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).

Notice: Only Gecko based browsers prior to FF4 support the multipart/mixed "server push" method used by this log reader to auto-update. Since you do not appear to use such a browser, this page will simply show the current log, and not automatically update.

#rockbox log for 2007-10-17

00:01:38amiconnPutting the whole decoder struct in IRAM doesn't help... at least not when sacrificing map_frame_overlap being in iram for that
00:02:19preglowjhMikeS: oy?
00:02:34amiconnpreglow: Would you happen to know which part of libmad profits least from using iram? (preferably a large part - need to free at least 1.2KB on X5 and 17.2KB on H300)
00:03:01preglowamiconn: no
00:03:10jhMikeSpreglow: ?yo
00:03:12preglowamiconn: 17.2 kb????
00:03:14preglowjhMikeS: had a look at fs #7972
00:03:34amiconnObviously the whole libmpeg2 decoder struct needs that much...
00:04:39jhMikeSI wonder why that's coming up suddenly
00:04:43 Join BigBambi__ [0] (n=Alex@
00:04:54keanuquestion (well, two) - once USB support is added, would the bootloader still boot the OF when the USB cable is attached, or only on <<|? Second - will SD(HC) cards be readable right when it's ready, or will that take a few more days/weekS?
00:05:24preglowkeanu: when usb support is added, we'll pretty much never boot of voluntarily
00:05:37keanupreglow, 'k
00:06:26Zagorkeanu: the card slot won't be supported in the first version
00:07:16bluebrotherZagor: will it be easy to extend it to the Ipods?
00:07:20keanuZagor, ok, thanks for confirming ;)
00:07:25jac0bif you want the original images (which are pretty big) just let me know
00:07:39preglowjhMikeS: i'll just leave that entry to you, should just be a quick commit anywa
00:07:58Zagorbluebrother: I should think so, yes
00:08:08bluebrothercool :)
00:08:08jhMikeSpreglow: now I'm wondering if that the _real_ problem
00:08:14 Join barrywardell [0] (
00:08:29preglowjhMikeS: i'm leaving it to you anyway :P
00:08:54 Join iamben [0] (
00:08:54jhMikeShehe...sure. I'm just saying.
00:09:12kugelhow's usb working?
00:09:59jhMikeSticks me off that GCC never warned about that occuring
00:10:03keanukugel, getting there ;)
00:10:14amiconnOk, I think I'll settle with just putting the idct block in iram. We can always improve further
00:10:23bluebrotherusb works by transmitting data over a cable :P
00:10:33keanukugel, as pointed out a few mins ago, the card slot won't work yet
00:10:35kugelhehe, nice one
00:10:37keanubluebrother, heh
00:10:51kugelkeanu: did i ask for that?! ;)
00:11:07keanukugel, you said how's it coming...
00:11:12bluebrotherit uses differential signals
00:11:36bluebrotherkeanu: he asked how it's _working_, hence my "explanation" ;-)
00:11:51keanubluebrother, I know ;)
00:12:03kugelguys, i meant, how is the progress of getting usb to work!?
00:12:09*amiconn awards jhMikeS the title "red delta king of the month" :\
00:12:31keanukugel, getting there ;)
00:12:38keanukugel, as pointed out a few mins ago, the card slot won't work yet
00:12:47preglowbut such pretty red deltas!
00:12:53bluebrotherkeanu: you're repeating yourself :)
00:12:59*preglow looks forward to preemptive multitasking :>:>
00:13:09keanubluebrother, I know ;) oh wait - even that'd be repeating...
00:13:18jhMikeSthose can't be right
00:13:45keanukugel, AFAIK, nobody except Z agor (no need to alert him...) has the source code
00:13:48 Quit qwm (Remote closed the connection)
00:13:48jhMikeSI moved some data out of IRAM, and some code into IRAM. I added nothing.
00:14:02 Join qwm [0] (
00:14:06bluebrotherkeanu: there's a patch in the tracker
00:14:22keanubluebrother, FS#?
00:14:35jhMikeSpreglow: that'll just burn my butt if if that's the real trouble SWP has been having. OTOH, I'll be happy as well to be able to use it. Will test.
00:15:07kugelkeanu: 7962
00:15:15preglowjhMikeS: i'll laugh myself to death if so :>
00:15:17keanukugel, thanks
00:15:58amiconnjhMikeS: Moving data out of iram means red delta (because dram usage increases), and moving code into iram also means red delta (calling code gets larger because calls to iram are longcalls)
00:16:21*bluebrother was too slow once again ...
00:16:27amiconnmpegplayer is very usable now on X5 :)
00:16:46jhMikeSamiconn: there's not that many calls. seems excessive.
00:18:14jhMikeSpreglow: I'll laugh too and then go whack those GCC guys over the head. :P
00:18:48preglowjhMikeS: they deserve that anyway
00:19:22 Quit WalterEgo (Read error: 104 (Connection reset by peer))
00:19:48jhMikeSat least then I can dispense with the weirdness. We'll see.
00:20:10preglowit would be weird if only those targets were affected anyway
00:20:18preglowi guess they could have a different arm core
00:21:22jhMikeSI think it's weird but that is listed as UNPREDICTABLE in the docs so I guess that mean it may or may not work too.
00:22:01preglowbut get testing! :P
00:22:02 Quit freqmod (Remote closed the connection)
00:22:03toffe82jac0b: nice page
00:22:21jhMikeSagh, and I have to resign from the red deltas and join the green ones on this one
00:23:04 Quit BigBambi_ (Read error: 113 (No route to host))
00:23:12amiconnWidescreen is sort-of usable on H300. Fullscreen still skips
00:23:42markunjac0b: thanks for the howto
00:24:42jhMikeSguess I'll know in a few minutes
00:26:27 Quit Nico_P (Remote closed the connection)
00:27:00barrywardellZagor: I can't remember if I mentioned it before, but FS #6494 might be of use to you. It's basically just the the /proc support ripped from the linux driver
00:28:40 Join BigMac [0] (
00:29:54Zagorbarrywardell: thanks, but that won't help me much now. I'm chiefly debugging request/response sequences at this point.
00:30:18barrywardellIn that case, feel free to close if it's not going to be useful any more
00:30:54Zagorlets keep it until the usb code goes into svn
00:32:58rasherIf anyone knowledgable about libxml, perl, xml could have a look at and and explain why I'm getting latin-1, I'd be ever so grateful
00:34:44jac0bmarkun: did you take a look?
00:36:05 Quit bluebrother ("sleep")
00:37:23Zagorrasher: have you read the perluniintro and perlunicode man pages? some useful info there.
00:37:49preglowjhMikeS: sooooo?
00:38:15rasherZagor: No.. I guess I should
00:38:42rasherI have a sneaking suspicion that perl is trying to be smart at me
00:39:14Zagoryou are probably right. a quick test is to add "use encoding 'utf8';" to the top of your script
00:40:10jhMikeSpreglow: nope, all lock. test_queue won't run now
00:40:11markundo we need a mi4 key to boot rockbox on a portalplayer device or just for disassembling the OF?
00:40:22rasherZagor: woo!
00:40:27rasherAnd also, boo
00:40:33preglowjhMikeS: won't run??
00:40:37markunjac0b: yes, I did
00:40:42jhMikeSjust freezes up
00:40:48barrywardellmarkun: for decrypting the of
00:41:07barrywardellso we can boot an encrypted of if we know the key
00:41:24barrywardelland we can decrypt it for disassembly too
00:41:25*jhMikeS hasn't need a paperclip for awhile for H10 and lost it
00:41:34preglowjhMikeS: as in worse than before?
00:42:08jhMikeSthe same I think. it runs a few seconds then freezes.
00:42:38jhMikeSI guess that verifies again that swp is actually broken
00:42:50preglowpatch should be commited anyway, yeah?
00:43:06saratogabarrywardell: did you write up the procedure for removing the sandisk bootloader on the e200 anywhere?
00:43:28jhMikeSI think so. Let me test e200
00:43:33barrywardellsaratoga: no, I can if you think it would be useful? It's very straightforward
00:43:47jhMikeSor did you test nano already?
00:44:10jhMikeSok, have to make for that
00:44:12markunbarrywardell: I was asking because of this:
00:44:18jhMikeSfor e200 that is
00:44:24saratogabarrywardell: well it didn't sound hard, but if I botch this, i do need to use recovery mode right?
00:44:34 Join Soap [0] (n=Soap@rockbox/staff/soap)
00:44:36barrywardellsaratoga: just build a bootloader, then run "sansapatcher -bl bootloader/bootloader.bin"
00:44:59barrywardelle200tool recover is needed if things go wrong
00:45:08barrywardellit happened to me and I was able to recover fine
00:45:08saratogabarrywardell: alright, sounds easy enough
00:45:10preglowjhMikeS: i figured it woldn't really change much
00:45:36jhMikeSit shouldn't change anything actually for any c code
00:46:21scorche|wsaratoga: heh...i totally forgot one could do that to the nanos (re: more storage)
00:46:21preglowif it tests fine on any of the targets that used to work, i don't think that small patch should matter on any of them
00:46:35scorche|wsaratoga: it added a new project to my list :)
00:46:47barrywardellmarkun: scanning for the key is easy with mi4code
00:46:53jhMikeSpreglow: if you want to commit feel free
00:47:11saratogascorche|w: yeah i know of one person who did, but soldering those chips looks really hard
00:47:17preglowjhMikeS: am in windows right now, can't
00:47:57scorche|wsaratoga: yeah, but i can do it if i take a few breaks...the key to that type of soldering is solder braid ;)
00:48:25saratogai never could
00:48:43scorche|wa good iron and tip helps too..
00:48:47saratogai remember screwing up moto HC11s in 40 pin DIP packages the size of my finger
00:48:48scorche|wand thin solder
00:49:36scorche|whrm...i wonder if the chips will allow me to boost it up to 16 gigs...
00:50:40 Join Mouser_X [0] (
00:51:02saratogabarrywardell: just to be clear, the bootloader.bin is the one that gets created in the bootloader folder?
00:51:07 Quit roolku ()
00:51:30markunbarrywardell: even if it's not known by mi4code you mean?
00:51:34barrywardelldon't use the mi4, use the bootloader.bin
00:51:45markundidn't find one
00:51:52barrywardellmarkun: yes, it can scan the bootloader for the key
00:52:18markunthere is non included with the firmware update. Should there be one on the player?
00:52:29barrywardellsaratoga: assuming nothing broke in the bootloader since I last looked (I'm thinking maybe the major scheduler changes?)
00:52:48saratogabarrywardell: nice to mention 1 second after i hit enter
00:53:12barrywardellsaratoga: you'll be able to recover anyway
00:53:20saratogaah well if i break it its karma for making fun of that guy who flashed his bootloader with a .patch file
00:53:44preglowsomeone did that? :PP
00:54:03saratogahis rom too, not bootloader i think
00:54:16scorche|wpreglow: our famous DWGR
00:54:19saratogathe bad one, whatever its called
00:54:22barrywardellhe should still be able to recover
00:54:29barrywardellthe i2c rom?
00:54:34saratogaalso, should sansapatcher take forever to do this? 2 minutes at 100% cpu
00:54:49barrywardellsaratoga: no, it was almost instant for me
00:54:59barrywardellsaratoga: make sure you build a new sansapatcher
00:55:09barrywardellold versions lock up like that
00:55:12saratogai downloaded the one in the manual
00:55:20saratogaits at "[INFO] e200 found - disk device 2", is it ok to kill it?
00:55:23barrywardellsaratoga: you need to build a newer version
00:55:30barrywardellsaratoga: yeah, kill it
00:55:52barrywardellsansapatcher seems to have a bug with unknown switches. it locks up like that
00:56:04barrywardelland I only added -bl a couple of days ago
00:56:08saratogais teh source included svn checkouts?
00:56:24barrywardellyou need to build a e200 and c200 bootloader
00:56:32barrywardellcopy the mi4 files to rbutil/sansapatcher
00:56:40barrywardellthen run make in rbutil/sansapatcher
00:57:02saratogai thought it was the .bin file?
00:57:22barrywardellmarkun: the bootloader is probably in flash, and the mi4 on the hdd
00:57:42barrywardellsaratoga: the mi4's are just needed for building sansapatcher
00:58:39preglowthis guy truly is the king of clueless
00:59:04markunbarrywardell: could we make a mi4 file to dump the flash?
00:59:08scorche|wpreglow: have you not encountered him before?
00:59:20saratogabarrywardell: are a lot of warnings normal?
00:59:30barrywardellon amd64?
00:59:42saratogasince they're all long long related
00:59:48 Quit ender` (" If the "black box" flight recorder is never damaged during a plane crash, why isn't the whole airplane made out of that stuf)
01:00:03barrywardellthey're all just printf warnings?
01:00:21scorche|wpreglow: well, for entertainment search for that name and DavidGWRawson (his old account that he deleted when he got banned)...and dont expect him to come into IRC, as he is banned permanently (he used too many names to search for ;)
01:00:24saratogadoesn't say, but they sound liek it given the mention of "format"
01:00:47jac0bpeace out to all
01:00:53barrywardellmarkun: yes, should be possible provided it accepts plaintext mi4s (it most likely does)
01:00:56 Quit jac0b ()
01:01:26barrywardellmarkun: IIRC the flash is remapped to 0x20000000, but better check that to be sure
01:01:59barrywardellsaratoga: you can just ignore them, those printfs are probably never even reached
01:02:29*scorche|w leaves for home and to go order a few memory chips for upgrading his nano...
01:03:08 Join Calcipher [0] (
01:03:09 Quit scorche|w ("CGI:IRC (EOF)")
01:03:18preglowjhMikeS: you're not going to commit it?
01:03:51 Quit jhMikeS (Nick collision from services.)
01:03:51CalcipherWhats happening all
01:03:53 Join _jhMikeS_ [0] (n=jethead7@rockbox/developer/jhMikeS)
01:04:45CalcipherI've just gotten back from work, and enjoyed my first on the field use of rockbox
01:04:48 Quit krazykit (Nick collision from services.)
01:05:02saratogabarrywardell: thanks a lot, works great
01:05:04 Join krazykit [0] (
01:05:07saratogaalso, damn this is fast
01:05:13CalcipherI have to say I'm impressed
01:05:33saratogamaybe 3 seconds to boot
01:05:52saratogatakes longer to spin up a disk on my 3g I think
01:05:56barrywardellsaratoga: I was working on adding the rockbox logo to the bootloader too. I think it would be a nice addition
01:06:05saratogayeah that would be cool
01:06:06CalcipherI use a sansa e280r, and am loving the experience, I also use extra large fonts and the voice menus
01:06:15barrywardellsaratoga: you wanna try the patch?
01:06:18preglowbarrywardell: logo in bootloader? why?
01:06:25preglowwe've cut that from all other bootloaders
01:06:36barrywardellbecause, otherwise we have just a blank screen
01:06:44barrywardellit would be compiled in
01:06:55saratogalogo makes sense on a flash player IMO, not like the disk access costs anything
01:06:57preglowwhy not do it like on ipod?
01:07:09preglowbootloader says nothing unless anything actually happens
01:07:19barrywardellbecause the ipod already has the apple logo shown
01:07:29barrywardellon the sansa, we'll just have a blank screen
01:07:33preglowfor how long?
01:07:37barrywardelland it won't cost anything
01:07:43saratogapreglow: about 1 second i guess
01:07:53preglowthen please make a new rockbox logo while you're at it :>
01:07:58saratogamore like 2 seconds
01:08:16saratogayes a low res logo would be nice
01:08:18CalcipherAnyone know if there is a 90 degree display rotate rotate capability using rockbox on a e200?
01:08:19saratogai want one for the default theme
01:08:26preglowthe current one really does lack style
01:08:43saratogaCalcipher: not going to work too well on a non square screen
01:08:57rasherLet me restate my praise for the logo used for the "H300 experimental build"
01:09:01 Quit BigMac ("Leaving")
01:09:08barrywardellyou also need to copy apps/bitmaps to bootloader
01:09:18Calcipherreally, I thought that it might be nice since the larger fonts wouldnt have to scroll as much
01:09:37saratogai can try it later though
01:09:37saratogabarrywardell: sorry, i'm not at the computer that can run sansapatcher anymore
01:09:42Calcipherlike only in the rockbox menus, for easier reading
01:09:54preglowrasher: it is pretty nice inded
01:10:10barrywardellpreglow: the logo would only show while booting, and it would be compiled in so there's no extra disk access
01:10:13*_jhMikeS_ 's editor whines about CREDITS and chars not in CP-1252 even if opened as UTF-8 now :\.
01:10:16amiconnsaratoga: The bootloader should either show the same logo as main rockbox, or no logo at all
01:10:23 Nick _jhMikeS_ is now known as jhMikeS (n=jethead7@rockbox/developer/jhMikeS)
01:10:23amiconnJumping logo would look strange
01:10:24CalcipherIt would kind of be like widescreen look
01:10:31preglowless cluttered, high contrast, sleek
01:11:22CalcipherI wish I could help, I really would put some time into this, since I'm one of the people who can benefit from all the accesibility features rockbox has
01:11:24rasherjhMikeS: it's times like these you wish they'd just created utf8 back when they were wiring up the first computers
01:11:47***Saving seen data "./dancer.seen"
01:11:57*barrywardell agrees with amiconn
01:11:58preglowrasher: think i'll try replacing the current one to find out how it looks
01:12:34*amiconn much prefers the current logo over any alternative suggestions so far
01:12:36jhMikeS:) ... only problem is the editor is not _supposed_ to be complaining because I'm telling it what it is.
01:12:47preglowi'm starting to hate the current one :/
01:12:56preglowit's just... cluttered
01:13:08*jhMikeS was going to say "too busy"
01:13:14rasherIt's not so much a logo as a collage
01:13:22saratogathe current logo will look interesting on the c200 anyway
01:13:26barrywardellthat h300 experimental build on is nice indeed
01:14:09saratogathe c200 has a really tiny screen IIRC
01:14:11jhMikeSsome simple design with few colors (maybe just two) would be nice.
01:14:20preglowsucky thing about the h300 experimental one: it contains small sized text
01:14:21amiconnsaratoga: Not as tiny as the archos screens
01:14:31preglowjhMikeS: agreed
01:14:36saratogayeah but the archos is B+W right?
01:14:45amiconnYeah, and?
01:14:48pixelmasaratoga: even harder...
01:14:54amiconnThe current logo doesn't need many colours
01:14:57saratogawe use the color logo on that?
01:14:58rasherpreglow: I pretty much mean only the square logo left of the text
01:15:07amiconnJust black, white, grey, and orange
01:15:21CalcipherAre there any builds I can test that would allow a screen rotate?
01:15:28Calcipheron a e200 series?
01:15:29 Quit kkurbjun ("leaving")
01:15:48CalcipherI'm sure some had to have tried it by now
01:15:50krazykitCalcipher, there was a patch for gigabeat, afaik
01:16:35*markun just wants to say he also likes the "H300 experimental build" logo
01:16:35CalcipherDo you know if that was maybe originally featured in the gigabeat player?
01:16:47 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
01:17:05pixelmasaratoga: it's prepared as black and white version (and aspect ratio 'adjusted')
01:17:11krazykitCalcipher, it was, actually, but the patch may be able to be converted for e200... though you'd need to make custom WPS's and such
01:17:16 Quit Toxicity999 ("Leaving")
01:17:33 Join Toxicity999 [0] (n=bryan@unaffiliated/Toxicity999)
01:17:38amiconnjhMikeS: Did you see my question earlier today?
01:17:49amiconn(or in fact already yesterday for me)
01:17:58CalcipherI just started working with rockbox last night, I still don't know what wps's do
01:18:14krazykitCalcipher, they make the What's Playing Screen look different
01:18:22Calcipherah I see
01:18:26saratogai still like the idea of a small text only "logo"
01:18:30jhMikeSamiconn: haven't travelled any logs yet
01:18:30saratogasomething liek that
01:18:52n1sCalcipher: to do that in a nice way, would be a _major_ task, almost all of rockbox' gui code relies on fixed width and height of the display
01:18:59CalcipherShiiet, If thats all it takes, I'd put in the time, when ever someone can port the gigabeat version for e200
01:19:06amiconnjhMikeS: Did you write mpegplayer idct_arm_c.c yourself, or is that something ported from libmpeg2 we didn't have before?
01:19:24amiconnI'm asking because I want to convert it to an .S file
01:19:33 Part hcs
01:19:39jhMikeSamiconn: it's of course based off their code and uses their shortcuts
01:20:15 Join donutman25 [0] (n=chatzill@
01:20:17preglowjhMikeS: you not going to commit that patch?
01:20:24jhMikeSI am
01:20:30Calcipherhmm, let me take a look at the screen sizes, hopefully they may be similar or atleast nearly proportionate
01:20:35n1sCalcipher: the gigabeat patch is basically a seperate build with width/height swapped and a hacked up bootloader allowing you to choose orientation on startup
01:20:40jhMikeScan't credit anyone though since my editor will mess up the file
01:20:49pixelmasaratoga: I wouldn't call that a logo (even less than our current one) - to me a logo is all about less text, more a sign
01:21:00preglowjhMikeS: what kind of crap editor is that?
01:21:11preglowrasher: tried on target now, the font of the main logo is too thin as well
01:21:18saratogaor maybe menu back drop
01:21:18saratogawell ok, mainly i just want to put the word "rockbox" somewhere in the default wps
01:21:21Calcipherthats a pretty cool feature
01:21:27jhMikeSTextPad - it does it even if I open as UFT-8.
01:21:41preglowit's not utf8
01:21:45preglowit's latin1, afaik
01:22:03preglowiso-8859-1 or wtf
01:22:08amiconnjhMikeS: But you wrote the actual asm?
01:22:16Calciphergigabeat LCD is 320x240
01:22:17jhMikeSThen it says "not in CP-1252" or whatnot. doesn't matter. I tried UTF-8 after the whining.
01:22:25jhMikeSamiconn: yes
01:22:34amiconnThen it's your code, imo
01:22:41Calcipherand the e200 series LCD is 220x176
01:23:11Calciphernumbers would def need tweaking
01:23:28preglowrasher: main problem with that logo is that i wastes too much space. i kind of like that, but i don't know if we can afford it if we are to use it on all our targets
01:23:31Calcipherhow do you guys usually go about deving GUI fixes
01:23:42Calcipherusing vm's?
01:24:12Calcipherrunning the emulator seems like a quick way to test the changes
01:24:18preglowrasher: perhaps we should just jot down a list of requirements and start a logo competition :V
01:24:42iambenpreglow: free rockbox as prize?
01:24:43jhMikeSa simple notan abstraction of a streaming BoS?
01:25:06rasherpreglow: what do you mean by wastes too much space?
01:25:08markunpreglow: it should be something that looks nice on a t-shirt and on a printing card as well :)
01:25:16markungreetnig card I mean
01:25:38preglowrasher: it doesn't really waste any more than the current one does...
01:25:44preglowrasher: too much empty space
01:25:57n1swe should hire a PR firm for the donation moneh!!!
01:26:22jhMikeSpreglow: I guess I should edit in something else and run dos2unix on the stupid thing to make sure. :\
01:26:43rasherpreglow: the one I mean is this:
01:27:18CalcipherI tried the rockbox e200 sim before I installed it
01:27:41markunrasher: also with these colours? I prefer the orange on black one
01:27:43CalcipherI still wouldnt know how to start editing the gigabeat UI specifics
01:27:58jhMikeSorange? eck
01:28:08n1sCalcipher: that's c code you'll have to deal with
01:28:18rashern1s: getting an actual graphic artist to jot something down might not be too awful, but paying for would probably be a bit over the top
01:28:19preglowrasher: why, that's not the one i was looking at :>
01:28:26preglowi'd like a logo with actual text, heh
01:28:35rashermarkun: I don't care much about the colours to be honest
01:28:59rasherpreglow: you can add "Rockbox" to the left of that, if you must, I suppose
01:29:14markunor to the right like it's now
01:29:18preglowthat cube thing is cool, thoigh
01:29:24rashereh, yes.. right
01:29:25jhMikeSjust a distinct icon - no words
01:29:42rasherPersonally I think a large version of that icon in the middle of the screen on bootup would look nice
01:29:53preglowit actually would
01:29:58preglowour own version of the apple :P
01:30:06markunand on the front of a t-shirt :)
01:30:10 Part toffe82
01:30:10jhMikeSa box with a "byte" out of it :P
01:30:20rasherOf course, it was drawn by a guy who seems to be awol, so that might be a problem
01:30:43zichowhat capacities are sansa e200 available in?
01:31:12 Join shnee [0] (
01:31:36iamben2G 4G 6G 8G
01:32:35zichoAh. Its a really good-looking machine, i must say. Way nicer than the ipod.
01:33:02iambeni enjoy mine
01:33:31preglowjhMikeS: you don't have any ipods, no?
01:34:04zichoiamben: i would too. although i enjoy my gigabeat =)
01:34:16jhMikeSpreglow: a 3g
01:34:44preglowjhMikeS: does that share the same clickwheel code as in the accel patch?
01:35:37jhMikeSI'm not sure. I'm ordering a FW card as soon as I find the deal I want. Haven't really done any dev on that device yet.
01:35:59preglowfw card?
01:36:09CalcipherIn the image on that site the rockbox logo is vertical on the lcd
01:36:12preglowahh, right, it needs that
01:36:34krazykitCalcipher, yes, an obviously bad photomanip
01:36:36pixelmajhMikeS: you said you were going to the shop today :P
01:36:37Calcipherdon't know if that was only the logo on start up or the whole gui
01:37:02Calcipherbut...but it would be so great
01:37:13 Quit barrywardell ()
01:37:15jhMikeSpixelma: I've been looking around. Haven't picked one though.
01:37:18preglowjhMikeS: i'm just really curious about the improvements you mention to the accel patch. i'd really like to this stuff to be pretty perfect when commited, people tend to bitch and whine if you alter their controls once they're used to it
01:37:47jhMikeSnoone bitched when I altered the sansa wheel. ;)
01:38:01 Quit Arathis ("Bye, bye")
01:38:12preglowwe're talking ipod users here :P
01:38:15jhMikeSpreglow: what improvements?
01:38:28preglowhysteresis, etc
01:38:35jhMikeSdid he do that?
01:38:42preglowyou mentioned some other stuff you'd do to a continuous controller like the ipods
01:38:46preglowdon't believe he did, no
01:38:55preglowi don't think he quite grasped what you were talking about
01:39:03preglowwell, yeah, not stepped like the sansa's is
01:39:06 Quit krazykit ("leaving")
01:39:14preglowcan't remember exactly what you said :/
01:39:14rasherWhatever happened to the magazine that was going to ship Rockbox on their cover-cd?
01:39:19jhMikeSso it reads an infinite number of positions?
01:39:22 Join krazykit [0] (
01:39:46preglowi don't think apple are that clever yet :D
01:39:57jhMikeSI thought it had 96 steps
01:40:02jhMikeSe200 has 48
01:40:29preglowit has 96 steps
01:40:41preglowperhaps i just misunderstood what you said, i can't even remember it too well
01:41:22*jhMikeS suggests Buschel find an e200 and just investigate the haptics on that :P (he like that word "haptics")
01:41:56*preglow visits
01:42:23*preglow comes back empty-handed
01:42:39*jhMikeS is at a loss to explain the "hysteresis" thing then. If you know what it means to a signal level, apply it to an angle.
01:42:58saratogajhMikeS: are you in some place thats cheaper then the cost of a firewire card to ship to from the USA?
01:43:18preglowhe's in the usa
01:43:38saratogacool, i have a few firewire cards, you can have one
01:44:00jhMikeSsaratoga: ah...I guess that's a good deal. :) Thanks.
01:44:22jhMikeSI doubt amazon or ebay or whatever can top that.
01:45:11saratogano problem, just message me with your address
01:45:59 Join advcomp2019_ [0] (n=advcomp2@
01:46:24Calcipherwho worked on the gigabeat patch?
01:46:52krazykitshould say on the tracker.
01:47:54 Quit n1s ()
01:47:59*preglow needs to bed
01:48:10jhMikeSnight night
01:48:48 Quit Zagor ("Client exiting")
01:48:58 Join animeloe [0] (n=animeloe@unaffiliated/animeloe)
01:51:37 Quit saratoga ("CGI:IRC")
01:54:41 Join Bill [0] (
01:55:01BillIs it possible to edit roms?
01:55:21zichoYou mean like, gameboy-roms?
01:56:27zichoYes i think it is. But i do not know how.
01:56:35CalcipherAs to my question about rotated display on the e200s, should I make a feature request? is that the way to go about this?
01:56:48krazykitBill, probably, but it's way offtopic here.
01:57:44Billi heard about tile editors, but i cant find a place to download any
01:57:48CalcipherI've seen dozens of hacked roms for practically every system, so definitely yes, and I have no clue how its done, and also zicho is right thats crazy off topic
01:58:07krazykitBill, it's offtopic. go ask google.
01:58:26zichoCalcipher: I wasnt the one who said it was OT =)
01:58:27Billive been asking google forever
01:58:31Billthat is why i am here.
01:58:41zichoWell, youve come to the wrong place
01:58:55Calcipherapparently I wasn't paying attention haha
01:59:57CalcipherI want to help with the TTS on the menus, I think we can add voice reading on some of the db category sub sections
02:00:11Calcipherartist album artis, ect
02:01:20CalcipherAlso I have many higher end tts voices running on my XP system, so I'd be willing to run some voice sets if someone can walk me through the process
02:01:29 Quit Bill (" Chat in IRC at school!")
02:01:31Calcipheror point me in the right direction atleast
02:01:45CalcipherI just don't know how to apply that to the ui
02:02:13 Join JdGordon [0] (
02:02:20Calcipheralso, is there a volume setting for the voice menu?
02:02:23 Quit n17ikh|Lappy (Connection timed out)
02:02:32CalcipherI couldn't find one
02:02:42 Quit advcomp2019 (Read error: 110 (Connection timed out))
02:03:02rasherI don't think it's interesting to add voice to some db sections. The strings depend entirely on the db-config file used
02:03:33Mouser_XApparently I'm way late on asking this (the build I was using was 3 or more months old), but why did controls on the Gigabeat get changed?
02:03:34Calcipherbasically its dependent on the player?
02:03:35rasherIt'd be more reasonable to create something like .talk clips for the db
02:03:50rasherCalcipher: it's dependent on the user's configuration
02:03:58krazykitMouser_X, to put it more in line with other targets. there's a thread in Announcements forum
02:04:32Mouser_XAh. I don't check the forums.
02:04:45 Quit linuxstb ("ChatZilla [Firefox]")
02:04:46Calcipherah so the user can congif what categories are included
02:05:09rasherCalcipher: not only which categories are included, but what they are called. The strings are completely configurable
02:05:17krazykitMouser_X, there was also a mailing list thread
02:05:25CalcipherI see
02:05:26Mouser_XI'm not on that eitehr.
02:05:43krazykitthen you're way the hell out of the loop :P
02:06:00CalcipherWell I was just trying to figure out how to make it easier to navigate for the blind
02:06:08Mouser_Xkrazykit: Most likely.
02:06:21Mouser_XAbout the only thing I check (at all) is what commits were made.
02:06:35Calcipherso generating a talk clip per user defined db sounds like a decent method
02:06:47 Quit spiorf (Remote closed the connection)
02:07:12rasherCalcipher: I think I pretty much have it figured out: 1) Create a client-side script that reads the db and the db-config, and creates .talk clips for all strings needed (menu entries and artist/album/track names), hooks in the db code to check for these clips at some predefined location
02:07:43krazykitMouser_X, guess you missed that commit :P
02:08:19Calciphersounds like a plan
02:08:45CalcipherDoes the tts system support any other synthesizers?
02:09:15rasherCalcipher: other than what?
02:09:38Calcipherwell I only saw espeak
02:09:43 Quit rjg ("leaving")
02:09:55Mouser_Xkrazykit: I read the comments. That doesn't mean I always understand what they say (and, since it was awhile ago, it's entirely possible I forgot about it, or did miss it.)
02:09:57Calcipherdoes that support sapi5 voices if theyre installed?
02:10:07rasherCalcipher: then that's because that's what you have installed. It supports festival, flite and another one
02:10:16rasherAnd sapi4/5 on cygwin
02:10:43amiconnSAPI4 is not really recommended though. SAPI5 is.
02:10:56CalcipherI'm on XP, and have only seen some of the talk clip generating features in rbutil
02:11:13Calcipherlet me take another look
02:11:19Calciphermaybe I can use the voices
02:11:29amiconnSAPI support is most probably not in rbutil
02:11:54rasheramiconn: I think it has something, but I don't know how it works.. probably you have to copy the vbs script somewhere
02:13:48Calcipheryour right
02:13:55CalcipherI only see espeak and flite
02:14:04JdGordondid zagor get usb going last night?
02:14:19CalcipherSo what would be an alternative
02:14:31rasherCalcipher: if you had sapi_voice.vbs somewhere, you could use SAPI5 as well, I think
02:14:46rasherBut I really don't know much about rbutil
02:15:00rasherAnd I should also see about getting .talk generation support in
02:15:58Calcipherhmm, where can I find that sapi5 java file?
02:16:19Calciphershouldn't that already be on my sytem
02:16:46amiconnHmm, I'm getting a warning in oggmalloc.c when compiling a sim that doesn't show up in the build table
02:16:54Calciphersystem* since I have updated sapi5 and commercial voice synthesizers
02:16:56amiconnThis is on cygwin, so host gcc is 3.4.4
02:17:03pixelmait does
02:17:06amiconnoggmalloc.c:8: warning: dereferencing type-punned pointer will break strict-aliasing rules
02:17:14rasherIt's not a java file.. and you can find it in the Rockbox source, or here:
02:17:14pixelmabut only for one buildserver
02:17:31pixelmait's the single warning
02:17:40amiconnOne of the rare cases where gcc 3.x throws a warning but 4.x doesn't
02:18:01Calcipheralso if it can help somehow, I have a tts setup for a linux based gps system, I believe its a tom tom set up
02:18:06 Quit qweru ("moo")
02:18:24amiconnMeh, didn't notice it in the build table
02:18:45pixelmaroolku wondered after his commit...
02:19:04Calcipherso point to that under rbutils options as the tts executable field?
02:19:11pixelmabut it appeared with the big commit of yesterday
02:20:12Calcipher...oh yeah vbs is visual basic...why did I think java...
02:21:39rasherCalcipher: I think that's how you'd do, but I don't really know for sure - I've never used it
02:22:21 Quit krazykit (Read error: 104 (Connection reset by peer))
02:22:33Calcipheralright, I'll let you know how it goes. what are you running?
02:23:02CalcipherI originally thought I was going to have to boot into ubuntu to do some rockboxin
02:23:39Mouser_Xmarkun: The Gigabeat remote doesn't work in plugins (specifically Mpegplayer). I recall you talking about this a few months back, when commited some changes to how the remote buttons were handled.
02:24:05Mouser_XHow easy would it be to get the remote working in Mpegplayer?
02:24:47 Join Bill [0] (
02:25:05 Join grndslm [0] (
02:25:16BillWould it be possible to add another codec? (hypothetically speaking, of course)
02:25:52Mouser_XOf course it would.
02:26:03Billhow hard?
02:26:06Mouser_XCheck the flyspray. There's a few on there that already work.
02:26:26Billlike an .avi codec?
02:26:38Mouser_XThat would be very hard.
02:26:56Mouser_XMostly because most players don't have enough CPU to decode video very well.
02:27:01 Join krazykit [0] (
02:27:08Mouser_XAlso, AVI is not a codec.
02:27:28Billwell i mean the codec to play an avi file
02:27:46Mouser_XAVI is a container format, it's not a specific set.
02:28:03Mouser_XIn other words, an AVI could be Divx, Xvid, or any number of other formats.
02:28:07Calcipherxvid divx? what do you mean?
02:28:16Billyes yes yes correct =)
02:28:28Mouser_XAVI files contain different formats. Xvid and Dvix are 2 different formats.
02:29:05 Join safetydan [0] (i=dc9d468b@rockbox/developer/safetydan)
02:29:15BillWhat other kind of codec patches are there?
02:29:36Mouser_XI have GBS (gameboy audio files) and MOD patches on my Gigabeat.
02:29:36 Quit jhMikeS (Nick collision from services.)
02:29:42 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
02:29:46Mouser_XThere's others, I think, but I don't use them.
02:29:47krazykitBill, i suggest checking the tracker
02:30:03Billi am
02:30:29Calcipherrasher, rbutil did not accept the script file under the executable field
02:30:43Billi just want a different video type codec because im having a little trouble converting my files to the right format
02:30:50CalcipherI guess it would need to be compiled into an executable
02:30:50 Quit Toxicity999 ("Leaving")
02:31:10 Join Toxicity999 [0] (n=bryan@unaffiliated/Toxicity999)
02:31:16rasherCalcipher: Then I don't know. As I said, I've never used rbutil to generate talk clips. I mostly know about the language voicefiles
02:31:37Calcipherthanks again though
02:31:44Mouser_XBill: What are you using to convert your videos?
02:31:50Mouser_XOr, how are you converting them?
02:31:58Mouser_XDid you check PluginMpegplayer?
02:32:08Mouser_XI use WinFF, and it works great.
02:32:13 Part DraX
02:32:20Calcipherwell is most of the dev work done on linux?
02:32:22Billim converting my files to MPEg using what ever the tutorial said
02:32:47Billthen i put the file into the folder on my ipod (j:\videos)
02:32:55Billthen when i play it says the video is 0:00
02:33:20Mouser_XOddly enough, I have had videos that crash other players (Media Player Classic, VLC) convert with WinFF. As such, I'd have to say it's pretty good.
02:33:34Calcipherbecause the voices I have for linux hq natural sounding sapi5 voices that may help whoever is working on these features
02:33:52rasherCalcipher: it probably is, yes, though almost everything used works on cygwin as well
02:36:29Calcipherhmm I guess I'll look into trying some of the tools using cygwin
02:37:42Calcipherwell if anyone has any ideas about my suggestions and questions I'll be online, afk
02:37:57 Quit Bill (" Chat in IRC at school!")
02:39:07SoapCalcipher, I saw earlier how you said you wanted to help...and are someone who can benefit from all the accessibility features. There is someone who goes by the IRC nick of ddalton...he is working on a series of patches he finds useful as a blind man. He is also always looking for feedback and testers of his modifications. You might want to get with him and see if there is any testing or help he needs.
02:39:38SoapFeedback even, I'm not sure he needs another coder - perhaps just someone to test and bounce ideas off of?
02:40:21Calcipherthat sounds good
02:41:25CalcipherI'd like to help, I'd also be interested in helping in any way I can with a port gigabeat patch for the e200
02:42:30Calcipherthank you for the info rasher, I'll keep an eye out for the name Soap, thanks
02:42:59Mouser_XCalcipher: I'm pretty sure that the Gigabeat rotation patch is largely built off of the iPod, which has a 320x240 resolution. In other words, I think much of their code derives from the iPod port.
02:43:08CalcipherI'll be afk for about an hour, so if anyone wants they can leave me a PM, thanks again
02:45:01CalcipherI thought the process would be a kind of find and replace when it came to certain positioning values...I wouldn't really know though, but again I'll be back later
02:46:25 Quit kugel (Read error: 104 (Connection reset by peer))
02:47:22 Join toffe82 [0] (
02:51:39amiconnPermanent rotation (i.e. a special build) shouldn't be difficult
02:52:05amiconnRuntime rotation, however, would require changing lots of the inner workings of rockbox
02:52:42amiconnThe sansa lcd is the same res as the H300 and iPod color, just rotated, so all the logos etc should be available for a landscape e200 build
02:53:47amiconnHowever, I wonder why people buy a DAP with a portrait display and then want to use it sideways, instead of buying a DAP with the desired display orientation in the first place...
02:56:43rasherDoes anyone know where I can find information on the database_[0-9]*.tcd files?
02:57:39 Join Gnu47 [0] (
02:58:04 Part pixelma
02:58:38safetydanrasher, what sort of information? AFAIK the only documentation is the code
02:59:20rashersafetydan: I just need to know enough to extract the tags out of the files (for generating .talk clips)
03:01:11JdGordonamiconn: i tihnk more people buy a dap for its audio abilites that its screen size
03:01:24rasherAnd the code is not terribly good documentation for someone like me
03:02:05Mouser_XI'd like a utility to use my Winamp Media Library tags (specifically song ratings) in Rockbox's DB... Obviously, to do this, one would have to know something about the format of each (Winamp can export its DB, as can Rockbox. Perhaps something can be done with that?)
03:03:46keanuanything new with USB? /me impatient ;)
03:04:28scorchekeanu: stop asking ;)
03:05:20keanuscorche, it's been like, three hours since I last asked ;)
03:05:27keanuthough that is rather short...
03:05:45scorcheyou will know when i happens by watching this channel and/or the commit log
03:06:12keanuso, I should stop asking, and wait for my IRC client to highlight "USB" ;)
03:06:53 Nick fxb is now known as fxb__ (
03:08:13safetydanrasher, the code isn't terribly good documentation for anyone unfortunately. Anyway, I believe each tcd file is just a list of every single tag value of a single type.
03:08:33safetydani.e. database_0.tcd is a list of every single song title, database_1.tcd is a list of every single album title, etc
03:09:08safetydanthere should be a entry count at the start of the file
03:11:50***Saving seen data "./dancer.seen"
03:13:44rasherYeah.. I'm wondering what the rest means
03:14:13 Nick advcomp2019_ is now known as advcomp2019 (n=advcomp2@
03:14:26safetydanthere's an offset to the first entry and maybe the maximum entry size (since i think they're fixed record width files)
03:14:37safetydanit's been a while since I looked at that code
03:15:04rasherHelpful all the same, thanks
03:15:21 Join karashata [0] (
03:17:15 Join J3TC- [0] (
03:19:28 Quit billytwowilly (Remote closed the connection)
03:26:52 Join iamben_ [0] (
03:32:30Lloreansafetydan: Actually, I'm about 95% sure he can do what he wants with the "format" string, which lets you sort by, among other things, numerical tags as long as they're supported in the database
03:32:41 Join kkurbjun [0] (
03:34:40Lloreansafetydan: That's why I'm having a problem understanding what he's wanting. The only thing I could think of that he wanted to do, but couldn't, was "Sort by an Album Year" rather than simply "Order the list by the Year tag" which as far as I can see is already doable.
03:35:30safetydanLlorean, I didn't think you could order the list by anything at the moment which was why people were asking for sort tags? But I don't use tagcache so I can't really know I guess.
03:35:52Lloreansafetydan: You can order the list by any supported tag.
03:36:01LloreanOr combination of supported tags using printf syntax
03:36:26LloreanPeople could already simply put an alternate form of "Title" in the Comment tag and support by that, I think.
03:36:37LloreanSort tags just give handy scratch tags that can be used in that format string universally
03:37:05 Quit iamben (Read error: 110 (Connection timed out))
03:37:39Lloreansafetydan: It seems you can even specify how many digits to give numerical tags so that you can correct 1, 2, .., 10, 11 track numbers
03:39:01rashersafetydan: two of my database files are identical, with only a difference in two bytes before each string, where one database has 0xffff, and another has some sort of value.. do you know what that might be?
03:39:24rasher_3 and _8 if that tells you anything
03:40:01safetydansafetydan, seems you're right. Or at least I think you're right since the tagnavi syntax isn't easy to read
03:41:03safetydanrasher, that's either an offset to the next tag or... well actually I have no idea
03:41:37rashersafetydan: those two bytes are not necessary to read the db at least - I have code working that reads out all strings from a database file
03:43:25safetydanrasher, if you look at the tagcache.c file and the tagfile_entry struct that might give you a clue
03:44:00rasherI guess it's the idx_id "Corresponding entry location in index file of not unique tags"
03:44:18rasherWhatever that means, it's not important for what I'm trying to do, I suppose
03:44:31Lloreansafetydan: The tagnavi syntax is a terrifying mess, I think
03:44:45LloreanI think the problem might be that you can't have nested custom format strings.
03:45:03safetydanrasher, no I'd say it's irrelevant. The tag_length should be all you need
03:45:07 Join billytwowilly [0] (
03:46:45rasherGuess I can safely ignore all of database_4 as well
03:47:38safetydanrasher, the tags_str array should give you a clue as to what each file is as well
03:48:00 Quit donutman25 ("ChatZilla [Firefox]")
03:48:30rasherYeah, that's what I'm looking at right now
03:48:45rasherI only seem to have database_0 through 8 though
03:49:46safetydanrasher, I think numeric tags are stored in a different file
03:50:45rasherOkay, well there are already mechanisms for speaking numbers
04:01:30keanuSince Zagor isn't even on right now, I take it USB won't be avail until tomorrow?
04:01:55Lloreankeanu: It'll be available when it's available.
04:02:45keanuLlorean, heh, ok
04:03:14 Join psycho_maniac [0] (
04:04:19 Join aliask [0] (
04:08:46 Join webguest75 [0] (i=47743139@gateway/web/cgi-irc/
04:11:07 Quit Mouser_X (Read error: 104 (Connection reset by peer))
04:11:08 Quit webguest75 (Client Quit)
04:13:02 Quit Seed (Nick collision from services.)
04:13:10 Join Seed [0] (
04:14:04 Join Mouser_X [0] (
04:14:59 Quit Thundercloud (Remote closed the connection)
04:24:02 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
04:24:11 Join miepchen^schlaf [0] (
04:30:47 Join RoC_MasterMind [0] (
04:31:07 Join ddalton [0] (
04:33:13Calcipherah so I saw above there is some hope, since the h300 has been rotated
04:33:55alienbiker99hm is this a bug, that hitting left whiel i nthe file view on the h300 sends you back to the rockbox menu instead of up a folder?
04:34:25alienbiker99hm now its not doing it
04:34:30CalcipherThe reason I bought my sansa e280 was because of the expandability, newly rockboxed support, and flash type storage versus hdd
04:35:30 Quit Gnu47 ("I think half 3 in the morning is bed time")
04:35:36CalcipherSo only after trying and enjoying the features and customizability of rockbox on a e200
04:35:47 Quit Lambuntu (Remote closed the connection)
04:36:16ddaltonalienbiker99: no not if you are in the root dir.
04:36:50alienbiker99i wasnt in the root directory, i was one below it
04:37:28Calcipherand using extra large fonts with a nice theme, did I realize how much screen space is sort of unused in some of the display area when, and that if it was displayed at 90 degrees it would spread the text longer and make it easier to read in larger fonts
04:38:25CalcipherI didn't just want to roll in like a jerk demanding features, I just thought it was a legitimate good suggestion, I really think it will look good with large fonts
04:38:38Mouser_Xalienbiker99: I've had something like that happen on my Gigabeat. For me though, I think it's just that I somehow touched the the pad lightly, but it was interpreted as a button press. In other words, it responds as if I pushed the button twice very quickly (which I might have, but didn't realize it.)
04:42:38ddaltonalienbiker99: what player?
04:43:00alienbiker99then it went away
04:43:19ddaltonso you pressed the left button
04:46:23 Join kugel [0] (i=kugel@unaffiliated/kugel)
04:46:25 Quit kugel (Client Quit)
04:46:27 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
04:46:58 Join SmokeDog [0] (
04:47:00 Nick SmokeDog is now known as Hukt0nWax (
04:47:08Hukt0nWaxi need some help ASAP
04:47:25 Quit RoC_MasterMind ("Leaving")
04:47:49Hukt0nWaxwhat happens if a friend loaded rockbox then took the files off of the ipod and gets an error saying can't load rockbox.ipod?
04:47:50LloreanGo ahead and simply ask, but please follow the guidelines.
04:48:08LloreanHukt0nWax: Just boot into disk mode, and follow the uninstallation instructions in the manual
04:48:12Hukt0nWaxand now the menu+select to reboot won't work
04:48:22LloreanMenu+Select will work. If it doesn't, you aren't doing it right.
04:48:50Hukt0nWaxhow can i be doing it wrong?
04:49:05LloreanI don't know, I can't see what you're doing or how steady you're holding your fingers
04:49:21LloreanBut holding Menu+Select works, and there's nothing in our code that can prevent it.
04:49:30Calcipherthat ain't easy to get if your not used to it
04:49:58 Quit Xerion_ (Read error: 104 (Connection reset by peer))
04:50:05Hukt0nWax<< sofa king we tod ed
04:50:16Hukt0nWaxi had to hold it for like 20 seconds
04:50:20Hukt0nWaxthanks :D
04:50:26Calcipherit was trial and error with my old nano, which I eventually smashed against a wall, first time I ever destroyed a piece of electronics out of rage but that goddamn thing deserved it
04:50:32 Join Xerion_ [0] (
04:50:36Hukt0nWax:D ncie
04:50:38Hukt0nWaxermmmm nice
04:51:13Calcipherit felt like some kind of just revenge...
04:51:14Mouser_XCalcipher: That would have been an awesome video...
04:52:04Calcipherhaha, I was genuinely infuriated, and I'm a pretty calm guy dealing with insane customers all day I'm used to things going south
04:52:30Calcipherbut that Ipod had every issue Ipods get all rolled in one, and a connectivity issue to boot
04:52:53CalcipherI was more worried about the wall
04:53:24Calcipherwait I think it was a mini
04:56:41CalcipherNow that I'm thoroughly enjoying my sansa e280 I will discard that little corpse
04:57:26Mouser_XIs the drive salvageable? It might be worth something (I don't know iPod types, so I have no idea if it would even have a drive...)
04:58:55CalcipherIt was a 4gb mini, which I wouldn't have bought, and DIDN'T
04:59:07Calciphermy friend gave it to me cause it was ackin' up
04:59:15 Join sin613 [0] (n=single@
04:59:32Calcipherit was the battery, so I went and got it replaced by Tekserv...buncha cocksuckers
04:59:52sin613is the framerate average in mpegplayer cumulative for the entire file being played back, or for the last x # of frames rendered?
04:59:59Calcipherthe battery still was garbage, and the player had connectivity issues
05:00:07krazykitplease take the offtopic talk to #rockbox-community
05:00:17 Quit billytwowilly (No route to host)
05:00:20Mouser_Xkrazykit: Beat me to it.
05:00:39Mouser_X(Sorry for instigating it.)
05:02:08sin613also, is there a logical explanation i'm missing as to why there are three boolean menu options in mpegplayer instead of two playback modes: benchmark and sync?
05:03:18 Quit karashata ("Leaving.")
05:03:49 Join karashata [0] (
05:04:27LloreanBecause having them as revealed bugs in both the frameskipping and the frame rate limiting code at different times
05:04:41 Quit Seed (Nick collision from services.)
05:04:47 Join Seedy [0] (
05:07:22 Join elinenbe [0] (
05:07:23sin613that last commit made a fairly massive difference on coldfire
05:07:48elinenbemarkun: how was the RockboxNYC meetup? Are you in NYC?
05:10:20 Join blargit [0] (
05:11:54***Saving seen data "./dancer.seen"
05:12:41 Join krazykit_ [0] (
05:13:04 Quit krazykit ("shutting down the desktop...stupid ext3 errors")
05:13:27 Nick krazykit_ is now known as krazykit (
05:15:15kkurbjunjhMikeS: are you around?
05:15:18 Quit bb (Nick collision from services.)
05:15:23 Join bb_ [0] (n=bb@unaffiliated/bb)
05:16:54 Join hannesd_ [0] (
05:18:34jhMikeSkkurbjun: yeah
05:19:27 Quit sin613 ("Leaving.")
05:19:35 Quit tictoc (Remote closed the connection)
05:19:36kkurbjunthe thread.c changes recently broke the mrobe build because of the definition of core_sleep.., the empty definition on line 609 does not match the prototype
05:20:08kkurbjunwhat are the consequences if core_sleep is left empty for a target?
05:20:28jhMikeSconstant furious spinning and so worse battery life
05:20:34 Quit hannesd (Read error: 110 (Connection timed out))
05:20:34 Nick hannesd_ is now known as hannesd (
05:20:48 Join Maha` [0] (i=tabac@gateway/gpg-tor/key-0xB9002659)
05:21:33jhMikeSWhy not implement it? It can't be that bad to do.
05:21:38kkurbjunok, but it won't break operation? Why are all these ifdefs in thread.c instead of being in their respective target trees?
05:21:58kkurbjunwell, I plan to, but I would rather spend my time getting the main build loading right now
05:22:06jhMikeSkkurbjun: it's just not been done yet. they should be though but inlining must be preserved.
05:22:18jhMikeSIt shouldn't break operation, no
05:23:03kkurbjunok, great, I was looking at the asm for the s3c.. and it looks like most of it would also apply to any other single core target correcT?
05:23:27jhMikeSI don't know about any other. It's based on the behavior of that particular chip.
05:24:31jhMikeSIf interrupts can stay disabled in cpsr and the core can still wake up, that's a simpler thing to do. s3c nor SH can do that so some timing tricks are used.
05:25:43jhMikeSboth CF and PP just leave them disabled in sr/cpsr and wake anyway
05:26:33kkurbjunhmm, what's waking the core?
05:26:46jhMikeSany interrupt on the core
05:27:00 Join stevenm [0] (n=stevenm@
05:27:41jhMikeSI should say "any unmasked interrupt" like the tick or DMA
05:27:42 Join n17ikh|Lappy [0] (
05:27:55 Join Seed [0] (
05:29:31kkurbjunah, I think I'm starting to see a bit how the s3c works.. what does setting the idle bit do?
05:29:38jhMikeSs3c sets the idle bit before reeabling the interrupts in cpsr _but_ it takes many cyces for the core to sleep so it gets past the msr instruction
05:30:01kkurbjunis that the idle on the SOC?
05:30:30jhMikeSyeah, just turns off the core clock
05:31:54kkurbjunhuh, I see, howd you find the timing on the idle bit? was that documented?
05:32:27Calciphernow I don't know how well you guys can pick apart the OF on the sansa players
05:32:29kkurbjunreally interesting too by the way
05:32:37jhMikeSthere's some pages about that in there.
05:33:12ddaltonwhat were your suggestions
05:33:39CalcipherI was thinking since so many people ask about the USB support, and the OF has this support is it possible rockbox can just read that part of the OF and implement the features?
05:34:01ddaltondon't know
05:34:07kkurbjunI definitely wouldn't have understood that if you hadn't explained it, so the PP chips can set an equivalent idle bit and will still restart the clock when an interrupt occurs even when they are masked?
05:34:08ddaltonwhat about accessibility?
05:34:11LloreanCalcipher: One of the biggest points of Rockbox is that it's not supposed to depend on or use the OF for things
05:34:13Calcipherwould that be harder than having it built right in to the softwa
05:34:23ddaltonread his message
05:34:44jhMikeSkkurbjun: the small delay loop was what was there and what the C-compiler produced ... I figured that was some aspect that shouldn't change.
05:34:59CalcipherAh I see
05:35:18Calcipherthen having it rely on foreign code would actually make it less stable
05:35:18LloreanCalcipher: Rockbox is intended as "replacement" firmware, meaning it should function even without the original firmware present at all when "complete"
05:35:23Calcipherlike in any situation
05:35:29kkurbjunjhMikeS: also do you know what, if any effect this has on battery life? Was this the same operation before your thread rework?
05:36:23jhMikeSkkurbjun: you can look at the diff in SVN. but it was: set the idle bit, loop, clear the idle bit after wake
05:36:29 Join teknofonix [0] (n=teknofon@unaffiliated/steve/x-00001)
05:36:32CalcipherI should have guessed as much since I was reading through all the old forum posts about double booting
05:37:37jhMikeSkkurbjun: It's really basically the same thing but with interrupts disabled so the waking list can be checked to abort the sleep. If that's not done, the CPU might wait 10000uS when a thread should be running _now_.
05:38:43teknofonixthe build for video ipod 30gb doesn't uncompress on my macbook
05:39:36kkurbjunI see, great stuff, thanks for taking the time to explain it, I'll be watching out for that once I get this build working :)
05:39:56Lloreanteknofonix: Try using a different tool to extract it then
05:41:07jhMikeSkkurbjun: good luck on that
05:42:35*jhMikeS goes back to writing a better mutual wake algorithm for PP
05:43:23 Quit Seedy (Read error: 110 (Connection timed out))
05:47:07 Join headphase [0] (
05:48:44stevenmDang, those pitch tables can get pretty big...
05:49:20stevenmOne day, the midi plugin will be reduced to 5000 tables and a single for loop.
05:50:38 Quit ddalton ("leaving")
06:06:34Calcipheranyone here familiar with tom tom linux?
06:07:46CalcipherI just want to figure out if the text to speech voices set up I have for a tom tom 510 can be of any use to you guys
06:08:04CalcipherI see an Install.bif file in one of its directories
06:08:23LloreanRockbox isn't Linux.
06:08:59Calcipheralso a
06:09:29Calcipheryeah I realize that, but it seems most of the dev work is being done in linux environment
06:10:10Calcipherwanted to see if any of those files sound familiar to the linux users
06:10:13scorcheCalcipher: what he was getting at, is this channel is about rockbox only
06:11:01Calcipherok, the reason I brought that up was to possibly help with the voice menu features development for rockbox
06:11:16LloreanEither you're talking about creating the pre-created voice files, in which case any TTS engine that can take a test string and output a wav can be adapted for use with Rockbox's speech generation routines, or you're talking about incorporating TTS directly into Rockbox, which requires a GPL licensed engine that can be converted to fixed point.
06:11:20Calcipherbut yeah I see your point on avoiding off topic stuff
06:11:24LloreanWhich requires not binaries, but source code.
06:12:45Calcipherfunny enough I think I saw a link for the sources, but I was refering more to using an existing external tts engine
06:13:18LloreanThat couldn't work, except as previously mentioned, using the pre-recorded clips as Rockbox currently does.
06:15:43Calcipheryeah, I just wanted to see if anyone doing voice clip related work was interested in using the engine I have here from the linux tom tom install
06:17:38scorchetypically, that is handy information to know up front ;)
06:18:05 Join [1]psycho_maniac [0] (
06:18:36 Quit psycho_maniac (Read error: 104 (Connection reset by peer))
06:18:38 Nick [1]psycho_maniac is now known as psycho_maniac (
06:18:59JdGordongpl violation settelemtns are crap... they got of scot free other than releasing their corse
06:19:25Calcipheryeah looks like it
06:20:06Calcipherthey donated something here and there, but that company I'm sure is doing great in sales mostly in Europe I notice, but even in the U.S.
06:20:35 Quit lazka (Remote closed the connection)
06:21:20Calcipherso if any linux user wants to give this tts package a look see and figure out if they can use it, feel free to contact me
06:21:56Calcipherand this is for rockbox development test purposes
06:22:27CalcipherI really appreciate all the work being put in to this system
06:24:11 Quit teknofonix (Client Quit)
06:29:17 Quit atsea-34 (Remote closed the connection)
06:43:58 Join atsea-34 [0] (i=atsea-@gateway/tor/x-715d375fcb22f399)
06:44:22Calciphernight folks
06:49:07 Part kkurbjun
06:49:13 Join tdb_ [0] (
06:50:13 Quit headphase (Read error: 110 (Connection timed out))
06:50:56 Quit blargit (Remote closed the connection)
06:51:08 Quit Mouser_X (Nick collision from services.)
06:51:25 Join Mouser_X [0] (
06:55:58 Join blargit [0] (
06:57:26 Quit blargit (Client Quit)
07:01:24 Join blargit [0] (
07:10:06 Quit tdb_ ("Leaving")
07:11:08 Join donutman25 [0] (n=chatzill@
07:11:27stevenmyay, FINALLY patches are scaled as they should be
07:11:58***Saving seen data "./dancer.seen"
07:14:39psycho_maniacwhat do you mean?
07:15:02stevenmjust a midi feature thing that's been eating at me for a year or so
07:15:37 Part toffe82
07:15:38stevenmone time I had it working.. but it made the load time insanely long due to extra computations
07:15:55stevenmnow it hits me.. there's a much faster way
07:15:58stevenmand it sounds like it's working
07:17:51 Quit aliask ("ChatZilla [Firefox]")
07:20:28 Join kkurbjun [0] (
07:20:34kkurbjunJdGordon: are you around?
07:21:25kkurbjunyou can boot the main build by copying apps/rockbox.bin to the root of the player as rockbox.mrboot with the mrobload bootloader and it will startup
07:21:44kkurbjuncan't do anything though since the keymap doesn't do anything right now
07:22:07kkurbjunhave to run, I'll talk later
07:22:11 Quit kkurbjun (Client Quit)
07:23:34 Join toffe82 [0] (
07:31:48 Join ender` [0] (
07:31:51 Quit stevenm ("Connection reset by beer")
07:35:41 Part toffe82
07:41:36 Quit jhulst ("Konversation terminated!")
07:42:36 Join rvvs89_ [0] (n=rvvs89@unaffiliated/rvvs89)
07:44:12 Quit rvvs89 (Nick collision from services.)
07:44:33 Quit ATravelingGeek (Read error: 110 (Connection timed out))
07:44:37 Nick rvvs89_ is now known as rvvs89 (n=rvvs89@unaffiliated/rvvs89)
07:45:34 Join ATravelingGeek [0] (n=ATG@pdpc/supporter/student/ATravelingGeek)
08:01:37 Join barrywardell [0] (
08:02:05 Quit pabs (Nick collision from services.)
08:02:27 Quit ATravelingGeek (Connection timed out)
08:02:55 Join ATravelingGeek [0] (n=ATG@pdpc/supporter/student/ATravelingGeek)
08:10:51 Join ramon8 [0] (
08:11:31 Join Rob222241 [0] (
08:15:06 Quit Maha` (Remote closed the connection)
08:15:31 Nick Tanuva|Zzz is now known as Tanuva (n=tanuva@
08:15:35 Join jumpatrain [0] (i=tabac@gateway/gpg-tor/key-0xB9002659)
08:15:49 Quit jhMikeS (Nick collision from services.)
08:15:55 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
08:21:18 Quit jumpatrain (Broken pipe)
08:21:47 Quit Llorean ("Leaving.")
08:23:30 Quit psycho_maniac (" HydraIRC -> <- The professional IRC Client :D")
08:28:21 Quit Rob2222 (Read error: 110 (Connection timed out))
08:29:18 Part safetydan
08:39:46 Join Maha` [0] (i=tabac@gateway/gpg-tor/key-0xB9002659)
08:46:32 Quit barrywardell ()
08:56:22 Quit BigBambi__ (Read error: 113 (No route to host))
09:07:46 Join Zagor [0] (n=bjorn@rockbox/developer/Zagor)
09:12:02***Saving seen data "./dancer.seen"
09:14:31 Join advcomp2019_ [0] (n=advcomp2@
09:23:55 Quit ToHellWithGA (Read error: 113 (No route to host))
09:24:22 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
09:31:00 Quit advcomp2019 (Read error: 110 (Connection timed out))
09:33:25 Join Llorean [0] (
09:39:22 Quit shnee ("Konversation terminated!")
09:42:47 Join advcomp2019 [0] (n=advcomp2@
09:51:59 Join pondlife [0] (n=Steve@rockbox/developer/pondlife)
09:53:34pondlifeamiconn: Good Morning, do you think Karim Boucher should be added to CREDITS for FS #5995?
09:54:18amiconnHmm, I'm undecided.
09:54:32amiconnNone of his code went in directly, I basically rewrote it.
09:54:45amiconnBut I used his implementation as a reference
09:55:06pondlifeNo problem.
09:56:01LloreanWell, "Credits" isn't, strictly speaking, the same as "List of Copyright Holders" so he may as well get credit for doing the initial work, right?
09:56:27amiconnYeah, we should probably add him.
09:57:29 Quit Mouser_X (Read error: 113 (No route to host))
09:58:19 Quit advcomp2019_ (Read error: 110 (Connection timed out))
09:58:22 Join homielowe [0] (
10:04:48GodEater_I don't even remember when I was added to Credits, or why...
10:04:56 Join safetydan [0] (n=safetyda@rockbox/developer/safetydan)
10:06:09safetydanwow, 2fps faster mpeg playback on h120
10:06:23safetydanthat's almost watchable
10:09:28amiconnNah, H1x0 is quite limited in its fps due to all the bit shuffling in the graylib
10:10:22amiconnOn X5, video playback is now smooth if there's not too much "action" in the video.
10:11:03Lloreanamiconn: What sort of framerate then?
10:11:30amiconnWith fullscreen it's around realtime on average
10:11:43LloreanWow, nice.
10:12:02LloreanDid the OF have video on the X5, or only the H300 had that?
10:12:05amiconnBut that's an average rate; we'd need more for completely smooth playback
10:12:27 Join spiorf [0] (
10:12:28amiconnAfaik both have video, but I never tried it on any OF
10:12:47amiconnIirc, OF does 10fps on H300 and 15fps on X5
10:12:53LloreanI was about to say the same thing
10:13:07pixelmaI remember seeing .avi files an a "video" directory on a stock X5 but never seen it running
10:13:21LloreanOf course, those were divx files, iirc.
10:13:29GodEater_I was about to ask what codec ;)
10:13:40LloreanIt's been a little hard to claim mpeg2 was "good for performance" when our player was still as slow or slower than the OF one
10:14:29amiconnWell, it's not anymore
10:14:57amiconnBut H300 still needs work to get it near realtime
10:16:36pixelmaGodEater: I just saw them listed in the browser.... :)
10:16:59amiconnAround 14fps fullscreen on H300 now
10:18:07LloreanSo it's almost up to "as fast as the OF on the Sansa", heh.
10:19:11 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
10:19:19amiconnI think there is more potential on coldfire. Someone needs to find the right spots to use iram for, and perhaps some more asm is needed
10:19:46 Quit daurn ("Cyas later...")
10:20:01 Nick bb_ is now known as bb (n=bb@unaffiliated/bb)
10:23:25safetydanooh, widescreen is nearly 14fps
10:25:59 Join obo [0] (n=obo@rockbox/developer/obo)
10:27:27 Join daurnimator [0] (i=daurn@unaffiliated/daurnimator)
10:35:12 Part pixelma
10:37:52 Quit daurnimator (Read error: 104 (Connection reset by peer))
10:47:05 Join kubiix [0] (
10:50:29GodEater_any comment on the compatibility (or lack thereof) of libpng's license with the GPL ?
10:53:32 Join barrywardell [0] (
10:54:12 Join davina [0] (
10:54:43ZagorGodEater: "This is a free software license, and compatible with the GPL." −− FSF
10:54:56GodEater_ah - must have missed that :)
10:54:58GodEater_thank you!
10:55:11Zagorwell I had to search a bit too :)
10:55:54 Quit pondlife (Read error: 104 (Connection reset by peer))
10:56:09 Join pondlife [0] (n=Steve@rockbox/developer/pondlife)
10:57:07LloreanI wonder how many licenses are compatible with GPLv3 yet.
10:57:33 Quit davina (Read error: 104 (Connection reset by peer))
11:00:52 Quit spiorf (Remote closed the connection)
11:01:19amiconnjhMikeS: Does the asm motion compensation have a significant effect on arm?
11:01:38amiconnI mean, it's only a subset of all the mc functions. Are those the ones used most?
11:02:32 Quit ramon8 (Read error: 110 (Connection timed out))
11:03:51jhMikeSamiconn: it seemed to be quite important. just asm change made quite a difference on gigabeat
11:04:24 Join davina [0] (
11:04:37amiconnDo you have a speedup percentage figure?
11:05:11GodEater_eww - libpng is full of floating point stuff
11:05:12jhMikeSIt was about 8fps or so I think. I don't recall the exact framerate at the time.
11:05:13amiconnI presume this works on frame data, i.e. data in dram, not iram?
11:05:46jhMikeSit moves the macroblocks around
11:06:07 Quit davina (Remote closed the connection)
11:06:23amiconnBtw, what is the first decoding stage in mpeg2? Is it huffman, like jpeg?
11:06:56jhMikeSI believe so. Not 100% sure.
11:09:29 Join davina [0] (
11:10:24 Join aliask [0] (
11:11:30jhMikeSidct_arm.S has a "$Id $" in it :)
11:12:06***Saving seen data "./dancer.seen"
11:12:10 Quit ashes (Remote closed the connection)
11:15:54 Join ashes [0] (n=ashes@2001:5c0:8fff:ffff:0:0:0:1a5)
11:16:42 Nick fxb__ is now known as fxb (
11:17:46 Quit pondlife (Read error: 110 (Connection timed out))
11:22:24 Quit idnar (Nick collision from services.)
11:22:26 Join idnar_ [0] (i=mithrand@unaffiliated/idnar)
11:32:20 Quit safetydan ("Leaving")
11:36:48 Nick Tanuva is now known as Tanuva|off (n=tanuva@
11:36:51 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
11:46:20 Join Philus [0] (i=9e274fdd@gateway/web/cgi-irc/
11:48:51 Join spiorf [0] (
11:49:36 Nick DiDjCodt is now known as djc_-_-_-_-_-_-_ (
11:50:55 Nick djc_-_-_-_-_-_-_ is now known as djc___-___-___-_ (
11:51:02 Nick djc___-___-___-_ is now known as DiDjCodt (
11:52:14PhilusWouldn't it be nice if you could toggle the backlight on and off by the flick of a switch? Like, if you toggle the keylock on and off within, say, 2 seconds it disables all backlight, and if you do it again it's back on..
11:53:18ZagorPhilus: have you seen the "backlight on hold" setting?
11:54:07Philusyes, isn't that just to disable all light while the keylock is on?
11:54:13Zagoryou can choose
11:54:21Zagori'm off to lunch
11:56:28LloreanZagor: I think he meant "hold on/off quickly turns the backlight on, permanently, until a second toggle of hold on/off"
11:57:05LloreanPhilus: A simple thing you can do is have two .cfg files in your root, one that permanently enables backlight, and one that sets it to a very short timeout so you can still access the menus but otherwise have it off
11:57:23PhilusI meant toggle the keylock twice to disable all backlight, and toggle twice again to go back to normal mode
11:59:16PhilusLlorean: that still means if I need to enable the backlight in total darkness (when we're in bed and the wife beats me if I turn the light on) I'd have to fiddle around in the menu to change settings
11:59:17 Join ramon8 [0] (
11:59:25amiconnjhMikeS: Weird - I *did* set svn:keywords ...
12:00:16rasherPhilus: I'm not sure it's Rockbox' duty to protect you against your wife. Why do you want to *completely* turn off backlight?
12:00:44Philusrasher: if ambient light is plentyful, I don't need it..
12:00:53Philuslike, daylight
12:01:06LloreanPhilus: Why not set the "short timeout" cfg before you turn the lights out, then?
12:02:25PhilusLlorean: so the lights go out faster?
12:02:37Philusi mean, the backlight.. err
12:02:46 Join Thundercloud [0] (
12:03:26LloreanPhilus: I just meant, why does the method have to involve "a switch or button combination"?
12:03:47 Join bluebrother [0] (i=9cjWVY2e@rockbox/staff/bluebrother)
12:03:58LloreanThe only situation where it's particularly problematic is if you forget to turn the backlight on before entering a dark area.
12:04:22LloreanAnd the solution to that is a zzzzzzzz.cfg in your root that simply enables the backlight again. :)
12:05:10 Nick fxb is now known as fxb__ (
12:08:37jhMikeSamiconn: is it just because it's "$Id $" and not "$Id$" ?
12:08:58amiconnEither that, or the order of keywords in svn:keywords matters
12:09:07amiconnI just changed both, now it works
12:09:15 Join Crash91 [0] (n=evil91@
12:09:32 Part karashata
12:10:43PhilusLlorean: more often than not, I don't need the backlight at all.. but when I do, fiddling with the menu would be more of a hassle than toggling the keylock twice. And it should ofcourse be optional
12:11:01 Quit donutman25 ("ChatZilla [Firefox]")
12:12:43 Quit homielowe (Read error: 110 (Connection timed out))
12:13:26LloreanPhilus: I never said "fiddle with the menu"
12:13:53PhilusLlorean: selecting a different .cfg means you have to go into the menu, no?
12:14:03LloreanYou have to go into the filetree.
12:14:05amiconnjhMikeS: Regarding motion compensation - is the destination always 32 bit aligned?
12:14:26LloreanBut if it's the last entry in the root, it's very very easy to navigate to blind without having to memorize lists or anything
12:15:09PhilusLlorean: I'll do that while I figure out how to implement it on the hold button, thanks for the advice :)
12:16:09jhMikeSamiconn: I don't know. I haven't gone into that very much.
12:16:25amiconnDidn't you write the arm code?
12:16:36jhMikeSIt should just be moving a block around each plane.
12:16:44jhMikeSI didn't write the motion comp code
12:17:40linuxstbamiconn: I think markun committed it, but he didn't write it either. I'm not sure where he got it from...
12:18:12 Quit Rob222241 (Read error: 104 (Connection reset by peer))
12:18:34 Join Rob2222 [0] (
12:19:17amiconnmplayer-w100 project, whatever that is...
12:20:15 Quit Crash91 ("Bye Bye!")
12:21:54linuxstbamiconn: Have you tried disabling the calls to lcd_yuv_blit() in mpegplayer to see what speed the decoder is running at? I tried that once for the ipod video, and IIRC it could only decode at about 22fps (fullscreen I think).
12:22:29amiconnNo, but it's a simple calculation now that I know the pure yuv framerates
12:23:11amiconnIt's like parallel resistors: 1/total_frame_rate = 1/decoder_frame_rate + 1/yuv_frame_rate
12:24:30 Quit Philus ("CGI:IRC")
12:24:32amiconnFor H300 fullscreen, that equation yields ~17fps for the decoder
12:25:16amiconnHence my desire to add some coldfire asm motion_comp code...
12:27:55 Join Arathis [0] (
12:31:57 Quit advcomp2019 (Read error: 110 (Connection timed out))
12:32:09 Quit spiorf (Remote closed the connection)
12:36:57markunelinenbe: I'm back from the US
12:37:52 Join karashata [0] (
12:39:45 Join qweru [0] (
12:41:20*amiconn thinks he now understands how the motion compensation code works :)
12:43:03 Quit J3TC- (Read error: 113 (No route to host))
12:47:46 Quit barrywardell ()
12:49:12amiconnI was puzzled about the _o, _x, _y and _xy variants; thought those have to do with the offset. But each x and y mean an additional movement by half a pixel
12:50:12amiconnAnd the destination is obviously guaranteed to be 32 bit aligned, otherwise the arm code would throw a data abort on PP
12:52:40preglowmotion compensation is bound to suck huge amounts of cpu on coldfire, i think
12:52:50preglowno data cache :/
12:53:18amiconnYeah, hence I want to put asm
12:53:41amiconnAsm can fetch larger amounts at once, making efficient use of burst accesses
12:53:55amiconnThe C code fetches and stores single bytes...
12:56:50amiconnUmm, that reminds me of something
12:57:14amiconnlinuxstb: How did you manage to seize the core's codec stack in test_codec?
12:57:57preglowsingle bytes? :PP
12:58:47preglowthat's a good point, can't the video codec use the codec stack?
12:59:04preglowshould work for now, at least
12:59:11amiconnEven better:
12:59:30amiconnRight now libmad is given a separate stack in mpeg player, put into iram
12:59:38amiconnAnd libmad is quite stack hungrry.
13:00:09amiconnIf we could seize the core's codec stack for libmad in mpegplayer, we could give the video thread a stack in iram as well
13:01:02amiconnOh, didn't notice that the video stack is already in iram
13:01:06preglowlibmad is very stack hungry, yeah, i think it puts entire frames in iram
13:01:12preglowan entire frame, at least
13:01:28amiconnBut still, seizing the core's codec stack for libmad would free quite some iram for other purposes
13:03:01preglowi don't really know what i think about reusing stacks either, that won't work when/if we go preemptive
13:03:27preglowbut could be a nice trick for now
13:05:34amiconnbah, preemptive :(
13:06:03 Nick Maha` is now known as tictoc (i=tabac@gateway/gpg-tor/key-0xB9002659)
13:06:03preglowthe way i see it, the code is becoming too complex to continue with the cooperative stuff
13:06:07preglowalso, it hides bugs
13:06:48preglowand keeps us from implementing features like proper realtime dsp (which i of course would like...)
13:10:12amiconnWhy is that?
13:10:20preglowwhich of them?
13:11:22amiconnI can't see why cooperative threading hides bugs, and how preemption would allow for realtime dsp
13:11:38amiconnI think the can-of-worms to benefit ratio of preemptive threading is rather bad
13:11:51preglowit hides bugs because some erronous thread communication might not malfunction until some yields() are added or shufflDed around
13:11:55preglowwhy is it a can of worms?
13:11:59amiconnRight now rockbox code is written for execution in a cooperative threading environment
13:12:08preglowyes, and it shouldn't
13:12:09***Saving seen data "./dancer.seen"
13:12:18preglowwriting that kind of code breaks when the yield() pattern isn't as you expect
13:12:33amiconnImo it should, because it's less complex
13:12:49preglowyou should keep thread communication to a minimum anyway, and do it properly when you do it
13:13:09amiconnPreemptive threading would need a ton of additional checks where you can now simply assume that nothing changes inbetween (because you're not yielding)
13:13:18preglowi think the preemptive scenario is less confusing than the cooperative one, at least with preemptive multitaskin, i know i can't trust _anything_ to work in the order i think
13:13:32 Join ddalton [0] (
13:13:36preglowwith cooperative, i don't really know, and worst case, it could be essentially preemptive, because some guy might come and shuffle yield()s around
13:13:56ddaltondoes anyone know what variable holds the value of the select item in the file tree?
13:14:12ddaltonSo if I am on item 5 from the top of the list what variale holds 5
13:14:36preglowamiconn: and it would allow us to do realtime dsp since we can assure that dsp gets run realtime
13:14:46preglowamiconn: we can't now, since we depend on people yielding properly
13:15:06amiconnYou would still have the boosting problem, that keeps us from going low-latency all the time atm
13:15:27preglowhmm, yeah, the boosting might give us an interesting problem
13:16:03amiconnThe most obvious example for preemtion badness are the iriver and iaudio remote lcd update functions (to me, because I wrote them)
13:16:06 Quit karashata ("Leaving.")
13:16:18preglowi wonder how a scheme with more clock freqs would work. trying to always choose one that gives reasonable load, more like powernowd on linux
13:16:24amiconnThese functions need to know the current cpu clock, because they have to adjust timing accordingly
13:17:11amiconnRight now, they can rely on that the clock *never* changes halfway through a transfer, because they don't yield, and calling them from interrupt context is forbidden
13:17:29amiconnBut with preemption, another thread could boost or unboost during a transfer...
13:17:31preglowyou would of course have means of making sure you're not interrupted
13:17:36preglowlike critical sectionms
13:17:42preglowwhich are dead easy to use
13:18:28preglowso that nothing comes and interrupts the ata driver during a transfer, for example
13:18:41amiconnUnboost would just slow down the transfer, but boost would speed it up - and the lcd contents gets garbled
13:18:58preglowevery os has tons of spots like that
13:19:28preglowas a matter of fact, they're the primary reason for linux' bad low latency behaviour...
13:19:41amiconnThat's not so much because the lcd controller couldn't follow, but due to the wire length. We simply can't push 10MHz over that thin wire to the remote, and expect it to work reliably
13:20:32preglowdual core support pretty much means we _have_ preemptive multitasking
13:20:45preglowwe can control that by selecting which threads go on which cores, of course
13:21:05preglowbut the problem might still come up and bite us
13:21:27preglowsimply because rockbox is getting big, and few people know every part of it well enough to know in what subtle ways threads might be talking to each other
13:22:08amiconnBloat is our problem, not the cooperative threading, imo
13:22:26preglowboth are problems, if you ask me
13:22:46preglowpreemptive threading won't even add much overhead in size, i'm quite sure
13:22:52preglowmost of the needed stuff is already in
13:22:58preglowthanks to dual core
13:23:05 Join mf0102 [0] (n=michi@
13:23:19amiconnDon't remind me of that :(
13:23:39preglowand if you're going to tell me we should just forget we have an extra core, you're just being silly
13:23:41amiconnI know that we need to utilize it, but I dislike it, completely
13:23:54preglowsure, you can dislike it all you want, it still exists
13:24:07preglowi don't like it myself, simply because i'm not very good at multi-threaded programming
13:24:14preglowbut i intend to learn that, heh
13:24:58amiconnI think that bugs in the threading would be even harder to find with preemptive threading than with cooperative therading
13:25:33*amiconn would really like to stay away from preemptive thread programming
13:26:25amiconnOne reason why I like rockbox coding, and why I don't code for linux or other os'es
13:26:41amiconn(other than simple things where I don't need to worry about multiple threads)
13:27:15preglowi think you'll find you don't really need to worry too much about it unless you write code that works with several threads
13:27:35preglowand why would the bugs be harder to find? i think they'll be simpler to find, since they appear consistently
13:27:43preglowwhich is not the case at all now
13:27:46preglowlook at playback.c
13:27:52preglowobscure bugs which come and go
13:28:27amiconnThread timing would be variable instead of fixed...
13:28:37preglowand execution order more random
13:28:54preglowthat'll contribute more
13:29:02amiconnHence bug which show up consistently in the cooperative model would be random with preemptive threading
13:29:23 Join karashata [0] (
13:29:27preglowwell, yeah, but then you're talking the ordinary kind of bug, not the threaded kind of bug
13:29:35preglowand the ordinary bug is still going to be simple to find
13:29:50preglowsince it'll always happen when the thread is run, and it'll be run sooner or later
13:30:30 Join advcomp2019 [0] (n=advcomp2@
13:30:33 Join spiorf [0] (
13:31:58 Quit tictoc (Remote closed the connection)
13:34:25 Join atsea-99 [0] (i=atsea-@gateway/tor/x-742eb7bf04210ed8)
13:37:56 Quit atsea-34 (Read error: 104 (Connection reset by peer))
13:38:50 Join jumpatrain [0] (i=tabac@gateway/gpg-tor/key-0xB9002659)
13:39:14 Quit ddalton ("leaving")
13:40:42 Join teknofonix [0] (n=teknofon@unaffiliated/steve/x-00001)
13:47:43 Part teknofonix
14:03:54 Join pondlife [0] (n=Steve@rockbox/developer/pondlife)
14:06:02 Part karashata
14:11:50 Join Crash91 [0] (n=evil91@
14:11:59 Join karashata [0] (
14:13:53*Crash91 hears the possibility of USB on sansa
14:14:03Crash91is it true?
14:14:17pondlifeZagor's working on it
14:14:39Crash91so theres been a lot of progress?
14:15:16pondlifeQuite a bit I think; you'll need to ask him though.
14:16:03Crash91\O/ yay!
14:16:21Crash91at least someone is now working on it =)
14:16:26ZagorCrash91: things are going forward nicely. basic usb comms is ready, working on the usb-storage protocol now.
14:17:18LloreanZagor: I won't pretend to know how these things work. Will software USB basically give direct access to the disk to the host PC, or will it basically have enough of Rockbox standing between it and the host to, say, hide folders (or show virtual folders, for example, MicroSD cards)?
14:17:21Zagoror, rather, the usb-storage protocol has revealed weaknesses in the basic comms that I'm working on
14:17:36Crash91Zagor: Yay! Thanks a lot for working on it =)
14:17:59ZagorLlorean: usb is on block level, so no hiding
14:18:38markunalso s/usb/ums/ ?
14:18:41pondlifeSo no transfer PC -> SD card?
14:18:42Zagorit also means we can't do anything with the disk during ums mode, or risk severe corruption
14:19:02Zagorpondlife: oh yes that will be possible, although not in the first iteration
14:19:13Zagormarkun: correct
14:20:18Zagorbasically ums is scsi-over-usb
14:20:28pondlifeAnyone mind if I commit I guess a larger tagnavi.config reduces audio buffer a bit?
14:21:21 Quit Nico_P (Remote closed the connection)
14:21:23markunZagor: the Gigabeat dock has a 4 port USB hub which connects to the usb device, host and ata-bridge of the gigabeat. No idea yet how we should select what to use.
14:21:27Lloreanpondlife: It should only affect people who've enabled Database anyway. I think we're still working toward a "perfect" tagnavi, but since I don't use it at all I don't know if that's a good addition. Sounds like it though
14:21:49 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
14:21:55 Join roolku [0] (
14:22:04preglowZagor: is there a mode in which rockbox could do the block level accesses?
14:22:14preglowor does the usb protocol only do things block level?
14:22:14Zagorpreglow: how do you mean?
14:22:54Zagorums (scsi) asks for "n blocks, starting at x"
14:22:57preglowZagor: something like the usb protocol working more like ftp than direct block access, for example
14:23:09linuxstbpreglow: I guess that's MTP
14:23:15preglowlinuxstb: yeah, makes sense
14:23:33preglowZagor: so to expose the sd card, you'll have to export it a separate device, i take it?
14:23:44LloreanBut if you could present to the host as "standard UMS", and put some sort of abstraction in, you could probably do some interesting things.
14:23:47Zagoryes. that's how the sansa OF does it too
14:23:50roolkupondlife: I would personally scrap the special "The" treatment, and the bugfix is nice to have
14:23:58preglowZagor: yeah, and i think most people will like it that way anyway
14:24:12linuxstbLlorean: What kind of abstraction. All UMS does is allow the host to read/write raw sectors
14:24:17Lloreanpondlife: I definitely think special treatment for "The" should be scrapped, but I'm in favour of sort tags which would negate a *need* for special treatment.
14:24:32*markun seconds that
14:24:33 Quit Crash91 ("Bye Bye!")
14:24:33pondliferoolku: OK, well I've committed with the "The"s for now, someone else can happily kill the "The"s.
14:24:35*preglow too
14:25:52Lloreanlinuxstb: I guess in that situation there's not too much you can do, can you. =/
14:25:55GodEater_linuxstb: but our USB code could lie about the sectors available
14:26:03markunpondlife: is there anyone interested in all the artists starting with "The" ?
14:26:07linuxstbGodEater_: Yes, like the E200R firmware does...
14:26:18GodEater_Llorean: why do you want this behaviour ?
14:26:34LloreanGodEater_: I don't have anything specific in mind, yet. Just curious about hypotheticals at the moment.
14:27:08*GodEater_ would be against any sort of weird data hiding via our usb code unless there was a *very* good reason for it
14:27:09pondlifemarkun: Not me, but it was in there, so I preserved it
14:27:46pondlifePersonally, I'm definitely for scrapping all of the "The" tests.
14:27:48Zagorwe can do anything we like with usb, just not with ums.
14:28:04Zagorums is the plain disk access protocol, and we should keep it that
14:28:05LloreanGodEater_: I was more wondering about MTP-like features, such as Rockbox scanning metadata for the database during transfers, or tracking dircache changes, or I dunno what really. But that doesn't seem so much in the picture.
14:28:43LloreanGodEater_: Basically, I was hoping that we could, without changing the apparent way UMS transfer works to the user, have Rockbox handle some housework while it's transferring.
14:28:45linuxstbZagor: Couldn't we implement extra "SCSI inquiry" commands - to for example provide rbutil with information about the running version of Rockbox, and what bootloader is running in flash?
14:29:11*linuxstb now wishes all our targets had software USB...
14:29:21*GodEater_ notes Zagor's todo list suddenly getting much longer :)
14:29:47Zagorlinuxstb: the better model for that I think is to add device-specific control messages. that is very easy to do.
14:29:49Calcipherooh, this is getting interesting heh
14:30:12 Join Crash91 [0] (n=evil91@
14:30:28linuxstbZagor: What a proprietory command to sync the clock? Would that fit in somewhere?
14:30:33Zagorthough I don't know enough about the usb api rbutil uses to say how easy that is on that side
14:30:36Zagorlinuxstb: sure
14:30:41CalcipherI'm gonna be real curious to see what you guys have done in the next day, good luck all of you, sounds great already
14:31:02LloreanCalcipher: Don't be impatient. It may not even be usable within a day, or several.
14:31:06markunlinuxstb: I've also googled for that a bit (but didn't find anything so far)
14:31:12linuxstbZagor: Sorry for distracting you, carry on with the basics ;)
14:31:35linuxstbmarkun: Yes, I don't think it's standard - I'm suggesting some form of proprietory extension.
14:32:06linuxstbItunes manages to sync the time on ipods via UMS for example, and I'm assuming they've just done something proprietory on top of UMS.
14:32:06 Quit karashata ("Leaving.")
14:33:25Zagoryeah the ums spec has openings for device-specific bulk commands too. but I think it's better to keep those commands on the device level (control) rather than linked to the storage protocol.
14:34:08GodEater_so basically you're saying you'd implement them outside of UMS itself ?
14:34:09Zagor("those": commands not immediately related to the file transfer)
14:34:25ZagorGodEater: time sync and such, yes
14:34:46*GodEater_ only has very rudimentary understanding of how usb works
14:35:13preglowlinuxstb: any idea why skip prev wouldn't work with wma?
14:35:15Zagordatabase-update-while-transferring requires a non-ums protocol, so that's not really an immediate issue imho
14:35:31preglownot at all
14:35:40GodEater_but I assume you basically connect the device, and your usb stack says "Hi, who are you, what do you do?" in the device control bit, and then when the device says "Hi, I'm an ipod, I do UMS", your launch your UMS sub code ?
14:35:54linuxstbpreglow: No...
14:36:16Lloreanpreglow, linuxstb: Isn't the problem because the first time you press it, it seeks to the beginning of the file?
14:36:26LloreanI seem to remember hearing that if you skip next, then immediately skip prev, it works
14:36:28preglowoh, right, that kind of prev
14:36:33preglowi thought he meant double click prev
14:36:34ZagorGodEater: no, an usb device can do many different things. it says "hi, I'm an usb device. talk to me on this control port. then I have these two storage ports for you. and these serial ports. and these...." etc
14:36:34preglowfor some reason
14:36:46LloreanThe second click won't work, because it never got to the beginning of the file on the first one, right?
14:36:53GodEater_Zagor: I was over simplifying :)
14:37:06preglowright, stock wma really, really ignores seeking
14:37:13ZagorGodEater: yeah but the point is an usb device isn't either ums or something else. it can be many things at once.
14:37:33linuxstbZagor: Ah, so for example we could have serial and UMS working simultaneously? (I can't think why though...)
14:37:34GodEater_I realise that sorry - I just picked UMS as one thing to say to avoid my line being enormous
14:37:36*preglow sees if there's a quick fix
14:37:39Lloreanpreglow: I'm kinda surprised that the first press actually uses seeking to get to 0, honestly.
14:37:57preglowLlorean: well, it's efficient
14:38:01GodEater_Zagor: for example, I don't imagine the USB protocol makes much use of the word "Hi" in reality ;)
14:38:03linuxstbpreglow: I think other codecs without seeking implement a special case to seek to the start of the file.
14:38:05preglowLlorean: no need to reinit the codec and all
14:38:08Zagorlinuxstb: yes we can. and if I don't fix my storage problems soon, I'm tempted to add serial capability for debugging...
14:38:14preglowlinuxstb: they do, i've removed several of them :)
14:38:25Lloreanpreglow: I'd just somehow assumed there'd be a "more efficient" special case or something already in place. That's all. :)
14:38:27linuxstbpreglow: ;) Maybe you need to add one for WMA...
14:38:35preglowi think i will
14:38:52preglowif i find a decent way
14:39:05*preglow wonders if saratoga has worked more on seeking
14:39:43linuxstbpreglow: Last time I spoke to him, he said it didn't have any time to work on it any more, but agreed to post what he had to flyspray. I don't think he ever did though...
14:39:58preglowi have some of it
14:40:04preglownot the newest, i think
14:40:08preglowworks semi-well but has bugs
14:41:38preglowif this works, it'll be a huge hack
14:42:21markunpreglow: can you post his code somewhere then?
14:42:36markunI could really use wma seeking
14:43:07 Join karashata [0] (
14:43:15 Join pixelma [0] (i=d5494db8@rockbox/staff/pixelma)
14:43:31 Quit ramon8 (Read error: 110 (Connection timed out))
14:44:35*preglow wants a make install for targets
14:46:08preglowlinuxstb: ok, i just do a goto next_track; if seek_time == 1 and it seems to work...
14:46:28preglowlinuxstb: would such a hack of the aeons be commitable?
14:46:29linuxstbpreglow: I just have some little scripts which do make && make zip && mount && unzip && umount
14:46:44preglowlinuxstb: problem is that make zip is dog slow
14:49:09preglowa simple cp should be heaps faster
14:49:22preglowlinuxstb: but yeah, should i commit that quick hack?
14:49:58linuxstbYou could script that - e.g. modify buildzip to build the files directly in a new location, rather than building a zip
14:50:11linuxstbpreglow: The "skip to 0" hack?
14:50:22linuxstbYes, I think it's useful.
14:50:47 Join nicktastic [0] (n=nick@unaffiliated/nicktastic)
14:50:48 Quit Crash91 ("Bye Bye!")
14:51:36linuxstbpreglow: In fact, it shouldn't be too hard to modify - it already builds a temporary .rockbox folder in your build directory, so that could be changed to build it anywhere.
14:52:15bluebrotherpreglow: might be nice for the sim too −− make install also creates a zip first IIRC
14:52:19linuxstbe.g. search and replace .rockbox with $ROCKBOX and set $ROCKBOX appropriately.
14:52:37 Quit crwl (Read error: 104 (Connection reset by peer))
14:52:45preglowi'll just mod a flag into buildzip
14:53:22linuxstbpreglow: You probably want to remove the "rm -fr .rockbox" line though...
14:54:33linuxstbAnd be careful with the search and replace - some instances of .rockbox need to stay...
14:54:35amiconnWee, asm motion compensation really helps on coldfire :D
14:54:45linuxstbamiconn: What are the numbers?
14:54:46preglowamiconn: i really, really expected it to :P
14:54:48preglowamiconn: how much?
14:54:56amiconn+5 fps fullscreen on X5, with just the copying version asm'ized
14:55:10preglowthat's bloody marvelous
14:55:37*preglow finds that is one of those delightful tools that don't explain their flags in a nice help screen
14:55:39amiconnAnd there are the various averaging versions which still wait for their conversion...
14:56:14amiconnI don't think I'll convert the ones which do destination averaging
14:56:32 Join crwl [0] (
14:56:45amiconnHmm, something is slightly wrong there
14:57:09amiconnSome blocks are messed up
14:57:35markunamiconn: maybe you should stop before you get better performance than the Gigabeat :)
14:58:29preglowahh, arm already has assembler
14:58:49amiconnYes, for copying (what I did for cf right now) and half-pixel x averaging
14:59:15amiconnNot for half-pixel y averaging, half-pixel x-y-averaging, and not for any destination averaging version
14:59:16*markun is pretending he never saw preglow's goto
14:59:30preglowmarkun: what goto? i don't understand what you're talking about
14:59:54markunpreglow: don't worry, I don't understand either what I was talking about :)
14:59:57preglowlinuxstb: the last wma related bug i can't be bothered to check out, need more test files
15:00:24J13:46 < preglow> linuxstb: ok, i just do a goto next_track;
15:00:32elinenbemarkun: are you living in NYC now? I saw the NYC meetup photos from a couple of days ago.
15:01:00markunelinenbe: no, I just made google pay for my vacation :)
15:01:10*preglow covers ears and sings
15:01:17markunand I'm back in the netherlands anyway
15:01:43elinenbeah... do you work for google?
15:02:07elinenbeI wish I'd have known about the meetup, as I am in New York.
15:02:10pixelmapreglow: what does covering ears help in IRC? :D
15:02:14markunno, I attended the google summer of code mentor summit (unlike that sining guy over there)
15:02:29GodEater_elinenbe: it wasn't a secret - it was advertised in the forums weeks in advance...
15:02:40elinenbepreglow: singing is offtopic.
15:02:45markunelinenbe: yes, too bad :(
15:02:58markunelinenbe: but they plan to have more meetings
15:03:21elinenbeGodEater: I'm not a forum regular −− just IRC (sometimes), and the mailing lists (which seem to be getting quite stale)
15:04:38 Join Crash91 [0] (n=evil91@
15:04:53 Join kugel [0] (i=kugel@unaffiliated/kugel)
15:05:56amiconnAh, found the bug... :)
15:06:31 Nick parafin|away is now known as parafin (
15:06:36amiconnI had the offset-1 routines and the offset-3 routines switched. Completely smooth now :)
15:07:04preglowstill same speedup?
15:07:26amiconnJust flipping 2 code paths for 16x16 and 8x8 blocks each
15:08:22preglowARCHOS=h120 <- i love these old remnants :D
15:08:26markunelinenbe: somehow I always thought you were german (are you?)
15:08:35preglowi always thought he was swedish :>
15:09:13 Nick idnar_ is now known as idnar (i=mithrand@unaffiliated/idnar)
15:09:38 Join Gnu47 [0] (
15:10:12amiconnNow, on to the 1/2 pixel x averaging...
15:12:10***Saving seen data "./dancer.seen"
15:13:10preglowmpeg2 doesn't have quarterpel averaging, no?
15:14:20amiconnI'm not sure what you mean
15:15:13 Quit parafin (Remote closed the connection)
15:15:27amiconnThe moion compensation moves a macroblock by a certain amount of pixels horizontally + vertically, and optionally by an additional 1/2 pixels horizontally, 1/2 pixels vertically, or both
15:15:47 Join parafin [0] (
15:15:48preglowthat's what i'm asking, it doesn't have 1/4 resolution?
15:15:50 Join daurnimator [0] (i=daurn@unaffiliated/daurnimator)
15:15:54amiconnCheck motion_comp.h for the C code - it's heavily macrofied
15:15:55preglowi think that got added in h264
15:16:18amiconnThe predict_X macros
15:17:19amiconn...and the MC_FUNC_x macros further down
15:19:06amiconnpreglow: It can also do the said motions, but instead of replacing the block at the destination, average the moved block with the destination
15:20:56 Quit pixelma ("bbl")
15:21:07amiconn16 fps fullscreen on H300 now
15:22:22preglowwhat are the screen dims on x5 and h300?
15:22:46Llorean160x128 and 220x176 iirc
15:22:59preglowis that the only reason for the difference in speed?
15:23:08elinenbe9fps -> 16 fps... great!
15:23:18LloreanAlmost double.
15:23:27preglowyeah, but that's with asm idct too
15:23:29preglowi think
15:23:40preglowstill very noticable :D
15:23:41elinenbeamiconn: can you see anywere else to squeak some speed out of it?
15:23:48preglowiram will be important
15:23:52amiconnIf I play the X5 fullscreen version on H300, it's equally fast
15:24:06markunthe video dimensions don't have to be a multiple of 16 for a long time, shall I change
15:24:26Lloreanmarkun: Isn't that a requirement of the mpeg standard?
15:24:28amiconnIt could be a bit faster due to the slightly simpler lcd_yuv_blit() if we would fiddle with the lcd waitstates between lcd_yuv_blit() and lcd_update()
15:24:58amiconnWhen boosted, lcd_update() needs 5 waitstates because it uses dma, but for lcd_yuv_blit(), 4 waitstates would be sufficient
15:25:02elinenbeamiconn: are these updates to the mpegplayer, or the LCD driver?
15:25:02markunLlorean: the blocks are 16x16, but the display resolution should just be a multiple of 2
15:26:07amiconnelinenbe: mpegplayer
15:26:36amiconnI committed lcd driver (and general core speedup) updates a couple of days ago though
15:26:37Lloreanmarkun: Is there any benefit to using a multiple of 16 (speed, etc?) Also, is there any likelihood there'll be increased compatibility problems if people start picking any random even number?
15:26:46 Quit Gnu47 (Nick collision from services.)
15:26:52elinenbeso, in the scenario a mpg4 codec is written, they won't benefit that... I got it.
15:26:53 Join Gnu47 [0] (
15:27:29Lloreanelinenbe: There's still no good reason for someone to write an mpeg4 codec.
15:28:21preglowsure there is
15:28:28preglowmpeg4 videos use less space
15:28:42*linuxstb would like to see more video codecs supported, just as we support many audio codecs
15:28:55amiconnlinuxstb: Why?
15:28:57Lloreanpreglow: I'm not so sure they will.
15:29:00markunLlorean: I don't think there would be any drawbacks.
15:29:00Avedoes it really make a diffence in low-resolution displays?
15:29:01LloreanThey'll use a little less space
15:29:02preglowLlorean: how come?
15:29:03linuxstbamiconn: Why not?
15:29:10amiconnVideo is different from audio in that you need to convert for the target anyway
15:29:26Lloreanpreglow: In terms of image quality, the screens are so small anyway that I'm not sure mpeg4 will get a significant bitrate advantage.
15:29:26preglowLlorean: there's a possibility you're right, but i would still like to see some numbers on it
15:29:34amiconn...and rockbox is a DAP firmware in the first place
15:29:42linuxstbNot always - 320x240 is a relatively common size for videos such as podcasts to be distributed in.
15:29:43preglowin the first place, yes
15:29:49preglowand now daps aren't only audio devices anymore
15:29:57Lloreanpreglow: As well, with the additional computational overhead, someone with mpeg2 video can just sacrifice framerate for image quality anyway.
15:30:05preglowvideos are distributes to be played on these devices
15:30:06amiconnlinuxstb: What are podcast? :>
15:30:09preglowso we should support these other formats
15:30:10markunif we can find a video codec which is better than mpeg2 in speed, size or quality I'm all for it
15:30:13LloreanI really wouldn't MIND other video codecs.
15:30:19LloreanI just don't think that it's likely to be an "improvement"
15:30:39preglowi don't really care about video, but why the hell not?
15:30:42LloreanWith the exception of maybe the Gigabeat, which might be able to play them at decent quality *and* framerate
15:30:54preglowLlorean: and let's face it, that's the way daps are going
15:30:59LloreanOh, I agree on that
15:31:03markunthe Gigabeat S might be capable of a lot more
15:31:06LloreanFor future targets, multiple video codecs will be a must
15:31:12LloreanI just don't think it's *yet* particularly beneficial
15:31:13preglowgigabeat s can do h264 while standing on its head
15:31:20linuxstbI'm not saying I want to work on other video codecs, but I don't think we should discourage it.
15:31:31Lloreanpreglow: h264 support would be nice, we could then support "iPod" encoded videos
15:31:37linuxstbpreglow: With an APE audio stream...
15:31:48 Join Benoitb [0] (
15:31:59Lloreanlinuxstb: I don't mean to discourage it, I just think I lot of people request it on a simple assumption that it *must* be better.
15:32:05preglowLlorean: those videos aren't drmed, are they?
15:32:26Lloreanpreglow: iTunes ones are. Podcasts that are pre-encoded, music videos that authors post to their websites, etc, usually don't tend to be.
15:32:28linuxstbpreglow: itunes-bought ones are.
15:32:45*linuxstb wonders how Llorean wrote such a long reply, quicker...
15:32:55preglowi wonder how much cpu they waste on decrypting...
15:33:07 Quit parafin ("So long and thanks for all the fish")
15:33:13 Join parafin [0] (
15:34:03linuxstbI don't know, but maybe mpeg4 encoders are more common then mpeg2, due to the proliferation of original firmwares using it.
15:34:38preglowlinuxstb: also assumes a .rockbox dir...
15:34:46Lloreanlinuxstb: Probably so. Most desktop video seems to be mpeg4, and then you can just pick a smaller resolution for your DAP
15:35:38LloreanAlso the iPod OF supports mpeg4 divx as well as h264/AVC I believe, and the PSP too.
15:35:52preglowit's the more widespread codec for this use
15:35:54preglowplain and simple
15:36:02 Join stewball`ghost [0] (n=WTFOMGBB@
15:36:10preglowyou don't see mpeC[C[Cg2 files spread around often unless you're looking at dvb or dvd
15:36:38Lloreanmpeg2 encodes right quick, though.
15:36:39 Quit aliask ("ChatZilla [Firefox]")
15:36:57preglowthe newer codecs are slow as hell
15:37:36 Quit ender` (Read error: 104 (Connection reset by peer))
15:37:52linuxstbBut anyway, no-one seems keen on actually implementing other codecs...
15:38:20LloreanWell there is the old xvid decoding patch in the tracker, right?
15:38:26 Part Benoitb ("Kopete 0.12.5 :")
15:38:33LloreanIf I recall it even decoded frames, there was just no display code etc yet?
15:38:36 Quit jhMikeS (Read error: 104 (Connection reset by peer))
15:39:03linuxstbThat's a nasty implementation though - IIRC it didn't even fit in the 512KB plugin buffer.
15:39:33linuxstbBut maybe it could be trimmed down...
15:39:55LloreanAnother basic advantage of mpeg4 divx/xvid support would be that preconverted videos for X5 and H300 wouldn't need to be remade.
15:40:32preglowlet's put the codecs in the core! :P
15:40:48linuxstbI'm sure tomasz will like that...
15:41:38 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
15:43:17preglowwtf, the arm idct has no multiplies
15:43:27markunbecause of the VGA resolution of the mrobe we were actually talking a bit about scalable font support during the GSoC summit :)
15:43:48 Join PaulJam [0] (
15:43:51markunI'm sure amiconn would love to have that in code :P
15:44:58 Join RaRe` [0] (
15:45:23preglowwe wouldn't want it in the core for our smaller targets anyway, so..
15:45:25 Nick Tanuva|off is now known as Tanuva (n=tanuva@
15:45:55linuxstbI would imagine the mrobe needing completely new UI widgets to work nicely.
15:47:35amiconnThe m:robe 500 is a strange dap anyway, imho
15:47:44LloreanWhat exactly would scalable fonts mean?
15:47:50amiconnThe touch ui...
15:47:50 Join Honkboy [0] (
15:47:51preglowthat we shouldn't port to it?
15:47:52LloreanIn my head it just means "the option to pick size along with font"
15:48:33linuxstbTouch UIs are sadly getting more common....
15:48:42markunLlorean: I was thinking about vector fonts like t1 or truetype/opentype
15:48:53linuxstbWhat's the spec of the CPU?
15:48:56LloreanAh, vector fonts.
15:49:06pondlifeCan anyone explain concisely why we shoudn't have special-casing for "The" in tagnavi.config? Llorean, maybe - you're eloquent.
15:49:29pondlifeI'm about to commit, but can't come up with a nice commit message.
15:49:41amiconn'The' is 100% language specific
15:50:33linuxstbmarkun: google tells me it's a 160MHz ARM926EJ-S
15:50:38 Part Honkboy ("Leaving.")
15:50:43Lloreanpondlife: In the tagnavi it's not *as* big a deal. But leaving it out could simply be explained as "Reduce clutter in the tagnavi with the expectation that special casing for 'The' will become unnecessary in the future, and can be cleared out to clean up the default interface" or something similar.
15:50:56*amiconn wants some extra data registers on cf :>
15:51:13pondlifeLlorean, amiconn: Thanks
15:51:14linuxstbmarkun: It's the same as the TCC77x - ARM v5 (but the TCC77x is only 120MHz)
15:51:17preglowamiconn: i want them on arm... cf has more
15:51:34amiconnarm has universal registers... cf does not
15:51:34Lloreanamiconn: Unfortunately, tagnavi is already english-specific, isn't it?
15:51:47preglowamiconn: ahh, like that, you can use addr regs plenty of places on cf anyway
15:52:02amiconnNot for shifting purposes...
15:52:09markunlinuxstb: doesn't sound too fast, but of course the DSP makes up for it
15:52:09preglowcorrect :/
15:52:55linuxstbmarkun: Yes, but it's an architecture better than the ARMs Rockbox runs on at the moment, and has extra DSP instructions (not including the DSP itself in the DM320)
15:53:40preglowmy, and have really bad command line argument handling
15:55:42preglowordering dependent and everything
15:55:53lostlogicpondlife: haha, plug your ears for a few days... :-P
15:55:56*preglow wishes people would use Getopt::Std
15:56:14pondlifelostlogic: Do you think people really use that much?
15:56:39pondlifeI'm happy to revert if pushed, I just though it was tidier this way.
15:57:13lostlogicthere are a lot of English speaking teenagers who will be unhappy −− at one point there was a big uproar wanting directory sorting that ignores "the"
15:57:39lostlogicpondlife: which doesn't make you wrong, it makes them...
15:58:01pondlifeWe never gave in on the sorting, and so now we are more consistent :)
15:58:42pondlifeBesides, by that argument I'd want to see The White Stripes appear under W, rather than having a The section, no?
15:59:04 Join scorche|w [0] (n=8dc5049d@
16:00:32 Quit Zagor ("Client exiting")
16:01:51 Join barrywardell [0] (
16:01:58 Quit RaRe (Read error: 110 (Connection timed out))
16:06:09JdGordonhey all
16:07:19 Join ender` [0] (
16:11:28 Join freqmod [0] (
16:14:32 Quit Crash91 ("Bye Bye!")
16:15:09*preglow has working make install
16:15:22preglownow, what to do with it :V
16:16:04markunwhere does it install to?
16:16:16preglowright now a hardcoded place in the makefile
16:16:24preglowbut really where you want to, i don't know how to pass options to a makefile
16:16:38preglowsomething like make install DEST=/media/H120/.rockbox would be nice
16:17:05linuxstbI'm not sure it belongs in the Makefile - as it's very specific to the user environment. Unless you add it to configure...
16:17:28preglowsure, i'd have to add it to configure to make it appear in every user environment
16:17:33bluebrotherwell, using an environment variable sound quite nice to me.
16:17:35linuxstbi.e. DEST=/media/H120/.rockbox ../tools/configure
16:17:48preglowsounds pretty clever
16:18:42preglowright now, all i've done is add to -c options to and which specify target, both default to .rockbox also doesn't actually build a zip if -c is specified
16:20:07linuxstbIs it much faster then building a zip?
16:20:19preglownot as much as i'd hoped
16:20:25preglowbut i'll do a quick time
16:22:04preglowlooks like it's actually slower, wtf
16:23:29preglowit's not much faster at all
16:23:32preglowi'll just scrap this entire attempt
16:24:09preglowit might actually be faster for sim builds, but i'm not jumping through hoops to make sim builds a couple of seconds faster to install
16:24:35 Join iamben [0] (
16:24:36 Quit iamben_ (Connection reset by peer)
16:24:43linuxstbI normally just cp whatever has changed anyway when using the sim
16:26:45preglowi do that for both target and sim, usually
16:28:13preglowlinuxstb: did you ever try out the hw eq removal patch?
16:30:53linuxstbNo. I'll do it now...
16:32:40linuxstbThe patch doesn't apply cleanly at all...
16:33:14linuxstbI don't have time to investigate now though.
16:34:09preglowbeh, not surprising
16:35:03preglowyou could just ask for a sync if it fails too miserably
16:38:02 Join webguest36 [0] (i=52985296@gateway/web/cgi-irc/
16:38:07 Join printhorizon [0] (
16:38:44 Join webguest78 [0] (i=52985296@gateway/web/cgi-irc/
16:38:52 Quit webguest36 (Client Quit)
16:39:10JdGordonlinuxstb: we'll know better on the weekend (when i get a chance to play), but with either big enough fonts, or a stylus the mrobe shouldt be too hard to control
16:39:13 Join proppy [0] (
16:39:41pondlifeJdGordon: How much of the code runs on the mrobe now?
16:39:47linuxstbJdGordon: You think the current UI is good enough then?
16:40:12JdGordonpondlife: we can run the main image, no buttons work so cant really test how well though
16:40:19JdGordonlinuxstb: we'll see... it ihnk it will be ok
16:40:24 Part qwm
16:40:50pondlifeJdGordon: So it'll display the settings menu, but that's it..
16:40:53proppyI reformated my HFS+ partition to FAT32 using gparted under linux,
16:41:04pondlifeSorry, I meant main menu
16:41:07JdGordonwith the default font its absolutly tiny :)
16:41:11proppynow ipodpatcher failed with the following error: [ERR] Partition layout is not an ipod
16:41:20linuxstbproppy: Did you do everything stated here?
16:41:31linuxstbproppy: It's not simply a matter of reformatting...
16:41:46 Quit karashata ("Leaving.")
16:42:10proppyoh ok
16:42:16proppyI wasn't aware of that wiki page
16:42:18proppythanks a lot
16:42:22 Quit JdGordon ("ZZzzzzz...")
16:43:00linuxstbThe manual links to it, so does ipodpatcher, and so does rbutil...
16:43:28proppylinuxstb: there is no 1go nano first generation partition table
16:43:44 Quit webguest78 ("CGI:IRC (Ping timeout)")
16:44:38proppylinuxstb: You're right the manual link it in Prerequisite section
16:45:34linuxstbThen you'll need to either find someone with a 1GB Nano who will donate their partition table, or find a Windows PC with itunes installed.
16:45:46linuxstbOr if you've confident with partition tables, you can try and create it manually.
16:46:27proppyI will try to study the difference between 4go and 2go
16:46:32linuxstbSee here -
16:47:00 Join karashata [0] (
16:50:31proppylinuxstb: I used the 2gb one
16:50:41proppylinuxstb: and just resized the partition with fdisk
16:50:59proppylinuxstb: now ./ipodpatcher recognized the partition table
16:51:31proppythat what advice btw :)
16:52:12proppycan I dump my partition table to upload it to ?
16:55:06PaulJamwow, nice work on the mpegplayer for coldfire. thanks amiconn.
17:03:24amiconnpreglow: The data register shortage isn't as bad as I feared; I can use an addr register for the loop counter, which means just one additional instruction
17:03:41amiconn(tst.l %a3, as subq.l #1, %a3 doesn't set flags)
17:11:13 Join n1s [0] (
17:12:11***Saving seen data "./dancer.seen"
17:12:33 Join desowin [0] (
17:13:38 Quit barrywardell ()
17:14:33preglowamiconn: yes, i've usually been able to work around in ways like that myself
17:17:06jumpatrainstill working on the mpc7 for rockbox?
17:17:44jumpatrainmusepack sv8 playback
17:17:48preglowjumpatrain: still working on? no one has started
17:18:03 Join sarixe [0] (
17:18:03bluebrotherI was thinking something like that too :)
17:18:12jumpatrainpreglow, ah. i thought for a time, weeks ago, you wanted to do that
17:18:34preglowjumpatrain: yeah, i did start, then immediately got highly bored
17:18:47jumpatrainyeah :)
17:18:51jumpatraingot that impression
17:18:56preglowmy tendency to work on stuff i won't use myself has gotten worse during the last year
17:19:54amiconnHmm. Can't tell for sure whether that halfpixel x averaging in asm has any effect on speed :|
17:20:00preglowworse meaning i don't do it as much :>
17:20:08preglowamiconn: because not much uses it?
17:20:19amiconnYeah, perhaps
17:20:31preglowfancy encoders should
17:21:27sarixeman, i had a really cool idea for a feature, but i completely forgot what it was
17:21:37sarixei'm so disappointed
17:21:41bluebrothersarixe: cool feature :)
17:22:37sarixeah, gtg though
17:22:40 Quit sarixe (Client Quit)
17:23:43printhorizonHey, I just noticed that the build of Rockbox I'm using supports cuesheets. I'm using ye olde Archos Jukebox Recorder v1, which just plays MP3s. Can cuesheets be used with MP3?
17:25:18printhorizonOkay... so I've got a cuesheet and an MP3 image in the same directory
17:25:26printhorizonI "play" the cuesheet
17:25:36proppylinuxstb: thanks for the wiki links they work like a charm
17:25:56bluebrotheryou need to enable cuesheet support first (and iirc you need to reboot after that) and play the mp3 file
17:26:11printhorizonOkay. Well, I did enable the cuesheet support and rebooted
17:26:21preglowprinthorizon: you play the mp3, not the cue
17:26:21bluebrotherboth files need to have the same filename except the file extension
17:26:32printhorizonIf I play the MP3 itself, I'm only seeing the MP3 as one big file
17:26:37preglowprinthorizon: after playing the mp3, you can play the cue. don't ask me why this is
17:26:42 Join FOAD_ [0] (
17:26:52preglowprinthorizon: what bluebrother said is important
17:27:18printhorizonOkay, I think the filenames were "artist - album.mp3" and "artist - album.mp3.cue"
17:27:22printhorizonProblem? :)
17:27:28printhorizonI see.
17:27:41printhorizonthe cuesheet should simple be "artist - album.cue" then?
17:27:44bluebrotherwell, not really a problem but you need to rename the cue to "artist - album.cue"
17:27:53printhorizonI'll try that
17:30:30printhorizonHmm... cuesheet support was not enabled. Weird. I know I did enable it.
17:31:13printhorizonRockbox doesn't seem to be saving my setting.
17:31:36rasherHrm.. is it not possible to use automatic update of Database on Sansa? The manual tells me to enable dircache, but that's nowhere to be found..
17:31:48 Join jac0b-work [0] (
17:31:58bluebrotherI don't know the details on the archos but make sure to actually shut down the player and not simply reset it
17:32:26rasherAlso, bluebrother.. rbutil is unable to detect my sansa, whereas sansapatcher deals just fine..
17:32:34rasherOn Windows.
17:32:41rasherIf you need me to test anything, shoot
17:32:47printhorizonI'm not too sure on the difference. What I'm doing is pressing "Off" once, it says something like "Press OFF again to shut down" and then I do so.
17:33:00bluebrotherwell, I don't have any idea what could be causing this right now.
17:33:15bluebrothermaybe Domonoky has −− he's working on windows.
17:33:24bluebrotherI guess you're not on vista?
17:33:28rasherNope, XP
17:33:48bluebrotherthere was a similar issue for ipodpatcher installation on xp.
17:33:52rasherRunning as an unpriviledged user, but running rbutil as admin, of course
17:34:54printhorizonHmm, could my problem be caused by extremely low batteries? Rockbox is saying to charge the battery when it is shutting down. :P
17:34:57pondlifebluebrother: Did you see my problem with talk file generation? It appeared to work, but didn't create them..
17:35:00*amiconn learned that he should make better use of asm macros
17:35:13bluebrotherpondlife: yes, I noticed it.
17:35:35amiconnLowers the likelihood of unnoticed typos....
17:35:35pondlifeI've been too busy to experiment much further, sadly.
17:35:44 Join ToHellWithGA [0] (
17:35:58printhorizonOkay, I got the cuesheet support enabled. Lots of thanks for the help.;
17:36:05printhorizonWorks like a charm! This is really cool.
17:36:11rasherSpeaking of talkfiles.. anyone up for hacking on adding voice-clips to Database? I've got the clip generation part going, but need to add something in the Database menu code, which I'm not sure I can handle..
17:36:43pondliferasher: What filenames are you using?
17:37:12rasherThe string to speak, sanitized to be FAT-friendly, and truncated at 100 chars.
17:37:22rasherI imagine putting them somewhere in .rockbox
17:37:56pondlifeAh, so it'll just do a lookup
17:38:15rasherYeah, should be fairly simple for someone in the know about the database code, hopefully
17:38:18pondlifeCouldn't that also apply to file browsing too?
17:38:43pondlifeMight save a little space as there would only need to be one clip for an album with a title track
17:38:48pondliferather than 2
17:38:49printhorizonHrm, I'm guessing the database might have some trouble with the MP3 images? Or is it smart enough to work with the cue points and display separate albums, tracks, etc.?
17:38:51 Quit FOAD (Read error: 110 (Connection timed out))
17:38:51 Nick FOAD_ is now known as FOAD (
17:39:22rasherpondlife: I suppose so, but I'm not sure how much is gained by that - plus you have then separated the talk clips from the files, which might not be wanted if people move their stuff around
17:39:45pondlifeI like the idea of a single system for both browsers though
17:39:59rasherThat's a good point as well..
17:40:16linuxstbprinthorizon: Rockbox just uses cuesheets as points to skip to within a track. The big file is still considered one track as far as the rest of Rockbox is concerned.
17:40:25rasherBut with the current system you can skip the string sanitizing, since you know the string to be a valid filename already
17:40:35rasherdon't know if that means anything in practice though, probably not
17:40:55 Join WalterEgo [0] (
17:41:07proppyI get many Prefetch/Illegal instruction error, is this a known problem ?
17:41:32rasherpondlife: Also, I'm sure we'll get angry fanmail if we break the vbs .talk file generating script
17:41:40printhorizonlinuxstb: Thanks. I suppose now I just need to weigh the pros and cons of using a single image and a cue sheet versus individual tracks.
17:42:13linuxstbprinthorizon: For Rockbox, individual tracks will work better - so you can shuffle, add individual tracks to playlists etc.
17:43:30linuxstbproppy: On some Nanos, yes. I'm not sure what the suggested solution/workaround is, but there are a few forum threads and a task on the bug tracker related to it.
17:43:57proppyoh ok let me search it
17:44:15rasherpondlife: I guess I'll upload a patch with the modifications and advertise for people to fix the database side of things
17:44:32printhorizonlinuxstb: I see your point. :) I rarely shuffle or use playlists but I do sometimes prefer to start with a song in the middle of the album, and that'd be easier with the individual songs...
17:44:59pondliferasher: I was thinking for people who use both file and database browsing it would be good not to need 2 sets of voice clips...
17:45:11pondlifeNot sure that came across clearly though :)
17:45:22rasherAh, very good point..
17:45:52linuxstbprinthorizon: The only uses I know of for cuesheets are 1) Making a perfect backup of a CD (but you're using MP3 anyway, so that's not relevant), and 2) For gapless playback with software that doesn't have gapless support (which Rockbox does have, if the mp3s are encoded/split correctly).
17:46:07rasherthat would roughly half the needed clips
17:46:26pondlifei.e. would be found for either browser
17:46:50rasherI suppose database support could come first, then unification afterwards
17:47:39pondlifeMight be free to support both types in the file browser after that.
17:47:58printhorizonlinuxstb: Right. I enjoy using FLAC with embedded cuesheets in foobar2000 on my computer, but I think individual tracks are perfectly fine for a DAP. Thanks for the advice... it's cool to know Rockbox does do this, though.
17:50:49 Join Lear [0] (i=chatzill@rockbox/developer/lear)
17:51:45 Join JRoT [0] (
17:55:43printhorizonHow about any personal recommendations on newer DAPs that Rockbox will work on? I'm gonna have to replace my Archos sometime, and I'd love to stick with Rockbox for the software.
17:56:03bluebrothercheck the BuyersGuide wiki page
17:56:40 Join Crash91 [0] (n=evil91@
17:56:54linuxstbI don't think any hard-disk based DAPs which Rockbox works on are sold new any more - you'll have to buy a used one. (but the video ipods may still be available)
17:57:04printhorizonI see.
17:58:14 Join styleism [0] (
17:58:27printhorizonI definitely want to get one that plays Ogg.
18:00:22 Quit Crash91 (Client Quit)
18:03:01 Quit PaulJam (".")
18:03:30 Quit atsea-99 (Read error: 104 (Connection reset by peer))
18:03:38 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
18:03:56 Join atsea-22 [0] (i=atsea-@gateway/tor/x-c51a2fef901cf737)
18:04:41 Quit jumpatrain (Remote closed the connection)
18:05:33linuxstbprinthorizon: All the non-Archos devices Rockbox runs on play the same audio formats - they all play Ogg vorbis.
18:05:46 Join jumpatrain [0] (i=tabac@gateway/gpg-tor/key-0xB9002659)
18:05:56rasherpondlife: if you're interested, I just submitted the script as FS #7984
18:06:01printhorizonlinuxstb: Oh, thanks, I was just about to ask where to find what formats were support on the various devices
18:09:44proppylinuxstb: update with my info
18:10:24proppylinuxstb: I can lend my ipod to a EU rockbox developer if they still need to
18:13:42 Quit Thundercloud (Remote closed the connection)
18:17:36printhorizonHeh, okay, I have one last question... has anyone had trouble with the navigation buttons on the Archos units losing their "clickiness"? The left and right buttons on mine still have quite a bit of "click" to them, but up and down actually down click at all... they work, but it doesn't feel like one is actually pressing them.
18:17:59printhorizonNot sure whether that's fixable without opening the unit up, but I thought it'd be worth asking
18:18:46 Join PaulJam [0] (
18:18:54 Join Thundercloud [0] (
18:19:33 Join pepie34 [0] (
18:19:46printhorizonoh, shoot, that was supposed to read "up and down actually DON'T click at all," not "up and down actually down click at all"
18:19:47 Quit pepie34 (Read error: 104 (Connection reset by peer))
18:20:35PaulJamhi, i just updated the build on my h300 and now in the wps the time and date informations often dissappear for a second. is this already known?
18:22:15 Join Konam [0] (
18:22:51Konamis there a way to change -or turn off- the rockbox presentation screen?
18:23:28 Join shnee [0] (
18:23:44PaulJamKonam: it is possible to change it, but this requires changing the sourcecode and compiling rockbox
18:24:28 Part printhorizon
18:24:37JRoTanyone knows why i don't see any album art on my sansa e260?
18:24:56rasherJRoT: Because Rockbox doesn't support Album Art.
18:25:25JRoTohw oke, cause a saw it by the screen shot of the themes
18:25:27JRoTbut THnx
18:25:59scorche|wJRoT: some themes are meant for unsupported builds which may have the album art patch in it, which is why you would be seeing that
18:26:02rasherThere is unofficial and unsupported album art funcitonality available through patching or running custom (unsupported builds)
18:26:40JRoTk i will keep it by the original rockbox
18:26:46JRoTWich is great
18:26:55 Quit mf0102 ("Verlassend")
18:27:20pixelmaPaulJam: interesting, I can see the time disappearing on my M5 too (build is r15153)
18:27:34 Join chris_mt [0] (
18:27:41pixelmacouldn't make out a scheme for it yet
18:27:44rasherWith any luck, proper album support shouldn't be too far off (few months or even weeks, perhaps - though we *don't* give out dates, so consider this my personal guess)
18:28:06*scorche|w tsks at rasher
18:28:25bluebrotherit was almost a timeline ;-)
18:28:29rasherWas not!
18:28:42PaulJampixelma: seems to be random, in the statusbar it happens too
18:29:03 Join kratonator [0] (i=544bb0fc@gateway/web/cgi-irc/
18:29:15 Nick fxb__ is now known as fxb (
18:29:16 Join Ambiquity [0] (i=d988c7f4@gateway/web/cgi-irc/
18:29:17bluebrotherLlorean: seen the display issue at ?
18:29:25pixelmaPaulJam: it's even visible in the statusbar... seems less likely to me though
18:29:31bluebrotherseems the forums don't deal with < and > gracefully
18:29:58 Join Addi2 [0] (
18:30:01scorche|wbluebrother: yeah...they dont..never have
18:31:09scorche|wfixed :)
18:31:17scorche|w...not really
18:31:49pixelmafun thing is that the let you think they work because the preview looks ok
18:32:47scorche|wwell, now they are fixed...
18:33:18bluebrotherisn't there a fix available for the forums software?
18:34:08 Quit davina (Read error: 104 (Connection reset by peer))
18:34:08scorche|wbluebrother: a full disabling of html code *should* fix it, but i am not sure
18:35:16preglowlinuxstb: as asf maintainer, do you expect #7981 to be related to the metadata parser itself?
18:35:49 Join davina [0] (
18:36:17 Quit obo ("bye")
18:38:03JRoTrasher you did release rockbox under a gnu license isn't it, so can't you then use the Album Art patch from a costum build?
18:38:24preglowrasher is responsible for releasing rockbox? :P
18:38:32rasherJRoT: Problem is the code is not how "we" want it. It's something of a hack
18:38:33bluebrotherhe is.
18:38:41rasherOh dear.
18:38:48bluebrotherblame rasher for the long no-release period!
18:38:52JRoTah oke thnx rasher
18:39:16rasherbluebrother: Hey, I was the one trying to come up with policies and inciting riots not too long ago!
18:39:51preglowanyone know what the mpegplay.c code in uisimulator/common is for?
18:39:51PaulJamis someone with a non coldfire target that has a rtc here?
18:39:52bluebrotheryou should come up with a release schedule! We want a release now!
18:39:55 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
18:39:55bluebrotherhehe, scnr.
18:39:58preglowPaulJam: yeah
18:40:05bluebrotherPaulJam: yes
18:40:17 Join kkurbjun [0] (n=kkurbjun@alamode.Mines.EDU)
18:40:52JRoTmp4 doesn't work isn't it?
18:40:55PaulJamcould you look if the problem that the time dissappears randomly for a second int the wps and in the statusbar happens on your player too?
18:40:55rasherpreglow: I believe there was once support for sound in the hwcodec targets.. perhaps that's it?
18:41:16rasherCompletely guessing here.. obviously
18:41:17preglowrasher: yeah, that might it. any plans to fix it?
18:41:17bluebrotherJRoT: mp4 what? Video? Audio?
18:41:45rasherpreglow: well it's been broken for years afaik, so it's probably not high priority
18:41:58preglowmy delete finger is itching :>
18:42:16bluebrotherJRoT: you might be interested in ... and mpegplayer needs mpeg2.
18:42:38 Quit Konam (Read error: 110 (Connection timed out))
18:42:58 Quit Arathis ("Bye, bye")
18:43:09rasherLlorean: why did you make power->stop in FS #7232?
18:43:19JRoTthnx blargit
18:43:22JRoT* bluebrother
18:43:36pixelmaPaulJam: I'll try soon on with a more recent on my c200 but r15145 behaved correctly
18:43:44bluebrotherPaulJam: the time display on my mini works fine here.
18:44:22bluebrotherbut I experienced an interesting behaviour when I flashed my h120 with an rtc build: the time showed −−:−− first but after a while showed an arbitrary value
18:44:23pixelmabluebrother: compare to your H100?
18:44:42amiconnlinuxstb: around now?
18:45:19*preglow wants simulator volume support
18:45:47amiconnIt's already there...
18:45:49 Join barrywardell [0] (
18:46:02bluebrotherpixelma: I had this issue when I flashed an RTC build to my h120 before I added the RTC chip. Now it's modded and the time works just fine.
18:46:53bluebrotherthe only issue that's left now is that it sometimes resumes playback after bootup even if the alarm didn't trigger. But I ahven't bothered enough to look into it.
18:46:53preglowamiconn: eh? why doesn't sim volume change, then?
18:47:13amiconnI don't know. Perhaps it depends on the sound driver or sth
18:47:16rasherpreglow: I blame Linux mixer hilarity
18:47:20proppyamiconn: ping
18:47:23amiconnJdGordon added it weeks ago
18:47:36preglowhahahah, nothing changes until you're _really_ low
18:47:42pixelmabluebrother: with a current build and watching it for a while?
18:47:46proppyamiconn: about do you want me to lend you a faulty ipod ?
18:47:48preglowi think i'll just go ahead and add software volume handling
18:47:56preglowone should never access the global mixer anyway
18:48:11n1samiconn: mpegplayer speed doubled on my h300 with your recent optimizations :-)
18:49:56 Quit Ambiquity ("CGI:IRC (EOF)")
18:50:09 Join webmind_ [0] (
18:50:09 Quit webmind (Read error: 104 (Connection reset by peer))
18:50:15bluebrotherpixelma: yes, I haven't noticed any issue with the current build.
18:50:39proppyanw I will drop a mail to -dev list
18:50:45 Nick fxb is now known as fxb__ (
18:52:16 Quit Lear ("ChatZilla [Firefox]")
18:52:45*bluebrother updates build on ipod
18:53:06 Join homielowe [0] (
18:54:13 Quit jumpatrain ("Leaving")
18:54:31 Quit stewball`ghost ()
18:54:40 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
18:55:12preglowrasher: according to sdl docs, the sdl volume does not affect any linux mixers
18:55:29 Quit BigBambi (Read error: 104 (Connection reset by peer))
18:55:46 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
18:55:49PaulJambluebrother: i'm writing a bugreport. which svn revision did you use on your mini?
18:55:57 Join Domonoky [0] (
18:56:36bluebrotherPaulJam: just replaced it with current svn, but my version wasn't older than two days.
18:56:39pixelmaPaulJam: I don't see this effect on my c200 with revision 15168
18:56:48bluebrotherwill keep an eye on it.
18:56:58PaulJamthank you.
18:57:13 Quit BigBambi (Read error: 104 (Connection reset by peer))
18:59:18rasherpreglow: sounds like it's just broken then
18:59:30preglowit passes the values it should
18:59:32 Join Lear [0] (i=chatzill@rockbox/developer/lear)
18:59:34preglowso the problem is with sdl
18:59:46rasherDomonoky: Any ideas about why rbutilqt on winxp doesn't detect my Sansa? sansapatcher works fine..
19:00:48Domonokyrasher: both autodetection and the bootloader install doesnt work with rbutil ?
19:01:01rasherDomonoky: Indeed
19:01:20bluebrotherbootloader installation will fail if ipodpatcher / sansapatcher doesn't detect the player
19:01:31Domonokyseems like a bug :-)
19:01:37bluebrotherso if detection doesn't work bootloader installation won't work too.
19:01:50bluebrotherDomonoky: there was a report about this in the wiki (urgh) for ipodpatcher too.
19:02:29*bluebrother needs to release a new rbutil in the next days before RL gets too consuming
19:03:11JRoTi was able to get the bootloader on with sansa patcher but not with rbutil
19:03:23bluebrotherJRoT: also on wxp?
19:04:53Domonokydo the patchers work with a linux version of rbutil ?
19:04:55 Join ilgufo [0] (
19:05:17bluebrotherDomonoky: I never had problems with ipodpatcher / rbutil's ipodpatcher on linux
19:05:33bluebrotherunder the prerequisite to have the appropriate permissions for raw disc access of course
19:05:56n1spreglow: volume in sim works fine here
19:06:44 Quit jac0b-work ("ChatZilla [Firefox]")
19:07:57 Join miepchen^schlaf [0] (
19:08:09 Quit freqmod ("Konversation terminated!")
19:08:53 Part pondlife ("Gone")
19:09:35 Join mf0102 [0] (n=michi@
19:09:40 Join mirak [0] (
19:12:10mirakcan ipods record with rockbox or is it impossible because of the hardware ?
19:12:13***Saving seen data "./dancer.seen"
19:13:07 Part karashata
19:13:16 Quit proppy ("Leaving.")
19:14:12kratonatorvoice recording is impossible because the ipods don't have a mic
19:14:38kratonatordo you mean that?
19:14:45amiconnPaulJam: I have a suspicion what might be going on (re the disappearing time)
19:14:59amiconnDo you play format(s) that require boosting from time to time?
19:15:05amiconn(e.g. ogg vorbis)
19:15:41PaulJami used mp3, but the wps has peakmeters
19:15:47 Join lee-qid [0] (
19:17:12nicktasticWhat is 'boosting'?
19:17:32bluebrotherincreasing the cpu clock
19:17:42PaulJamin the audio debug menu i don't see boosting, so at least when showing the browser it shouldn't boost, and there the time dissappears fom the statusbar from time to time.
19:17:44nicktasticI see, thanks
19:19:04*Domonoky ha absolutly no idea why the patchers in rbutil would fail on WinXP.. but i also cant test them, so its difficult to debug..
19:19:56amiconnPaulJam: If you make it constantly boosted, the time should never disappear. Somehow I didn't notice this effect earlier :\
19:20:21 Join karashata [0] (
19:21:25rasherDomonoky: well, I can test, but I doubt I could help any more than that
19:21:47amiconnThe disappearing time means i2c read errors...
19:21:48PaulJamYes, when i increase the boost counter it doen't happen
19:21:48 Join obo [0] (n=obo@rockbox/developer/obo)
19:22:14Domonokyrasher, you could step through a detection with a debugger and see why it doesnt find it.. :-) ( big task)
19:23:18 Quit homielowe (Read error: 110 (Connection timed out))
19:25:01rasherNo thanks..
19:26:23 Quit mf0102 ("Verlassend")
19:28:12 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
19:28:31linuxstbpreglow: Yes, 7981 looks like an ASF parser issue. I'll post a comment asking for a sample file.
19:28:33 Join Crash91 [0] (n=evil91@
19:29:31 Quit BigBambi (Client Quit)
19:29:47 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
19:30:29preglowlinuxstb: excellent
19:31:15linuxstbmirak: Some ipods can record in mono via the right channel of the headphone socket, others can record via the line-in pins hidden in the dock connectors, others don't have any recording hardware at all.
19:31:27linuxstbamiconn: I'm around now...
19:32:05amiconnlinuxstb: I'd like to put some extra arrays into IRAM in mpegplayer for a few extra fps
19:32:23amiconnI know that you implemented seizing of the codec stack in test_codec
19:32:45amiconnWould be nice if mpegplayer would do the same for its audio thread stack, saving 9KB of IRAM...
19:32:59 Join darkless [0] (
19:33:01amiconnRight now the extra arrays only fit on X5, but not on H300
19:33:10amiconn(the ones in vlc.h)
19:33:23*preglow wonders if there's some way of coding codec plugins so that mpegplayer might reuse them
19:34:52 Join lazka [0] (
19:35:07linuxstbpreglow: That would mean no iram left for mpegplayer itself to use.
19:35:22preglowyes, extremely good point
19:35:40preglowhow does logf work with the sim again? i would have thought it all got written on console
19:37:04 Join webguest97 [0] (i=50d81efc@gateway/web/cgi-irc/
19:37:09linuxstbamiconn: Yes, I've always thought mpegplayer should/could do that.
19:38:19 Quit webguest97 (Client Quit)
19:38:23 Join webguest97 [0] (i=50d81efc@gateway/web/cgi-irc/
19:39:08amiconnPutting all those const arrays into iram gives ~10% speedup on cf
19:40:22webguest97any news from Zagor? :)
19:40:42linuxstbwebguest97: (the IRC logs...)
19:41:15 Join rotator [0] (n=e@rockbox/developer/rotator)
19:45:38rasherSkip frames in mpegplayer doesn't seem to maintain A/V sync - is this known?'
19:46:19linuxstbYou mean turning it off?
19:46:38rasherNo, turning it on doesn't help maintaining a/v sync
19:47:34linuxstbYou need both Skip Frames and Limit FPS to maintain A/V sync. But if mpegplayer can't decode in realtime (and drop enough frames), then you won't get A/V sync either.
19:47:51amiconnAnd the audio sample rate must be correct
19:48:05preglowwould you even want to _not_ skip frames?
19:48:11preglowor limit fps, for that sake
19:48:24amiconnOnly for testing....
19:48:34preglowthen why are they part of ordinary builds?
19:48:51rasherlinuxstb: I guess it's just too slow then, and limit fps isn't enough
19:48:52webguest97linuxstb: yeah.. Only hope that he will not be hit by a bus :)
19:49:50linuxstbpreglow: I guess just because mpegplayer has slowly become usable...
19:50:31preglowshould consider #ifdeffing them away, if only because they might confuse users
19:50:32 Join moos [0] (
19:51:28rasherThen there's barely any point in having the menu at all
19:52:51linuxstbButchers... ;)
19:52:59preglowi'm quite sure we'll end up having a menu, but if there's no need for one...
19:53:25moosamiconn: about this disapearing time; here I noticed it when the HD is working
19:53:26rasherSlasheri: if you get the chance, could you have a look at possibly implementing the needed stuff for FS #7984 (talk clips in Database) to work?
19:53:46preglowSlasheri: plus fixing dircache needlessly reiniting itself in the foreground all the time
19:54:08amiconnpreglow: I like the menu, for testing both speed and appearance
19:54:24rasheramiconn: testing appearance?
19:54:37amiconnI think it would be sufficient to have _one_ switch, could be called "test mode"
19:54:38preglowamiconn: good, as long as the menu has some reason for being, it should be there. but i don't think stuff primarily meant for testing should be compiled in by default
19:54:52amiconn...which would disable frame skipping, frame limiting, and enable fps display
19:55:12amiconnrasher: Check how good/bad it looks with the current frameskipping
19:55:48amiconnIf the start menu stays, the setting could be put there, getting rid of the second menu
19:55:54 Quit Crash91 ("Bye Bye!")
19:55:55 Quit Domonoky ("Trillian (")
19:56:01preglowi sure hope the start menu doesn't stay :/
19:56:06rasherPlease no
19:56:20amiconnpreglow: I consider mpegplayer work-in-progress
19:56:21rasherOnly if there's a bookmark for that file, surely
19:56:47preglowamiconn: i consider rockbox work in progress, doesn't change much, users somehow expect it to work
19:57:35preglowperhaps some kind of TEST_BUILD #define would be cool
19:57:47linuxstbI agree with amiconn - I would like to see the options unified in a "test mode" setting. Users shouldn't then be confused, and it's still convenient for devs.
19:58:01preglowwell, it certainly is better than the current thing
19:58:09linuxstb(and for users who we ask to make benchmarks)
19:58:25pixelmaI was surprised that there are now behind a "display options" menu entry... when did that get in?
19:58:37linuxstbThere's a display options menu?
19:58:46rasherThere's a display options menu.
19:58:48kkurbjunamiconn: were you working on a resume feature for the 1st-3rd gen ipods?
19:58:49amiconnAsk roolku...
19:59:23amiconnkkurbjun: I need to, but only collected information so far :/
19:59:35amiconn..and made some proof-of-concept tests
19:59:35preglowlinuxstb: thanks for test_codec, btw, i'm hunting a bug that happens 5 minutes into a file and without speed test, this would have been seriously boring :P
19:59:50linuxstbHmm, no Rockbox header and (C)/license info in mpeg_settings.[ch]
20:00:18 Quit BigBambi ("Please insert girder")
20:00:33 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
20:01:05kkurbjunwhat were you thinking in terms of a way to suspend rockbox? Have a function that suspends operation, and resumes, but doesn't go through shutdown?
20:01:18Learpreglow: re dircache foreground scan, he is aware of one issue, and has ideas on a fix at least (occurs next boot if you start playback too soon in some cases).
20:01:40preglowLear: goodie
20:01:49preglowgeh, it seems saratoga has reformatted some of the wma code
20:01:54preglowthat won't exactly aid merging
20:02:05linuxstbreformatted compared to what?
20:02:24linuxstbWell, ffmpeg's wma codec has changed a lot since the Rockbox fork.
20:02:38 Join bertrik [0] (
20:02:45preglowdoesn't look like it to me, looking at svn log
20:02:59preglowi assume he synced the code during gsoc
20:03:03linuxstbDidn't saratoga carry on marsdaddy's work, rather than starting with the current ffmpeg? So it's a couple of years old.
20:03:04preglowhe said he did
20:03:08amiconnkkurbjun: First of all, the shutdown needs to be reworked so that it's message based. Every thread needs to do its own husekeeping, and no thread should kill itself
20:03:31preglowlinuxstb: yeah, but i did tell him to resync, and he said he did
20:03:36amiconnThe ipods have suspend-to-ram, so rockbox would then just continue to run when woken up
20:03:40preglowlinuxstb: might have been more manual than i thought, though
20:03:52amiconnAnd btw, I will *only* do this for 1st/2nd gen
20:04:08amiconn...since those have *no* hardware shutdown
20:04:19preglowlinuxstb: you think it would be worth trying to edit it back to ffmpeg's looks?
20:05:25 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
20:05:40n1samiconn: I think you could save a small amount of iram by moving the three tables in header.c to dram, two of them are only used in mpeg2_idct_init() which ifaict is only called once and one is only used in mpeg2_header_sequence() so it could be put on the stack if it really needs to be in iram.
20:05:43preglowlinuxstb: yeah, he can't have meant a complete resync, some of the type names are different and everything
20:05:50 Join Konam [0] (
20:06:16Konamis there a way to change -or turn off- the rockbox presentation screen?
20:06:23kkurbjungigabeats have a suspend also, we could cut the power to the ram when we shutdown, but it's not being done right now. What housekeeping were you thinking for the threads?
20:06:37n1sbtw those two other tables should be const
20:06:44kkurbjunamiconn, you don't think it should be an option for players that have the hardware to do it?
20:06:51amiconnn1s: I need ~2.5KB to put the tables in vlc.h into iram
20:07:09linuxstbpreglow: Yes, it's probably worthwhile. Are there any known issues left with the decoder?
20:07:20 Quit atsea-22 (Read error: 104 (Connection reset by peer))
20:07:22n1swell, that will give you a whooping 256B ;-)
20:07:22amiconnkkurbjun: I see no point in wasting battery for suspend if there is true hw shutdown
20:07:27linuxstbKonam: What do you mean by "presentation screen" ?
20:07:33preglowlinuxstb: well, it crashes on my only wma file
20:07:39preglowlinuxstb: five minutes into the bastard
20:07:49amiconnn1s: Haha. Did you check how much spare iram we have atm on H300 and H1x0 mpegplayer?
20:08:01amiconnAround 160 bytes iirc
20:08:15Konamlinuxstb the rockbox screen that's showed before presenting the main menu
20:08:17n1sooh, that's a tight fit indeed
20:08:53kkurbjunamiconn: for me it would be useful since I regularly pause my player for 10-20 minutes, the gigabeat is refreshing ram anyway and it would be nice not to have to restart everytime spinning up the disk unnecessarily, but saving battery power
20:08:53amiconn156 bytes to be precise
20:09:18amiconnkkurbjun: Eh? Does the gigabeat not have true shutdown?
20:09:50kkurbjunit could be shtdown further, but we don't cut the ram refresh when it's shutdown
20:10:17amiconnWhy not?
20:10:38kkurbjunbut I could see it as an option for people that would like a fast restart.. I don't think anyone's noticed the battery drain
20:11:34amiconnThe 1st/2nd gen ipods not having true shutdown is in fact nasty. You need to recharge every ~4 weeks even if you don't use it at all
20:11:54rasherWell done indeed, Apple
20:12:07preglowi wonder why sim says it opens "config.cfg" each time i change the volume...
20:12:11amiconnNot on my 1st gen atm though - it's lying on my table with one battery wire cut for measurements
20:12:27amiconnWhen not measuring, I leave those wires open of course :)
20:12:34kkurbjunthat is a bit annoying, I've never tested to see how long the gigabeat can sit refressing ram
20:12:53 Quit spiorf (Remote closed the connection)
20:12:58amiconnpreglow: Sim doesn't cache config writes - it writes right away
20:13:05Konamlinuxstb the screen shows the rockbox logo: I want to change it or turn it off
20:13:32rasherKonam: the answer you were given last time still holds
20:14:07preglowbah, i botched the wma seek fix
20:14:10rasher<PaulJam> Konam: it is possible to change it, but this requires changing the sourcecode and compiling rockbox
20:14:12linuxstbamiconn: You probably have 2.5KB on the main stack.
20:14:39Konamrasher what answer? I asked before here but my machine crashed before I get that answer
20:15:14Konamcan you repeat it to me rasher?
20:15:24rasherKonam: Then you could have checked the irc logs.
20:15:30rasherI just did...
20:16:03linuxstbamiconn: There used to be a 4.5KB (IIRC) libmad buffer on the main stack, but it was removed because the stack was overflowing in some situations.
20:16:08amiconnlinuxstb: Those tables are compiled into slice.o from vlc.h. I can't see how I would put those on the stack without major changes in slice.c ...
20:16:21Konamok, thanks rasher
20:16:34preglowlinuxstb: btw, think you mpegplay.c code for uisim can be deleted now, or will it be used again?
20:16:41kkurbjunif you have some pointers on how you invision the implementation though I would be interested in helping.
20:16:45 Quit Konam ("Leaving.")
20:16:45 Quit krazykit ("bah work")
20:17:33linuxstbpreglow: Was that me?
20:17:38preglowlinuxstb: it has your copyright
20:18:04amiconnkkurbjun: The shutdown and the suspend should work message based. Pressing the power off button combo would send either SYS_POWEROFF or SYS_SUSPEND (depending on target)
20:18:37amiconnBoth messages would cause all threads *except the ata thread* do their housekeeping and signal completion when done
20:18:37 Quit shnee ("Konversation terminated!")
20:18:51kkurbjunahh, I see, similar to usb events?
20:18:54amiconnWhen all threads are ready, the ata thread starts to do its housekeeping
20:19:10linuxstbpreglow: I guess it can...
20:19:19amiconnAnd finally, when the ata thread is done, the power is cut or the suspend function is entered
20:20:02amiconnWhen waking up from suspend, a SYS_WAKEUP message would be sent that tells all threads that they can continue
20:20:24linuxstbpreglow: I'm doing it now...
20:20:31kkurbjuncould it be a SYS_SUSPEND message universsally, and then the target handles it appropraitely if it supports suspend and the user wants it to suspend (instead of shutdown)?
20:20:49preglowwrong window :/
20:20:52kkurbjunor are you completely against the idea of supporting suspend on targets that have a hardware shutdown?
20:21:02amiconnProblem with that approach is that (1) since I want the powermgmt thread to be the "master" of all this, it needs to be restructured, so that the main loop is outmost, and the function(s) that are called every minute are called from there
20:21:38amiconnThis would have the additional advantage that the target specific code (e.g. software controlled charging for recorder v1) can be easier separated into target tree
20:22:31amiconnAnd (2) since there will be a couple more messages, either all threads need to handle them, adding quite some code to threads which don't need to care, or we need some sort of "event registration"
20:23:33linuxstbpreglow: It's gone.
20:23:33amiconnAnd of course, USB housekeeping should be the same. Right now it's hit-and-miss whether the ata thread is finished before some other thread that still has some data to save...
20:25:02 Quit kratonator ("CGI:IRC (EOF)")
20:27:01kkurbjunhmm, I'll have to look into the code a bit more so that I can get a grasp of what all that entails, your plan seem very complete though
20:31:28 Nick smurf is now known as MatthiasUrlichs (n=smurf@debian/developer/smurf)
20:34:30preglowlinuxstb: cool
20:34:41preglowlinuxstb: seems like ffmpeg current just ignores the thing that makes me error now
20:36:52rasherMan, I can't even figure out where the Database menus are handled.. guess I should just give up
20:37:09pixelmapreglow: had to read your last commit message a few times before understanding it (which nouns belonged together) :)
20:37:17preglowpixelma: sorry about that :P
20:38:09 Join merbanan [0] (n=banan@
20:38:35amiconnThe bug that works :>
20:40:54 Join MethoS- [0] (
20:43:14 Quit Rob2222 ()
20:43:30 Quit Gnu47 (Nick collision from services.)
20:43:38 Join Gnu47 [0] (
20:44:11Slasherirasher: those are handled in tagtree.c
20:45:07rasherSlasheri: did you see what I'm wanting to do?
20:46:22Slasherirasher: not sure about that yet :)
20:47:24Slasheriah, looks interesting
20:48:09 Join atsea-22 [0] (i=atsea-@gateway/tor/x-ff14e054a3cb8b96)
20:49:01rasherSlasheri: if you could help out (preferably with code - this is really over my head), that'd be great
20:49:03 Quit jhulst ("Konversation terminated!")
20:49:37Slasherirasher: sure, what kind of knowledge do you need? =)
20:51:37rasherWell... more or less a pointer to where I'd need to look, to detect when a new line is selected...
20:51:59 Quit chris_mt (Read error: 104 (Connection reset by peer))
20:52:33Slasherihmm, new line is probably handled in tree.c somewhere
20:52:41Slasheridon't ask me, i have no better idea either :D
20:54:20preglowooooouch, wmadeci.c uses 64x64 -> 64 divides
20:54:23preglowsounds fast ...
20:55:35rasherSlasheri: does it use dirbrowse?
20:56:26 Nick MatthiasUrlichs is now known as smurf (
20:56:27 Quit lazka (Remote closed the connection)
20:56:27amiconnpreglow: very ;)
20:56:43preglowi can imagine arm _loves_ those
20:57:22 Quit Galois (Read error: 104 (Connection reset by peer))
20:57:31Slasherirasher: yes, most likely it does
20:58:37amiconnpreglow: Somewhere in the decoding loop or just for setup?
20:58:45preglowamiconn: in decode_block
20:58:56markunkkurbjun: any idea how much mA we would gain by turning off RAM? Right now we use 0.71mA in sleep mode
20:58:57preglowdoesn't get run too often, though
20:59:41amiconnmarkun: I think that's what the ram needs in self refresh
21:00:07markunwow. so perhaps we would be very close to 0
21:00:09amiconnThe ipod draws ~1.2mA in suspend, that's ram self refresh plus the PP running at 32kHz
21:01:31amiconn<10uA is what I would expect from a half-decent, current dap
21:02:19rasherI really can't find heads nor tails of this..
21:02:34preglowamiconn: that's not enough for ram by far
21:02:52amiconnI mean when powered off
21:03:00preglowpowered off != suspend
21:03:03amiconnNo ram self-refresh
21:03:10|Rain|I've not measured its power consumption, but the beta 1.03.x sansa OF also suspends instead of turning off completely (though it'll still power off if you hold down the power button for long enough)
21:03:14amiconnI want poweroff, not suspend
21:03:37preglowi want both
21:03:44preglowi like the apple scheme, really
21:03:50preglowsuspend when idling
21:03:56preglowthen power off after an hour or so
21:04:08preglowwhen i press power off, i expect it to power off, of course
21:04:36markunpreglow: I like suspend/resume, but an option to shutdown would be great too
21:06:19preglowyeah, like i say
21:06:28 Join spiorf [0] (
21:07:54amiconnI'd rather focus on fast boot than suspend
21:08:18preglowi don't think i would
21:08:25preglowsuspend sounds like much more fun
21:12:14***Saving seen data "./dancer.seen"
21:12:22 Join nardul [0] (
21:12:54 Join sarixe [0] (
21:13:06sarixehey, i'm getting some ugly compile-time errors
21:13:16sarixei'll pastebin
21:13:21amiconnpreglow: Suspend is for the lazy devs who don't want to optimise boot speed ;)
21:13:44XavierGrdoes anyone know a copy command for linux that will not prompt me to overwrite files?
21:13:51rasherXavierGr: cp -f?
21:13:55nardulDoes anyone know if there's development going on for the 5'th generation Ipod CPU usage / battery lfie?
21:13:55XavierGrtried that
21:13:56preglowsounds to me like making suspend work isn't exactly any easier than optimizing boot speed
21:13:59*amiconn thinks he found the reason for the pcf i2c glitches
21:14:05rasherXavierGr: rather, stop using cp -i
21:14:30rasherXavierGr: it's probably aliased... "alias cp"?
21:14:49XavierGrmaybe let me check
21:15:00XavierGryeah it is :)
21:15:07sarixei usually just have a cron'd atuomatic build script running
21:15:08XavierGrdidn't have a clue
21:15:25sarixebut i noticed that my build version didn't change on bootup
21:15:28zesupplying -f should've overwridden it anyway though, afaik
21:15:31rasherimho, -f should still have overwritten even when using -i, but I guess that's debatable
21:15:49sarixeso i ran the compile manually, and it gave me what you see in the pastebin
21:15:58XavierGrthose kind of preset aliases are quite annoying (though they try to play it safe)
21:17:56XavierGrkkurbjun: ah that's why when I turn on my Gigabeat F after some days it is empty even after a full charge since then?
21:23:23sarixeuh, anyone know why i'm getting these compiler errors?
21:25:18kkurbjunXavierGr: that could be the RAM refresh running
21:25:38XavierGrso the gigabeat never turns completely down?
21:25:40markunin theory it would be half full after 24 days
21:26:10markunif it was fully charged when you turned it off
21:26:26rasherSlasheri: any way the database could include/exclude tracks based on the presence of the removable media they are on? So if I remove the microsd card, database won't show songs found on that card
21:26:40markunkkurbjun: do you have a ampere meter connected to your gigabeat?
21:26:47XavierGrcharge it, leave it there off (after I turned it on once) then after a month (or half a month) it was nearly empty)
21:27:03markunXavierGr: yes, that's normal
21:27:04amiconnrasher: The old volume id idea for the database...
21:27:13Slasherirasher: that's not yet possible..
21:27:37rasherSlasheri: Yeah, I sortof figured that - I meant it it is relatively feasible
21:28:11rasherBecause it would be rather neat
21:28:12 Quit sarixe ("(EE) Failed to load "quit" module")
21:28:32Slasheriwell, there could be an extra tag for the volume info or a separate db on the removable media
21:28:47kkurbjunmarkun, no I don't have a setup to measure it, all of my stuff is still in storage
21:29:06markunforgot about that
21:29:38kkurbjunlooks like I might be getting it soon though, might move out of storage this weekend :)
21:30:53kkurbjunwe could shutdown properly until a suspend feature is properly implemented..
21:31:32rasherSlasheri: having it as simply a tag would mean less work on insert/remove, but slightly more work for each search, I assume?
21:31:59markunkkurbjun: yes, that's what I was thinking
21:32:05pixelmarasher: you have an e200 now?
21:32:13rasher(cpu-wise, rather than code)
21:32:20markunbtw, do you ever get the battery empty warning when it's still full and you want to turn on the gigabeat?
21:32:27rasherpixelma: Yeah, I sold my h120 and had a sansa flown in
21:32:51markunmaybe something wrong with the way we turn off the ADC?
21:32:59Slasherirasher: somewhat, yes. and more work for update too if there is no media present during the update
21:33:03kkurbjunit would break the fast resume, but I guess enabling ram refesh could be added to teh patch. I've never seen that
21:33:35 Join sarixe [0] (
21:33:45sarixesorry, my router had to be reset
21:33:45kkurbjundo you just have to try turning it on again?
21:33:48rasherHrm.. yes.. I guess having the db on the card makes most sense - since it's more likely to be up to date then - especially if you have multiple cards
21:34:04markunI haven't notice it for a while
21:34:05Learsarixe: make clean
21:34:14pixelmarasher: I'd say database should handle a simple manual "update now" with removeable data first :\ (at least it didn't work correctly on my Ondio and jhMikeS confirmed on his e200 a while ago)
21:34:17markunmaybe it got fixed by accident
21:35:46pixelmarasher: try initialising the database with inserted card, then remove it and hit "update now"
21:37:01sarixeLear : thanks
21:37:48rasherpixelma: looks correct..
21:38:15 Quit Seed (Nick collision from services.)
21:38:18rasheralmost - a single leftover from the card
21:38:21 Join Seedy [0] (
21:38:23 Quit sarixe (Client Quit)
21:38:32rasherWhich is also quite odd
21:38:58|Rain|I think I've had that problem without an SD card
21:39:12|Rain|but I haven't put much time into figuring out the details
21:39:19pixelmafor me there was no progress at all in rebuilding the database but I'll try again
21:40:48 Join toffe82 [0] (n=chatzill@
21:44:00 Quit nardul (Remote closed the connection)
21:44:15 Quit obo ("bye")
21:46:03 Join Seed [0] (
21:48:30 Quit desowin ("use linux")
21:48:37pixelmait's still the same (statusbar constantly shows "disk" activity and "View database info" says progress:-1)
21:48:44 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
21:49:07rasherpixelma: I guess it just depends on other factors
21:49:37pixelmasuch as?
21:49:43rasherNo idea
21:50:14rasherBut it happens for you and jhMikeS, and not for me, so it must be more complicated than it simply not working
21:50:50 Quit styleism (Read error: 110 (Connection timed out))
21:56:01amiconnIt seems that buffered writes might not be that good an idea for the gpio area :\
21:56:25amiconnI can get the pcf driver to work properly by sprinkling it with nops
21:56:32amiconn(which sync the pipeline)
21:57:06amiconnBut the question is whether there might be similar dangling problems in other drivers
21:57:48amiconnI better disable buffered writes for gpio, even though that means that the iriver/iaudio remote lcd drivers will be back to their old speed
21:58:14amiconnMain lcd will retain its speedup, as I won't disable buffered writes for it
22:00:24webguest97Zagor: Tell us how you're doing! :)
22:00:43keanuwebguest97, Zagor isn't on ;)
22:00:58barrywardellwebguest97: I'm sure he'll let us know when he has made more progress
22:01:43webguest97I know, I'm just eager
22:02:35n1swell, constantly asking for progress doesn't magically speed it up :-/
22:02:45 Quit Seedy (Read error: 110 (Connection timed out))
22:02:49 Quit ilgufo ("So Long, and Thanks For All the Fish -")
22:02:52barrywardellif anything, it slows him down
22:04:03 Join styleism [0] (
22:04:12 Quit styleism (Client Quit)
22:07:06 Quit Toxicity999 (Read error: 104 (Connection reset by peer))
22:08:02 Nick Tanuva is now known as Tanuva|Zzz (n=tanuva@
22:10:47 Join Toxicity999 [0] (n=bryan@unaffiliated/Toxicity999)
22:11:02 Quit Seed (Read error: 110 (Connection timed out))
22:12:38 Join Domonoky [0] (
22:14:41 Join Seed [0] (
22:15:39 Quit amiconn (Nick collision from services.)
22:15:45 Join amiconn [0] (n=jens@rockbox/developer/amiconn)
22:21:06 Quit webguest97 ("CGI:IRC")
22:27:50 Quit karashata ("Leaving.")
22:29:29 Join webguest29 [0] (i=45f66aca@gateway/web/cgi-irc/
22:30:02webguest29Hi, is anyone there?
22:30:26markunlots of people
22:30:50webguest29Markun, can you help me on how to play games using rockbox?
22:31:07markundepends, I might be able to point you in the right direction
22:31:17markundo you have any question in particular?
22:31:44webguest29Well, I'm trying to run a .GBA type file on it, but it doesn't run when i click it on the menu
22:31:50webguest29Is there a plugin or something?
22:32:08markunno, .gba is not supported
22:32:18webguest29only .WAD?
22:32:35markunwebguest29: .gbc is supported for example
22:33:02markunalso, we have a few native games:
22:33:09kkurbjunwebguest29: wad and gba are completely unrelated, wads are for doom, gba is for a gameboy advance emulator
22:33:18 Quit merbanan (Remote closed the connection)
22:33:26kkurbjunwe have a gameboy/gameboy color emulator
22:33:47markunand someone is working on NES
22:33:48webguest29How would i run a gb/gbc emulator on rockbox?
22:33:55markunjust click on the file
22:34:04markunlike you tried with the .gba file
22:34:04kkurbjunwhich includes .gb and .gbc, I think it also include .sgb but I've never tried it
22:34:26webguest29Any specific folder i have to put it in?
22:34:41|Rain|just something you can find in the file browser
22:34:52markunwebguest29: nope
22:35:04webguest29and running .WADs?
22:35:19kkurbjunthey have to be in a specific folder
22:35:22markunwebguest29: a bit more difficult:
22:35:50webguest29is sandisk/games/doom/addons right?
22:35:53markunwebguest29: but if you like games a lot, maybe you should look for a handheld games console
22:36:06markunPSP or GP32X
22:36:54markunGP2X or GP32, my bad
22:37:26webguest29ah so running gb/gbc games is just a one click thing?
22:37:47markunwebguest29: just try it, if you still have problems just ask again
22:37:49*amiconn stared on his H300's status bar clock for 10 minutes now
22:37:55amiconnNo more glitching
22:38:19webguest29Any websites you recommend for the ROMs?
22:38:28markunno, we don't recommend roms here
22:38:50webguest29After i find a rom i just copy and paste it into the sansa drive?
22:39:06markunI think we've answered your questions
22:39:25 Quit Lear ("ChatZilla [Firefox]")
22:39:34markun(at least this one)
22:40:00webguest29well ima go find a rom and get back with you if it works or not
22:41:27 Join krazykit [0] (
22:42:35 Quit nicktastic ("Leaving")
22:42:43toffe82markun: did you see for the battery of the F in the S ?
22:43:19markunI don't understand the question
22:45:57amiconnPaulJam: Just committed the fix for the disappearing time...
22:46:02webguest29uh markun
22:46:08webguest29how do you exit from the gb/gbc game?
22:46:33markunwebguest29: depends on the player. Check the rockbox manual, do you know where to find it?
22:46:33PaulJamamiconn: thanks :)
22:46:51JRoTgive me a sec will loook for you
22:47:17webguest29my player is a sansa e250 and no i don't.
22:47:46|Rain|try a short press of the power button, and
22:47:54JRoTpuss the power button webguest29
22:48:08JRoTyou also can save and load in there
22:48:19webguest29Thanks lRainl and JRoT
22:49:57JRoTanyone know if someone is working on using the sansa as msc under rockbox
22:50:15|Rain|brand new code to do that is due Real Soon Now
22:50:27JRoTthnx |Rain|
22:50:51n1syou forgot (tm) after that!
22:51:09|Rain|I violate trademarks with gusto!
22:51:12krazykitJRoT, yeah, it's being worked on.
22:52:09 Nick parafin is now known as parafin|away (
22:52:29JRoTthats really nice, cause thats the only thing i'm missing on rockbox(tm)
22:53:20|Rain|a lot of people are very anxious to see it happen
22:53:36krazykitat the rate it's being coded, it appears that it'll happen soon. don't worry, there will be announcements about it.
22:53:45 Join ompaul [0] (n=ompaul@gnewsense/friend/ompaul)
22:54:28 Quit blargit (No route to host)
22:54:33webguest29Does anyone know what [ C ],[ ! ], ( J ) means after the title of the Rom?
22:55:02bertrikCan anyone have a look at my patch for a bug I found in the playback control menu?
22:55:15PaulJamit means you didn't make them yourself and they are illegal
22:55:23 Join Zagor [0] (
22:55:47webguest29pauljam, thats to me right?
22:56:00 Join billytwowilly [0] (
22:56:09webguest29Are they still playable?
22:56:28n1swebguest29: please check the irv guidelines linked in the topic
22:56:40 Join donutman25 [0] (n=dagni@
22:59:34 Join petur2 [0] (n=petur@rockbox/developer/petur)
23:00:09JRoTanyway to show pictures in landscape mode like it does with video? (sansa)
23:00:44markunautorotation would be nice
23:01:18webguest29awh man, as i hold down my button to boot it up with rockbox, it doesn't work anymore, any suggestions?
23:02:30JRoTdon't hold it, just press power
23:02:58 Part smurf ("♩♬♯♪♩♭♫♮♩")
23:03:31webguest29it boots to my sansa's main menu either way
23:04:37 Join ramon8 [0] (
23:04:40 Quit WalterEgo ()
23:09:15 Nick petur2 is now known as petur (n=petur@rockbox/developer/petur)
23:09:15 Quit webguest29 ("CGI:IRC (EOF)")
23:12:19***Saving seen data "./dancer.seen"
23:18:17 Quit darkless ("Leaving")
23:23:08 Join sup [0] (i=1000@
23:24:13 Join BonBonTheJon [0] (n=bonbonth@
23:24:38 Join midgey [0] (
23:27:24 Part Domonoky
23:30:50 Quit hannesd ("Client suicide")
23:31:54 Join Galois [0] (
23:32:20 Quit midgey ()
23:37:16 Quit n1s ()
23:38:02 Quit bertrik ("bye")
23:43:54miraklinuxstb: is that random or dependent of the generation ?
23:44:47*preglow wonders if most wma files use noise coding
23:45:09mirakamiconn: thanks for implementing the coldfire idct
23:45:48mirakamiconn: it seems it's possible to optimise it, somebody says that. maybe you changed some stuffs
23:46:06courtcpreglow: probably for low bitrates.
23:46:20amiconnI optimised whatever I found, without changing the basic principle
23:46:23 Join Rob2222 [0] (
23:46:43amiconnIt's basically a reimplementation of your idea, with all-asm functions
23:47:51 Quit ramon8 (Read error: 110 (Connection timed out))
23:49:06 Join safetydan [0] (i=dc9d468b@rockbox/developer/safetydan)
23:51:46preglowsafetydan: yo, could you sync your remove hw eq patch?
23:51:48courtcpreglow: you can see when libavcodec/ffmpeg uses noise coding starting on line 133 of
23:52:39courtclooks like just for very low bitrates.
23:53:00preglowcourtc: pretty low bitrates
23:53:12preglowok, so i need some low bitrate wma files
23:55:22safetydanpreglow, not until I get home unfortunately and I just started work so that's at least eight hours away
23:55:33preglowsafetydan: no rush, mate
23:56:28safetydanpreglow, you planning on making it work?
23:58:21preglowsafetydan: do you have a video, btw?
23:58:38preglowsafetydan: i'm planning on commiting it and adapting the cutoff freq fucntionality to other targets

Previous day | Next day