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

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

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

#rockbox log for 2005-07-17

00:06:26 Quit Harpy (Read error: 110 (Connection timed out))
00:09:57 Join webguest87 [0] (
00:10:09webguest87Hello all
00:11:08webguest87a quick question for the Bagder administrator :)
00:12:04webguest87why in the CVS buil table, the 2 colums of the Gmini port still apear ?
00:12:19Bagderbecause the columns are based on more data than what is shown
00:12:21webguest87i know it's not important
00:12:52webguest87are will always here?
00:13:19webguest87i know this port is in "stand by"
00:14:25BagderI would call it abandoned
00:14:48webguest87:) anyone worked on it since long time?
00:15:00Bagderbeen months since
00:15:31webguest87there was just one dev worked on it?
00:16:08webguest87difficult to port Rockbox alone, no?
00:16:20Bagderhis computer broke down
00:16:31Bagderand he couldn't afford a new one
00:16:48webguest87not good for him
00:17:09webguest87personaly i'm iriver user
00:17:21webguest87one very happy user :)
00:18:18webguest87i read in the logs of today maybe Slasheri'll go to imlement paek meter things
00:18:27webguest87it's very good
00:19:31webguest87i love your Rockbox fw :)
00:20:25webguest87and i love his "association" with the GPL concept
00:20:45webguest87not money way
00:20:54webguest87and free to code
00:22:39webguest87Bagder: think you in the futur (soon i hope), it'll be possible to have got a beep option like the original fw or most of music players?
00:22:52webguest87it's not a request just question :)
00:23:00Bagderwhat would that do?
00:23:17BagderI mean, when would it beep?
00:23:28Bagderbut yes, surely we can make it beep
00:23:39webguest87when pushed a key
00:24:08BagderI always disable such stuff
00:24:13webguest87like mostly player music
00:24:14Bagderon every device I own
00:24:44webguest87you don't like it?
00:25:03BagderI hate it
00:25:22webguest87:), scuse for this evocation in your brain :)
00:25:35Bagderwell, that's what options are for
00:25:52webguest87but it be very usual for lot of people i assume
00:26:24webguest87think you it'll be dificult to add to options?
00:27:08BagderI would guess that the most difficult part is to find someone who'll write the code for the feature
00:27:58webguest87yes i assume, just if one dev work on it, needed lot of codes?
00:28:08webguest87lot of time?
00:28:16webguest87for him
00:28:31Bagderthat would depend on the coder of course
00:28:40webguest87:) of course
00:29:24webguest87for example, you, you are an experimented hackers
00:29:37Bagderno, it would not take very long
00:30:45Bagdersleep time
00:30:58webguest87have a good night
00:31:40webguest87thanks for your answers
00:32:44webguest87good night to all too :)
00:32:48 Quit webguest87 ("CGI:IRC")
00:34:35 Quit matsl (Remote closed the connection)
00:37:55 Join matsl [0] (
01:01:45 Join CBM-away [0] (
01:09:12***Saving seen data "./dancer.seen"
01:37:48XavierGrwhen a static variable is declared on a a1.c file does it get recognised if called by a a2.c file (included the a1.h)?
01:39:04XavierGralso if someone could tell me where the tree initialization occurs when running rockbox I would be obligued
02:26:01 Quit matsl (Remote closed the connection)
02:32:02 Quit Moos (" Like's GUI? Then try HydraIRC -> <-")
02:41:10XavierGrgood night all! I have to sleep!
02:41:13 Quit XavierGr ()
03:07:53 Quit Sucka ("a bird in the bush is worth two in your house")
03:09:16***Saving seen data "./dancer.seen"
03:10:21 Join textchimp [0] (
03:15:20 Join ashridah [0] (
03:28:29 Quit Coldtoast (Read error: 110 (Connection timed out))
03:54:14 Quit hicks (Remote closed the connection)
03:54:40 Quit cYmen_ ("zZz")
03:57:39 Nick CBM-away is now known as CheeseBurgerMan (
04:05:34 Join QT_ [0] (as@area51.users.madwifi)
04:16:03 Quit QT (Read error: 110 (Connection timed out))
04:16:03 Quit textchimp (Read error: 104 (Connection reset by peer))
04:24:02 Quit CheeseBurgerMan ()
04:24:35 Join CheeseBurgerMan [0] (
04:43:05 Nick CheeseBurgerMan is now known as CBM (
05:09:18***Saving seen data "./dancer.seen"
05:24:56 Nick CBM is now known as CheeseBurgerMan (
05:33:37 Nick CheeseBurgerMan is now known as CBM-AFK (
05:36:28 Nick CBM-AFK is now known as CheeseBurgerMan (
07:09:20***No seen item changed, no save performed.
08:19:41 Quit pike (Read error: 110 (Connection timed out))
08:27:59 Join pilled [0] (
08:37:53 Quit pill (Read error: 110 (Connection timed out))
08:37:53 Nick pilled is now known as pill (
08:52:39 Join Coldtoast [0] (
08:53:59 Nick CheeseBurgerMan is now known as CBM-away (
09:09:24***Saving seen data "./dancer.seen"
09:35:25 Join pike [0] (
09:42:30 Join Harpy [0] (
11:05:32 Join Lear [0] (
11:09:28***No seen item changed, no save performed.
11:09:40 Quit edx ()
11:13:28 Join cYmen [0] (
11:24:23 Join matsl [0] (
11:30:47 Join edx [0] (
11:31:10 Quit cYmen (Read error: 104 (Connection reset by peer))
11:31:13 Join cYmen [0] (
11:37:34SlasheriHmm, my current implementation of the peak meter looks more like a spectrum meter..
11:38:28LearToo fast?
11:39:38Slasherinope, it meters one fixed frequency.. Maybe something like 4 kHz
11:39:50Slasherii will try to fix that
11:40:01Slasherifor example it igrones bass totally
11:40:34LearAren't you just looking at peak sample values?
11:41:10Slasherihmm, we dont have peak values but we have to calculate tehm?
11:41:54 Quit cYmen (Read error: 104 (Connection reset by peer))
11:41:54LearHow else would you do it?
11:43:19SlasheriHmm.. currently i try to find min and max sample values between a specified period and then calculate average peak value for a longer period
11:45:39LearDoesn't the peak meter do some averaging as well?
11:45:51Slasherimaybe i should first find the bias value and then base the peak readings on that
11:46:00Slasherithen it might work better
11:46:39Slasherimaybe, but it takes samples only ~20 times per second
11:46:53Slasheriso we have to calculate some average also
11:49:42Slashericurrent implementation works well for higher frequencies but its more an eq analyzer than a peak meter
11:50:06 Join cYmen [0] (
11:50:15Slasherii will try that different bias thing when i get to home
12:07:23 Quit Seed (Nick collision from services.)
12:07:30 Join Seed [0] (
12:25:03 Join LinusN [0] (
12:29:35 Join austriancoder [0] (
12:48:05 Join Moos [0] (
12:48:18MoosGood Day all
12:50:39austriancoderLinusN: we should maybe rename i2c-h100 to i2c-coldfire?
12:53:14LinusNone of those days, yes
12:54:38MoosHmm, anyone of you 2, planed to improve the i2c patch? :)
12:54:50LinusNone of those days, yes :-)
12:54:55austriancoder g
12:55:17austriancoderif i have the bdm in my hands and got it wokring
12:55:51LinusNreally? cool!
12:56:09LinusNwas it hard to find out where to connect it?
12:56:34LinusNoh... "if i"... :-)
12:56:53*LinusN cleans his glasses
12:57:08austriancoderjep.. i am looking where to connect the bdm today
13:01:22LinusNshould be pretty straightforward, as the cpu pins are visible
13:02:29austriancoderi hope that the palces to connect are marked
13:03:58LinusNyou still haven't scanned the pcb's?
13:05:04LinusNreboot, brb
13:05:07 Part LinusN
13:06:41LearInteresting... At least since September, the ID3 screen hasn't shown all information. Doesn't anyone use it? :)
13:07:08 Join LinusN [0] (
13:07:58austriancoderi have scanned it, but as the pcb is not laying straight on the scanner it gets blurred
13:08:03MoosLear: all informations?
13:08:18LinusNaustriancoder: aha
13:08:31LearYeah, some info, like bitrate, frequency and path aren't possible to display.
13:08:48Moosi see it
13:09:05austriancoderLinusN: i must look if i could get better results
13:09:21LinusNwould be good, as we all could see the pcb's
13:09:30***Saving seen data "./dancer.seen"
13:10:16 Part LinusN
13:11:28 Join LinusN [0] (
13:11:33austriancoderand everybody could help to find the bdm connectors ;)
13:12:52austriancoderits time to setup my whole laptop as linux-only-zone
13:13:34austriancodertime to go
13:13:36austriancodersee you
13:13:40 Quit austriancoder ("CGI:IRC 0.5.4 (2004/01/29)")
13:14:09 Part LinusN
13:14:18 Quit edx (Read error: 110 (Connection timed out))
13:15:37 Join LinusN [0] (
13:19:59 Part LinusN
13:21:17 Join LinusN [0] (
13:22:02 Join hicks [0] (
13:28:47 Nick Febs is now known as Febs_away (
13:32:56 Join Tangleding [0] (
13:34:03MoosHi crazy Tang
13:34:53TangledingHi Moos
13:35:04TangledingBack to Brest finaly
13:38:23 Join ghostiger [0] (~ghostiger@13e7c2047dcb9e70.session.tor)
13:51:11Tangledinghello Rbx :)
13:51:26TangledingA x
13:51:37TangledingI wasn't on IRC chan for a while
14:00:06TangledingCongrate for latest progresses...
14:00:13TangledingVery good work
14:00:50Tangledingjust waiting for remote displaying patch to be comitted in CVS
14:01:03Tangledingthen i'll leave original iRiver fw :)
14:05:04 Quit ghostiger ("Leaving")
14:18:59 Part LinusN
14:26:25 Join LinusN [0] (
14:28:01Tangledinghi linus
14:28:07Tangledinghow are you?
14:30:48LinusNfine i guess
14:37:28Tangledingwant to gongrate you
14:37:38Tangledingfor the H110 bootloader
14:37:52Tangledingvery nice work
14:56:05 Join edx [0] (
15:00:16TangledingOkay i've to go
15:00:41Tangledingi will make a donation at spetember/october
15:00:52LinusNwee! thx
15:00:54Tangledingsince i've suceeded in my exam
15:01:11Tangledingand i'll work now and get payd
15:02:43TangledingCheers :)
15:03:52 Quit Tangleding ("Chatzilla 0.9.68a [Firefox 1.0.4/20050511]")
15:09:31***Saving seen data "./dancer.seen"
15:40:09 Quit cYmen (Read error: 104 (Connection reset by peer))
15:40:40 Join firefox_i686 [0] (
15:44:38 Part firefox_i686 ("faites vous plaisir ----->")
15:56:23 Quit matsl (Remote closed the connection)
15:56:24 Quit Coldtoast (Read error: 104 (Connection reset by peer))
15:59:51 Join matsl [0] (
16:00:48 Join Coldtoast [0] (
16:05:50 Quit matsl (Remote closed the connection)
16:08:28 Join cYmen_ [0] (
16:13:10 Join asdsd____ [0] (
16:14:44 Part asdsd____
16:35:20 Quit ashridah ("Leaving")
17:01:24 Join BBub_ [0] (
17:02:40SlasheriHmm, if you find any bugs with playback, please tell me as soon as possible. I am only few days here and then month away
17:06:46LearSlasheri: the end of each track seem to be trimmed by a few seconds...
17:06:46crwlonly three weeks away ;)
17:07:24Slashericrwl :)
17:07:33SlasheriLear: Hmm, with any codec or mp3 only?
17:08:29Slasherii think this is mp3 problem..
17:09:34***Saving seen data "./dancer.seen"
17:10:41LearI mainly play MP3, but I've notived it for oggs as well, when testing the resampler. It started happening within the last couple of days or so...
17:11:50SlasheriLinusN: btw, the optical out is working with current builds. Beam is only lower intensity while not playing anything
17:12:20SlasheriLear: oh, i will check that
17:12:20LinusNi assumed he meant while playing
17:12:44LinusNafaik, it's been working for quite some time
17:14:30SlasheriLear: thanks, there is something wrong, ogg playback was not gapless when i just tried with a sine wave
17:15:45LearGood, easy to reproduce. :) Crossfade was off when I noticed it, btw.
17:18:08 Quit BBub (Read error: 110 (Connection timed out))
17:18:08 Nick BBub_ is now known as BBub (
17:23:44SlasheriLear: that was easy to fix, i will commit soon :)
17:26:52Coldtoastabout 3 days ago I was listenign to tracks and instead of tracks crossfading, I got the start of the next song, a burst of the last song then the rest of the next song. Ithink it was after buffering then changing to a different dir, listening for a few secs then skipping to the next track
17:27:29Coldtoastso kind of a "worst case scenario" for playback
17:27:50Moosi experimented this too few days ago :(
17:28:02crwli've got stuff like that too, and i don't use crossfading
17:28:03Coldtoastonce it buffered again tho, it was fine
17:28:33crwlnot buffering enough initially or something (this only happens sometimes when changing tracks to something that isn't buffered yet)
17:33:32BBubSlasheri: i found a problem on flac, latest build, it makes a short noise on the beginning of a track
17:33:37 Join XavierGr [0] (
17:33:45BBubmaybe its the tag as rockbox doesnt read it
17:34:00SlasheriBBub: Hmm, i will try that too
17:34:05Coldtoastheh. I've queued every single MP3 I have in Winamp then enabled Shuffle and now listening to tracks I've never even heard before
17:34:42BBubthx Slasheri ;)
17:49:21*LinusN cut the lcd updating time from 1.8ms to 1ms
17:49:42LinusNin 120mhz
17:53:18LinusNand from 3ms to 2.5ms in 48mhz
17:54:55 Join Tangleding [0] (
17:55:15TangledingIs there slashery?
17:55:26Tangledingslasheri sorry
17:56:24SlasheriTangleding: hi, yes i am :)
17:56:36TangledingHow are you?
17:57:22Slasherifine thx ;) just trying to catch the last bugs before i leave for few weeks :)
17:57:37Tangledingabout the gapless fix
17:57:48Tangledingit's conceerning ogg i guess?
17:57:56Tangledingno lame?
17:59:24Slasheriall codecs had the problem, but it doesn't mean lame would be totally gapless now
17:59:50Slasheriat least other codecs should be true gapless
18:00:01Tangledingah okay
18:00:26Tangledinglame is really diffrent from others concerning gapless issue?
18:00:55Slasheriyes it's because mp3s are not designe to be gapless..
18:01:58Tangledingin fact i've discussed with Gabriel Bouvigne
18:02:09Tangleding(a Lame developer)
18:02:19Tangledingand according to him
18:02:33Tangledingthe mainly différence was a "standadisation" one
18:02:36LearLAME-encoded stuff can be gapless, but it is something tacked on later.
18:02:48Tangledingsince there are many encoders
18:02:58Tangledingunlike Vorbis for exemple
18:03:38Tangledinghe told me that thje vorbis gapless feature was using same method thanrecent lame (with fixed delay)
18:03:43XavierGrso its pretty much impossible to see gapless to non-lame mp3s right?
18:03:56Tangledingi guess so
18:04:01Tangledingwould be necessary
18:04:23Tangledingto set the delay according to each encoder
18:04:27SlasheriXavierGr: we may have the dsp to detect and cut silence in future
18:04:39LearYes, because the decoder doesn't know how much the source was padded. Encoder delay isn't fixed either.
18:05:01Tangledingoit's fixed for lame anyway
18:05:20Tangledinghave you got contact with some of the lame developers?
18:06:02LearProblem is, how to not detect/cut silence not at the very end...
18:08:00XavierGrLinusN: Do you happen to from where the tree.c code is executed? I found the tree_init function called by main.c but this isnt most likely
18:08:13XavierGr^happen to know
18:08:17TangledingA cool feature about silence remover
18:08:38Tangledingshould allow to detect long silence not only at the file end
18:08:49Tangleding(for kinda bonus track ;) )
18:09:00LinusNXavierGr: app_main() calls init() and then browse_root()
18:11:28XavierGrso if I make a remote-tree.c identical to tree.c (cahnged lcd_ functions to remote_lcd) and call remote_tree_init() and remote_browse_root() (changed respectively) I can get the remote working right? (then toss up the clutter code)
18:12:56LearExcept having two browsers active at the same time sounds like a bad idea... The remote should mirror a clipped version of the main display, IMO.
18:13:37LinusNLear: agreed
18:14:24 Quit CBM-away (Read error: 54 (Connection reset by peer))
18:14:27XavierGrWell amiconn suggested that we use a saperate thread for each. Like creating remote-tree.c remote-wps-display.c e.t.c
18:14:39Tangledinghow do the iRiver fw proceeds?
18:15:28LearThat might work, but it could also be confusing (for the user), and wouldn't there be synchronization issues?
18:18:11XavierGrwell I started a quick hack by replicating every lcd_ function to lcd_remote in the tree.c but encountered a dead end. Where a a function returns the height of the screen. I cant find a way to return 2 values, 1 for the main unit and for the remote. So I thought a different .c file would be better.
18:18:59XavierGrwhat do you suggest?
18:19:38Learwhat functions do you refer to?
18:19:58XavierGryou mean the lcd_ ones?
18:20:15XavierGror the replicate_height()?
18:20:42Learthe ones you had problems with
18:23:13XavierGrin tree.c line 239 recalc_screen_height. int height, is returned with the height value of the screen (LCD_HEIGHT) but then the remote screen isnt displayed correctly (text is hidden in the bottom part)
18:23:14Learhm... separating tree into a model and a view sounds like one way of doing it. I.e., tree (and also the wps) maintains a model over what to display, then call on functions to draw that model, once for the main display, and once for the remote, if present.
18:24:28XavierGrso if I change height = REMOTE_LCD_SCREEN the remote is displayed correctly but the main is not
18:24:32Learyes, you need to maintain two values (using arguments to return the value or indicate what you want, different functions,etc)
18:26:03XavierGrso what to do. 1.replicate and adjust 2.make another file for the remote
18:26:40Lear3. split out the rendering code into two cases, main and remote
18:27:15XavierGrin the same tree.c?
18:27:40Learcould be, but that isn't a requirement...
18:28:03Learthe point is to logically separate the rendering from the rest.
18:28:54XavierGra complete rewrite as I see it... I will have to understand each and every line of the code... and I am newb. nevertheless i will try.
18:30:01Learno, not a complete rewrite. Lots of restructuring, yes...
18:30:57XavierGrto rewrite tree.c for the screen height is tricky though. I tried many times but the rendering functions must be run twice for every display.
18:32:11Learno, once per display should be enough. :)
18:32:38XavierGryeah my bad I meant twice, once for each display
18:33:21Learbut there is no other way, because the screens need to be rendered differently.
18:34:06Learand the easy way to handle that is to put the rendering in separate functions.
18:34:52XavierGrthough amiconn said clearly to me to make a different thread, not just a quick replicate hack, but then I will have to make a huge restructural work in order not to duplicate unnecessary code...
18:35:53XavierGrso I could make a remote-tree.c which has the rendering functions for the remote and then call them from the tree.c if there is a remote?
18:35:59Learpersonally, I don't see the need for a separate thread... I guess the restructuring needed is why nobody has cared enough just yet. :)
18:36:38XavierGrLinusN: What do you think?
18:36:55Learyes, something like that. Only I'd always call them, and let them return immediately if there is no remote.
18:37:43LinusNXavierGr: i agree with Lear
18:38:32XavierGrbut what if the code in tree.c is not saperated with rendering/setup functions. It is most likely that they are mixed up.
18:39:08XavierGrI will have to dig up a little to be sure
18:39:18Learthe code sure isn't separated today, but it needs to be...
18:40:43XavierGrPitty because I quite sure that I will have troubles to saperate them or even likely to understand them...
18:41:06Learto a large degree, I think what is needed is a "render list" function, that based on the loaded directory, cursor position and an internal state (the top line) draws that list.
18:42:51XavierGrgod the remote thing is the only thing that keeps me away from using rockbox. I am dying to make something for it.
18:44:03LearA few additional things are needed though,e.g., the "create playlist" splash. Some could problably be made with splash() (which perhaps also should know how to handle an additional display).
18:44:57XavierGranother question: a static is seen without change in a single .c file, right? What if I want to use a static variable from another file?
18:45:28LearI'd start by going through all rendering (lcd calls) and try to group them logically. As in, render tree, create playlist...
18:46:56XavierGrtruth is that not all functions in tree.c has and lcd rendering function.
18:46:59LearA static is only visible within that module, yes. If you want access from another file, use accessor functions. dsp_stereo_mode() is a good example. It returns the current stereo mode, which is maintained within dsp.c.
18:47:28Learwell, then perhaps they don't need to be changed at all?
18:47:44XavierGrbut it has some huge ones that do many things.
18:47:50Leari.e., there's no rendering to break out.
18:48:10XavierGralso they counter react with other files for various tasks.
18:48:43Learbut those things can be grouped into a limited number of "functions"...
18:49:57XavierGr7 functions have at least one lcd rendering function
18:50:33XavierGrand dirbrowse is HUGE also there are alot of ifdes which need to be removed for the remote.
18:52:42LearBut most of that is non-render logic, afaict. The "model" part I mentioned before.
18:54:28XavierGrhow can I return 2 values from a function? Maybe as a void functions which alters 2 static values?
18:55:48Learon a tangent, shouldn't the lcd_puts_scroll code be common to all targets? And wasn't it, once upon a time... :)
18:56:14Learpass in a pointer, and set the value through that pointer,
18:57:40Learsee e.g. get_tag, it sets flags that way
18:57:51Lear(in wps-display.c, that is)
18:59:15XavierGrok thanks. I will think of this a little bit and I will see what I can do...
19:03:12XavierGrso Slasheri are you going on vacation?
19:08:11 Join Sucka [0] (
19:09:37***Saving seen data "./dancer.seen"
19:12:44SlasheriXavierGr: no.. i have to go to the mandatory (civil) military service.. :D
19:15:03XavierGrfor how long?
19:16:30Slasheri13 months and one month (3 weeks in fact) without net access
19:17:02Slasheribut i can do other things during that 12 month period too :)
19:17:30XavierGrso we will see you again after 13 months!!!!!
19:17:43Slasherihehe :D
19:17:52XavierGrwe will surely miss you.... :(
19:18:00Slasheriafter one month could be more realistic i hope ;)
19:18:28XavierGrbut you said 12 months?
19:18:41Slasheriyes, the service lasts 13 months
19:19:01XavierGryou will have some breaks?
19:19:06Slasheribut during that last 12 month period i have the net access and can do other things too if i have time :)
19:19:22thegeekdo as I did
19:19:24thegeekget lucky
19:19:29thegeekand get out after 7 months
19:19:33Slasherihmm :D
19:19:40thegeekbecause of bad economy
19:19:47thegeekthey just had to dump some of us
19:19:47Slasherihehe, nice ;D
19:19:50XavierGrhow old are Slasheri?
19:19:55SlasheriXavierGr: 20 :)
19:20:24XavierGrand going to military so soon? I am 21 and I will go in 2011!
19:20:48Slasherioh :) i just wanted to go there as soon as possible.. then i don't need to worry about it later :)
19:20:49IsmoSlasheri: inspired by your latest commit I browsed the playback.c a bit. Is the line 1349 correct or should there be some if statement before that?
19:21:04SlasheriIsmo: Hmm, i will check that
19:21:59IsmoI haven't looked how the pcm stuff works, but that line looks like a bit too alone :)
19:22:01SlasheriIsmo: yes, that's correct. If we were filling the buffer, AUDIO_PLAY event will start the crossfade
19:22:08Slasheriand if not, then it's started there
19:23:03Slasheriin fact codec_track_changed() could be on that statement too but it's outside it because in every case we want to update the wps immediately
19:23:33XavierGrand what country are you from Slasheri?
19:23:41SlasheriXavierGr: from Finland :)
19:24:00XavierGrI thought that is was voluntarily there.
19:24:19Slasherithis is the about only EU country that still has mandatory mil service.. :/
19:24:37XavierGrwell count Greece in that too
19:24:39thegeekdont forget norway
19:24:42thegeekor sweden
19:24:47thegeekoh well
19:24:50thegeeksweden then
19:25:26XavierGrits 12 months (i think) mandatory here too.
19:25:55XavierGrI remember the elders calling 36 months in the military! :x
19:27:10Slasherihuh, that's sick :D
19:30:50 Join Stryke` [0] (
19:33:42Stryke`small bug report: the patch that was added to stop scrolling at the bottom of a list when holding it down does not work in the Debug menu
19:34:41LinusNStryke`: iirc, that patch was only for the browser, not for menus
19:35:19Stryke`oh, any reason it would not be applied there?
19:35:53LinusNno particular reason
19:36:19 Join stripwax [0] (
19:43:17XavierGrhi stripwax!
19:43:45crashdanyone in london fancy a beer ?
19:44:15stripwaxcrashd - i'd love one, but I've just got back from a four hour drive from the midlands .. way too hot to do anything right now
19:44:26stripwaxXavierGr - hey dude
19:44:41crashdim sunburnt from sitting up on hampstead heath yesterday
19:44:59stripwaxcrashd hehe. i'm sunburnt from driving round back from the midlands! ;-D
19:45:04crashd: )
19:45:25 Join webguest72 [0] (
19:45:33webguest72hello folks
19:45:45webguest72a litle question please
19:46:54*stripwax slaps himself for running old plugins in rockbox sim and wondering why it kept crashing..
19:46:55webguest72there is a patch for i2c in the tracker, and apear anyone want to work on it, it's normal? :)
19:47:10webguest72the FM it's very usual :)
19:47:16stripwax"usual" ?
19:47:46webguest72"important" for much people
19:48:21webguest72it's not in your plan?
19:48:51webguest72Linus: i'm sure you can do it easyly,no?
19:50:35webguest72Linus:i noticed in the log that you have adapted the code a bit for i2c for irivers, have you planed something please? :)
19:51:44stripwaxwebguest72 - which patch are you referring to?
19:52:16 Quit webguest72 ("CGI:IRC (EOF)")
19:52:26 Join webguest72 [0] (
19:53:18 Quit west-acre ("—I-n-v-i-s-i-o-n— 2.0 Build 3515")
19:53:24stripwaxwebguest72 - you'll have to ask ac about that one...
19:53:32 Join solex_ [0] (
19:53:56webguest72but he deserted this because he working in iaudio now
19:54:08webguest72and alone Linus can help us
19:54:41 Join bagawk [0] (~lee@bagawk.user)
19:54:45webguest72Linus: nope?
19:55:23LinusNi optimized the i2c bit rate
19:55:45webguest72have you planed to aplly the ac patch?
19:56:06webguest72wonderfull :)
19:56:21webguest72it's easy for you, no?
19:56:34LinusNwhat do you mean?
19:57:16webguest72it's more easy for you to do this work than one other dev, because Rockbox is your baby, no?
19:57:38LinusNhehe, rockbox is our baby, not mine
19:57:44Coldtoastit's open source so many ppl have written code for it
19:57:44stripwaxwebguest72 - any of the devs could apply ac's patch, i expect
19:57:57Coldtoastit's not like LinusN could know ALL the code intimately
19:58:30webguest72i just want to know he was the must experimented with it
19:58:53LinusNwas that english? :-)
19:59:19webguest72scuse me for my english :)
19:59:32 Join spiralout [0] (
19:59:34LinusNsince i have hooked up my logic analyzer, it feels natural to add the i2c reading, since i can easily verify it
19:59:56LinusNright now i'm investigating the s/pdif output
20:00:11stripwaxLinusN- neat - what's to investigate there?
20:00:33LinusNsome people have problems with it, their dac refuses to play it
20:01:28webguest72Linus: scuse me for this stupid questions and thanks, we hope you'll can add i2c things for FM soon if you can :)
20:01:37LinusNwebguest72: don't worry
20:01:58webguest72have a long life to Rockbox :)
20:02:11webguest72bye all
20:02:15 Quit webguest72 ("CGI:IRC")
20:03:11LinusNi have a theory why spdif doesnt work, but i don't have a toslink cable to test with :-)
20:03:51LinusNdinner time
20:03:57LinusNcu in a while
20:04:10 Part LinusN
20:04:11 Quit solex (Read error: 110 (Connection timed out))
20:09:05stripwaxAnyone want a patch for sokoban so it caches leveldata?
20:11:20BBuba small bug: on the h110 rockbox always shows 276mb of free space no matter how much space is really free
20:11:51stripwaxhehe, neat!
20:15:29stripwaxMy sokoban cache patch is here -
20:25:29 Join hicks_ [0] (
20:40:25 Quit hicks (Read error: 110 (Connection timed out))
20:42:23 Quit Tangleding ("Chatzilla 0.9.68a [Firefox 1.0.4/20050511]")
20:43:04 Join webguest80 [0] (
20:43:51webguest80BBub: What if you press the joystick while on the "Disk info: Free" screen? (this will rescan free space)
20:43:52HClwhere's preglow when you need him
20:44:02 Quit Maxime ()
20:44:35 Quit bagawk ("Leaving")
20:46:52 Join muesli- [0] (
20:50:32XavierGrhave to go see you all later!
20:50:35 Quit XavierGr ()
20:51:37muesli-dont you understand "re" !?
20:52:39 Join Psy-Dead [0] (
20:53:20Coldtoasta lot of ppl don't understand anything less than a TLA
20:57:23 Quit Seed (Read error: 104 (Connection reset by peer))
20:57:27ColdtoastThree Letter Abbreviation
20:57:40muesli-and ppl?
20:58:17Coldtoastyou in the UK? or Australia?
20:58:30muesli-neither nor..germany ;)
20:58:39muesli-but have been in oz last year
20:58:48Coldtoastdid you like it there?
20:59:55muesli-it was in october and exprected to be warm but it was lousy cold ;-) but oz is nice..wished to had more time
20:59:56 Join Seed [0] (
21:00:48muesli-are you from down under?
21:01:57ColdtoastI am
21:02:04Coldtoastwhere the heck WERE you in Australia? heh
21:02:31muesli-just in sydney to see the coat hanger (spelling!?)
21:02:57muesli-whats yr location?
21:04:39 Join dj-death [0] (
21:04:42dj-deathHi all
21:04:54dj-deathI'm new to rockbox
21:05:06ColdtoastI'm actually in Tasmania
21:05:07dj-deathI just got it on my iriver ihp120
21:05:19dj-deathit works great ;)
21:05:44dj-deathand has some really cool feature
21:06:12muesli-tasmania sounded great by brochure. and where do you live most of the time?
21:06:21 Join LinusN [0] (
21:09:02muesli-high LinusN
21:09:38***Saving seen data "./dancer.seen"
21:10:43dj-deathhas someone problem at the end of file when playing on iriver ?
21:11:04LinusNdj-death: try the latest bleeding edge
21:11:34dj-deathI mean when listening a file (mp3/ogg) rockbox jump to the next track 2/3 seconds before the end of the current file
21:11:50LinusNyes, try the latest bleeding edge
21:12:00dj-deathbleeding edge ?
21:12:07dj-deathyou mean daily build ?
21:12:12LinusNthe very bottom of the daily build page
21:13:09muesli-isnt it this one?
21:13:20Stryke`those are daily builds
21:13:39muesli-where do i find the bleedings?
21:13:41Stryke`thats bleeding edge
21:13:50webguest80<LinusN> the very bottom of the daily build page
21:14:10dj-deathok upgrade done
21:14:14dj-deathlet's try
21:14:43muesli-now its the upper one
21:16:21dj-deathLinusN: works fine thx a lot
21:16:31 Part dj-death ("bye")
21:19:18SlasheriLinusN: Hmm, would you like to look at the peak meter code if i commit it soon? at some degree it works but it has quite a strange behaviour
21:19:52LinusNSlasheri: strange?
21:20:53Slasherii think the scaling might not be right and for some reason i have to "invert" the calculated average to get the right reading. You will find more from the code :)
21:29:10SlasheriLinusN: committed, please take a look if you have time
21:29:41Slasherii really don't know how that peak average should be calculated so i just tried _something_ =)
21:32:37 Quit crashd ("leaving")
21:32:48 Join memmem [0] (
21:33:39 Join crashd [0] (
21:36:01Coldtoastsorry dude. got distracted
21:36:16Coldtoastcity called Launceston muesli-
21:36:53muesli-mmh..never heard of it..which state?
21:38:52muesli-mmh..didnt made it far in the south..
21:38:55muesli-next time ;)
21:39:35HClwhat a horrid url
21:39:49Coldtoastyeah. google link. sorry
21:39:55Coldtoastanyway. that's where I live
21:40:14HCllooks cloudy ;x
21:40:29Coldtoastheh. it is. especially in Winter
21:40:38Coldtoastblue skies in Summer tho
21:43:49muesli-yeah, gorgious nature..thats what i heard about tasmania
21:43:58muesli-g'Day HCl btw
22:01:41muesli-l8er mates
22:05:57 Join Yasmine_ [0] (
22:06:01 Part Yasmine_
22:17:34 Join ep0ch_ [0] (
22:18:45ep0ch_hey, anyone else noticed that all of sudden 128 kbps mp3 is now realtime with no CPU boost?
22:19:11ep0ch_i could swear it needed boost before...
22:19:14ColdtoastI have no 128kbit MP3s on mine, so no. But great to hear
22:22:08stripwaxSlasheri - is that pcm peakmeter basically the same approach as I had or does it precalculate volume levels?
22:22:53ep0ch_oh wierd *some* 128 kbps are realtime, and some arent.
22:23:15 Join webguest04 [0] (
22:23:21 Quit webguest04 (Client Quit)
22:23:57Slasheristripwax: hmm, it tries to calculate some average volume from one chunk of 512 bytes. But i don't even exactly know what it does ;) what kind of approach you have?
22:24:41ep0ch_the peakmeter is wierd... even quiet parts of the music show full peak
22:25:02Slasheriyes, it's kind weird.. try using linear scale instead of logarithmic
22:25:09 Quit muesli- (Read error: 110 (Connection timed out))
22:25:42ep0ch_Slasheri: oh thanks for all the recent playback fixes :)
22:26:08Slasherinp, one more fix is coming (i found just a bug in the playback code)
22:26:36ep0ch_hmm i just changed to "linear" peakmeter and the music just hung :s
22:26:58Slasheridid it crash or just hung?
22:27:09ep0ch_just got silence
22:27:12ep0ch_no sound
22:27:39ep0ch_i'll try it again, see if it was the peakmeter
22:28:37Slasherihmm, is the peak meter disabled if wps don't show it? i hope it is..
22:28:56ep0ch_nah i dont think it was changing the peakmeter that broke the sound.. did it again and it worked...
22:29:46Slasheriif you have logf enabled, take a dump and try to see what happened
22:30:08ep0ch_not enabled
22:30:21ep0ch_i'll try and reproduce
22:30:49Coldtoastsorry. telling him to "take a dump" is childishly amusing
22:35:33stripwaxSlasheri - yes, that was my approach too but I remember Linus saying it was quite strange so I guess I thought an approved approach would have to be different ;-)
22:36:01Slasheriah, interesting :)
22:36:55stripwaxSlasheri - how do you work around something like SAR0 changing while you're reading the buffers? or the remaining buffer size being less than your averaging window?
22:37:31stripwaxSlasheri - i think my patch was here
22:38:26ep0ch_so basically you average a block of samples, divide by the bit depth and mutiply by the number of pixels wide the peak meter is, or something?
22:39:20stripwaxep0ch_ - sort of, the recorder/peakmeter.c code just expects a number between 0 and 0x7fff
22:40:02stripwaxso yeah, we average a bunch of sample data, and obtain a left and right 'volume' value between 0 and 32767
22:40:32ep0ch_sample data is signed?
22:41:20stripwaxep0ch_ - would that make any difference? volume would still be between and 0 and 32767
22:41:54ep0ch_if you have sample values, -255 and 255 average is 0
22:42:01ep0ch_when it should really be 255 for the peakmeter
22:42:11ep0ch_i.e. the average of 255 and 255
22:43:01ep0ch_so you should ignore the sign if sample data uses it?
22:43:16ep0ch_bah i'll look at some C code and see if i understand it :)
22:43:21stripwaxep0ch_ - ah, understood. yes, to derive the sample data obviously we need to know what 'zero value' means.
22:43:45stripwaxI think I assumed sample data was 16 bit signed in my code, dunno what Slasheri does
22:44:45stripwaxHmm - interesting, he's assuming sample data is unsigned, but only deriving a volume value from samples with value > 32768. weird?
22:45:51ep0ch_well half the values wont be calculated then
22:46:15stripwaxwell, they'll be calculated but wrong ;-)
22:46:43ep0ch_should average the whole lot and add 32768 to the total maybe
22:46:47 Join matsl [0] (
22:46:58ep0ch_bah i dunno
22:47:19stripwaxep0ch_ no, don't add 32768 - just take absolute value of each sample, and average that.
22:47:49ep0ch_oh yeah of course
22:50:45ep0ch_bah playback just stopped again for no reason about 2/3 of the way through an MP3. i'll put logging on from now on
22:51:07ep0ch_was quite enjoying the tune too ;)
22:51:19stripwaxooh, logging ? where/what's that?
22:52:29Bagderenable it with the configure script
22:53:01*stripwax never even knew. neat
22:53:19Bagderlogs on the remote lcd
22:53:33Bagderand offers a "dump logf buffer" function
22:53:45*stripwax has no idea where his remote is .. :-(
22:53:53ep0ch_i need the remote???
22:53:57Bagderyou can browse the log on the main unit too
22:54:01ep0ch_mines still in pieces :s
22:54:04Bagderep0ch_: only if you want to see the log live
22:54:45ep0ch_if i want it on the main screen can i change the wps to display it?
22:54:51stripwaxBagder - ah, cool
22:54:57Bagderep0ch_: no
22:54:58ep0ch_theres plenty of space...
22:55:02 Join webguest69 [0] (
22:55:15Bagderep0ch_: unless you add the code for it ;-)
22:55:33ep0ch_give me a couple of years to get up to speed with rockbox code
22:56:04BagderI'll expect your patch by noon tomorrow B-]
22:56:19stripwaxheh heh
22:59:54ep0ch_stripwax: back to peakmeter, if the data is unsigned taking the average of absolute values is pointless
23:01:11stripwaxep0ch_ - audio data is 16 bit.... if data is unsigned then 'zero' is 32768. so subtract 32768. and then it becomes signed. so you need to take absolute values...
23:01:12Learepoch_: pcm data is signed.
23:01:31Learat least in the iRiver... :)
23:01:32stripwax(or, equivalently, invert data if sample is > 32768)
23:02:23memmemNo, the result of inverting will be off by one.
23:03:23stripwaxmemmem - by invert I meant 65536-x
23:03:24 Quit webguest69 ("CGI:IRC (EOF)")
23:04:25memmemOK, PCM data uses two's complement representation for negative numbers.
23:04:45stripwaxi.e. signed shorts
23:05:08stripwaxand not unsigned shorts, which is what Slasheri's pcm vu meter code seems to use..
23:05:25memmemSigned shorts on architectures that use two's complement arithmetic ;-)
23:05:36BBub[20:43:54] <webguest80> BBub: What if you press the joystick while on the "Disk info: Free" screen? (this will rescan free space) <- that works
23:05:52BBubi thought it would do it by itself
23:06:08Bagderdo what by itself?
23:06:17BBubrefresh the free space
23:06:22Bagderwhy would it?
23:06:31Bagderit is time and battery consuming
23:06:51BBubi only assumed it because the original firmware does it
23:07:05BagderI didn't even know it could do that ;-)
23:07:14stripwaxBBub the original firmware loads the entire directory structure on startup, which is time and battery consuming ;-)
23:08:12webguest80This is unrelated, I believe
23:08:22stripwaxBagder - how does rockbox fw know if/when file system changes have been made in usb mode
23:08:24BBubi dont really need it anyways ;)
23:08:31webguest80The problem lies in either iriver firmware, rockbox or your OS not updating the fsinfo block
23:08:37Bagderstripwax: it doesn't really
23:08:51 Join amiconn [0] (
23:09:23Bagderwebguest80: rockbox updates it
23:09:31Bagderwindows usually does not
23:09:42***Saving seen data "./dancer.seen"
23:09:54stripwaxwebguest80 - the problem lies in the fact that the OS has no way of knowing what exactly happens when the harddrive is in usb passthrough mode
23:09:55Bagderafter all, they only made up FAT and wrote the spec
23:09:58webguest80Indeed, I have a beef with Windows for lying to me about free space on more than one occasion
23:10:28webguest80stripwax: not at all, the OS should update the FAT record
23:10:33Bagderstripwax: but the fsinfo stuff should be updated by the OS
23:10:52 Quit Psy-Dead ()
23:12:21stripwaxBagder/webguest80 - oh right - I misunderstood. yeah, you're right
23:14:21memmemAny chance that rockbox will be able to record MP3 at 256 kbit/s or 320 kbit/s on the iAUDIO X5 (as does the original firmware which makes heavy use of EMAC instructions)?
23:14:46Bagdermemmem: we have no good mp3 encoder that runs fast
23:15:22Bagderso currently we don't see that coming any time soon
23:15:34memmemDitto for OGG?
23:15:43Bagderyes afaik
23:15:44BBubbetter chances for recording in flac/wavpack ?
23:15:53memmemFLAC? shorten?
23:15:54BagderBBub: yes indeed
23:16:01BBuboh, that sounds nice
23:16:28memmemThe NoDo page says that WAV recording won't be done but that probably applies to the Archos devices only.
23:16:47Bagderwav recording will most likely be done first
23:16:57Bagderand it even already is somewhat implmented
23:17:14webguest80Wavpack isn't too far away either, it seems
23:18:16 Quit webguest80 ("CGI:IRC")
23:22:24BBubis the waveform-switch in the recording-menu used to switch between raw pcm and pcm in a wave-container?
23:24:41amiconnBagder: The gmini builds completely fell out the table, yet the 2 columns are still there...
23:25:11 Quit Stryke` (Read error: 104 (Connection reset by peer))
23:25:44Bagderyes, since they're still around
23:25:52Bagderthe columns are based on more data than what is shown
23:30:25 Part LinusN
23:32:22 Join CheeseBurgerMan [0] (
23:34:58BagderI'll go away for a about a week starting now
23:37:24amiconnNo localisation v2 commit before that? ;)
23:37:33Bagderno ;-/
23:37:58amiconnI have an experimental plugin showing 13 shades of grey on iriver :)
23:38:12Bagderdoes it look good?
23:38:24amiconnYes, way less flicker than on archos
23:38:32Bagdercool indeed
23:38:44amiconn...because it flicks between 2 adjacent native grey levels, not black & white
23:39:03stripwaxamiconn - that doesn't require constant CPU boost or anything silly does it?
23:39:25amiconnThe experimental thing is without random pixel pattern shift, so more levels would flicker
23:39:41amiconnI hope to support 97 levels of grey in the actual lib
23:40:13amiconnstripwax: Yes it requires constant boost, unfortunately. That's why it will be plugin only, like on the archos
23:40:20stripwaxfairy nuff
23:40:28Rickhmm :o
23:40:36amiconnIt's not a performance problem, but a problem of constant timer period
23:40:39Rickhas rockboy been updated to support grayscale yet?
23:40:52Bagdersee ya in a week
23:40:53amiconnYes, 97 shades.
23:40:55 Quit Bagder ("Off to search for that connect-resetting peer guy!")
23:41:02 Join muesli- [0] (
23:41:51amiconnThe method quickly switches between several bitplanes. The maximum number is limited to 32 to allow storing the pattern in a longword for quick access
23:42:20amiconnOn archos, this allows 33 shades of grey, with the lcd being b&w only
23:42:37amiconn((2-1) * 32 + 1)
23:42:53amiconnThe iriver lcd already provides 4 shades natively, so
23:43:01Rickyummy yummy :)
23:43:04amiconn(4-1) * 32 + 1 = 97
23:43:55memmemThe X5 display supports 2^18 colors (and I confirmed that the firmware transfers 2^18 bits per pixel to the LCD controller).
23:44:07amiconnNice :)
23:44:24RickX5 display?
23:44:34amiconnBtw, this allows less shades of (pure) grey than my grayscale lib will support on H1x0 :-P
23:44:53memmemamiconn: The pixel data is not shifted left and is sent as two 9 bit words.
23:44:57Rickwhat is X5 display? H3x0?
23:45:04amiconniAudio X5
23:45:11RickNever heard of it :P
23:45:25amiconnmemmem: Any news concerning the display controller?
23:45:44amiconnI still think it is quite similar to the H3x0 lcd...
23:45:58memmemamiconn: I found enough to be able to initialize it (though I don't know what those commands actually do) and to write pixel data.
23:46:57amiconnWe should still try to find the type, complete with a datasheet
23:47:12Rickthe x5 looks slick
23:47:17memmemOf course. There are probably also some I/O pins involved.
23:47:20amiconnThat will allow us to use some tricks the original firmware doesn't
23:47:35memmemsuch as? (simple example) the display flip
23:47:47memmemWhat's that?
23:47:54Rickflipping the display
23:47:57Rickso its upside down
23:48:13amiconnAll pixel based units have a mode to use the unit upside down
23:48:26memmemThat could also be done in software. It's used rotated 90 degrees anyway.
23:48:35 Quit Coldtoast ("Peace and Protection 4.22")
23:48:41Rickwouldn't hardware be faster though?
23:48:48amiconnThat's useful from time to time, and since it uses a hardware feature of the controller, it's simple...
23:49:43amiconnmemmem: Doing it in software would be quite a bit more complex
23:49:51 Join yngwi [0] (
23:50:00memmemIf all drawing operations go through a display buffer there's no difference. Just subtract 160 instead of adding 160.
23:50:28stripwaxmemmem - does that mean an extra test for EVERY pixel operation? or does it mean two functions where one would suffice? :-p
23:50:29amiconnThey go through a display buffer with the same layout as the internal buffer of the lcd
23:50:44amiconn...and it's not as simple
23:50:57Rickright...because then you would have to rearrange data
23:51:11amiconn(1) Flipping the line blocks is simple, as ou said just invert the line sequence
23:51:12stripwaxok, just submitted another useless patch - this one for Cube plugin lets you rotate the cube manually when it's paused
23:51:15memmemThe iAUDIO firmware's buffer has a different format than expected by the LCED controller.
23:51:38amiconn(2) Flipping the lines itself is also simple, just read from the end decrementing the pointer
23:52:00amiconn(3) BUT the hard thing is flipping the pixels within a line block
23:52:06memmemI think we don't want to use 2 9-bit words per pixel with one color channel split between two words...
23:52:28memmemJust use a 256-byte table.
23:52:46memmem...for b/w graphics.
23:53:08memmemHowever, we need 18 bits per pixel unless we find out how to select a different mode.
23:53:16amiconnI wouldn't use a palette mode for high/truecolour displays
23:53:33amiconnI would just use 18 bit mode and use 3 bytes per pixel
23:53:44memmemI was talking about a table for flipping the bits inside a byte for b/w displays.
23:53:57stripwaxamiconn - but think of the Plasma-style/color cycling effects you could do in a palettised mode! :-p
23:54:11Rickstripwax: that would be fun :)
23:54:15stripwaxmemmem right or just do it in hardware..
23:54:30memmemThe iAUDIO firmware uses 3 bytes per pixel in RAM and converts to the format expected by the LCD controller while copying the data to the LCD controller.
23:54:32 Quit Harpy (Read error: 110 (Connection timed out))
23:54:39stripwaxmemmem wow, that sounds hideous
23:55:09amiconnMaybe the controller does only allow that format, but I doubt it
23:55:24Rickhave you identified what lcd it uses?
23:55:53amiconnA bit of conversion is probably necessary anyway for colour lcds
23:56:57amiconnmemmem: There are more efficient ways to swap bits than using a table...
23:57:29memmemAll the bits (7<->0, 6<->1, ...)? Pray tell me...
23:58:05amiconnYou can use a 3-staged shift approach, and that can work on 4 bytes at once, i.e. longwords
23:58:13amiconnI tried it, it work nicely
23:58:29memmemFor 8-bit bytes, a table is better.

Previous day | Next day