to *have* any voice
06:18:48Paul_The_NerdBut it boots fine
06:19:01 Join speacial_ed [0] (
06:19:28Paul_The_NerdAnd it freezes when you try to start playback
06:19:40lostlogicPaul_The_Nerd: ... that's not what preglow nor someone else were reporting about nano as of today
06:19:51lostlogicermh, it wouldn't play audio at all if it had a voice file on it or something
06:19:54Paul_The_Nerdlostlogic: Well it's what I'm reporting as of "I just built it, and copied it over"
06:20:02Paul_The_NerdIt doesn't play audio at all
06:20:08lostlogicok, yeah, that's the bug I meant
06:20:28Paul_The_NerdIt freezes like the freeze I described earlier, with time at 0:00, and audio still adjustable, backlight responds, but you can't leave the WPS
06:20:41Paul_The_NerdOr no, volume not adjustable
06:20:44*Paul_The_Nerd swears it was a second ago
06:20:48Paul_The_NerdBut the backlight at least responds.
06:21:07 Join Zoide777 [0] (
06:21:08lostlogicis a thread race condition
06:21:14 Join InteliWasp [0] (
06:21:24lostlogicprobably codec thread and voice codec thread are ina deadlock
06:22:01lostlogicsigh, my bug list doesn't seem to shrink, because I keep refactoring things.
06:22:28Galoisearlier today, when my ipod froze with the voice file, the volume was adjustable the first time you tried to play a song
06:22:37Galoisthen if you tried to play another song it all went to hell
06:22:58pirokoholy crap track changes are instantaneous now...
06:23:01Zoide777amiconn: hi, any news on grayscale lib?
06:23:10lostlogicpiroko: I hope you approve :)
06:23:35pirokouh... yeah! except that my ipod locks up after about two minutes now... :(
06:24:02lostlogicimproved track changes were the original reason behind my refactoring of crossfade today
06:24:23lostlogicpiroko: wonder if it could be the crossfade bug... is it on a track change?
06:24:51pirokoit was shortly after a track change. and if I reboot, rockbox freezes at the rockbox screen
06:25:00pirokoi have to flip hold to reset settings to get it running again
06:25:09lostlogicthe world hates you.
06:25:19pirokoI know
06:25:25lostlogic(I just have no idea what would cause that)
06:25:28pirokoI wrecked my dad's car today...
06:25:34lostlogicaww, you serious!?
06:25:35lostlogicyou ok?
06:25:36DBUGEnqueued KICK lostlogic
06:25:42InteliWaspcan someone help me figure out if i have found a bug or not?
06:25:47pirokoyeah. i backed up into a concrete wall...
06:25:48lostlogicInteliWasp: sure, what is it?
06:26:03lostlogicpiroko: yuck −− that's gotta be an ego killer :-P
06:26:03pirokohappy to see that someone cares :)
06:26:12 Quit nownot ("I-n-v-i-s-i-o-n 2.0 Build 3515")
06:26:14pirokoi don't have one...
06:26:41Paul_The_NerdI was in a very minor wreck Thursday. No damage to the cars, but that's just because I was incredibly lucky there was a car in front of me to keep me from sliding on into an intersection. =/
06:26:55Paul_The_NerdI would like to say that Austin roads are WAY too oily.
06:27:28InteliWaspi have an ipod 4G, when i plug in a usb cable that is connected to a charger, the ipod goes into disk mode the resets, then back into disk mode and so on...
06:27:32lostlogicahh... everyone should get bicycles.
06:27:57Galoishey I get a nice new bug that I haven't seen before
06:27:57pirokoInteliWasp: hold menu in rockbox, then plug in the cable
06:28:17pirokolostlogic: the bug hit on a track change this time
06:28:18Galoisfast forward on a song from the file browser is supposed to play the song right?
06:28:20InteliWaspok i'll do that next time, thanks!
06:28:28lostlogicInteliWasp: sorta bug yes. probably rockbox is missdetecting the difference between a charger connected to a USB cable and a PC connected to a USB cable, so the ipod disk mode puts it in that cycle.
06:28:36Galoisit doesn't anymore. It just brings up WPS for a second and then goes back to tree view
06:28:41Paul_The_Nerdlostlogic: Actually, if I lived in the city I would. Austin is great for bicycling, but there's not a safe way for me to get from where I am to where I could start bicycling safely.
06:28:48pirokoGalois: I've gotten that bug before
06:28:52lostlogicPaul_The_Nerd: *nod*
06:29:04pirokoGalois: It usually happens when I select the same track I'm listening to
06:29:13Paul_The_Nerdlostlogic: That, and it seems there's a never ending war between the two wheeled and the four wheeled folk of austin.
06:29:14GaloisI got it while changing tracks too
06:29:24lostlogicGalois: piroko: there's a bug currently that selecting a track from the browser doesn't work if there's one playing.
06:29:24pirokoGalois: yup, got that too
06:29:27lostlogicsee the wiki.
06:29:37lostlogicPaul_The_Nerd: that's true of every city.
06:29:41pirokoand now rockbox is completely unusable for me because of this new bug... I wanna cry...
06:29:43Galoisthe SELECT button works, just not the right arrow
06:29:48lostlogicPaul_The_Nerd: I was in a pretty bad bike-car crash last weekend.
06:30:11lostlogicpiroko: turn off crossfade for now, will probably go bye bye... I'll be working on it.
06:30:22pirokoit was off :P
06:30:32lostlogicpiroko: oh... wtf do yo udo to that poor thing :(?
06:30:39pirokoplay music?
06:30:41lostlogicGalois: oh!? *tries it*
06:31:13pirokolostlogic: and I *have* to reset the settings to get it to boot again
06:31:14lostlogicselect works just as notly as ff.
06:31:33lostlogicpiroko: :-\
06:31:37 Join dj-fu [0] (
06:31:39pirokolostlogic: :(
06:31:42lostlogicpiroko: wonder what changed for you.
06:31:49pirokono idea...
06:32:12pirokothe only thing I did differently from the bleeding source was apply the sid patch
06:32:17pirokoand it patched cleanly...
06:32:18Paul_The_NerdGalois: Select didn't work for me when I tried an hour ago.
06:32:23pirokoand I wasn't listening to sids...
06:32:40lostlogicbut any way, I must get laundry from dryer and go to sleep. Sorry for the bugs, will see what can be done tomorrow ;)
06:32:59Galoisgreat, now select doesn't work anymore
06:33:48pirokolostlogic: ok. goodnight
06:33:59pirokolostlogic: sweet dreams ;)
06:35:09pirokoi should probably go to bed myself...
06:35:34pirokoI'm waiting for my 500mhz laptop to compile the newest linux kernel... it takes a while...
06:35:59ManojWhat is the recommended bit rate people us for ogg files?
06:36:16 Quit dj-fu (Client Quit)
06:36:20RoC_MMif your used to 128kpbs mp3s use quality 3
06:36:22CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
06:36:22*piroko uses q3
06:36:30*lostlogic uses q7 and goes to sleep.
06:36:32RoC_MMif your used to 160-192 kpbs mp3s then pick quality 5
06:36:36lostlogicg'night all
06:36:39Paul_The_Nerd <−−- See, I KNEW this would come back around.
06:36:47pirokolostlogic: goodnight!
06:36:52RoC_MMyou should not pick a bitrate, because different songs are different levels and business and require different bitrates
06:37:08pirokoq3 is said to be transparent to the ear
06:37:20RoC_MMthe ogg encoder is smart, you just say "quality 5", and then it looks and sees what bitrate is needed to accurately reproduce the sound
06:37:29 Quit InteliWasp ("Leaving")
06:37:32RoC_MMI'm sorry
06:37:37 Quit Zoide777 ("CGI:IRC 0.5.7 (2005/06/19)")
06:37:39RoC_MMyou misspelled "q5 or q6"
06:37:49Galoisguruboolez says q2 ogg beats 128k mp3 for rock
06:37:55Galoisfor classical you need q3 ogg
06:38:24 Join dj-fu [0] (
06:38:24RoC_MMmaybe people are satisfied with q2 and q3 oggs for music...especially when you have small earbud headphones
06:38:35RoC_MMsee if you can tell a difference between the WAV and the q3 ogg
06:38:42RoC_MMif you can't, then its' transparent for you
06:38:45GaloisI'm satisfied with q-1 on isolation headphones, but I can tell the difference easily
06:38:51RoC_MMif you can actually tell the difference, then use q5.
06:38:54*piroko notes that his ipod locks up playing sids as well as every other music format :(
06:39:01*piroko cries
06:39:40pirokois there any chance of this being a battery problem?
06:41:32RoC_MMWhat is problem piroko ?
06:44:38 Quit dj-fu ("Leaving")
06:45:53 Quit RotAtoR ("zzzzzzzzz")
06:45:58Manojwell, I don't have anywhere close to 60GB of music, so I guess -q 6 it is.
06:47:15RoC_MMq6 will sound splendid
06:47:43 Join dj-fu [0] (
06:47:51Paul_The_NerdIf space is no concern, why not FLAC?
06:48:56RoC_MMor those other crazy new formats, wavpack, ....whatever the others are...
06:49:14RoC_MMone thing I decided
06:49:42RoC_MMis that since FLAC isn't as widespread as ogg (outside of rockbox), that instead of using flac, I'd use ogg vorbis quality 100 (about 300-500bitrate)
06:49:50RoC_MMquality 10 is pretty transparent
06:50:00RoC_MMand you get all the support of ogg vorbis
06:50:13RoC_MMand only a smidgen of loss that you'll never ever hear
06:50:32Paul_The_NerdYeah, unless there are any encoder errors, or codec flaw style artifacts. :-P
06:50:57*RoC_MM points out those are issues on all formats, including flac.
06:51:10Paul_The_NerdI like flac because I can transcode a selection of my music to fit any size player and format requirement in a batch operation overnight anyway.
06:51:47RoC_MMnever choose a lossy format and regret it
06:51:48Paul_The_NerdRoC_MM: They're not an issue with FLAC unless the encoder itself creates an improper FLAC file. You don't get artifacts with the codec itself.
06:51:56RoC_MMflac gives you that ability indeed
06:52:15RoC_MMI see.
06:52:48 Join TCK [0] (
06:52:49Paul_The_NerdA .flac file when extracted should create a bit-perfect identical stream to the original WAV. If it doesn't, either the encoder or decoder has a flaw.
06:53:44Paul_The_NerdBecause of the nature of lossy encoding though, even at higher qualities it's possible for a lossy encoder to make a mistake and generate an audible artifact.
06:54:01RoC_MMTrue. But if your wanting to pick the newest format with the most compatibility it would be Ogg and not FLAC I think.
06:54:09RoC_MMYes yes yes.
06:54:29Paul_The_NerdWell, if you were sure you never wanted to either transcode (from lossless) or re-encode again, you'd pick MP3.
06:54:35RoC_MMBut I have high respect for FLAC
06:54:52Paul_The_NerdBut yeah, Ogg is pretty good generally. Not as widespread as would be nice, but certainly more than FLAC.
06:55:54RoC_MMOne thing to realize
06:57:04 Quit hardeep ("[BX] Hrm... I wonder if I paid this month's electr...EOF From client")
06:57:08 Quit webguest08 ("CGI:IRC (EOF)")
06:57:36dj-fuwow, crossfeed rocks
06:57:38Galoisflac is middle of the pack as far as lossless codecs go, but it has the major advantage of being an open format
06:57:47dj-fui turned it on to see if it was good, played without skipping, sounsd really good
06:58:04RoC_MMI for example have decided that *all* new music that I have control over the encoding is going to be ogg vorbis and not that even though ogg vorbis is less mainstream than's a better codec. Mp3 will not be around forever, and ogg vorbis is a newer and better technology. If I choose Ogg Vorbis, I create demand, and demand creates support. But all my applications support ogg out of the box already (cd recorder, portable music devic
06:58:04RoC_MMe, ripper, sw music player).
06:58:36GaloisRoC_MM, I go a step further: I refuse to patronize music that I do not control
06:58:48Galoisit follows that all of my music is in ogg/flac
06:59:15RoC_MMI refuse to give someone uncontrolled access to my money if they won't give me uncontrolled access to what they give me.
06:59:43Manojwell. the difference in size between a 128bps ogg and an 192 bps off is an additional 30-40% disk spacem in my quick and dirty tests. Flac takes it up to 700%
06:59:49Paul_The_NerdAll of my CDs get backed up to FLAC on a large external HD so I can really easily batch-transcode them to whatever I want as my space needs change.
06:59:50HotfusionI can careless what format what plays as long as there is music playing that's all I need
07:00:06Paul_The_NerdManoj: It was being discussed in a "space is no object" situation though
07:00:22 Join Rondom [0] (n=Rondom@
07:00:31Hotfusionif rockbox get's better with battery consumpsion and improved flac and ogg I may convert my songs
07:00:33pirokoRoC_MM: sorry for the delay. there's a big lockup bug in the 4g ipod now...
07:00:39ManojPaul_The_Nerd: well, a60GB ipod is what I have, so space isn't exactly no object (and I think my question kicked off the discussion)
07:00:48RoC_MMManoj, a 128 kbps file using ogg will sound OK if your used to a 128 bitrate mp3. However, MP3 does not get "CD-quality" until 256Kbps..128 was never a good sounding bitrate. Ogg Vorbis gets "CD-quality" aruond q5 or q6.
07:01:19Paul_The_NerdManoj: Well you said you had nowhere close to 60gb of music, and then I said "Well, if space is no object..."
07:01:27HotfusionI try to get mp3'ss at 160 but I can't tell teh difference
07:01:31Manoj60GB is enough for me not to compromise on sound quality my ears can discern (though I am not sure I can hear the difference between q 3 and q 6, but I can afford to be vain :)
07:01:32Hotfusionsome can I guess
07:01:39RoC_MMHotfusion, don't ever transcode your songs if they are alreadyi n a compressed format...this would result in more loss of quality...simply choose to get new music in your new format.
07:01:54*piroko is just anal and uses 320 mp3
07:02:08Paul_The_NerdRoC_MM: I disagree. A 192kbps mp3 or an alt-preset-standard LAME VBR one can be considered CD quality to most people unless they're making a direct comparison to the CD, and even then not always.
07:02:44pirokoactually, i switched over to lame vbr ~192 for a while before switching to ogg
07:03:24Galoisdidn't lame drop the −−alt part years ago? It's just −−preset-standard now right
07:03:33RoC_MMPaul_The_Nerd, I think 192kbps is a good bitrate for MP3, but I we are obviously not talking about subjective feelings, since if that were the case, 128 would be OK for a lot of people. I am referecing documents I have read. I would think no one would disagree with 256 being transparent either.
07:03:58GaloisRoC_MM, 192 VBR means the hard parts are encoded at 256 and up
07:04:04Paul_The_NerdGalois: Well, it's habit to type the -alt to some people. As far as I know the encoders do the same thing now if you do it with or without (just to keep from breaking compatibility with scripts/etc)
07:04:32RoC_MMPaul_The_Nerd, if you want CD-quality, I'd say "256", but if your just are being normal, my mp3s are 192, as you suggest.
07:04:49dj-fuanyone know of a music player on linux which does crossfeed?
07:04:53RoC_MMAll the better Galois, for I use 192.
07:04:55dj-fui know there is a plugin for foobar on 'doze
07:05:05RoC_MMXMMS may have one somewhere dj-fu
07:05:07Paul_The_NerdRoC_MM: Well, the thing is it's all subjective. You can't draw a hard line at where "Transparent" is because people could create test tracks where 320kbps MP3s can be ABXed against a CD successfully simply because of weaknesses in the format.
07:05:11dj-funod, ty
07:05:21dj-fuxmms2 will be out soon, that'll be wicked
07:05:34Paul_The_NerdSo you have to go by either "Where transparent is for me" or "Where transparent is to the average listener"
07:06:15RoC_MMPaul_The_Nerd, I have no experience with ABX...but I think I'm rather done saying all that I think.
07:06:45 Quit Shadowarrior13 ("( :: NoNameScript 4.01 :: )")
07:06:51Galoislame −−preset-standard is transparent to most people and averages 192, with wide variations
07:07:06 Join mymomthelush [0] (
07:07:17Paul_The_NerdI'd say it averages about ~210 in my experience, but that could just be the nature of my music collection
07:08:04dpassen1Its even higher for most of my music
07:08:35RoC_MMhumans picking bitrate is stupid too
07:08:48RoC_MM−−preset-standard and Ogg's quality numbers are very good.
07:09:02Paul_The_NerdPicking a bitrate gives you a much more consistent time/size ration
07:09:21Paul_The_NerdIt's nice if you have a codec that can do bitrate averaging well.
07:09:28dj-fuanyone have a clue how long it'd take me to convert 10gb worth of mp3 to ogg?
07:09:39Paul_The_NerdYou shouldn't
07:09:40RoC_MMI care more about quality than if all my mp3s are the exact same size and bitrate.
07:09:56dj-fuyeah i know, converting lossless to lossless = loss
07:10:02RoC_MMPeople pick the bitrate because of the perceived quality, not because they want "exactly 1.5 meg per minute"
07:10:08Paul_The_NerdWell when you have 1gig of space to fit them on, sometimes it's nice to be able to know exactly how many minutes of audio you can put on it. :)
07:10:16Paul_The_Nerddj-fu: So why bother?
07:10:29dj-fuapparently ogg are better than mp3
07:10:32Paul_The_NerdRoC_MM: Well see, that's a mistake, unless you're doing the maximum bitrate of the codec.
07:10:46Galoisan mp3 converted to ogg will not be better than the mp3
07:11:00Paul_The_Nerddj-fu: In what way? I mean, if you're converting from MP3 to Ogg, then you have all the flaws of MP3 + All the flaws of Ogg, which ends up being worse than either
07:11:13RoC_MMPaul_The_Nerd, you have too much words to say at this hour EST.
07:11:18 Join mightybrick [0] (
07:11:21Galoisit's worse than adding up the flaws, because the flaws interact in bad ways
07:11:37Paul_The_NerdGalois: Well yeah, but adding them up is the simplest way to make it very clear. :)
07:11:41dj-fuic, nod
07:11:46pirokodj-fu: listen to mp3 at 64kbps compared to mp3 at the same bitrate. it's amazing...
07:11:47 Quit mymomthelush ("Trillian (")
07:11:57dj-fuI don't doubt it
07:12:09RoC_MMdj-fu, don't transcode, just relax with the ones you have an decide that *new tracks* will be in the new format.
07:12:10Paul_The_NerdRoC_MM: I had an afternoon nap. It's unfortunate, but I'll be up for many hours yet, getting things done.
07:12:30dj-fuI don't rip stuff, that's the problem, haven't bought audio for years
07:12:37 Join mymomthelush [0] (
07:12:55Paul_The_NerdThat would be your problem then
07:12:55pirokodj-fu: i meant ogg, sorry. it's late...
07:12:57RoC_MMGood music stores let you pick the format you want..
07:13:13dj-futhat's true
07:14:52 Quit mymomthelush (Client Quit)
07:15:49 Join mymomthelush [0] (
07:16:17 Quit piroko ("telinit 0")
07:16:29 Quit Nibbler (Read error: 110 (Connection timed out))
07:16:58 Join Nibbler [0] (
07:17:36 Quit mymomthelush (Client Quit)
07:18:23 Join mymomthelush [0] (
07:18:24 Join chi [0] (n=chi@tor/session/external/x-71ce6844fffed580)
07:19:41 Join piroko [0] (
07:20:45 Quit ^BeN^ (Read error: 110 (Connection timed out))
07:22:10 Quit speacial_ed ("Bye")
07:23:21 Quit piroko (Client Quit)
07:23:24 Quit chi- (Remote closed the connection)
07:23:55 Quit mymomthelush ("Trillian (")
07:25:08RoC_MMHave the 4G builds stabilized?
07:25:24 Join mymomthelush [0] (
07:25:45***Saving seen data "./dancer.seen"
07:26:09Hotfusionis may day here yet
07:26:32Hotfusion2 days till I get my 1st mp3 player
07:26:42Hotfusionsad isn't it
07:26:47RoC_MMwho says 4g will be stable on may 1
07:27:34 Quit mymomthelush (Client Quit)
07:28:50 Join mymomthelush [0] (
07:29:21 Part mymomthelush
07:31:26 Join mymomthelush [0] (
07:32:54Paul_The_NerdYeah, there's no guarantee 4Gs will be stable with 3.0
07:33:01Paul_The_Nerd3.0 is the iRiver release
07:33:32 Quit mymomthelush (Client Quit)
07:34:18RoC_MMit will be cool when rockbox has an exe installer for the ipod targets
07:36:49 Join mymomthelush [0] (
07:36:49 Quit RoC_MM ("Leaving")
07:36:58 Quit mymomthelush (Client Quit)
07:37:40 Quit dj-fu (Read error: 104 (Connection reset by peer))
07:39:17 Join mymomthelush [0] (
07:45:45 Join LinusN [0] (n=linus@rockbox/developer/LinusN)
07:46:09Hotfusionso I guess 3.0 won't matter all for the iaudio x5 series
07:47:07Paul_The_NerdWell, it'll matter a little bit to all series
07:47:14Paul_The_NerdThere are bugs that affect all platforms
07:49:21Hotfusioni heard 3.0 will help battery life. Knowing the iaudio is 10% done in that regard I hope it's better with 3.0 version
07:49:33Manojhoo. stuff cross compiled
07:49:41Paul_The_NerdDon't expect a magical change overnight
07:51:01ManojI keep getting hit by this failure to sync
07:51:09Manojand corrupted file systems
07:51:17Paul_The_NerdFailure to synce?
07:52:53Manojyeah. I'm back to doing sync. sync. sync. like it is back in the 70's again
07:53:25 Join whatboutbob [0] (
07:53:31Paul_The_NerdI have no idea what you're talking about.
07:53:39LinusNthe 10% for power handling on the IaudioPort page has nothing to do with battery life
07:54:20LinusNit has to do with implementing functions to control the charging etc
07:54:45Manojwhen I omunt -t vfat /dev/sdX2 /media/ipod, write stuff to /media/ipod/, and umount without sync'ing, I get file system corruption on the ipod
07:55:53Paul_The_NerdDid your iPod not do this before, or something?
07:56:04Manojwell, this is a new thing for me
07:56:18ManojI only started playing with the ipod today
07:57:17Galoisfor what it's worth, I've never had sync problems
07:57:23 Quit Sinbios ("If the definition of a klutz is someone who doesn't have eyes on their ass, then yes, I suppose I am a klutz.")
07:58:39LinusNManoj: linux?
08:00:34Manojah. This time I had the proper number of incantations of sync
08:00:59ManojGreen5g is working perfectly
08:01:43 Quit Rondom (Read error: 104 (Connection reset by peer))
08:02:21LinusNManoj: are you running a recent kernel version?
08:03:23GaloisI almost never bother to invoke sync
08:04:24Manojso not all that recent (I think we are up to 2.6.17-rc2 now)
08:05:08Manojwell, it is not a big deal −− I used to be very used to the triple sync back in the day
08:05:20LinusNthat's recent enough
08:05:22GaloisI got you beat, I use kernel-2.4.21-40.EL
08:05:39Galoisand kernel-2.6.9-34.EL on the laptop
08:06:10Manojheh. I work with SELinux, so bleeding edge kernels are de riguer
08:06:49Galoisto each his own. I prefer to stay with something that has known bugs
08:08:03 Quit TCK (Read error: 104 (Connection reset by peer))
08:11:23 Quit mymomthelush (Read error: 110 (Connection timed out))
08:11:57Hotfusionlinus do you think 3.0 could help with battery. I assume we'll know by testing when it's released
08:11:58 Join ender` [0] (i=ychat@
08:12:02Hotfusionfor the iaudio
08:12:53 Join speacial_ed [0] (
08:13:06Paul_The_NerdHotfusion: All 3.0 is the cumulative effort of what's been committed up until that day. Any performance improvements in general will help with battery, and any hardware optimizations that happen to coincide with hardware on the x5.
08:13:23Paul_The_NerdIt's not like 3.0 is already done or anything.
08:13:49Galois3.0 isn't even targeting ipods, much less iaudio
08:22:05HotfusionI'll try it anyway :) It's only going to get better
08:25:33 Quit speacial_ed (Read error: 104 (Connection reset by peer))
08:28:53 Join Acksaw [0] (
08:30:01LinusNthe iaudio is not a priority right now, but most of the optimizations for 3.0 will benefit iaudio as well
08:30:44LinusNthere are a few obvious x5-only optimizations to do, but i don't have time for that right now
08:30:45Hotfusionyeah that's cool
08:30:55Hotfusionjust curious is all
08:33:54blindxwith rockbox loaded, does my iPod charge, or do i have to load the original firmware to do that?
08:34:14Paul_The_Nerdblindx: It should charge fine if it's in disk mode.
08:34:38blindxdoesn't it auto-put-itself-in-disc-mode when you plug it in?
08:34:50Paul_The_NerdYes, but you can force it not to
08:34:57blindxi see.
08:35:07Hotfusionsaw a creative player yesterday. Smaller than I thought. Didn't care for it much heh
08:35:19Paul_The_Nerdblindx: Since I
08:35:19Hotfusionthe new one that they came out with
08:35:32Paul_The_Nerdblindx: Since I'm assuming you aren't forcing it to not go into disk mode, then yes it should charge fine
08:35:42blindxhaha ok. thanks
08:39:29 Join aliask [0] (
08:40:14 Join midkay [0] (n=midkay@rockbox/developer/midkay)
08:42:52amiconngood morning
08:44:46 Join mymomthelush [0] (
08:44:54 Quit mymomthelush (Client Quit)
08:45:48 Join mymomthelush [0] (
08:50:55 Quit blindx ("Leaving")
08:51:21 Quit JdGordon (Read error: 104 (Connection reset by peer))
08:51:26scorchelet's go!
08:53:05 Join speacial_ed [0] (
08:55:50 Join JdGordon [0] (
08:57:20whatboutbobLinusN: Just showing my ignorance again, but why can't gain be offered w/ digital recording? (just curious)
08:57:39LinusNbecause it's digital
08:57:45whatboutbobso it does go A > D > A?
08:58:20LinusNwe could of course let the cpu apply a digital gain, but that would put some strain on the cpu
08:59:21whatboutbobthe reason i ask is some preamps apply a minimum gain that's fairly hefty (eg +17dB)...I was hoping it would be possible to attenuate that somewhat.
08:59:49 Join blindx [0] (
09:00:02blindxThe only thing I don't like about rockbox is the lack of a "sleep" function.
09:00:11 Join petur [0] (
09:00:13LinusNblindx: please explain
09:00:23blindxWell, I have an iPod
09:00:38Paul_The_NerdLinusN: iPods don't power off. They suspend. Basically, drain the battery slowly for 3 days before actually shutting down, but boot *really* fast.
09:00:53blindxyeah that
09:00:58blindxbeat me to it
09:01:16blindxbut sans the "3 days" portion
09:01:20LinusNok, so the sleep timer is disabled in the ipod builds?
09:01:23blindxI've let mine go unplugged for a week at a time.
09:01:34*LinusN suddenly understands
09:02:07blindxyeah, so when i wanna plug it in in the car, which i often do, i have to grab it, wait for it to boot up, then find my song.
09:02:10Paul_The_Nerdblindx: Well, at some point it parks a copy of the ram contents onto the boot partition, and powers off fully. Not 100% sure when that point is though.
09:02:30blindxyeah, every now and then, i've had that happen
09:02:32blindxbut not very often
09:02:51blindxas opposed to the original firmware... i could 'sleep' it, and have it wake up to exactly where it was when it went into sleep mode.
09:03:00blindxand quickly.
09:03:10 Join linuxstb_ [0] (
09:03:10blindxbut other than that, rockbox is great
09:04:29blindxI also had a problem.. I was playing a song, and it turned off after a while. I guess the power-off timer doesn't care if you're playing a song or not?
09:05:12 Quit speacial_ed ("Bye")
09:06:06blindxAh crap. I'm gonna be late to class.
09:06:12blindxTalk to you all later :D
09:10:08amiconnLinusN: Does the X5 LCD remote have a backlight?
09:10:11midkayPaul_The_Nerd, the point at which the iPod actually powers off is within like 10 hours..
09:10:29midkayquite possibly it's 12.
09:10:29LinusNamiconn: yes
09:10:40Paul_The_Nerdmidkay: Odd. Somewhere I read it was more than a day.
09:11:07 Join speacial_ed [0] (
09:11:33midkayPaul_The_Nerd, hm, i remember a number of occasions where simply leaving it off overnight and using it rather early the next day required me to wait for it to start up.
09:12:05Paul_The_Nerdmidkay: Dunno. I could easily be wrong. Wouldn't be the first time. :)
09:12:19midkayPaul_The_Nerd, haha. :)
09:12:36amiconnLinusN: Do you know the lcd controller in the X5 remote?
09:12:48LinusNno i don't
09:13:05 Quit JdGordon (Read error: 104 (Connection reset by peer))
09:14:42 Quit whatboutbob ("CGI:IRC (EOF)")
09:15:25 Join JdGordon [0] (
09:15:58 Quit speacial_ed ("Bye")
09:17:06 Quit linuxstb__ (Read error: 110 (Connection timed out))
09:21:26peturLinusN: have some interesting usb behaviour: at work, connecting via the hub on my PC display sometimes makes my h340 stay in usb mode when disconnected. After some time, the text goes away and I'm left with a blank screen
09:21:59 Quit mymomthelush ("Trillian (")
09:22:02peturattaching it again doesn't mount it again
09:22:12peturit just freezes hard
09:23:02peturmaybe I should blame DELL as the lcd screen is from them ;)
09:25:16LinusNis usb charging enabled?
09:25:48***Saving seen data "./dancer.seen"
09:26:30peturin retailos?
09:27:23peturyuck... alright, will boot it
09:27:42peturplease wait :)
09:29:08peturwtf, the display is going nuts in retailos
09:29:20peturcontrast is wrong completely
09:29:57peturusb charging is on in retailos
09:30:05LinusNtry disabling it
09:31:43 Nick linuxstb_ is now known as linuxstb (
09:31:53peturweird, retailos is completely going nuts: some settings aren't selected at all and when you try to change them it starts writing radio buttons outside the settings dialog...
09:32:13peturas if we wrote into their eeprom?
09:32:35peturwhat does usb conn mode do?
09:32:59LinusNyou mean the hub/no hub option?
09:33:24peturit was set to 'none of the above' :/
09:34:07LinusNi think it selects if the ISD300 is to draw its power from the usb or from the battery
09:34:24LinusNsomething we don't control
09:36:33peturno difference
09:36:46peturwill now try it with conn mode set to hub
09:38:59peturhmmm... looks better (tried 3 times, no freeze anymore
09:39:27LinusNtime to find out how to do that ourselves then
09:39:52LinusNsomething for the usb charging patch
09:50:10linuxstbpreglow: Have you tested the libmad optimisations on your Nano? I noticed the comments talk about 32-bit memory accesses for the bit read functions, so I'm hoping that doesn't cause data aborts.
09:51:06 Quit Xerion (" ")
09:56:54 Join einhirn [0] (
10:00:00markunlinuxstb: Shonky cracked the gigabeat's encryption! Now we can start the real work :)
10:02:48 Join mikearthur [0] (
10:03:22LinusN0x19751217 - looks like a birthday
10:03:40markunyes, it could be
10:03:54LinusNChristian Hack, yeah right
10:04:09markunI even think it's his real name
10:04:25LinusNcool if it is
10:04:29crashdit's destiny
10:04:59LinusNplenty of Christian Hack out there, according to google
10:06:10 Join bagawk [0] (
10:08:11 Quit darkless (Client Quit)
10:08:33 Join SereRokR [0] (
10:12:07linuxstbmarkun: Yes, I read that in the logs. Congratulations.
10:14:23markunAh yes, I didn't read the logs yet, only saw it in #gigabeat
10:17:34 Quit bagawk_ (Read error: 110 (Connection timed out))
10:20:51linuxstbmarkun: Just looking at the firmware update files now. Seems that FWIMG00.DAT is an gzipped initrd.bin, FWIMG01.DAT is the kernel, and FWIMG02.DAT is the root filesystem...
10:21:37 Quit SereR0KR (Read error: 110 (Connection timed out))
10:32:13 Part Paul_The_Nerd
10:42:05preglowlinuxstb: no, just tried on the iriver
10:45:48preglowLinusN: where did you find the description of the sample rate field of the EBU1RCVCCHANNEL1 register? i can't seem to find it
10:47:37LinusNit's in the s/pdif spec
10:49:39preglowah, right
10:50:10preglowdoesn't spdif support 96khz as well, thouvh?
10:51:33LinusNi suppose, but the s/pdif docs i have don't mention it
10:51:35preglowyes, found out now
10:51:54preglowit's probably the double rate flag
10:52:27LinusNwhich one is that?
10:52:39preglowthe on right before the sample rate
10:52:52preglowcalled DRS
10:53:29 Join Lynx_ [0] (
10:54:01LinusNi see
10:54:04preglowi have no way of testing that
10:55:04LinusNneither have i
10:57:28 Join darkless [0] (
11:01:08preglowamiconn should have, but he needs a tos/jack adapter
11:06:04 Join damaki_ [0] (
11:08:09Slasherihmm, i could probably test it
11:08:24Slasheriat least i have the equipment with 96 kHz s/pdif in/out
11:14:00preglowright now?
11:14:36amiconnpetur, LinusN: The hub/no hub option probably selects the maximum charging current the LX2201 can draw from usb
11:15:07Slasheripreglow: not yet until i get to home.. but in the evening :)
11:16:31preglowSlasheri: ok, patch is at when you have time
11:16:47preglowSlasheri: if it works, the resulting file will probably be tagged with the wrong sample rate, though
11:16:53preglowbut it should play back really slowe
11:17:44Slasheripreglow: nice, i will try that
11:19:58preglowLinusN: and yeah, spdif monitoring doesn't seem to be able to clock correctly in the iriver firmware either
11:20:12preglowLinusN: right now i just use the uda iis clock, since i can't think of anything better
11:20:22 Join TeaSea [0] (
11:21:36 Quit damaki__ (Read error: 110 (Connection timed out))
11:25:50***Saving seen data "./dancer.seen"
11:27:30 Join tvelocity [0] (
11:27:58 Join whatboutbob [0] (
11:34:17whatboutbobpreglow: thrice now i've attempted to record optical for a while ( >1hr). i return to find the iriver off and a corrupt file.
11:34:45whatboutbobnot sure what has been happening but i should be able to monitor it now (have been busy).
11:36:13whatboutbobtwice the file has been ~900meg but the headers must be stuffed cos when viewed they appear to be about 3 seconds of silence.
11:36:24whatboutbobthe third time the file was 0bytes.
11:37:14 Join jochen [0] (
11:37:30peturwhataboutbob: the file gets synced on every write so the length should be ok, but the header isn't written until the file is closed
11:37:33jochengood morning/afternoon/evening/whatever. ;)
11:38:03jochenlostlogic: are you around? i've also got that track skipping problem.
11:38:13peturwhataboutbob: battery troubles?
11:38:38peturwhatboutbob, sorry ;)
11:39:11whatboutbobpetur: nah...thought it may have been hdd troubles tho but i tested 3 long recordings w/ mic & line in with no trouble.
11:39:25whatboutbobno worries peter ;-P
11:40:15preglowwhatboutbob: it would be interesting to see at what point it shuts down
11:40:21preglowi haven't got enough free space to record that long, heh
11:40:26 Join Paul_The_Nerd [0] (
11:40:39whatboutbobyeah...I'm going to monitor it now.
11:41:13whatboutbob900 meg...about how long's that work out to? tis about 10meg/minute isn't it?
11:41:25preglowdepends on the rate
11:41:34 Quit darkless (Read error: 113 (No route to host))
11:41:51petur44kHz stereo gives about 90 minuts (bit more)
11:41:52preglowbut yeah, a bit over 10 megs a minute, then
11:42:53preglowoh, and btw, recording _should_ support more than 16 bits (i can't test it) now, but the wav routine will still just write 16 bits
11:43:23preglowand i'm really not inclined to code anymore on this now
11:43:29whatboutbobso its just sampled down to 16 bits?
11:43:36preglowthe last bits are just ignored
11:43:51preglowlike i said, i can't test whether it works
11:43:54 Part mightybrick
11:44:01preglowso someone else needs to do that part
11:44:18 Part Paul_The_Nerd
11:45:09 Join darkless [0] (
11:45:53whatboutbobpreglow: i *think* my ad20 (what i'm testing on) outputs slightly higher than 16bits...but i can't recall where i read the specifics so i can't confirm just atm.
11:46:05 Join pondlife [0] (n=pondlife@
11:46:26preglowwhatboutbob: and i don't think it's possible for me to detect it anyway
11:46:37preglowwhich brings up a nice problem, i guess
11:46:47preglowwe might need a new option for iriver spdif recording
11:49:09linuxstbYou mean bit-depth?
11:49:35preglowlike i said, i don't know of any way to detect bit depth from the spdif interface
11:49:44linuxstbWe'll need that for analogue recording on the ipods as well (I hope).
11:51:09preglowit looks like we always receive twenty bits from the spdif interface
11:51:51 Join mymomthelush [0] (
11:51:53linuxstbYes - I think I read somewhere that spdif is always 20-bits, even when transmitting 16-bit data.
11:52:05linuxstb(it's just padded)
11:52:06 Quit mymomthelush (Client Quit)
11:52:14preglowbut i wonder how 24 bits is supported
11:52:16preglowi don't think we can anyway
11:52:35linuxstbFrom wikipedia: "S/P-DIF is meant to be used for transmitting 20 bit audio data streams plus other related information. It can be reduced to 16 bits by zero padding or to 24 bits by renouncing the extra information."
11:53:03preglowi don't understand that last...
11:53:08linuxstbNor do I...
11:54:06linuxstbBut isn't the Coldfire IIS limited to 20-bits?
11:54:16whatboutbobcan anyone make any sense of this: (re: my unit's output)
11:54:51whatboutbobok...i think it means it outputs 20bits but only 16.5 is meaningful...
11:55:06pondlifeIt's the wrong way round, surely..
11:55:21preglowlinuxstb: yes, we can't use anything more
11:55:28preglowlinuxstb: the internal audio bus is 20 bits wide
11:55:54preglowpondlife: it's 16 bit
11:56:19whatboutbobpreglow: oh, do you happen to know if its possible to apply negative gain w/ optical?
11:56:35 Quit jochen ("Verlassend")
11:56:44LinusNyes it is
11:56:52pondlifeI mean the wikipedia entry - you'd zero pad to obtain 24 bits and truncate to get 16
11:57:00LinusNbut we have no code for it
11:57:06whatboutbobthanks Linus
11:57:13preglowwhatboutbob: of course it's possible, it just doesn't make sense
11:57:43preglowwhen i'm recording spdif, i sure as hell want the bit perfect input
11:58:04preglowpondlife: you can't zero pad to 24 bits when the sample size is max 20
11:58:18whatboutbobHehe...there's a method to my madness. For recording live music a lot of preamps have a large minimum gain...sometimes too large.
11:59:14whatboutbob(eg +17-20dB)
11:59:27linuxstbIsn't applying gain to the digital input the same (but less reliable) as applying gain to the whole recording in post-production?
11:59:48 Join Xerion [0] (
11:59:57whatboutbobi'm talking about negative gain (attenuation?).
12:00:14amiconnattenuation for digital makes no sense at all
12:00:17linuxstbSo am I - it doesn't matter what kind of gain it is, it's still a digital gain.
12:00:24LinusNwhatboutbob: but you will still attenuate after-the-fact
12:00:34LinusNthe distortion is already there
12:00:48amiconnEither the input signal is low enough, then we surely want bit-perfect recording
12:01:11amiconnIf the input signal is clipping, digital attenuation won't fix that
12:01:35whatboutbobAh...ok...I wasn't thinking it through. Of course the distortion would already be there.
12:01:50preglowthe only thing you do when decreasing the amplitude of the digital signal, is throwing away bits
12:02:05preglowso you really should keep it as you recorded it until you have a good reason to postprocess it
12:02:49amiconnApplying positive gain might make sense - if we apply it before truncating to 16bit
12:03:30preglowsure, but let's not think about that
12:03:43preglowif you want more bits, you record with 24 bit wavs
12:03:49preglowkiss, etc
12:04:10LinusNand we currently only use the 16-bit fifos anyway
12:04:26amiconnIs the recording dma capable of handling >16bits?
12:04:47amiconnOkay, then I say forgot about 24bit...
12:05:00pondlifeWould you consider dither if truncating 20 to 16 bits? Or is that likely to use too much CPU?
12:05:28LinusNwell, i haven't found a good way to program the dma for the 20-bit fifos
12:06:25LinusNa second glance at the data sheet proves me right
12:06:35LinusN"It is possible to use DMA to transfer data to/from the FIFOs in the audio interface module. Only PDIR2 and
12:06:35LinusNPDOR3 registers support DMA, as others need more than 1 long-word to transfer data to/from the FIFO
12:06:35LinusNand cannot be used with DMA operation."
12:08:55preglowhmm? don't we use dma with PDIR1 now?
12:10:13LinusNPDIRx and PDORx are the fifos, to which any audio channel can be routed
12:10:25 Join damaki [0] (
12:11:10preglowso we actually use pdor3 for playback
12:11:23preglowi thought the fifos were coupled to the interfaces
12:11:25preglowbut that can't be, no
12:11:48preglowbut anywho, we should support non-dma modes some time as well
12:12:15preglowwell, it would be nice with 20 bit recording
12:12:22preglowand playback, i guess, even though i don't care much about that
12:12:40LinusNat least for lossless
12:13:36preglowwhich reminds me we still don't support dithering
12:14:00preglowperhaps i'll fix it when i overhaul dsp.c a bit
12:14:50amiconnImho 20bit playback makes no sense - considering the bad snr of the uda1380...
12:16:22preglowi still say we should give it a try. measuring audio quality from data sheets has always been a bad idea
12:16:37preglowplus, spdif out supports more than 16 bits perfectly
12:17:15amiconnAh, yes, s/pdif out
12:18:08amiconnI'm not judging the uda from the datasheet. If I can hear white noise from the headphone amp even at my usual listening volume, the snr can't be good
12:21:09preglowno, i don't think it's very good either
12:21:11preglowbut still worth a shot
12:21:28preglowof course, we'll have some trouble in the fact that our pcm buffer is 16 bit
12:24:52 Quit damaki_ (Read error: 110 (Connection timed out))
12:26:43 Quit darkless (Read error: 113 (No route to host))
12:30:11 Join muesli__ [0] (n=muesli_t@
12:30:30preglowwhich i think we should just make 32 bit anyway, of course :>
12:31:01 Join darkless [0] (
12:38:53 Join Moos [0] (
12:39:44 Join romanos_melodos [0] (
12:40:38romanos_melodoshas anyone any idea if rockbox is going to support video in next releases for ipod nano?
12:41:15Zagorromanos_melodos: not in next release, no. (3.0)
12:42:22romanos_melodosanyway.. can i just use linuxipod and rockbox together?
12:44:09 Join RedBreva^ [0] (
12:44:41romanos_melodospaul_the_nerd where are you?????
12:45:08 Quit Seed (Nick collision from services.)
12:45:10 Quit romanos_melodos ()
12:45:16 Join Seed [0] (
12:45:30RedBreva^Mornin all
12:48:42 Quit Seed (Nick collision from services.)
12:48:44 Join Seedy [0] (
12:49:57RedBreva^I am still strugling to get this 'Cross Compiler' stuff sorted... but I am getting closer. I have managed to make the sh1 tool chain, but am unsure which bits I need to repeat for the arm and m68k platforms... I assume I need to rebuild binutils with target="arm-elf" etc. but do I need to clean out the old sh1-elf build first, if so how?
12:53:24 Quit yobesoom (Read error: 110 (Connection timed out))
12:55:05peturRedBreva^: you do a make install after building, then start again for the next target in a clean (other) directory
12:55:34 Quit Xerion (" ")
12:59:38 Join Xerion [0] (
13:02:50RedBreva^So I have to do 3 separate binutils... binutils_sh1, binutils_arm binutils_m68k for example?
13:03:50peturlooks like it, yup
13:04:55RedBreva^Are the binutils needed to create the final firmware, or just to build the cross compiler?
13:05:25RedBreva^that is, can they be deleted after all the compilers are built?
13:05:41peturyou must make install them as well
13:06:36peturI only build one target, got bored with the long wait and downloaded the precompiled packages of the other targets ;)
13:06:57RedBreva^So the build folder can be deleted once the associated gcc/gdb is built?
13:07:37RedBreva^It's more a case of trying to understand it at the moment, rather than actually wanting to be able to build multiple platforms
13:08:09peturI think so, yes - but I'm in no way a knowledgable person in this matter
13:08:20RedBreva^But I do intend to get a Nano 4GB this week to add to the H140, so I will want multiple platform support soon
13:08:47peturdoing step 3 of is the easiest way...
13:09:35linuxstbYes, you can delete the build directory after you've done "make install".
13:09:58RedBreva^Did that in Cygwin, but I am trying to get my laptop working with SuSe10
13:10:22RedBreva^linuxstb: Ahh, Ok - thanks
13:11:56linuxstbI have all three toolchains installed, and I just installed them (you need both binutils and gcc for each target CPU) with different prefixes passed to configure - I use /usr/local/sh1, /usr/local/arm and /usr/local/m68k. You then need prefix/bin in your PATH for each.
13:14:23 Quit Seedy (Connection timed out)
13:15:35RedBreva^so having run configure with target="" etc for sh1, make and make install... can I reran configure for the next target, and I get errors, so should I delet the binutils and re-extract the archive?
13:16:15RedBreva^damn, my spelling is crap today!
13:16:26LinusNRedBreva^: don't run configure in the same dir as the source files
13:17:44RedBreva^That's what I am trying to follow, But I am finding it extreamly confusing... :(
13:18:07LinusNwhat parts are confusing?
13:18:42RedBreva^So I have binutils.gz, which I extract to a binutils folder... Is this not where the configure is run from?
13:18:53LinusNnot according to the instructions
13:18:58RedBreva^Names shortened to save typing mistakes ;-)
13:19:05LinusNsee "Create the build tree"
13:20:12RedBreva^Yep... I am in home/me/build/binutils running ../../configure isn't that running the configure from the source files?
13:21:06LinusNat that point, you are not in the source tree, since you are in an empty dir
13:21:17LinusNso that is perfectly correct
13:21:44RedBreva^OK, it's a terminology issue I think
13:21:59LinusNthat means that the home/me/binutils-blabla directory is untouched
13:22:27LinusNand you can just delete everything in the build/binutils dir and reconfigure
13:22:58LinusNno need to extract the binutils tar file again
13:23:01RedBreva^rm binutils/*.* ?
13:23:14LinusNrm -rf build/binutils/*
13:23:38RedBreva^It's starting to make a bit more sense now
13:23:53LinusNalso note that you can use the same binutils version for all targets, but you need different gcc versions for them all
13:24:28RedBreva^GCC 3.4.6 (coldfire) 3.3.6 (Sh1) and 4.0.2 (arm)?
13:24:43RedBreva^hehehe. getting there
13:24:54RedBreva^Thank you for your patience
13:25:51***Saving seen data "./dancer.seen"
13:28:59amiconnLinusN: I hope to get gcc 3.4.6 working for sh, together with using -Os for the core
13:29:09amiconnWill cut down binary size by ~5% ...
13:29:50RedBreva^OK, time to go to the Gym - thanks again
13:29:59 Quit RedBreva^ ("CGI:IRC")
13:30:06amiconnStrangely enough, binary_size(sh-elf-gcc-3.3.x, -O) > binary_size(sh-elf-gcc-3.4.x, -O) and binary_size(sh-elf-gcc-3.3.x, -O2) > binary_size(sh-elf-gcc-3.4.x, -O2)
13:30:18amiconnErm, wrong
13:32:51preglowLinusN: i take it we'll drop the extra wps patches, then?
13:33:07LinusNpersonally, i want them in
13:33:26LinusNi wish i had taken the time to include them earlier
13:33:32amiconnbinary_size(sh-elf-gcc-3.3.x, -O) < binary_size(sh-elf-gcc-3.4.x, -O) and binary_size(sh-elf-gcc-3.3.x, -O2) < binary_size(sh-elf-gcc-3.4.x, -O2)
13:33:53amiconnbut binary_size(sh-elf-gcc-3.3.x, -Os) > binary_size(sh-elf-gcc-3.4.x, -Os)
13:34:16preglowyeah, i'd like them too, but time is growing a bit short
13:34:32preglownot that we'll meet the date anyway
13:34:43LinusNi'm pretty sure we won't make it untim may 1
13:35:18linuxstbLinusN: Do those WPS-related patches need much work?
13:35:26LinusNi don't think so
13:36:36*amiconn thinks the already present hacks need cleanup instead of introducing even more hacks
13:37:01amiconnThe recording screen adaption to the remote (including peakmeter) is a big hack...
13:43:04LinusNamiconn: i agree that the wps patches don't have priority, but i still think they add good value
13:43:56LinusNi can't remember saying that i wanted the recording screen patch
13:46:21 Join _FireFly_ [0] (
13:46:35preglowwhat did it do?
13:46:43preglowi think the current recording screen is pretty decent
13:48:55LinusNtry it with the remote control :-)
13:49:06preglowworks fine
13:49:27preglowsample rate isn't there, but that's that
13:51:02amiconnIt works, but it's an ugly hack
13:51:21amiconnE.g. peakmeter must be same height and postion
13:51:40amiconnThat's why peakmeter on remote only works for recording, and not in the remote wps...
13:51:50LinusNmost of the features in rockbox started out as ugly hacks
13:52:01preglownothing wrong with a hack
13:52:04preglowwe'll always have them
13:52:14preglowwe should try to get rid of them, of course
13:52:17LinusNthat's the whole spirit of "release early, release often"
13:57:43amiconnLinusN: A wiki page describing the target tree concept would be nice...
13:58:09preglowamiconn: how is it a hack, btw? only thing i can see is that it's a bit tight on space
13:58:31preglowthat being a problem on the remote doesn't exactly qualify the attempt as a hack
13:58:56amiconnCheck the source...
13:59:17Moospreglow: speaking about hacks, what's about the musepack one? :)
13:59:27amiconnThe peakmeter needs to be made a proper gui widget, but that's only one problem...
13:59:33preglowMoos: it doesn't work on ipod, and i can't say i'm thrilled at debugging it
13:59:45Mooshehe : )
13:59:59preglowso it went further down my list
14:00:11 Join akaidiot [0] (
14:00:12Mooslet's hope the author of the patch will try to debug it himself then
14:00:17linuxstbpreglow: What's the ipod problem? data aborts, or does it just not work?
14:00:24preglowlinuxstb: just not work, all tracks are skipped
14:00:32preglowlinuxstb: works in sim, so endian should be alright
14:00:34preglowin short, i have no idea
14:00:44Moosuch :(
14:00:47linuxstbIf the endianness is OK, then maybe signed/unsigned chars.
14:00:53preglowand i'll try to restrict myself to 3.0 critical stuff for the time i have
14:01:00preglowlinuxstb: ahhhh, that i didn't think about
14:02:05amiconnMaybe we need to disable aac for 3.0
14:02:21linuxstbI was just thinking the same thing. But it works OK on the ipods.
14:02:32linuxstb(at least, my single 128kbps AAC file works)
14:02:50preglow3.0 isn't for ipods
14:02:52preglowand yes
14:02:56Moospreglow: did you spoke with the author of the musepack patch yet?
14:02:58preglowwe'll disable that if nothing more happens
14:02:59preglowMoos: no
14:03:05amiconnAny news about ffmpeg aac?
14:03:13linuxstbI'm just saying we should only disable it for Coldfire targets.
14:03:13preglowamiconn: yeah, it's in the works
14:03:29preglowlinuxstb: well, that's the only swcodec target we have for 3.0
14:03:43amiconnlinuxstb: The disabling would be only for release anyway, to be re-enabled for cvs builds afterwards
14:03:47preglowi see no need to disable it in cvs
14:03:59lostlogicpondlife: ping?
14:04:42Mooslostlogic: hi, crossfade completly broken+still negative values in debug audio, in case...
14:04:50pondlifelostlogic: hi
14:05:18lostlogicpondlife: if you have a voice file present during boot up, music doesn't play for you at all on H300?
14:05:39linuxstbamiconn: There seems to be activity in the ffmpeg AAC decoder, an initial patch was posted to the ffmpeg-devel mailing list, but it's still incomplete, and it's floating-point only at this stage. But I believe they are planning to implement a fixed-point version as well.
14:05:51lostlogicMoos: totally borked => freezes? Always?
14:06:03 Join robin__ [0] (
14:06:04lostlogicMoos: how do you create the negative values in debug audio?
14:06:17Mooslostlogic: here yep :(, 4sec out/in
14:06:19robin__Hi, does the gameboy emulator already work on the Ipod Nano ?
14:06:38MoosI stoped playback go to debug and -421 value : )
14:06:39pondlifelostlogic: I updated to the current (still) CVS build; H300 would no longer boot (just locked up at logo). I then tried clearing settings and deleting playlist_control (using RetailOS), but no improvement. Renaming english.voice allows me to boot.
14:06:41lostlogicMoos: all coldfire targets seem to have problems :(
14:06:48Moosuch :(
14:07:07lostlogicpondlife: ugh, thanks.
14:07:10robin__Hi, does the gameboy emulator already work on the Ipod Nano ?
14:07:26linuxstbrobin__: No.
14:07:28 Part robin__
14:07:46lostlogicMoos: playback stopped having strange values *shrug*
14:07:50pondlifelostlogic: I can now boot, but it crashes whenever I try to play.
14:08:12amiconnlostlogic: Not all. My H340 works fine with current cvs, apart from the single hang last night...
14:08:22pondlifelostlogic: Will be at lunch shortly, will attempt to disable voice and/or crossfeed to see if I can work out what's up.
14:08:25lostlogicamiconn: you don't crossfade, correct?
14:08:40pondlifeamiconn: Try enabling it...?
14:08:48Mooslostlogic: I usually don't skip with Crossfade Off, but I tried since Crossfade broken, and noticed som strange artifacts betwen tracks when buffer full? is this one behaviour related to your last changes?
14:08:58lostlogicno need, it will break coldfire targets with crossfade.
14:09:13Moos-1st ?
14:09:28lostlogicMoos: my latest changes should have eliminated a lot of skipping track artifacts.
14:09:49Moosohh, ok never tried w/o crossfade before
14:10:11Moosit seems still few of them then :P
14:10:22lostlogicdefine these artifacts.
14:10:41 Quit mirak (Read error: 110 (Connection timed out))
14:10:54Moosahah, my poor english will not help me, let's try :)
14:11:10amiconnlostlogic: Sometimes I also get these artifacts.
14:11:23amiconnA short chirp before the actual skip
14:11:25Mooslostlogic: some few ms of strange noises like deformed songs
14:12:37lostlogicis it on transitions from mp3 by chance?
14:13:08Mooshere tried with mp3's yes
14:13:30pondlifelostlogic: Have just disabled crossfade, but still crashing. Maybe it's beeps or voice that's causing my problem?
14:13:33Moosand happened when buffuer already filled
14:13:55lostlogicpondlife: disabled voice and moved away the voice file
14:14:06lostlogicMoos: did it happen with other audio formats as well?
14:14:22Mooslet me try...
14:14:43Moosjust mp3/mpc/ogg here
14:15:33 Quit akaidiota (Read error: 110 (Connection timed out))
14:16:38 Join zigford [0] (
14:17:35Mooswow, my X5 just refused to switch mpc−−>mp3
14:17:56lostlogicpondlife: both beeps and crossfade now use a similar mixing strategy, so it's possible that if one freezes on coldfire both will. still don't see why there would be all this trouble on coldfire and not on ipod with these changes.
14:18:05pondlifelostlogic: Beeps are ok
14:18:09 Join Kohlrabi [0] (
14:18:45amiconnlostlogic: mp3->mp3, as 99% of my music
14:18:48pondlifelostlogic: Have just got my H300 here and cleared settings. Enabling voiced directories (.talk clip) seems to be the problem
14:19:34amiconnpondlife: I always use .talk clips in the car...
14:20:14amiconnMy voice settings for in-car: Menus: enabled, Dirs: .talk clip, Files: Number
14:20:27Mooslostlogic: it sounds a bit like the last mp3 frame read is distorded during few milliseconds before to skip
14:20:34pondlifeamiconn: Are they working for you at the moment?
14:21:13amiconnCrossfade: off, beeps: off
14:22:28lostlogicMoos: yes, that's what I was semi-expecting, might be due to my 8 bytes-hack for mad.
14:22:29amiconnlostlogic: A further hint towards a race thing: I am using different (faster) ata timings
14:22:32pondlifeOK, I have crossfade:off , beeps: off and only voicing dirs (not menus or files). I then shut down and restart (having saved config). There is then hard disk access for loading the .talk clip, but nothing is played.
14:22:49Mooslostlogic: ok
14:23:49pondlifeThen I go into a folder and play an MP3 - it does a "soft" crash - i.e. backlight timer works, but no progress/playback and cannot get out of WPS
14:24:16lostlogicpondlife: that's the same thing nano is doing
14:24:18pondlifeAnd cannot shutdown.
14:24:25Mooslostlogic: and confirmed always negative debug audio values at stop
14:24:29pondlifeHDD icon was on initially, but is off now
14:24:36lostlogicpondlife: if you disable voice it will probably work
14:24:41lostlogicMoos: at stop, no concern for now.
14:25:43Moosit seems you that will be hard to fix all those issues before to come here :( and before release
14:26:06*Moos speptic about one release at 1st May :)
14:26:20 Join akaidiota [0] (
14:26:47lostlogicI am also skeptical, and not only because of my work.
14:26:58Moosindeed :(
14:27:35Mooslet's hope this last april week will be more bug killing : )
14:28:25*Moos is frustrated to canot help on this
14:28:31preglownah, too much needs doing
14:29:05Moospreglow: optimistic?
14:29:51preglowMoos: not at all
14:30:01preglowlinuxstb: char is unsigned on arm?
14:30:52pondlifelostlogic: OK, voice clips disabled has resolved that crash. But if I then re-enable crossfade, I can make a hard crash that way - but you probably knew that..!
14:31:01lostlogicyes, as noted on the wiki.
14:31:14pondlifeSo it works best if both voice and crossfade are disabled. Beeps seem to be safe though
14:31:16lostlogicI wonder if they are the same race somehow.
14:31:24preglowMoos: lucky you, i just made musepack seeking work on arm
14:31:46Moospreglow: wee \o/ lucky *us* : )
14:31:55pondlifelostlogic: Hope so.
14:32:10preglowi'll just chat a bit with the musepack people, then we'll see
14:32:11pondlifelostlogic: Actually, whilst beeps are not crashing, they are also not happening every time
14:32:40Moospreglow: that what I wanted to suggest you but didn't know how to ask you this :)
14:33:34preglowtheir irc server seems to be down
14:33:34pondlifeI'm not really au fait with debugging yet (except on the sim), so not sure how more I can help - please suggest if you think I can :)
14:33:47Moosin case you don't have link at hand
14:34:19preglowread it, but i prefer to see if i can contact them on irc first
14:34:24preglowi'm not too big on forums
14:34:35linuxstbpreglow: I think so - unsigned on ARM, signed on x86/Coldfire etc
14:34:43Moosthat what I said him in my post, to come here to speak
14:35:30 Join skwad [0] (
14:35:57preglowMoos: i don't exactly think this patch is watertight
14:36:26preglowafter seeking to near the end of a track, the track just never ends
14:37:42MoosI'm sure if you advise him a bit, the things can evolute a bit
14:37:58MoosI don't know if he is still working on it or not
14:38:36Moosor maybe you can do this yourself by any luck? :)
14:38:43preglowhe probably knows what's wrong
14:39:15Moosgood, he is then still working on it, at least let's hope so
14:39:36 Join _Lucretia [0] (
14:39:47preglowbut seeking does indeed work
14:39:52 Quit _Lucretia_ (Read error: 110 (Connection timed out))
14:40:00preglowlostlogic: any idea why the seek bar snaps back when seeking again?
14:40:05preglowor has that behaviour never gone away?
14:40:15Moospreglow: that is a good point for release, no?
14:40:28preglowbut it doesn't work correctly yet
14:40:45preglowi'll commit it if he fixes it
14:40:59preglowi haven't got time to look more into it right now
14:41:07Moosno problem
14:41:10preglowi'll leave some feedback for him on his patch
14:41:12lostlogicpreglow: didn't I fix that yesterday?
14:41:24Moospreglow: it's already a good thing
14:41:57 Nick _Lucretia is now known as _Lucretia_ (
14:43:00preglowlostlogic: perhaps :>
14:43:04lostlogicI just committed something which may or may not improve the crossfade situation, if some coldfire user could test crossfade the same way it was freezing now.
14:43:05Moosyep, fixed here
14:43:27Mooslostlogic: I will
14:43:43 Quit akaidiot (Read error: 110 (Connection timed out))
14:43:44pondlifelostlogic: so will I
14:43:50lostlogicpreglow: the reason was that the codecs would call a position update after a seek complete, but before decoding a new frame.
14:44:22lostlogicpreglow: er not after a seek complete, but after a seek_time had been set, but before a seek complete, so I made the playback ignore position updates during seek.
14:44:55preglowMoos: ehh, no tracks end with this patch
14:45:02preglowMoos: they just get stuck at the end
14:45:08Moospreglow: uch :(
14:45:52Mooslot of things to point to snowgoon then :)
14:47:02lostlogicw00t! I've just gotten the race -> playback acts stupid. maybe logf will help me now.
14:49:20preglowMoos: no, not much. only that
14:49:39Mooscool !
14:50:36pondlifelostlogic: Well crossfade still crashes here, hope this new find is the one!
14:50:42pondlifei.e. with your last commit
14:51:03lostlogicpondlife: latest one is not erlated to crossfade problem
14:51:07lostlogic(latest find)
14:51:15lostlogicit's related to the voice startup race
14:51:43pondlifeSo, crossfade is still unhappy :(
14:51:54preglowwhatboutbob: any news on the recording, then?
14:52:35 Join akaidiot [0] (
14:54:15lostlogicpondlife: I still haven't managed to reproduce the crossfade one here as I don't have a coldfire target to play with... makes it very difficult to hunt :(
14:54:34Mooslostlogic: still broken :(
14:54:40pondlifeCan I offer any logging stuff as assistance?
14:55:06pondlifeIt's totally reproducible - just skipping to next track with crossfade locks up hard
14:55:49lostlogicif either of you have an LCD remote that you can logf on and could get the last few lines of the logf from it freezing, that might be helpful.
14:55:55pondlifeAh - H300 sim gives a memory access fault at the same point...
14:55:58pondlifeI can look into that
14:56:16lostlogicpondlife: *nod* I coudln't get the H3x0 sim working right today
14:56:33lostlogicpondlife: if you have a chance, and could test exactly teh same in an ipod 5g sim, that would also be useful.
14:56:43lostlogicif the sims are different then ... I just don't even know :0\
14:57:42 Join chi- [0] (n=chi@tor/session/external/x-90c11c2de0af7920)
14:57:50 Quit chi (Remote closed the connection)
14:58:35lostlogicw00t! found my race, I'm 90% sure
14:58:55lostlogicpreglow: You were getting that nano won't play shit yesterday, right? Test latest CVS if you get a chance.
14:59:13pondlife5G = video, right?
14:59:13preglowlostlogic: i'm on my way out to wash my bike, will do later
14:59:23lostlogicpondlife: yeah
14:59:40lostlogicpreglow: no rush, my commit definitely fixed some bugs, just want to check that it fixed that one too. thanks.
14:59:51amiconnlostlogic: With a secondary pcm buffer for voice, would it be possible to mix in the dma isr?
15:00:16lostlogicamiconn: I've noticed _STEEP_ performance penalties for every instruction in the ISR, so I'd be hesitant.
15:00:23amiconn(mix one buffer ahead from what is currently playing)
15:00:28 Quit whatboutbob ("CGI:IRC (Ping timeout)")
15:00:41amiconnlostlogic: Oh?
15:01:00*amiconn can't imagine why mixing in the isr would be more demanding tan mixing in a thread
15:01:20lostlogicamiconn: I mean we could try it and see if what I was seeing was just flukey, but we're talking elimination of 1 or 2 instructions would be noticeable on the audio debug screen
15:01:56amiconnHmm, that's on ipod, right?
15:02:02lostlogicno, that was when I was on H3x0
15:02:17amiconnHow large is one dma chunk?
15:02:23lostlogic32k give or take
15:02:42lostlogicseems like they average 34k or some such
15:06:40pondlifeHmm - ipod 5g sim gives divide by zero error immediately!
15:06:59LinusNi can't see how the performance in the dma isr of the h300 could be critical
15:07:05 Quit akaidiota (Read error: 110 (Connection timed out))
15:11:26pondlifeShouldn't that CVS build be complete by now...?
15:11:50 Part LinusN
15:12:13goffathis is probably a known bug, but songs are locking about 2 seconds before the end of the track
15:12:23goffai have latest cvs
15:12:33goffawas experiencing this all weekend with earlier builds too
15:13:04peturwhen just playing them?
15:13:14ravonI got that too.
15:13:31ravonAnd some songs choke for 1-2s every now and then.
15:14:12 Join webguest78 [0] (
15:14:21 Join mborus [0] (
15:15:51goffapetur: yes
15:16:15goffahaven't had the songs hickup for 1-2 seconds though.. it just usually stops playback
15:16:19goffatowards the end
15:16:30goffaif i hit next track, it'll play next track though
15:17:22goffaat first i thought it was only mpc.. but it seems to happen on some oggs too... seems random though
15:18:17peturI listened to a bunch of mp3 on the way to work this morning - no problems
15:18:48goffayeah.. haven't had probs with mp3
15:18:58goffabut then i deleted a lot of them this weekend before i left
15:19:04goffareplaced them with mpc and ogg
15:19:05 Join Seed [0] (
15:19:16goffadon't know what percentage i'm at of mp3... probably 20
15:21:01 Join egotrippen [0] (
15:21:22 Quit jeremy ("-> office")
15:22:35egotrippenthis isn't really important but i wanted to ask
15:22:46egotrippenwas crossfade turned off on song skipping?
15:25:52 Join whatboutbob [0] (
15:25:54***Saving seen data "./dancer.seen"
15:25:54 Join akaidiota [0] (
15:26:34amiconnBagder, t0mas: Build hanging...
15:27:12lostlogicegotrippen: no, but it could be buggy.
15:27:20lostlogicgoffa: that's quite weird, I've not seen that
15:27:45Mooslostlogic: crossfader still broken even with last changes
15:27:49lostlogicamiconn: if Linus says it shouldn't be critical, we could try it
15:27:54lostlogicMoos: that change wasn't meant to change crossfade.
15:28:01Moosoh ok
15:28:12lostlogicshould fix voice
15:28:27 Quit _FireFly_ ("Leaving")
15:28:38egotrippenlostlogic: ok. i just noticed it wasn't, but it wasn't a "real" test
15:28:43whatboutbobpreglow: bah, well, my attempt to get a long recording to screw up failed dismally. I got bored after two and a half hours.
15:28:45egotrippenjust wanted to ask because i liked that feature
15:29:00goffalostlogic: doesn't seem to happen with random off...
15:29:18lostlogicgoffa: I always shuffle −− you don't use crossfade, do you?
15:29:19goffabut the problem seems to happen randomly :)
15:29:25goffano crossfade
15:29:26whatboutbobDon't know why the other 3 long recordings didn't work out. I'll put it down to the battery in my preamp until someone proves otherwise.
15:30:53whatboutbobpreglow: the only issue I have to report thus far is that,about 1 in 10 attempts to record hangs at 8kb.
15:30:55lostlogicegotrippen: it sounded to me like it tried to work, but didn't quite, you on the latest CVS?
15:31:48egotrippeni made a build like 8 last night
15:31:55lostlogicegotrippen: yeah, definitely works for me, try the latest.
15:32:16lostlogicegotrippen: what platform are you on, btw?
15:32:46lostlogicwait... it doesn't freeze up for you like it does for pondlife and moos and 123309812 other people !?
15:32:49lostlogicI hate computers.
15:33:03egotrippenlemme make a build, i'll check out anything you need me to
15:33:05lostlogica bunch of people are having crossfade just plain freeze up on coldfire targets.
15:33:18egotrippeni had a freeze last night, but i think crossfade was off
15:33:22whatboutbobpreglow: oh, if it is the battery in my pre-amp playing up, would that explain why the headers weren't being saved for the files?
15:33:26egotrippeni jsut reset it and it's been working since then
15:33:36Moosegotrippen: try wuth
15:33:43Mooswith very last build
15:33:47lostlogicunfortunately, I can't stick around for testing right now, must go to working, but I'll read the log when I get there.
15:33:57 Quit tvelocity ("Ex-Chat")
15:34:02 Quit mikearthur (Remote closed the connection)
15:34:15egotrippengimme a sec to dl the bleeding edge
15:34:15goffaso... things are more stable.. with crossfade?
15:34:56egotrippeni usually compile my own, but my two tiny patches shouldn't affect anything
15:35:19whatboutbobif that's the case, then it would be nice if the recording could shut down gracefully when a signal is no longer being received.
15:36:22egotrippenok, what do you want me to do?
15:36:53lostlogicegotrippen: just turn on crossfade, crossfade mode, and make a manual track transition that would normally crossfade.
15:37:01 Quit akaidiot (Read error: 110 (Connection timed out))
15:37:32egotrippenyeah, it crashed
15:37:51lostlogicalright, you're not an anomaly. what are the two patches you use, btw?
15:38:21egotrippenone to gwps.h to change the 'navi' button functions to the 'rec' button in WPS
15:38:40egotrippenand one to filetree.c to display files with '..' but not '.' so i can see all my albums
15:38:48egotrippeni have two or three that start with elipses
15:39:16whatboutbob...for my next trick (just for giggles really) i'm going to compare rockbox f/w optical recording with rockbox optical.
15:39:17lostlogicI still don't get how it can _always_ break coldfire and _never_ break ipoop. but now to work, back in a bit.
15:39:49egotrippenthanks a lot for the quick answer just now. every experience with the RB devs has been great
15:39:56egotrippenreally, thanks
15:41:01whatboutbobpreglow: thanks again for working on digital recording. I particularly appreciate your efforts because I understand you have very little interest in recording but still took one for the team. :-)
15:42:07amiconnwhatboutbob: rockbox optical should always be identical to rockbox optical ;)
15:42:34goffastill waiting for recording on x5..
15:42:52goffanot a big priority... but i do like to record my drumming from time to time to see what i'm doing wrong
15:43:01 Quit egotrippen ("CGI:IRC (EOF)")
15:43:03goffaor hear i should say :)
15:43:17 Join akaidiot [0] (
15:43:46 Quit akaidiota (Read error: 110 (Connection timed out))
15:45:43preglowwhatboutbob: ARGH!! it still has hangs?
15:49:58 Quit Kohlrabi ("Fast alle Menschen sind Regenwürmer")
15:53:48 Quit aliask ("bye")
15:56:46 Join XavierGr [0] (
15:56:56XavierGrHello All!
15:57:44pondlifeCan somebody give the build servers a kick...?
15:58:15peturah XavierGr...
15:58:41peturI'm unable to connect to the share of vmware, any ideas?
15:59:13peturcould it be because my pc is in a domain?
16:02:19preglowmborus: here?
16:02:40 Join Flub [0] (
16:05:07amiconnpetur: Check the networking setting of the vm. To make sure network acces host<->guest works, it needs to be either 'NAT' or 'Host only'
16:06:37mboruspreglow, I'm on a lunch break
16:06:47amiconnWith 'Bridged', the vm needs an IP from the same range as your host's lan ip in order to make the access work
16:06:49peturit's NAT and I can cvs update ok, but i can't see the share :(
16:06:51mborusI just posted some spdif tests in the forum
16:07:27preglowmborus: i know
16:07:41preglowmborus: i just wondered why the hell the satelite recording doesn't save the wav as 48khz
16:07:46preglowmborus: it should, i've tested it myself
16:08:04peturamiconn: and host-only does the same
16:09:02mborusDon't know. Maybe it's because I did some 44.1 digital recordings and then tried 48 kHz without rebooting.
16:09:11preglowmborus: should still work
16:09:38preglowmborus: check in the spdif debug screen and see if the sample rate is set correctly there from the satellite source
16:09:46mboruspreglow I need to get back to work - I will do some more tests later tonite
16:11:10preglowi really, really don't have any more time for recording now
16:11:26preglowthe fact that i also have to guess what to do doesn't ease anything
16:11:56mborusdebug shows 44.1.
16:12:09 Join akaidiota [0] (
16:12:52preglowmborus: then that's the error
16:12:56preglowmborus: it lies about its sample rate
16:13:15preglowhmm, or no
16:13:24preglowi think the sample rate detected is estimated from the actual sample rate used
16:13:28pregloware you 100% sure it's 48khz?
16:15:39mborusThe signal is 48kHz. For sure. Unless the satellite receiver changes that internally. I'm back from work in aprox 4 hours and will check that possibility.
16:16:03Slasheripreglow: hmm, now back home.. i could try the patch soon
16:16:12preglowSlasheri: cool
16:16:25preglowmborus: in any case there's absolutely nothing i can do about it
16:16:30mborusSorry, need to go now.
16:16:33XavierGryou type \\debian and it won't connect?
16:16:36 Quit einhirn (Read error: 104 (Connection reset by peer))
16:16:51 Quit mborus ("CGI:IRC (EOF)")
16:17:00amiconnSlasheri: Found my tagcache test results in the log?
16:17:01XavierGrhmm try run->cmd
16:17:08XavierGrping debian
16:17:30peturunknown host
16:17:31Slasheriamiconn: yes, seems pretty good. I will try to figure also the bugs soon
16:17:51linuxstbI've never seen any broadcasts which are not 48KHz, but the DVB specs do allow for 32KHz and 44.1KHz, so in theory it's possible...
16:18:06XavierGrpetur: while running the VMware image, right?
16:18:18peturhehe, sure
16:18:24amiconnOn H1x0 and Ondio, I observed less. resp. no commit speedup. Seems like the disk/media is the limiting factor there
16:18:24preglowlinuxstb: still doesn't change the fact that coldfire thinks the sample rate is 44.1khz
16:18:35preglowlinuxstb: the value i read is estimated from the incoming audio clock
16:18:40amiconnThe speedup on archos is remarkable...
16:18:56XavierGrpetur: does ssh work?
16:19:21petureuh.. linux noob here
16:20:09goffaso... if the coldfire seems to hose up quite often with new builds, etc... why do devs seem to prefer it, or is that sarcasm? just asking
16:20:16 Join romanos_melodos [0] (
16:20:49 Quit zigford (Read error: 110 (Connection timed out))
16:20:56goffa<- owns a cold fire
16:21:08goffaprobably going to get another one if i win my auction
16:21:29peturXavierGr: I can ssh debian in eterm and that connects fine
16:21:34romanos_melodoshowdie! Anyone knows where can i find some rockbox games except from the ones allready included?
16:21:43peturif that's what you meant
16:21:46lostlogicgoffa: I don't think coldfire is more hosed than other targets...
16:22:03XavierGrpetur: what the ethernet box says in VMware window? (top of the window)
16:22:05lostlogicgoffa: I do prefer the m68k assembly to arm though (probably from lack of arm knowledge)
16:22:24lostlogicI'm sure I break coldfire more, because I don't have a coldfire device to test on currently.
16:22:26peturalso tried host-only
16:22:28XavierGrand it says connected?
16:22:31romanos_melodosi hate vmware\
16:22:45goffaah... not being critical either, just asking
16:23:02peturcan even cvs update
16:23:13romanos_melodosanyone know where else can i find some rockbox games?
16:23:29XavierGrpetur: do you have a windows ssh client? (though I doubt it will see the machine if the ping failed)
16:23:34romanos_melodosi wanna play pacman , and it's on the pdf but not on the rockbox
16:23:47 Quit akaidiot (Read error: 110 (Connection timed out))
16:23:55peturXavierGr: no
16:24:33peturXavierGr: I'm wondering if I rebooted since installing vmware, maybe I should...
16:24:56romanos_melodosi go to : browse plug-ins, pacbox and it says no roms ....
16:24:57XavierGryou should.
16:25:08XavierGrbecause the ethernet adapter maybe isn't configured
16:25:13lostlogicBagder: t0mas: in case you didn't see, the build server seems stuck.
16:25:24goffajust locked again on an mpc
16:25:30XavierGrromanos_melodos: you need the right files to download
16:25:39romanos_melodoswhere can i find them?
16:25:47goffanext got it out
16:25:51XavierGrSearch Google: we can't distribute them
16:26:17peturXavierGr: will do that tonight then - payed work now... and seriously, this W2K dev box sometimes isn't rebooted in months :D
16:26:42XavierGrromanos_melodos: read the wiki page for the exact list of files needed.
16:26:47XavierGrpetur: ok
16:27:01XavierGrpetur: I just hope it will work...
16:27:38peturI'm not sure - it means I can do RB work here too.... :P
16:28:03XavierGrnice :) How the USB OTG goes?
16:29:26 Join Kyomi [0] (
16:30:22XavierGrpetur: Control Panel -> Network Connections.
16:30:41XavierGrDo you see: VMware Network Adapter VMnet1 and VMware Network Adapter VMnet8?
16:30:46KyomiI think my battery might be dying
16:30:59XavierGr(maybe it will be a little different than that)
16:31:43preglowKyomi: my deepest sympathies
16:32:23 Join akaidiot [0] (
16:33:14peturXavierGr: both present and enabled
16:33:33lostlogicgoffa: had you tried to seek within the mpc that it broke at the end of?
16:33:34XavierGrstrange :(
16:34:01peturno worries, I'll reboot the box tonight
16:34:08XavierGrpetur maybe it needs special configuring for your case. A reboot should make it clear. Feel free to ask me next time.
16:34:13Kyomipreglow: This isn't normal is it?
16:34:30KyomiOver the scale of about 2.5 hours
16:34:37preglowwhat unit?
16:34:51preglowmight be
16:34:55preglowbut 3.59 is a wee bit low
16:35:17KyomiI had it fully charged using the original f/w
16:35:45KyomiKnow any place that I could get a new battery? Maybe a longer lasting one?
16:35:54preglowhmm, not anymore, no
16:35:55preglowprobably ebay
16:36:05lostlogicKyomi: I already told you, ebay, iopd 1/2 gen, switch the leads.
16:36:16KyomiI didn't see that
16:36:24KyomiThere isn't a place that sells them?
16:36:33preglowapparently not too many anymore
16:36:38lostlogicalternately, you could buy one of my 3 for a reasonable donation to rockbox.
16:36:46preglowebays probably cheaper anyway
16:36:56Kyomilostlogic: How do you mean?
16:37:00Kyomilostlogic: And what are they?
16:37:01amiconnKyomi: There's still an issue with rockbox on H300 drawing more power than it should
16:37:20lostlogicKyomi: I have 2 original H300 batteries and one replacement that I got off of ebay but I broke both of my H300s, so ...
16:37:21Kyomiamiconn: So it may not be dying then? Just running out fast?
16:37:28amiconnYes I think so
16:37:39lostlogicKyomi: if it's 2.5 hours @ 3.59, your battery is definitely on its later legs.
16:37:42KyomiDoes using a 24-bit wps matter?
16:37:43lostlogicat least IMO
16:37:44 Join Paul_The_Nerd [0] (
16:38:00KyomiI'm using the Trinity one... I dont think that would do much
16:38:13Mikachu you could check how long it lasts on the stock firmware
16:38:17lostlogicPaul_The_Nerd: budy, pal, why the heck is coldfire and not arm broken for crossfade? :)
16:38:18Kyomiamiconn: So... I should fully charge it and use the original f/w for a day?
16:38:48amiconnlostlogic: Several people reported the battery running empty on H300 much faster than in retailos, sometimes only 4 hours
16:38:49preglowwell, that would show us if it's rockbox or the battery that sucks rod
16:39:19lostlogicamiconn: ah, didn't realize it was that extreme in some cases.
16:39:32preglowamiconn: damn, that's nasty
16:39:49preglowwe've got pretty much the same situation on hd ipods
16:40:08KyomiI usually just use the usb charging function till I get hoome
16:40:11amiconnI didn't measure myself, but H300 rockbox battery runtime definitely shorter than mini2g rockbox battery runtime
16:40:29amiconnThat's in spite of iriver promising 16 hours with the stock battery
16:40:51Kyomiamiconn: well I think I did have that with stock firmware
16:41:00KyomiI could go like 2 days without it needing to be recharged
16:41:32 Join bluebrother [0] (
16:41:46 Quit darkless (Client Quit)
16:41:51 Quit skwad ("Parti")
16:42:27amiconnBagder? t0mas?
16:42:33Paul_The_Nerdamiconn: Several of those people though turned out to have been playing doom, and stuff like that.
16:42:59Paul_The_Nerdlostlogic: Because "Coldfire" is a cool enough name that *something* had to go horribly wrong elsewhere to compensate.
16:43:00 Quit akaidiota (Read error: 110 (Connection timed out))
16:44:23preglowlets hope apple chooses something spectacularily uncool for their next ipod, then
16:44:29preglowsomething so uncool it's got docs, even
16:44:50lostlogicI bet PP doesn't release docs so they don't get laughed at.
16:44:59crashdheh, maybe they'll build it on an array of hideously overclocked zx80's preglow
16:45:03preglowwouldn't surprise me
16:45:15preglowcrashd: that would be the logical choice, yes
16:45:20crashdimho, anyway
16:45:26preglowmodulating the output of a sid chip for pcm playback
16:45:27crashd:> make emulating the gb on rockbox simple as hell tho
16:47:37preglowwell, ok, i guess i'll check out tonis libmad improvements on arm now, then
16:50:08 Quit romanos_melodos ()
16:51:33 Join _Lucretia [0] (
16:53:29Paul_The_Nerdlostlogic: Regarding Crossfade, I'm not entirely sure it's *crossfade* that's actually broken. With that same build, sometimes just trying to start music resulted in the same sort of freeze (with crossfade set to OFF) as you'd get when crossfade caused it to freeze.
16:54:05Paul_The_Nerdlostlogic: There were also some crazy hangs on startup that I finally resolved by getting in and disabling dircache. I think there's something else entirely wrong on Coldfire maybe.
16:54:59 Quit webguest78 ("CGI:IRC (EOF)")
16:55:17Paul_The_NerdFunny thing is, when it said "Settings Cleared" it'd still hang, but the *next* boot would work
16:55:40preglowlostlogic: seems playback with voice file works again now, gr8
16:56:17Slasherihmm, now host system is configured at 96 kHz but can't directly playback anything from computer.. so i will have to play something and try to record :) 192k seems to be available also, but i think that's for adat
16:56:43preglowSlasheri: spdif != adat
16:56:51Slasheriyes, i know
16:56:52preglowso if you've got adat, you need to set it to spdif mode first
16:56:54preglowif you can
16:57:05Slasherithe outputs can be configured as adat (default) or s/pdif
16:57:10lostlogicpreglow: cool, love when solving 1 race solves multiple problems.
16:57:20preglowhole SHIFT
16:57:28preglowi just broke my bloody ears
16:57:47 Quit _Lucretia_ (Connection timed out)
16:57:47lostlogicPaul_The_Nerd: let me know if you have more non-crossfade freezes of that same sort, 't would be interesting.
16:57:48 Join webguest44 [0] (
16:58:05preglowSlasheri: i've never heard about 192khz adat
16:58:22preglowSlasheri: what sound card is this, btw?
16:58:31Slasheripreglow: i think that is multiple adat channels combined (the whole bandwidth)
16:58:37amiconnlostlogic: Dircache is another hint towards the race. I always have dircache disabled
16:58:41SlasheriE-MU 1820m
16:58:52Paul_The_Nerdlostlogic: Yeah, these were all using the build I'd been trying with last night so it doesn't have that pcmbuf.c change from today
16:58:54preglowSlasheri: if you combined all eight channels you could possibly have 192khz stereo, yes
16:58:54 Quit webguest44 (Client Quit)
16:59:32preglowtoni's libmad patches don't work well on ipod
16:59:44preglownote to self: lower volume when trying out patches
16:59:55crashdpreglow: is this a performance patch on arm?
17:00:01preglowwell, it should work for arm too
17:01:31preglowyeah, the huffdecode breaks on arm
17:01:46preglowhopefully an endian issue
17:01:49Slasheripreglow: hmm, should the peakmeters work while recording?
17:02:07Slasherior should i do something to set the 92k sample rate for iriver?
17:02:20preglowSlasheri: they should
17:02:28Slasherithey do nothing :/
17:02:28preglowSlasheri: for optical input, you should do nothing
17:02:35preglowSlasheri: try a lower sample rate, then
17:02:42Slasherihmm.. i will try with other sample rates also
17:02:53preglowit actually seems aac doesn't work anymore now
17:02:55preglowwhat triggered that?
17:04:34 Join webguest78 [0] (
17:06:27 Quit webguest78 (Client Quit)
17:06:38 Join RedBreva [0] (
17:06:43 Join hardeep [0] (i=hardeeps@SDF.LONESTAR.ORG)
17:07:04preglow#define swap(x) (x) ........
17:07:07*preglow slaps toni
17:07:21Paul_The_NerdThat's an actual line?
17:08:01Slasheripreglow: hmm, should i have s/pdif format as s/pdif or aes/ebu?
17:08:28preglowaes/ebu isn't strictly compatible
17:08:29RedBrevaIs the Nano playing audio reliably ATM? I am getting one latter this week , and would be very pleased if I *NEVER* need to run iTunes...
17:08:58linuxstbRedBreva: Lots of people in this channel never use itunes with their ipods...
17:09:11Paul_The_NerdRedBreva: I have a nano, and never use Itunes with it
17:10:11RedBrevaThat's great....I just havn't been paying too much attention to the current state of iPod's, but having been promised I can have one to use at the gym I thought I better check ;-)
17:11:20preglowfile reading sure is slow on nano
17:11:22preglowfor being flash based
17:11:49Paul_The_NerdIs that a driver issue, or just an "What the hell did Apple do to make it work like this" issue?
17:12:07 Quit JdGordon (Read error: 104 (Connection reset by peer))
17:12:20preglowamiconn: around 5 percent shaved off with antonius' huffdecode patch (libmad)
17:12:36Slasheripreglow: hmm, did you test the patch with any sample rates yet?
17:13:00Slasherieither i haven't configured the output properly yet (very possible), or it just don't work at any sample rate now
17:13:08 Quit Kyomi (Read error: 104 (Connection reset by peer))
17:13:27preglowSlasheri: i've tested it with 44.1 and 48, which are the only ones i have access to
17:13:38preglowSlasheri: try it in the iriver firmware
17:13:44preglowSlasheri: or no... you haven't got that, have you
17:13:59Slasheriinteresting, then it might be a problem with my setup.. just a moment, i will try more :)
17:14:12Slasherihehe, yeah. that would be little difficult ;)
17:15:30amiconnpreglow: where did you find that swap(x) define?
17:16:06preglowamiconn: in bit.c
17:16:09whatboutbobpreglow: sorry, i can't figure out how to replicate that hang at 8kb. will keep trying.
17:16:18preglowamiconn: in antonius' libmad patch
17:16:49preglowwhatboutbob: at least i can't make it happen anymore
17:17:18amiconnmp3 is big endian?
17:17:25whatboutbobpreglow: it happened 4 times today for me :-(
17:18:13preglowamiconn: it's pretty much endian neutral, everything is described as bit fields
17:18:14linuxstbamiconn: I thought mpeg audio was just a stream of bits, so I suppose that counts as big-endian.
17:18:22preglowyeah, probably
17:18:44preglowyeah, that would be big endian, wouldn't it
17:18:52linuxstbSo s/swap/betoh32/ ?
17:19:00preglowi just did swap32
17:19:09preglowbut yeah
17:19:12preglowthat should probably do
17:19:15whatboutbobamiconn: by god you were right! rockbox optical recording is *exactly* the same as rockbox optical recording. nice call. :-)
17:19:31linuxstbpreglow: So that was the only libmad problem on ARM?
17:19:36preglowlinuxstb: yes
17:19:43amiconnpreglow: Yes I think it should be betoh32() (or betoh16() maybe)
17:20:08amiconnwhatboutbob: I only replied to your msg..
17:20:12preglowamiconn: btw, boost on iriver.mp3 went from 72ish% to 68ish% with the huffman decode patch
17:20:37linuxstbWhich target?
17:20:40amiconnHmm, did you try the other parts too?
17:20:49preglowthe gain is much, much larger on coldfire
17:20:51whatboutbobamiconn: yeah i know...i should just stop typing.
17:20:53amiconn(of course the imdct part is coldfire only)+
17:20:54preglow320kbps mp3 is unboosted there now
17:21:04preglowamiconn: i don't think the other parts are changed for arm
17:21:22amiconnBagder! t0mas!!
17:22:35amiconnpreglow: Ah, yes. Then _someone_ need to do something similar for arm ;)
17:22:46preglowamiconn: if i only had a clear idea of how...
17:23:49 Quit Zagor ("Client exiting")
17:24:01 Quit lostnihilist ("Leaving")
17:25:49preglowwould be neat if some other people could test his patch too
17:25:54preglowi haven't got the biggest mp3 collection ever
17:25:57***Saving seen data "./dancer.seen"
17:26:10crashdpreglow: what you need testing?
17:27:09crashdah, you wan it testing on iriver targets?
17:27:22preglowall targets
17:27:42crashdahright, ill make a build once par is finished
17:27:54Paul_The_NerdDon't forget the small swap(x) fix for ipods, right?
17:28:07Slasheripreglow: hmm, i think iriver might not work as it should.. for example i can't stop the recording at all
17:28:15Slasheriand it seems to to write anything to disk
17:28:28Slasherineed to reset to shutdown the unit..
17:29:00preglowSlasheri: did it stick at 8kb written?
17:29:44 Quit carini (Remote closed the connection)
17:29:57Slasheriit doesn't work with any sample rate :/
17:30:13preglowif i have spdif source, i seem to hang
17:30:19preglowhaven't tried too many times
17:30:33preglowalso, if an spdif source is present, you should hear it and see it on the peak meter
17:30:41preglowso you might have misconfigured spdif or something
17:30:45preglowat least other people could record
17:30:58Slasheriwith other sources it seems to work normally
17:30:58lostlogicbut seriously, where are bagder and t0mas when you need them?
17:31:13Slasheriah, i will try if i could hear it
17:31:28 Join solexx [0] (
17:31:38Slasheripreglow: might there be issues with the consumer/professional mode?
17:31:48preglowSlasheri: i believe it takes both
17:31:49Slasherii think my sound card puts out the professional format
17:31:56preglowSlasheri: at least it can send both
17:32:03Slasheriah, hmm
17:32:25preglowthe only difference between professionsl and consumer, afaik, is a bit and sturdier clock accuracy margins
17:34:02amiconnpreglow: Could you add your fix for arm to the patch?
17:35:57goffaanyone know how x5 60gb compares to ipod 60gb in terms of battery life?
17:36:10goffaand how does that compare to h340
17:36:45Slasheripreglow: hmm, do you have a direct url for the patch so i could check again it's the correct version?
17:37:15 Join webguest90 [0] (
17:37:23preglowamiconn: need to try it first
17:37:41Slasheriseems to be correct..
17:38:06preglowSlasheri: have you checked with the spdif debug screen to see if it actually picks up a signal?
17:38:18Slasherinow i got the peakmeters work
17:38:26Slasheriat least 48 kHz works ok
17:38:48whatboutbobslasheri: you may need to reset settings to get digital recording working...and make sure optical out is on (not sure if preglow's mentioned that yet)
17:38:51preglowSlasheri: how?
17:38:58Slasherii think my cable might be bad.. it has low signal level, or iriver is not too sensitive..
17:39:03preglowwhatboutbob: you know, i might not have mentioned that...
17:39:24Slasherihmm, now the signal lost again
17:39:50Slasherii will try with another cable
17:40:19preglowbut yeah, like whatboutbob says, optical out needs to be enabled, apparently
17:40:39preglowamiconn: also, his patch breaks a couple of sine sample wavs i have
17:40:50preglowamiconn: which might be an indication he's doing something subtly wrong in imdct
17:41:15whatboutbobpreglow: ^"apparently"? I haven't actually tested that...just took your word for it. :)
17:41:36preglowwhatboutbob: oh, i have
17:41:49preglowa couple of times
17:41:59preglowi wasted an hour wondering why nothing i did worked
17:43:03whatboutbobhehe...ok...i just confirmed it anyways.
17:44:07preglowSlasheri: at least all the three top fields on the screen should say OK when it receives a valid signal, afak
17:44:10whatboutbobi notice the 'source' doesn't stick when restarting the iriver. defaults to mic.
17:44:11preglowafaik, yes
17:44:23preglowwhatboutbob: oh? mine seems to remember it sometimes
17:44:47Slasheripreglow: ah, if iriver loses the signal, it wont resync until reboot
17:44:49 Join ashridah [0] (
17:45:05preglowSlasheri: oh???
17:45:27Slasheripreglow: yep.. when i removed the cable and plugged back in, peak meters did no longer work
17:45:32Slasheribut reboot fixes that
17:45:42preglowSlasheri: works just fine and dandy here
17:45:45preglowSlasheri: over and over again
17:45:59Slasherihmm, interesting
17:46:13lostlogichardeep: any interest in looking at (or helping me look at) why a Manual skip that should change directories doesn't, but automatic skip that should change directories does?
17:46:21Slasheriit is enough just to re-enter the recording screen
17:46:33preglowSlasheri: now i just killed my peak meters too
17:46:39preglowSlasheri: i love how this isn't consistent
17:46:46Slasheripreglow: oh, in fact only peak meters stop working
17:46:51Slasherimusic still plays correctly
17:46:55goffamy locking seems to be limited to mpc... ogg didn't do it today after i updated to latest cvs
17:46:56preglowSlasheri: if peak meters work, then recording will work too
17:47:00preglowSlasheri: if not, then recording has broken
17:47:11preglowSlasheri: the monitoring is just an echo function the coldfire chip has
17:47:27Slasherii will try now other sample rates..
17:47:35preglowi see that reentering the recording screen fixes the problem now
17:47:47preglowso i need to do some more initing when i change source, apparently...
17:47:56lostlogicgoffa: odd. wiki it if you get a chance... or bother preglow :)
17:48:21preglowand when resyncing to spdif.....
17:48:48Slasheriit also seems that iriver is not at all too sensitive.. with the other cable, it wont work at all
17:49:01Slasherithere is just some periodical noise coming out
17:49:06whatboutbobpreglow: come to think of it...i'm sure i've seen the source stick to digital before too...but I can't get it to work now...
17:49:14 Quit webguest90 ("CGI:IRC (EOF)")
17:49:27 Quit RedBreva ("CGI:IRC")
17:49:54 Quit solexx_ (Read error: 110 (Connection timed out))
17:50:04preglowSlasheri: would be pretty interesting to know if this also happened with the iriver firmware, heh, but i guess you don't have flash routines handy? ;)
17:50:32Slasheripreglow: well, i would need the full flash dump from a unit with iriver firmware..
17:50:40Slasherithen i could flash that image over to it :)
17:50:53 Quit ashridah ("Leaving")
17:51:05Slasheriof course i no longer have the original rom dump.. :D
17:51:44whatboutbobpreglow: did you see mborus' comments on the forum? Seems a couple of his recordings failed after ~14minutes.
17:51:44preglowi think a flash dump shouldn't be too hard to get
17:51:46Slasherii accidentaly overwrote it a long time ago
17:51:58Slasheripreglow: it should be available directly from the debug menu
17:52:05preglowi could make one
17:52:11Slasherihehe, nice :)
17:52:19preglowh120, thought, but i guess it doesn't matter
17:52:26Slasheriyeah, it should work
17:53:44amiconnSlasheri: I could provide a dump of my H140 later
17:53:55amiconn(but I guess it really doesn't matter)
17:54:11Slasherih120 dump should be just fine
17:54:19 Join ashridah [0] (
17:54:29Slasheripreglow: thanks :) i will try to reflash soon
17:54:44preglowmd5sum: 3bff190db5f1eea25c24175cab11aa5a /media/usbdisk/internal_rom_000000-1FFFFF.bin
17:54:51Slasheripreglow: 403, forbidden
17:55:02preglowgherhg, gimme a sec
17:55:43Slasheriyau, now got it
17:56:20XavierGrSlasheri: Please inform us if you manage to reflash ok. This sounds very thrilling! (and dangerous)
17:56:43SlasheriXavierGr: hehe, i will :) trying very soon
17:56:43whatboutbobpreglow: if there's no signal the recording freezes @ 8kb.
17:56:50Slasherijust checking the flash routines first
17:57:18preglowwhatboutbob: i know
17:57:33preglowwhatboutbob: at least i know what the problem is now
17:57:48preglowwhatboutbob: what worries me is that it might also happen during ordinary recording if the signal is bad
17:58:18whatboutbobyeah...i'm wondering if that's what's happening with the 'freezing' recordings...
17:58:28preglowme too
17:58:31Slasheripreglow: both s/pdif and aes/ebu output formats seems to work fine
17:58:43preglowSlasheri: yeah, i found out they're compatible after all
17:58:57preglowSlasheri: but what about 96khz?
17:59:04Slasheriwell, trying it now :)
18:00:47 Quit _Lucretia (Read error: 110 (Connection timed out))
18:00:53Slasheripreglow: works fine, playing guitar @ 96 kHz and audio sounding correct and peak meters work
18:00:56whatboutbobpreglow: perhaps is there some way to put signal detection in so that if it does drop-out it pauses the recording or we can at least know for sure what's going on.
18:01:34preglowSlasheri: how about recording?
18:01:53 Join swampcow [0] (
18:01:54Slasheripreglow: it worked too, i will try to listen to the results now
18:02:10Slasherihowever, now even the serious cable had serious troubles with that high frequency
18:02:13preglowwhatboutbob: well, i need to find out how
18:02:26 Join _Lucretia [0] (
18:02:31preglowSlasheri: sounds like some dodgy cables you have there
18:02:33Slasherii have to hold it at excact position not to lose the audio
18:02:38Slasheriyep :/
18:02:46preglowboth of mine works great
18:02:46XavierGrSlasheri: Where on Rockbox?
18:02:52preglowthough i know optical cables can be dodgy
18:03:01XavierGr^(insert comma)
18:03:11SlasheriXavierGr: hmm, what did you mean? :)
18:03:24 Part vmx_ ("Have fun!")
18:03:34whatboutbobslasheri: the iriver optical input can be a bit finicky about certain cables. I had all sorts of issues with one that worked perfectly on other devices but just didn't like the iriver.
18:03:36DBUGEnqueued KICK XavierGr
18:03:36XavierGr[19:00] ,04Slasheri,99: ,05preglow,99: works fine, playing guitar @ 96 kHz and audio sounding correct and peak meters work,99 <- Rockbox or iRiver firmware?
18:03:52SlasheriXavierGr: ah, rockbox :)
18:04:06XavierGrAh So you didn't reflash. Nice
18:04:08Slasherii will try with iriver firmware soon as i get it installed first
18:04:14Slasherii will now
18:05:03whatboutbobAnyways, i've got to get up in 4 hours so I'm off to bed. G'night all.
18:05:03CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
18:05:03*XavierGr crosses fingers!
18:05:10preglowthe coldfire manual seems to suggest reclocking the cpu as a solution to adapting to spdif clock rates
18:05:17preglowwhatboutbob: night
18:06:08Slasherieh.. now i got only white noise when i try to playback any file :D
18:06:14Slasherieven rebooted the unit
18:06:23Slasheriah, my mistake..
18:06:28Slasheriforgot to update the codecs
18:06:55Slasherihard to developt with a windows machine as my main workstation is still broken :/
18:07:10 Join dpassen2 [0] (
18:09:25amiconnSlasheri: Install the vmware player and use the dev vm image...
18:09:33Slasheripreglow: yeah, playback works fine, only with wrong sample rate
18:09:35 Quit petur ("reboot for vmware test")
18:10:25Slasheriamiconn: hmm, that might be interesting to try
18:10:53preglowSlasheri: cool
18:11:03amiconnBagder!! t0mas!!!
18:14:00preglowis the h1x0 coldfire a 160bga package?
18:15:26Slasherilostlogic: hmm, the recent cvs seems to crash when crossfade should kick in
18:15:31 Join KlrSpz [0] (n=klrSpz@
18:16:08preglowSlasheri: i think he knows about it, heh
18:16:14Slasherihehe, ok
18:16:50KlrSpzlostlogic, which ver of gcc/glibc did you say to use for ipod(arm) ?
18:17:16amiconnlostlogic: What's the problem with detecting end-of-playlist?
18:17:32Slasherinow i have patched the firmware plugin.. maybe i should just try now
18:23:32Slasheriok, trying now.. hopefully it doesn't crash :)
18:24:28*XavierGr is spooked!
18:27:10Slasherigrrr! entered firmware flash plugin and tagcache started building on background or something like that.. huh, it didn't crash
18:27:18Slasherichecking the rom dump
18:28:08Slasheriwell, "checking".. i forgot usb mode doesn't work rockbox at the moment :/ so just rebooting then
18:28:37Slasheriiriver 1.65E starting :D
18:28:40XavierGrOk "Flasheri" is back on the market!!!
18:28:43lostlogicamiconn: how would it properly be detected? What condition indicates that both the current track is the last track in the playlist, and the stop being processed is automatic not manual.
18:28:49Slasheriuh, it's so ugly
18:28:53Slasheribut seems to work
18:28:56preglowyes, yes it is
18:28:59lostlogicSlasheri: yeah, the crossfade thing is very odd, because it breaks on coldfire, but not arm :(
18:29:20lostlogicKlrSpz: there is no glibc for the target, just a C compiler is needed, and binutils, 4.0.3
18:30:26KlrSpzu mean gcc 4.0.3?
18:30:34KlrSpzshould i just grab the core or the whole deal?
18:30:40lostlogicjust core, you only need a C compiler
18:30:41Slasheripreglow: anyway, thanks for the rom image. now i can start changing the eeprom content and figure out if there are any safe bits to change without affecting original fw
18:31:22amiconnlostlogic: If you're trying to advance the playlist to the next track, the playlist code should tell you that the playlist ended.
18:31:30KlrSpzthanks, sorry if i'm bothering ya, never set up a cross compiler before and i'm going between the tut on the wiki and a gentoo page, but the gentoo one is proving to be a pos
18:31:35preglowi wonder why recording _hangs_ when recording hasn't been inited correctly, somehow
18:31:38amiconn(during playback, not during buffering)
18:32:00amiconnIt worked with the old playback code, and it works on archos, so there must be a way...
18:32:01Slasheriweird.. it seems iriver fw still remembered the last file i had played months ago with the original fw
18:32:31 Quit SereRokR ("XChat Aqua")
18:32:39lostlogicamiconn: hmm... maybe the end of playlist var should be set by the advances, not by the buffering peeks, indeed... still hafta program something so that the playback code knows the difference between an automatic stop and a manual stop.
18:33:10preglowamiconn: stray unhandled interrupts on coldfire shouldn't be a problem, no?
18:33:12lostlogicamiconn: oh sure, it's possible, jsut doesn't work now ;)
18:33:33amiconnpreglow: Stray interrupts should call UIE()
18:33:41preglowamiconn: yes, exactly
18:34:09preglowperhaps i should try enabling some interrupts and see if some interesting ones get called when recording hangs
18:37:27Slasheripreglow: eh.. how in earth can i change the recording source to optical in iriver fw? it always reverts back to the int mic..
18:37:46 Join speacial_ed [0] (
18:38:11 Join matsl [0] (
18:38:14Slasheriah, now found it
18:38:49preglowyeah, it's tricky
18:39:22Slasheripreglow: it's also as sensitive
18:40:00preglowcan you hang it, though?
18:40:24Slasheriwell, it doesn't have a peak meters so no.. :)
18:40:44 Join darkless [0] (n=darkless@
18:41:06 Quit muesli__ (Read error: 110 (Connection timed out))
18:43:19 Join muesli__ [0] (n=muesli_t@
18:48:16 Quit speacial_ed ("Bye")
18:48:18preglowtime for food
18:50:13goffawonder if lostlogic has any mpcs
18:54:14 Join new5guser [0] (
18:55:37 Join Hansmaulwurf [0] (
18:56:12 Quit hardeep ("Read error: 3.1415926535 (Excessive pi)")
18:57:50lostlogicgoffa: I don't.
18:58:07lostlogicgoffa: nor do I want any until I figure out crossfades major malfunction on coldfire.
19:03:07 Join obo [0] (
19:04:05 Quit muesli__ (Read error: 110 (Connection timed out))
19:06:48 Join Lear [0] (
19:06:55amiconnBagder!!! t0mas!!!!
19:07:00KlrSpzlostlogic, do i need to mess with the newlib stuff still?
19:07:21Paul_The_NerdThe newlib stuff is for the SH1 compile only I believe
19:07:59lostlogicKlrSpz: That's only for gcc 3.3
19:08:04lostlogicyeah, that too
19:08:05amiconnEven only for sh-elf-gcc 3.3.x
19:08:05KlrSpzk, cool
19:08:12KlrSpzwhat's SH1?
19:08:19lostlogicarchos processor
19:08:28KlrSpzah ok
19:08:28 Join Lucretia_ [0] (
19:09:09 Join muesli__ [0] (n=muesli_t@
19:09:43lostlogicanyone else think it's a sad state of gcc that we have 3 architectures, and the best gcc series is different for each of the three? not only taht, but the current release series doesn't work for any of them!?
19:10:31Paul_The_NerdActually didn't a recent test show that SH1 could have a smaller binary with one of the 4.0 or 4.1 series?
19:11:05 Quit KlrSpz ()
19:11:07lostlogicI think that was a -O to -Os comparison
19:11:20lostlogicbuy -Os comes closer to working on the newer series
19:11:51lostlogicPaul_The_Nerd: you don't happen to have an LCD remote, do you?
19:12:00Paul_The_NerdFor H120 I do.
19:12:06Paul_The_NerdIt's in a sorry condition, but I *think* it works still
19:12:26 Quit markun ("leaving")
19:12:29lostlogicPaul_The_Nerd: do you think you could make a logf build and cause that crossfade crash on your 120 and see what the last couple lines of logf are?
19:13:40 Quit _Lucretia (Read error: 113 (No route to host))
19:13:54Learlostlogic: Had a playback hang earlier today. Pressed play to resume, quickly realized it was for the wrong playlist and pressed stop. A bookmark was created (or so the screen said) but it never exited the WPS. Pressing stop for a couple of seconds had no effect.
19:14:45lostlogicLear: yeah, I think that's on the wiki... something to do with a stop event before the codec initializes, I think.
19:14:54Paul_The_Nerdlostlogic: Sure, gimme a minute to dig up the remote. I was in the process of trying to open it (and failing miserably) so it's somewhere in my "workspace"
19:15:05amiconnPaul_The_Nerd: With -Os, gcc 3.4.x creates smaller binaries than 3.3.x (but they don't run yet)
19:15:13amiconnWe cannot use any 4.x gcc for SH1
19:15:16lostlogicfood, back soon.
19:15:20Learlostlogic: well, I did hear a bit of sound, I think...
19:15:29amiconn(due to a silly design decision of the gcc team)
19:15:52amiconnI've opened a bug report for that, but they prefer to ignore it, as usual...
19:16:00lostlogicLear: hmm, interesting. I definitely need to look into it, because that's one of the bugs that actually effects me... stupid coldfire crossfade first though ;)
19:16:34Paul_The_NerdI just need "(L)ogf" and no other developer options, right?
19:18:22 Join Rondom [0] (n=Rondom@
19:19:01Learlostlogic: Or maybe it hadn't started playing music yet; it was a vorbis track, so it was likely still seeking. Can't easily reproduce it at least.
19:19:04 Join mirak [0] (
19:19:16 Join _gmrD00d [0] (n=klrSpz@
19:20:04 Quit new5guser ("CGI:IRC")
19:20:07 Join new5guser [0] (
19:21:34Paul_The_NerdApparently a logf build inflates the size of Rockboy so that it doesn't fit in the plugin memory any more?
19:23:13linuxstbOn which target?
19:23:22Paul_The_Nerdlinuxstb: h120
19:23:47Paul_The_Nerd"Region PLUGIN_IRAM is full"
19:23:50 Join markun [0] (
19:24:43Paul_The_NerdWhich isn't the same thing.
19:25:14HansmaulwurfPaul_The_Nerd: does this mean rockboy doesnt work?
19:25:17goffalostlogic: was just asking, because that's where the freezes happen on playback in latest cvs
19:25:34Hansmaulwurfcoz with the latest bleeding edge it works on my H120
19:25:41preglowlostlogic: i'm using gcc 4.1 for arm as we speak
19:25:43preglowand have been for a long time
19:26:00***Saving seen data "./dancer.seen"
19:26:24Paul_The_NerdHansmaulwurf: Naah, this is a completely different situation.
19:27:42Paul_The_Nerdlinuxstb: I assume the easiest "quick fix" for me is to just disable building Rockboy while I run this test?
19:27:42 Join JBGood [0] (i=Johnq@
19:27:45 Quit muesli__ (Read error: 110 (Connection timed out))
19:31:17_gmrD00dhow are ya'll testing your binaries? just copying back to the device?
19:31:45 Join muesli__ [0] (n=muesli_t@
19:32:07 Join Ribs [0] (
19:32:36 Quit JBGood25 (Read error: 110 (Connection timed out))
19:32:46 Join ep0ch [0] (n=ep0ch@
19:33:04Paul_The_Nerd_gmrD00d: Depends on what's being tested, but yeah quite often.
19:35:02 Nick _gmrD00d is now known as KlrSpz (n=klrSpz@
19:35:32ep0chanything urgently need testing? Or else i'll do some boost ratio testing for the libmad patch on ihp and 5g and also test the iram patch from linuxtsb
19:35:45Paul_The_NerdAnyone know where it tells Rockbox to build rockboy?
19:37:47Learplugins/SOURCES or something?
19:38:24Paul_The_NerdIn plugins/SOURCES it's just the overlay for archos
19:38:33Paul_The_NerdIt *used* to be in plugins/MAKEFILE
19:38:40Paul_The_NerdNotice the past tense
19:39:02LearThough for rockboy it should be plugins/SUBDIRS..
19:39:13Paul_The_NerdOh, there's a SUBDIRS file now?
19:39:28Paul_The_NerdI didn't see that
19:39:54 Join JBGood25 [0] (i=Johnq@
19:42:19 Join _Lucretia [0] (n=munkee@
19:42:19 Quit dpassen2 ("CGI:IRC")
19:43:34preglowep0ch: you don't really need to test the libmad patch
19:43:37preglowep0ch: it's already done
19:43:53 Quit Lucretia_ (Read error: 110 (Connection timed out))
19:43:54preglowep0ch: what you can do is just try playing around with it. it doesn't seem to work 100% correctly here
19:43:57preglowespecially when changing files
19:44:48ep0chany idea if it had an effect of ipod performance?
19:44:51preglowit did
19:44:57preglowaround 5% less boost at 320kbps
19:45:15ep0chthats promising
19:45:16preglowthat's with only the huffdecode part, i don't think the others will help at all
19:45:43ep0chyour optical recording stuff been tested?
19:47:56preglowwith varying results, as usuakl
19:48:31 Join xmixahlx [0] (n=xmixahlx@
19:48:49 Join muesliiiii [0] (n=muesli_t@
19:50:36 Quit muesli__ (Read error: 110 (Connection timed out))
19:52:31ep0chpreglow: what did you mean by the libmad patch not working 100%?
19:56:35preglowep0ch: like i said, it just hangs when i switch tracks some times
19:59:05obodoes ata_disk_is_active work okay on nanos? I can't remember if thats the problem the battery bench had...
20:07:09*Paul_The_Nerd sighs.
20:07:12 Quit ashridah ("Leaving")
20:07:37Paul_The_NerdI don't understand the "can we trick it into" mentality. "Can Rockbox be tricked into playing AAC+ files?" "Can we trick it into playing a game with a video in it?"
20:08:32 Quit JBGood (Connection timed out)
20:08:36 Join linuxstb_ [0] (
20:10:50 Join SereR0KR [0] (
20:14:29Hansmaulwurflol Paul_The_Nerd, just discovered the thread and though the same
20:15:10 Join JonSenior [0] (
20:17:14JonSeniorLinuxstb: If you're around... further to my battery queries the other day, I've just fitted a 2200mAh battery to my H300. Seems to be working so far (Still charging, but it hasn't gone bang!). Do I need to do anything to cause Rockbox to recognise this, or will the battery voltages be enough for it. I'm guessing so, but I just want to make sure!
20:18:52 Quit akaidiot ("( :: NoNameScript 4.02 :: )")
20:19:31 Join OnkelJonas2 [0] (
20:20:21 Join Steveoo [0] (
20:20:33 Quit JBGood25 (Read error: 110 (Connection timed out))
20:20:42 Join Kohlrabi [0] (
20:20:45Steveooyo guys i need help with my archos
20:20:52Steveoocan anyone help
20:21:16 Join scott666 [0] (
20:21:44Steveoocan the av480 be downgraded?
20:22:19 Quit scott666 (Client Quit)
20:23:33 Part OnkelJonas2
20:23:37 Quit Paul_The_Nerd ("Leaving.")
20:23:58ep0chi've just done a grep for ICODE_ and it looks like libmad doesn't use ICODE, is this right?
20:24:12preglowno, i don't believe it does
20:24:17Steveoocan the av480 be downgraded?
20:24:43ep0chwhich may explain why linuxstb didn't get much of a performance change
20:24:44preglowSteveoo: what makes you think we know?
20:24:56ep0chso i'll do a test on tremor
20:24:57Steveooi was on a website
20:25:03Steveooand it said come here
20:25:18preglowwe're not an archos support channel
20:25:32Steveoook dont get your knickers in a twist i only asked
20:25:50preglowi'm just telling you we're not an archos support channel
20:25:55preglowthat doesn't qualify as knicker-twisint
20:25:57preglowtwisting too
20:26:26Steveooa simple no would of done
20:26:40 Quit _Lucretia (Connection timed out)
20:27:13preglowSteveoo: i don't know if it's yes or no
20:27:22preglowand i tend to not answer when i don't know the answer
20:27:31Steveoofairy muff
20:27:46Steveooso what channel is this
20:27:47linuxstb_ep0ch: That's a good point about icode in libmad. The same applies to vorbis (on ARM, icode is used on Coldfire).
20:28:44Steveooicode in libmad is down to the flux capacitor using the giggle pin on the over ride switch
20:28:52ep0chany codec or dsp i can use to see the effect it may have?
20:29:25amiconnBagder!!!! t0mas!!!!!
20:29:46t0masbuild hang
20:29:48linuxstb_FLAC uses icode, but that's probably 0% boost already. Maybe Wavpack or ALAC...
20:29:53preglowlinuxstb_: then why'd you get a percent difference in performance?
20:29:53amiconnt0mas: Builds hanging since four hours ago...
20:30:04*t0mas takes his big process-stungun
20:30:05amiconnActually 5 hours
20:30:20LearMultiple explamation marks, a sure sign of a desperate mind. :)
20:30:38ep0chpreglow: 1% could be down to testing error
20:30:52amiconnLear: One more for each call...
20:31:10ep0chi think SID has icode... so maybe i'll use that
20:31:12preglowif you call load enough on freenode, the recipients get a telepathic message
20:31:48markunSteveoo: maybe try the #archopen channel
20:32:26amiconnpreglow: Well, even if the codec in question doesn't use icode, iirc linuxstb disabled icode globally, i.e. also in the core
20:32:47preglowamiconn: yeah, but i don't think we use it too many places
20:32:50amiconnSo I'd still expect a performance difference
20:32:56preglowfor ipod, that is
20:33:40amiconnWell, some low level functions are in iram on both sh1 and coldfire, and so I'd expect them to be in iram on arm as well
20:33:47t0maskilled and restarted
20:34:28amiconnt0mas:Thanks. It's till annoying that only very few people can use the force and are able to do that
20:34:43t0masyes it is
20:35:04t0masbut I shouldn't be nessesary to use that power at all :)
20:35:11 Join jbauman [0] (i=Johnq@
20:35:25amiconnEither it should get more stable (i.e. auto-recover from a hang, or better not hang in the first place), or we need more people being able to use the force
20:36:00ep0chok i think i'll do a flac test and put eq on so i get some boosting, that way i'll know if icode has an effect of flac or not
20:36:34 Quit Kohlrabi (Read error: 104 (Connection reset by peer))
20:38:39 Join _Lucretia_ [0] (n=munkee@
20:38:41amiconnep0ch: You could also make a special build with a lower default frequency. I think 22MHz will give some boost even with flac
20:39:35 Join Kohlrabi [0] (
20:39:59 Join hotwire [0] (
20:41:56 Quit JonSenior ("CGI:IRC (EOF)")
20:42:12lostlogicpreglow: oh, it's just warnings in 4.1, not problems?
20:42:25ep0chhaving an ATA benchmark plugin and mem speed plugin would be useful
20:43:42ep0chbolox forgot to umount
20:43:44amiconnI do have an ata benchmark (or rather, file benchmark), and a test plugin for memset() and memcpy()
20:44:36amiconnBoth need some polishing to be generally useful (right now I tend to shuffle around some hardcoded values depending on what I want to test)
20:46:14 Join StrathAFK [0] (
20:46:36preglowlostlogic: just warnings
20:47:14preglowamiconn: still better than nothing
20:47:37amiconnep0ch: Guess how I measured the improvements achieved by my asm optimised ata transfer routines?
20:47:50goffaugh... i hope mpc playback gets fixed again soon.. i may have to go back to friday's build
20:47:56 Join mborus [0] (
20:48:39preglowgoffa: what's wrong with it?
20:48:44preglowno one has touched mpc for ages
20:48:49preglowand i just tried it anyway, it worked fine
20:49:31mboruspreglow - I'm back and I checked the satellite signal. It's really 48 kHz and the digital recordings the Iriver made show 44.1 and play too slow.
20:49:45goffawell... since saturday mpcs hang about 2 secs before the end of the track
20:49:53goffaand won't advance until i hit next
20:49:55preglowmborus: try recording with the original firmware
20:50:14preglowgoffa: ooooh, really? i thought that was a bug in the patch i just tried
20:50:33goffai don't think so... i did compile myself
20:50:39amiconnlostlogic: Referring to the mailing list, why/how does low latency mode kill voice?
20:50:42goffaand i switched 2 variables in tree.h
20:50:47mborusOK. BTW, I set the recordings sample rate to 48 kHz just for fun. Didn't make a difference.
20:50:50goffabut i don't think that would affect mpc
20:51:04preglowmborus: setting the sample rate doesn't do shit
20:51:05 Quit Rondom ("I'm leaving on a jetplane, don't know when I'll be back again...")
20:51:13 Join ep0ch| [0] (n=ep0ch@
20:51:15preglowmborus: the sample rate is deduced from the spdif signal
20:51:38preglowgoffa: well hot damn, you're right
20:51:45preglowit's gotta be a playback system thing, then
20:51:47 Nick ep0ch| is now known as ep0ch_ (n=ep0ch@
20:51:53goffayeah.. that's what i'm thinking
20:52:01goffaalthough there was an update to all codecs
20:52:09goffano wait, i updated before then
20:53:24goffaprobably is playback.c
20:53:26Leargoffa: that's probably it, it included what appears to be a bad change in mpc: sometimes the return value might not be set...
20:54:00LearOr not, I spoke too soon. :)
20:54:22goffalooking at changelog
20:55:15mboruspreglow: Original Firmware records in 48 kHz
20:56:09preglowdoes it now
20:56:09LearHm... If mpc tends to end tracks with a decode error, the change will probably cause problems...
20:56:12goffai think i was in here
20:56:13goffa2006 22 Apr 15:13lostlogicapps/codecs/mpa.c 1.44 Comment and make pretty the hack
20:56:17goffa2006 22 Apr 14:53lostlogicapps/playback.c 1.295 Fix warning
20:56:20goffa2006 22 Apr 14:48lostlogicapps/codecs/mpa.c 1.43 Replace one hack with a better hack
20:57:08preglowmborus: and the spdif debug screen says 44.1khz as well for that source?
20:57:14goffaand that was after the reworking of voice to the codecs
20:57:17goffaso i dunno
20:57:40Leargoffa: but mpa.c doesn't decode mpc...
20:58:33preglowLear: you can see no good reason to keep the chunked approach dsp.c uses in processing now, no?
20:58:43 Join JBGood25 [0] (
20:58:45 Quit jbauman (Connection timed out)
20:58:52goffabut... the updates before then
20:58:58goffawould apply to current
20:59:20goffaand the one before that is the big 2006 22 Apr 14:40
20:59:30goffaso it could have been voice codec or playbac.c
20:59:41goffai couldn't correct it if i knew
21:00:24goffajust trying to narrow it down to when to make it easier for whoever fixes to find it.. probably lostlogic or preglow
21:00:40 Quit Strath (Read error: 110 (Connection timed out))
21:00:52mboruspreglow: the spdif debug screen said that when I tested it a few hours ago. I'm running another test recording now, afterwards I'll check again.
21:01:19Learpreglow: if it is okay to modify the buffers passed to dsp.c, I guess I don't see any good reason.
21:01:43preglowi can't see why it shouldn't be ok
21:01:45lostlogicamiconn: not sure, I just assumed he was right about that −− didn't actually look into it.
21:01:54preglowLear: it's not like they're reused in the codecs anyway
21:02:14preglowand i think they're reused in some situations right now anyway
21:03:03 Quit ep0ch (Read error: 110 (Connection timed out))
21:03:11amiconnpreglow: Do you have a solution for upsampling?
21:03:28 Nick ep0ch_ is now known as ep0ch (n=ep0ch@
21:03:34preglowamiconn: no, that'll still have to use a temp buffer
21:03:45 Join speacial_ed [0] (
21:04:34lostlogicI must be missing something −− what's wrong with upsampling? the code asks the DSP how big of a buffer it'll need to do its thign and then gets that much from pcmbuf...
21:05:17preglowlostlogic: you can't do it in-place
21:05:49lostlogicbut you have the input buffer from the codec, and the output buffer from the pcmbuffer... which should be sized correctly...?
21:06:31lostlogicor does resampling happen before trimming to 16bit?
21:06:40lostlogicI guess it would.
21:07:11Learcorrect, resample is done before, so the size is wrong.
21:08:02preglowresampling happens at the very start
21:08:17LearNo, gain is applied first.
21:08:22preglowit's worth considering removing resampling to either last or first spot, depending on whether we down or upsample
21:08:35 Quit _Lucretia_ (Connection timed out)
21:09:09 Join _Lucretia_ [0] (
21:09:25preglowif we downsample, it's better to do it at the start, if we upsample, at the end
21:09:31amiconnpreglow: In case you do (sounds like a good idea) I think you need to adjust the eq coeffs
21:09:38preglowamiconn: yes we do
21:09:45preglowamiconn: and everything else that depends on the sfreq
21:09:52Learbut I guess it would be possible to use pcmbuf as resample buffer, cause it should be possible to request x bytes, but only use y (where y < x)...
21:10:06preglowLear: the resample buffer should be iram
21:10:08preglowas it currently is
21:10:08amiconnpreglow: Ah, yes, like eg. the crossfeed delay...
21:10:19Learah, yes, that too...
21:10:28preglowamiconn: yeap
21:10:53lostlogicLear: yes, that would be possible −− the dsp_output_size function is used to determine how much to get from the pcmbuf, and then the retval from teh dsp_process is how much is 'committed' back to the buffer.
21:11:02preglowamiconn: the crossfeed might vary a little from sample rate to sample rate, though, since the delay time probably won't fit in an integer number of samples for all sample rates
21:11:18preglowlostlogic: btw, there's currently a bug in dsp_input_size and dsp_output_size
21:11:37preglowi notice the sim spits a warning at me all the time when the pitch ratio is above 100%
21:11:43preglowi suspect what is wrong too
21:12:22preglowi don't think they take the initial value of the delta into account
21:12:31preglowehh, position, i mean
21:12:38preglowso it might be off with one sample from time to time
21:12:59lostlogicpreglow: mm, nod, shouldn't cause any major trauma
21:13:02 Quit mirak (Connection timed out)
21:13:24lostlogicpreglow: actually I think I removed the error if the input size happens to grow...
21:13:35preglowmight be, i haven't checked for a while
21:13:45lostlogicpreglow: because it's not really an error condition, just means that we got extra space from the pcmbuffer that we won't end up committing
21:14:22amiconnpreglow: Doing the resampling depending on the resampling direction has another advantage - since the dsp code needs to be adjusted for handling varying frequencies, it will make it easier to use different pcm frequencies on platforms where this is possible (ipod)
21:14:35preglowamiconn: yup, that'd be nice too
21:14:49preglowamiconn: we can also use different frequencies on the iriver
21:15:03amiconnWe'll then need a mechanism in the pcm buffering to mark frequency changes
21:15:12RedBreva_Aghhhh! gcc 3.4.6 is failing to compile
21:15:13amiconnpreglow: Ah, yes, 44.1/22.05/11.025
21:15:34amiconnEven the ipods will differ
21:15:56preglowamiconn: we should probably always use the smallest sample rate we can what sample frequencies they can support
21:16:03preglowunless something says we shouldn't
21:16:26amiconnI would suggest to use the next-higher possible rate for intermediate ones
21:16:43RedBreva_libgcc and stmp-multilib are failing
21:17:10Learredbreva_: target?
21:17:17amiconnIirc the WM8731 (in fact it's a WM8711) can only handle 8/32/44.1/48 (and perhaps higher)
21:17:20lostlogicshouldn't be very difficult to have the pcmbuffer be samplerate aware, and stop / setup / resume playback on a sample rate change, by putting the sample rate of a chunk on that chunk, and comparing it to the previous.
21:18:45amiconnpreglow: Or would it be better to stick to integer upsampling factors whenever possible?
21:19:18Learredbreva_: and you've applied the config file patch?
21:19:19amiconnE.g. on mini2g, upsample 16kHz to 32kHz, but 22.05kHz to 44.1kHz?
21:20:06lostlogichmm, only hard part would be what thread to run and how to trigger the pcm_play_data call on sample rate change.
21:20:27RedBreva_I tried the 3.4.4 but it fails with a file not found... this is after all 3.4.6 ;(
21:21:46Learwell, it is easy to do manually. Open gcc/config/m68k/t-m68kelf file and modify the MULTILIB_OPTIONS line, like the patch shows.
21:21:54amiconnRedBreva_: cygwin?
21:22:02RedBreva_no Suse10
21:22:12amiconnHmm, strange
21:23:20Learhm... any release notes for 3.4.6? Can't find any info on
21:24:26 Quit mborus ("CGI:IRC (Ping timeout)")
21:25:01 Quit speacial_ed ("Bye")
21:25:32lostlogicLear: even the NEWS file in the gcc-03.4.6 package only mentions 3.4.5
21:25:57ep0chlinuxstb_: with libflac + 5eqs and icode enabled boost = 30%, without icode = 31%
21:26:03***Saving seen data "./dancer.seen"
21:26:16 Join Kyomi [0] (
21:26:22LearThe releases page mention 3.4.6, but that's pretty much the only place I've found so far.
21:26:44KyomiAlright.. probably not the place to ask this... but.. how do you view playlists in stock h320 firmware?
21:28:12 Quit Steveoo ()
21:28:20*amiconn didn't use the original firmware long enough to answer that question
21:29:22amiconnThe original firmware is good for flashing the rockbox bootloader enabled version, and for usb host mode (which is extremely cumbersome to use)
21:29:23RedBreva_I edited the file as per the patch, but still the same failure :(
21:30:23amiconnRedBreva_: Did you add the bin/ dir of your binutils installation to the path before building gcc?
21:30:37amiconnOtherwise it will of course fail building the multilibs...
21:30:48 Join Gargamale [0] (
21:31:00RedBreva_yep, i successfully built sh1
21:31:18lostlogicany iriver/iaudio users with an LCD remote and who can build a logf build around?
21:31:20GargamaleI successfully spelled success
21:31:37Learredbreva_: you might need make clean (or even re-run configure) for the change to "bite"...
21:31:39amiconnRedBreva_: You could still have forgotten to add the m68k binutils dir to the path
21:31:47amiconn(unless you
21:31:56amiconn've installed both in the same path)
21:32:00Learlostlogic: yes...
21:32:28lostlogicLear: you had the playback freeze on crossfade bug?
21:32:54Learamiconn: hm.., can different target builds of gcc share the same −−prefix safely?
21:33:00RedBreva_whoops!! Back later.... <sneaks off queitly to add the new binutils dir>
21:33:03Learlostlogic: never use crossfade, so I've no idea. :)
21:33:21lostlogicLear: want to try to recreate it with a couple extra logging lines?
21:33:46preglowLear: not recommended
21:33:56LearI guess I could, yes - as long as it doesn't take too long (like an hour or more... :)
21:33:58preglowLear: that is, it looks like it might work
21:34:03lostlogicLear: 10m
21:34:12Learpreglow: good, then I did it the right way. :)
21:34:24 Join ts-x [0] (
21:34:31lostlogicI use the same prefix for all of my targets, FWIW
21:34:34lostlogic if (fade_out_chunk == NULL)
21:34:34lostlogic logf("NULL fade out chunk");
21:35:07amiconnLear: Hmm, not sure, perhaps not
21:35:16Learlostlogic: so I add that where?
21:35:32lostlogicLear: put those two lines at pcmbuf.c:463 and 491, logf it and see what you get in the logf (if possible, put a 1 and a 0 in the two logfs so we can tell which one goes null)
21:35:48ts-xKyomi: AFAIK you can't view the contents of playlists in the stock firmware
21:35:56linuxstb_ep0ch: Thanks for testing. Which ipod do you have?
21:36:00lostlogicI'm pretty sure that one of the two must be going null in order for this to break.
21:36:34Kyomits-x: According to the manual you can using the A-B button during playback standby mode
21:36:40Learlostlogic: both line numbers before the change? :)
21:36:42KyomiBut it doesn't like me and wont do that :/
21:36:45amiconnlostlogic: Interesting... I checked the sub-dirs, and while most seem not to clash, the share/locale/ dirs do...
21:36:51lostlogicLear: first before change, second after first change.
21:37:02 Join mborus [0] (
21:37:14Learin the while in the first case?
21:37:17lostlogicamiconn: but doese that matter? :)
21:37:25lostlogicLear: yeah, last in the while first, and last in the if in the second.
21:37:27ts-xKyomi: To actually view a list of playlists, yes. But to view their contents, no.
21:37:35 Quit linuxstb_ ("CGI:IRC")
21:37:40 Join linuxstb_ [0] (
21:37:47 Join linuxstb__ [0] (
21:38:22mboruspreglow - the S/PFID analyzer shows "Frequency 0 (44.1 kHz)" no matter if a cable is connected or not. When I connect the cable, SYM goes OK, that
21:38:35mboruspreglow - the S/PFID analyzer shows "Frequency 0 (44.1 kHz)" no matter if a cable is connected or not. When I connect the cable, SYM goes OK, that's all that happens
21:39:52lostlogicLear: then just turn on crossfade: always/0/5/1/5/crossfade and do a track skip and it should freeze, and let me know what logfs you get.
21:41:05preglowmborus: trouble is that's all i've got to go by
21:41:10Kyomits-x: I dont wanna view their contents
21:41:13LearLostlogic: Building on Cygwin, so it takes a little while... :)
21:41:14preglowmborus: if that says 44.1khz, then that's what i get
21:41:21amiconnlostlogic: fyi, while there is an lcd remote for the iaudio, it is not yet supported in rockbox
21:41:29KyomiI just wanna play the playlists because I think my battery dying is just he H300 problem with rockbox
21:41:43KyomiSo I'm testing a full battery on stock firmware
21:41:50lostlogicamiconn: ah, ty
21:42:07amiconnLear: testing on H1x0 or H300?
21:42:11lostlogicKyomi: you can play them with I think the ab button from the main screen...
21:42:13ep0chlinuxstb_: 5g, anything else you think is worth testing?
21:42:13 Quit new5guser ("CGI:IRC 0.5.7 (2005/06/19)")
21:42:18lostlogicKyomi: thought you wanted to _view_ them.
21:42:28amiconnOkay, so maybe I should test on H340...
21:42:38amiconn(in case you don't get the hang)
21:42:41lostlogicamiconn: I think it'll be the same, but sure :)
21:43:55Kyomilostlogic: All it comes up with is a lil thing in the bottom right that says "Set |>"
21:44:03KyomiI know I can do it
21:44:07KyomiBecause i got there before
21:44:14 Join Sinbios [0] (n=Sinbios@
21:44:19Gargamaleep0ch: 5G iPod?
21:44:43lostlogicKyomi: that's the best I can remember, also didn't use the iriver firmware for long.
21:45:14ts-xKyomi: I just tried it, it works here with AB. Maybe your playlists are corrupted?
21:45:31ep0chGargamale: yes, iPod Video
21:45:42amiconnKyomi: Did you build these playlists with rockbox?
21:45:50Gargamaleep0ch: Any word on the playback stuttering or crashing while you scroll bug?
21:45:54 Join dpassen [0] (n=dpassen1@resnet-236-163.resnet.UMBC.EDU)
21:46:29ep0chGargamale: well i try not to scroll too much
21:47:15 Join actionshrimp [0] (
21:47:31ep0chGargamale: i think its something to do with the fact that the screen takes a long time to update and the scroll wheel eats cpu
21:47:32GargamaleBut its a known bug
21:47:38GargamaleAnd peeps are workin on it, eh?
21:48:39lostlogicGargamale: it is not explicitly a bug really... it's performance...
21:48:56GargamaleWell, I guess my point is you guys are aware this behavior is happening
21:49:05GargamaleAnd are working toward correcting it
21:49:25GargamaleMakin sure its not like one of those things only a few people experience
21:49:26ep0chall aspects of the ipod are underperforming
21:49:27preglowcorrecting it is a huge job
21:49:30 Quit ts-x ("CGI:IRC")
21:49:37Gargamalepreglow: I figured as such
21:49:37lostlogicvaguely... by making the system perform better, it will get better...
21:49:54GargamaleSo optimize that code, biotches
21:50:01 Join ts-x [0] (
21:50:03GargamaleI have no clue what Is peak of, pay me no mind
21:50:09GargamaleI speak, even
21:50:30amiconnOh noes...that reminds me. I should go writing a mono bitmap/text performance test plugin
21:50:50amiconnThen someone would need to run that on the colour ipods
21:50:56lostlogicamiconn: that would probably help scrolling on the 16bit LCDs a fair bit eh?
21:51:01GargamaleHow does one exit a rock w/o holding menu/play
21:51:10amiconnI'll do the same on H300 of course
21:51:15lostlogicGargamale: depends on the plugin
21:51:26GargamaleJPG viewer
21:51:30Gargamaleand TXT viewer
21:51:52GargamaleEverytime I view album art or an album review txt I have to reset
21:52:00GargamaleCuz I cant figure out how to exit
21:52:05Gargamale5G iPod, obviously
21:52:15RibsGargamale: middle button?
21:52:23Ribsor maybe menu
21:52:25GargamaleDoes nothing
21:52:29GargamaleTried every button
21:52:36RibsRockbox's controls are somewhat confusing at first
21:52:38Ribshrm, that's off
21:52:42Ribsodd, rather
21:52:47 Quit linuxstb (Read error: 110 (Connection timed out))
21:53:23lostlogicGargamale: select + play or select + menu short presses?
21:53:37Gargamale"Why is she Black Debbie"
21:53:41GargamaleDidnt try that lostlogic
21:53:43KlrSpzi thought they were odd at once, but they are 2nd nature now
21:53:48amiconnSelect+Menu is standard on ipod for leaving plugins
21:53:56amiconnJust Menu in some simple plugins
21:54:39KlrSpznow all i gotta do is set up cvs
21:54:53LearThat reminds me, Should fix jewels so that A-B brings up the menu rather than stop... :) (On H1x0 at least.)
21:55:08amiconnYes, Stop should leave
21:56:12mboruspreglow - I just realized something else. The satellite receiver has done something with the signal: It has changed the volume. So it might have removed other important info from the s/pdif data.
21:56:15amiconnHmm, there was a jewels related commit yesterday... I should check the current behaviour
21:57:06 Quit Ribs ("Leaving")
21:57:45preglowmborus: if the iriver firmware understands it, so should we
21:57:50LearHm, looks like it added exit on A-B, if I read the defines right. :)
21:59:03mboruspreglow - The other problem annoys me more. When I boot fresh and then start a recording with no digital signal present, the Iriver freezes.
21:59:04LearAh, must've been for ipods only...
21:59:22amiconnLear: Any result from your logf build?
21:59:30 Quit dpassen1 (Read error: 110 (Connection timed out))
21:59:35mborusIt does not freeze after some successful digital recording.
21:59:49LearJust finished, kind of - rockboy didn't build, as somebody mentioned...
22:00:01LearFinished building, that is.
22:02:43preglowmborus: i know, i don't how to detect a valid signal yet
22:05:10Learlostlogic: immediate freeze, logf:s in reverse order:
22:05:17LearNULL fade out chunk
22:05:37 Quit linuxstb_ ("CGI:IRC")
22:05:53Leartce: <path to file>
22:05:59 Quit Seed (Nick collision from services.)
22:06:01LearNew track loaded
22:06:05 Join Seed [0] (
22:06:23LearLet me know if you want more. :)
22:06:25amiconnpreglow: How does the s/pdif analyser do it? It shows SYM: OK with a valid input signal as mentioned earlier by mborus
22:10:11preglowgood point
22:10:12Learlostlogic: same freeze without track skip even
22:11:20amiconnHmm, I should visit a certain electronics shop tomorrow
22:11:32amiconnToslink-3.5mm adapter
22:11:46 Join ep0ch| [0] (
22:13:08preglowamiconn: good idea
22:13:17 Join linuxstb_ [0] (
22:13:21amiconnEUR 2.50
22:13:29 Nick ep0ch| is now known as ep0ch_ (
22:13:29preglownot too much, heh
22:14:17 Quit ep0ch (Connection timed out)
22:14:50 Quit SereR0KR ("XChat Aqua")
22:15:02preglowamiconn: mno, it seems to show sym when it's got a carrier
22:15:08preglowval: is blank when nothing is sent
22:15:11preglownow let me press play...
22:15:26preglowyep, then it popped on
22:16:58preglowamiconn: stopping audio or even disconnecting while it's recording does nothing
22:16:59preglowjust silence
22:17:10preglowthen i wonder why the hell i can't initiate playback with no source
22:17:13RedBreva_W00T! All platforms, sims and manuals compiled.... thanks to all those who have kept me pointed in the right direction today...
22:18:02preglowamiconn: now i just recorded with no source plugged in...
22:18:11 Join petur [0] (
22:18:14preglowamiconn: this is what i like about this. it's impossible to find out what i'm doing wrong
22:18:27preglowi'll just go ahead and enable all the interrupts i can find...
22:19:20 Join Bou-divx [0] (
22:20:16lostlogicLear: ah, did you put something to identify which of the NULLs that was?
22:20:34lostlogicLear: at least that's a step in the right direction, now if only I could tell _how_ it's getting null.
22:22:02Learlostlogic: no, you didn't say anything about *that* :)
22:22:20 Join webguest33 [0] (
22:22:51 Quit Kyomi (Read error: 104 (Connection reset by peer))
22:23:05Learlink is obviously null
22:23:09Learbut I guess you realized that... :)
22:23:21lostlogic14:35 < lostlogic> Lear: put those two lines at pcmbuf.c:463 and 491, logf it and see what you get in the logf (if possible, put a 1 and a 0 in the two logfs so we can tell which one goes null)
22:23:23*amiconn fires up vmware
22:23:51 Quit webguest33 (Client Quit)
22:24:05lostlogicwait, it must be in the second one, because if you have a zero fade out delay the first shouldn't be happening.
22:24:34lostlogicor, no, I told you to use a 1 fade out delay, ugh.
22:24:40Learlostlogic: Hm.. 8 hours off...
22:25:31lostlogicLear: where do you hail from?
22:25:50LearNo, seven, it was an hour ago..
22:27:03lostlogica swede, and you ditched devcon!? :-P
22:27:16 Quit linuxstb__ (Read error: 110 (Connection timed out))
22:27:26preglowhe is weak and squishy!
22:27:34Learyep... a bit sad in a way...
22:27:48Learpreglow: look who's complaining! ;)
22:28:54lostlogicugh, more importantly, how the heck can one platform have a NULL link there and the other not??? insanity.
22:29:12preglowwhat're we talking about?
22:29:30preglowi'm guessing gcc lies at the bottom of this...
22:29:46sharpei hate spiders..
22:30:09lostlogicpreglow: pcmbuf.c near line 489, how platform makes a difference...
22:30:17Learlostlogic: second
22:30:42pondlifelostlogic: If you can remember a few weeks ago ;-), Did you ever find out why the pointer to last track data was NULL, or did you just remove the dependencies on it?
22:30:46lostlogicLear: k, that's what I figured.
22:31:08lostlogicpondlife: don't remember context... sorry...
22:31:31 Join midgey34 [0] (
22:31:36lostlogicis resume on startup still broken?
22:31:46pondlifeThere was a crash on H300 which was caused by an unexpected NULL ptr - sounds familiar..
22:32:17lostlogicplayback.c vs. pcmbuf.c
22:32:50lostlogicis resume on startup still broken, btw?
22:33:32 Join Toni1 [0] (
22:33:39sharpehmm... i've a slight idea, anyone know the minimum requirements for an install of windows 95/98, or even 3.1? linux would defeat the purpose of this, therefore it's not included in my idea...
22:33:59lostlogicpentium 100 or something for 95
22:34:32sharpehmm, that's probably out of the question then, however it may still work...
22:35:00lostlogicused to run 3.1 on a 486 dx 33 or similar
22:35:12Toni1preglow, can you send me a sine mp3 which fails in my imdct patch please?
22:35:15petursharpe: dos 3.2 was the most stable os they ever made :)
22:35:36sharpei liked dos 7.1, or was it 7.2...
22:35:41sharpeone of the sevents.
22:35:46sharpeer, sevens.
22:36:22preglowToni1: sure
22:36:25preglowToni1: gimme a sec
22:36:28sharpewhich however wasn't really released on it's own...
22:36:35petursharpe: if you really want to run Win9x, use 98SE, the most stable of the pack
22:36:57Toni1preglow, thanks, I will investigate then tomorrow.
22:37:14sharpepetur: i know, i've got it
22:37:22preglowToni1: and
22:37:22petursharpe: better still, look at this:
22:37:32preglowToni1: when i tried on ipod, it sounded disjointed and cut up
22:37:45preglowToni1: btw, have you done anything with the default routines in imdct and synth_full?
22:37:49sharpepetur: heheh
22:37:53preglowToni1: or did you just rearrange the old code?
22:37:53Toni1preglow, so the asm part is ok?
22:38:05 Join hardeep [0] (i=hardeeps@otaku.freeshell.ORG)
22:38:13preglowToni1: the asm part is ok as far as i can see, but i haven't studied it to validate it's correct
22:38:27Toni1preglow, yes, I reordered in the same way as done to the asm routine
22:38:34 Join romanos_melodos [0] (
22:38:44preglowToni1: ah, ok, so we might even have some speedup on arm as well, then
22:39:03Toni1hope so:)
22:39:09preglowToni1: oh, and yeah, please supply a working swap() for little endian as well
22:39:16lostlogicsomeone should definitely buy Toni an ipod, or a collection of OGGs :)
22:39:28romanos_melodoshi guys! I got a small question. Does anyone knows how can i use both rockbox and ipodlinux on my nano?
22:39:36Toni1preglow: ok, will do.
22:40:20preglowToni1: i think just defining swap(x) to be betoh32(x) should work fine
22:40:52dwihnoToni1's working on mad optimizations?
22:40:55preglowdwihno: indeed
22:41:02lostlogicromanos_melodos: ... install them both...?
22:41:04preglowand fine work it is
22:41:27Toni1You don't have permission to access /~thomj/rockbox/gaplessinus1.mp3 on this server.
22:41:31Toni1You don't have permission to access /~thomj/rockbox/gaplessinus1.mp3 on this server.
22:41:33preglowToni1: i'll fix now
22:41:34 Join damaki_ [0] (
22:41:39sharpebut i just had a great idea, after having to install an older version of windows on a computer, and from the c64 emulator i'm working on... an x86 emulator for rockbox! heheh...
22:41:42Toni1preglow: thanks
22:41:56lostlogicsharpe: you're insane.
22:42:09sharpelostlogic: why do you say that? ;)
22:42:11preglowToni1: try now
22:42:12ep0ch_320kbps mp3 + crossfeed = 0% boost :D
22:42:20preglowep0ch_: holy smokes
22:42:38*preglow finds out he's great after all!
22:42:44Toni1preglow, better now.
22:42:47*preglow kicks arm
22:42:57Learep0ch_: iriver or ipod?
22:43:04ep0ch_ohh maybe 0.1% boost
22:43:09ep0ch_iriver of course :)
22:43:24preglowromanos_melodos: you just install both of them
22:43:26preglowromanos_melodos: that's that
22:43:34lostlogic... and I'd already said that :-\
22:43:42 Quit Toni1 ("CGI:IRC (EOF)")
22:43:42romanos_melodoshow do i do that?
22:43:54romanos_melodosand which one first?
22:44:09dwihnopreglow: you know how well Toni's opts affects ipod performance?