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 2005-06-07

00:00:07MoosCamarocongates for Slasheri and Linus for this ;)
00:00:12elinenbeHCl: congrat's on the database!
00:02:33 Quit ghostiger (Remote closed the connection)
00:03:50Plugh_just ran songdb.jar
00:03:52Plugh_nice work
00:04:33Plugh_although I don't like the idea of needing something installed on the host system to make it work (JRE)
00:04:52[solid]doesn't really work for me :(
00:04:53Plugh_the perl.exe+dll in the drive was nice
00:06:06[solid]Exception in thread "main" java.lang.UnsupportedClassVersionError: SongDB (Unsupported major.minor version 49.0)
00:06:16ashridahit wouldn't be too hard to manage the same thing if gcj is available on cygwin i guess
00:06:27ashridah[solid]: rofl. are you using java 1.4?
00:06:41[solid]ashridah: i'm using blackdown...
00:06:53Bagderjava hell (tm) :-)
00:06:54[solid]yeah, 1.4
00:07:03*ashridah hits HCl
00:07:37[solid]Java(TM) 2 Runtime Environment, Standard Edition (build Blackdown-1.4.2-01) to be precise
00:07:52ashridahBagder: it's got nothing to do with java hell, and lots to do with HCl using java 1.5 to compile it and not turning on a backward compatibility flag
00:08:08[solid]i might be running it wrong tho - is java -classpath SongDB.jar SongDB ok?
00:08:29ashridah[solid]: no, it's quite literally because HCl's compiled it with java 1.5
00:08:37[solid]oh, ok :)
00:09:10spiraloutwhere can i find the songdb.jar??
00:09:23ashridahspiralout: it's on the wiki
00:09:28ashridahunder 'tags database'
00:09:34[solid]i might give a shot a java 1.5 tho... azureus will be happier probably
00:10:09[solid]it just sucks that on gentoo it's hardmasked and you also need to download it by yourself
00:10:47cYmenazureus is d4 p41n anyway :)
00:10:50*ashridah shrugs
00:12:04ashridah[solid]: as long as HCl hasn't used java 1.5isms, he should just be able to recompile it with 1.4
00:12:14ashridah(and if that's the case, we can probably create a .exe out of it anyway)
00:12:42[solid]i'd be pretty cool with doing that... where can i fetch the source?
00:12:51ashridahyou can't.
00:12:58ashridahso far as i know, it didn't get released.
00:13:22[solid]slightly-less-open source, eh?:)
00:13:54[solid]i'll just sit back and wait, no rush after all ^^
00:14:47ashridahbut yeah, hit HCl when he wakes up and tell him to use javac -source 1.4 -target 1.4 when compiling :)
00:17:58 Quit spiralout (Read error: 131 (Connection reset by peer))
00:18:04 Join asdsd____ [0] (
00:18:32 Part asdsd____
00:18:49 Quit Aison ("( :: NoNameScript 3.72 :: )")
00:19:31 Quit crashd ("leaving")
00:20:43 Join crashd [0] (
00:20:54 Quit crashd (Client Quit)
00:21:36 Join preglow [0] (
00:22:06preglowpleait should be rewritten in a decent language :P
00:22:23crwllike python :P
00:23:10preglownah, we wouldnt want to bore the programmer to death
00:23:24 Quit Sucka` ("a bird in the bush is worth two in your house")
00:23:49preglowsaw pedro's post on the mailing list, and i can't say i hear any difference myself
00:28:18[solid]that could be verified easily once the optical output is working
00:30:00[solid]iriver might have used that low accuracy thingy, like rockbox did before vorbis decoding got optimised, no?
00:31:40 Quit Zoom2 ("CGI:IRC (EOF)")
00:31:47preglowthey might
00:31:51preglowbut then they'd have to be really stupid
00:32:18 Quit cYmen ("zZz")
00:32:24preglowwell, it's really easy to test
00:32:34preglowjust dump an ogg to wav with vorbis2wav
00:32:46preglowthen do fast switched between that and the ogg in the iriver fw
00:32:57preglowyou should be able to hear a difference then, but i can't
00:33:23[solid]tbh i have experienced some flaws with the original firmware - but it never happened with my own encoded files, there was this kind of hiss going on
00:33:33t0mashm... maybe find somebody with really good recording equipment?
00:33:42t0masand record both... then analize on a pc...
00:33:52crwljust record the iriver fw version with an optical in
00:34:06[solid]i always suspected it was older oggenc versions... actually i think i'll test some now
00:34:08preglowahaha, i have never _analized_ audio before
00:34:11[solid]intriguing that is:)
00:34:18t0mascrwl: bad idea...
00:34:36crwlt0mas, why?
00:34:38t0masyou have to record both with a line in, or both optical..
00:34:44t0masnot one optical and one line in
00:34:52t0masas that's a quality difference too...
00:34:58crwlwell you can make rockbox dump to wav
00:34:58t0masoh wait...
00:35:07t0masyes, realised that a few seconds later ;)
00:35:26preglowt0mas: you can really just record with line out
00:35:39preglowt0mas: the audio chip is the same for both rockbox and iriver fw
00:36:08preglowso shouldn't matter too much
00:36:16t0maspreglow: yes, but then record both with line out... not one with optical and one with line out
00:36:22t0masthat was my point...
00:36:24preglowt0mas: of course
00:36:40t0masbut the optical out for iriver and wav for rockbox seems ok to me too...
00:37:23 Quit matsl (Remote closed the connection)
00:38:23 Join matsl [0] (
00:40:48[solid]rockbox spoiled me... the boot time is so short in comparison that i nowadays think it takes ages for the original fw to boot
00:40:57 Quit matsl (Remote closed the connection)
00:41:13t0masbut I like the bootload usb mode
00:41:20t0masbecause that saves me a lot of waiting :)
00:41:39[solid]it simply ownz
00:41:56 Join matsl [0] (
00:42:53Stryke`bootload mode is ridiculously fast, i doubt it everytime
00:47:06[solid]hm.. i don't hear no difference either
00:48:34[solid]and that's a nice piano part with vocals (and some distorted guitar later on)
00:52:00 Join iRiverowner [0] (
00:52:27 Quit matsl (Remote closed the connection)
00:53:07iRiverownerat last there is sound for iriverbox!
00:53:22 Join matsl [0] (
00:53:22t0masrockbox ;)
00:53:45iRiverownerIts good but the machine hangs during playback, I can deduce more work is needed
00:54:26t0masyes, it would be usefull to know when it hangs...
00:54:29preglowfor what files?
00:54:31t0masand with what version..
00:54:33t0masand what file...
00:54:39iRiverownerSince the iriver is software driven, im just wondering what codec is used for MP3 playback
00:54:42iRiverownerhanging is random
00:54:58t0masiRiverowner: no it's not... it's not designed by Microsoft...
00:55:23iRiverowneri didn't say anything was designed by Microsoft
00:55:43 Quit zeeeeeeee ("Miranda IM! Smaller, Faster, Easier.")
00:55:52iRiverowneri was asking about the codec used for MP3 playback on the iriver in Rockbox
00:56:18t0masso the MAD codec?
00:56:21iRiverownerokay cool
00:56:38crwlbut where's the coder part in MAD?
00:56:39iRiverownerIm wondering if AAC, WMA, MP3Pro will be supported as well
00:56:49t0masif someone writes a codec... they will
00:57:00t0masbut WMA isn't open... so don't hold your breath
00:57:05iRiverownerMP3Pro would be useful
00:57:12[solid]it isn't open either
00:57:19crwli don't really like the way decoders are called codecs
00:57:38 Part MoosCamaro
00:57:47iRiverownereasy reference i suppose
00:57:51iRiverowneri have an H120 by the way
00:58:33iRiverownerhi HCI
00:58:58[solid]HCl: i was supposed to tell you to compile songdb with -source 1.4 -target 1.4 :)
00:58:58preglowcrwl: they might very well end up as proper codecs
00:59:03HClashridah: no, the source is included in the .jar :)
00:59:07HCl[solid]: i saw :p
00:59:08preglowcrwl: a lot of the libs we use have both decoders and encoders
00:59:13preglowcrwl: well, some, at least
00:59:24HCl[solid]: and i will, but after i get java on my linux machine, windows makes a horrible dev environment
00:59:25ashridahHCl: ah, that explains the jump in size then
00:59:27ashridahi didn't think to look, tbh :)
00:59:29[solid]okay, /me goes to figure out how to compile it then
00:59:32HClashridah: ;x no, not entirely
00:59:37iRiverownerIs there a codec required for recording since the archos is hardware done and the iriver is software done as we know?
00:59:45HClashridah: second backend for tagreading :/ :/ ugly solution, but it'll do for now.
00:59:46ashridahHCl: if it's in there, i can do it instead.
01:00:05HClits mostly that you need to do javac `find -name '*.java'`
01:00:09HCland i can't do that on windows
01:00:12t0masiRiverowner: yes recording on iriver is software too
01:00:12 Quit iRiverowner ("CGI:IRC (EOF)")
01:00:13HCland i hate having to do every dir manually
01:00:24HClyou'll bump into some "assert" stuff
01:00:26ashridahHCl: good god man, why aren't you using ant?!
01:00:28HClthat won't compile for 1.4
01:00:39HClyou can just get them out, it should still work.
01:00:58HClwhats ant?
01:01:02*[solid] tries
01:01:22[solid]and god bless bootloader usb mode one more time xD
01:01:38HClit will save your ass if you bricked your player
01:02:11HClfortunately, our brick count is still 0 :)
01:02:23ashridahHCl: actually, no you don't run into assert stuff
01:02:32HClit did here
01:02:37elinenbequick question... how do you boot directly into the iriver firmware?
01:02:41ashridahnot if you compile it with 1.5 using the -target/-source flags, it seems
01:02:41HClfor the .de something stuff
01:02:47HClah.. maybe..
01:02:47ashridahelinenbe: hold record
01:02:53HCli haven't tried that
01:03:00ashridahHCl: how do you rebuild the .jar?
01:03:01HCli compiled with actual 1.4, not knowing those options existed
01:03:05ashridahi tried, but it bitched about the manifest
01:03:12[solid]i did run into them
01:03:16[solid]but only two
01:03:27[solid]both in entagged/audioformats/asf/data/
01:03:31[solid]lines 86 and 101
01:03:33HCltype jar
01:03:37HClthen use the bottom example
01:03:41HClits exactly what i use
01:03:44ashridah[solid]: you're using a 1.4 jdk
01:04:20[solid]woo it looks like it's workin' ^^
01:04:24HCl :)
01:04:35t0masgood night :)
01:04:38ashridahHCl: but there's no manifest in it...
01:04:43t0masI'm off to sleep :)
01:04:47HClashridah: there is, in the META-INFO dir or so
01:04:51HClits called MANIFEST.MF
01:04:52HCli think
01:05:33ashridahyeah. HCl well, it works with 1.4 now
01:05:40ashridahwant me to upload it as a replacement SongDB.jar?
01:05:48HClyes, please :)
01:05:54[solid]and so it built a database!
01:05:58HCl :)
01:06:03HCltry the latest cvs with it please
01:06:09HCli added v3 support to the searchengine and databox
01:06:24HCli still have to implement some planned features..
01:06:32ashridahi think i've already been using it
01:06:40HCli committed it earlier tonight
01:07:08[solid]workin :D
01:07:10[solid]so cool
01:07:16HCl :)
01:08:48ashridahhow do you remove an attachment in the wiki?
01:09:02preglownew tremor commit!
01:09:05HClclick attachment
01:09:13HClthen click the "action" next to the
01:09:24HClto update it
01:10:04 Join lostlogic [0] (
01:10:56ashridahwell, it's on its way. doesn't help that i'm on dialup
01:10:59HCl :)
01:11:09HCli'm glad at least some people are able to use it
01:11:16HClits loads and loads better than the perl variant
01:11:24ashridahhrm. i think i've noticed Slasheri's oddity with vorbis as well
01:11:58ashridaha couple of files that had warbling artifacts in them on iriver (that i assumed was a crap encoding job by me) don't seem to have them in rockbox.
01:12:31preglowi also i've left a bug in libmad, btw
01:12:39preglowso don't be surprised if some clear sounding mp3 has artifacts
01:12:40ashridahHCl: i'll look into it tomorrow, but we may also be able to create one with gcj on windows/cygwin too. just not sure if it will require a giant libgcj to go with it
01:13:12HCli tried gcj
01:13:17ashridahHCl: if i can work out how to make a restricted libgcj, we can probably get it down to a small standalone util
01:13:20HClit fails on stringbuilder
01:13:26HClgcj doesn't have that class included yet..
01:13:27ashridahah, hrm.
01:13:42HClwe'd have to rewrite the code to not use it
01:13:45HClor implement our own.
01:13:46 Join lostlogic_ [0] (
01:13:49 Quit matsl (Remote closed the connection)
01:14:19ashridahHCl: which part of the code uses stringbuilder?
01:14:33HClnot sure :/
01:14:46HClpart of the ogg decoder, i think...
01:14:50HClogg tag decoder, tha tis.
01:14:51HClthat is*
01:15:35ashridahdo you possibly mean StringBuffer?
01:15:52HCleh, i don't think so, but i'm not 100% sure
01:16:10ashridahbecause there's definently no reference to StringBuilder :)
01:16:19HClits an internal thing, i think.
01:16:21HCli can try again
01:16:24HClhold on
01:16:32ashridahi'll look at it tomorrow. i've got an exam to deal with today
01:17:32ashridahHCl: SongDB.jar updated in wiki now
01:17:37HClde/jarnbjo/ogg/ error: class 'java.lang.StringBuilder' has no method named 'append' matching signature '(Ljava/lang/String;)Ljava/lang/StringBuilder;'
01:17:53HCl :)
01:18:43 Join matsl [0] (
01:20:07***Saving seen data "./dancer.seen"
01:25:57ashridahi assume you're working on a different version?
01:26:13HCli used the one of the wiki
01:26:20HCljust gcj SongDB.jar
01:26:23ashridahbecause line 233 is just a simple System.out.println
01:26:32HClstringbuilder is an internal class
01:26:34HClthat java uses
01:27:00ashridahah, okay. i'll have to poke at it more later. also need to kick your project a bit until it compiles nicely in eclipse :)
01:27:09HCl ;p
01:27:34 Quit Seed (Read error: 145 (Connection timed out))
01:27:42ashridahnevetheless, there's a working SongDB up for 1.4 and above atm
01:27:55HCl :)
01:28:18 Quit Harpy (Read error: 60 (Operation timed out))
01:29:04preglowwhat the hell
01:29:14preglowsuddenly vorbis2wav says 'file too large' for every file i try
01:29:20HCl :X
01:29:22HClnice optimizations :X
01:29:41HCli hope you didn't commit it yet o-o
01:29:58preglowi'm not doing anything
01:30:17preglowjust testing a couple of ideas for pedro's last commit
01:30:21preglowplayback works fine
01:30:24preglowbut not vorbis2wav
01:31:04 Quit matsl (Remote closed the connection)
01:32:49ashridahpreglow: current CVS works fine on my oggs
01:33:04ashridahand at 150% like he said
01:34:21 Quit lostlogic (Read error: 110 (Connection timed out))
01:35:46preglowtremor really got faster
01:36:38 Join Seed [0] (
01:37:18thegeeknewest jamiroquai album rocks
01:37:40preglow'round 185% realtime here now, with no disk write, and 160kbps vbr
01:39:23Stryke`-q5 ?
01:39:40 Join elinenbe_ [0] (
01:39:41 Quit elinenbe (Read error: 54 (Connection reset by peer))
01:39:46 Nick elinenbe_ is now known as elinenbe (
01:45:08preglow-q6, isn't it?
01:45:09preglowi don't remember
01:45:10elinenbepreglow: what speed is the CPU at when it is running those oggs?
01:45:35preglowwe'll never see 180% realtime at 49mhz, heh
01:45:36thegeekK:\>java SongDB.jar
01:45:36thegeekException in thread "main" java.lang.NoClassDefFoundError: SongDB/jar
01:45:44thegeekanyone got a tip?
01:45:55elinenbewhat is the max speed the CPU safely runs, and the max speed period?
01:46:00Stryke`are there any benchmarks of libmusepack yet?
01:46:29 Join xen` [0] (
01:46:33[solid]thegeek: java -classpath SongDB.jar SongDB
01:47:13thegeekwhy is that needed?
01:47:20ashridah[solid]: the last bit should be unnecessary
01:47:26ashridahjava -jar SongDB.jar
01:47:28ashridahshould just work
01:47:40thegeekit does
01:47:47thegeekI'm a java noob
01:47:57[solid]oh... /me too
01:48:07[solid]didn't know bout the -jar option d'oh
01:49:18HClelinenbe: 120mhz, 140mhz
01:49:44*HCl goes to sleep
01:53:04 Quit preglow ("of")
01:56:58 Join Zoom2 [0] (
02:05:32 Quit Seed (Nick collision from services.)
02:05:39 Join Seed [0] (
02:08:02 Join LinusN [0] (
02:09:19 Join bg_ [0] (
02:09:49 Part Pomyk
02:09:51thegeekwhy do I get unsupported database v2
02:09:54bg_I'm having some issues getting rockbox to boot on my 140, wondering if anyone could help me
02:09:54thegeekusing the latest cvs
02:10:01thegeekand songdb.jar
02:10:04HClcause the database format has been upgraded.
02:10:10thegeekyeah I got that;)
02:10:10HClhuh o.o.
02:10:20HClsongdb.jar should *not* generate a v2 database.
02:10:31thegeekperhaps I have to delete the old db first
02:10:38thegeekI did run songdb in the root
02:10:42thegeekbut perhaps it did not overwrite
02:10:44HClyou need to move the file
02:10:47HClfrom the root to .rockbox
02:10:49ashridahrockbox.id3db needs to be in .rockbox
02:11:01thegeek(feels noobish YET again)
02:11:34 Quit tvelocity ("Leaving")
02:11:52HClbut. i'm asleep.
02:11:52*HCl resumes sleeping.
02:11:54thegeeklooks very nice:)
02:12:52 Quit bg_ (Client Quit)
02:15:21 Quit ashridah ("gone")
02:17:08 Join TCK [0] (
02:20:59 Join ehntoo [0] (
02:24:47 Quit elinenbe (" HydraIRC -> <- IRC for those that like to be different")
02:28:39 Part LinusN
02:29:34Plugh_missed him by a minute
02:41:10 Quit Seed (Nick collision from services.)
02:41:18 Join Seed [0] (
03:15:11 Quit Stryke` (Read error: 60 (Operation timed out))
03:20:11***Saving seen data "./dancer.seen"
03:30:24 Join amiconn_ [0] (
03:46:03 Join tvelocity [0] (
03:48:38 Quit amiconn (Read error: 110 (Connection timed out))
03:48:38 Nick amiconn_ is now known as amiconn (
03:59:42*ehntoo just downloaded the bleeding edge rockbox build
04:00:03ehntoowith wps this time, even.
04:01:06ehntoomaybe it's just me... but I think this sounds even better than the default fw...
04:01:19ehntoothe eq is better without all those sound filters
04:06:00 Join QT_ [0] (as@area51.users.madwifi)
04:08:07 Quit bill2or3 (Read error: 145 (Connection timed out))
04:17:03 Quit QT (No route to host)
04:42:12 Join Stryke` [0] (
04:54:03 Join psycho_maniac420 [0] (
04:56:30psycho_maniac420how is everyone
05:00:11 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
05:04:51 Part psycho_maniac420
05:20:12***Saving seen data "./dancer.seen"
05:32:44 Quit Zoom2 ("CGI:IRC (EOF)")
05:47:50 Join BTKDaImMaikata [0] (~Bager@
06:07:20 Join bill2or3 [0] (
06:08:19 Quit Bager (Read error: 110 (Connection timed out))
06:15:53 Quit BTKDaImMaikata (Read error: 60 (Operation timed out))
06:35:40 Join Strath [0] (
06:36:49 Quit pabs (Read error: 113 (No route to host))
06:57:52 Join cheriff [0] (davem666@wagner.orchestra.cse.unsw.EDU.AU)
07:00:50 Quit StrathAFK (Read error: 110 (Connection timed out))
07:00:58 Join Bager [0] (~Bager@
07:04:38 Quit courtc (Read error: 110 (Connection timed out))
07:05:13 Join courtc [0] (
07:07:43 Join matsl [0] (
07:16:45 Quit matsl (Remote closed the connection)
07:17:49 Join matsl [0] (
07:19:07 Join webguest29 [0] (
07:20:16***Saving seen data "./dancer.seen"
07:53:25Bagermorning, guys
08:05:10 Join noclue2 [0] (
08:08:57 Quit ehntoo (Read error: 60 (Operation timed out))
08:10:57 Join LinusN [0] (
08:13:31 Join IpodMiNi6GiG [0] (
08:13:33 Part IpodMiNi6GiG
08:13:58matsltime for work, cu l8r
08:14:11 Quit matsl (Remote closed the connection)
08:24:15Slasherimorning :)
08:24:29Slasherifinally i fixed the ogg problem, committing now..
08:26:06LinusNnice, what was it?
08:26:21LinusNmorning amiconn
08:26:33LinusNinteresting discussion in the ml
08:26:45Slasherivorbis codec was not freeing dynamically allocated memory and it was overflowing to file buffer. Now calling codec_init on every new track
08:27:13LinusNogg is so lame when it comes to memory handling
08:27:33LinusNdynamic memory is the devils tool
08:27:57LinusNthat's why we don't have dynamic memory in rockbox
08:28:16 Join luminerd [0] (
08:28:24luminerdHey, what's rockbox?
08:28:30LinusNluminerd: lol
08:28:48LinusNluminerd: how did you find this channel?
08:28:56luminerdLinusN, was bored and did a /list
08:28:56Slasheriyep, that is very sensible because cpu has no mmu or virtual memory
08:29:15LinusNSlasheri: exactly
08:29:42*luminerd checks it out
08:29:51*LinusN gets some coffee
08:30:11oxmanwhere i can see the change log of the h120 firmware ?
08:30:16oxmani see a new version 0607 of the firmware
08:30:23oxmanbut i dunno the changement
08:30:31 Join Zoom2 [0] (
08:31:41luminerdah, stuff for windows hm?
08:31:42Slasherioh no, cvs commit opened vi editor.. can i quit that without committing?
08:31:56luminerdSo bassically a replacement to WinMediaPlayer?
08:32:44Slasherihuh, it worked :D
08:33:07luminerdoh, no...avail. for linux too I see
08:33:24luminerdsorry, first thing I saw was .zips and .exes, figured windows software haha
08:33:35LinusNluminerd: not linux either
08:33:50LinusNfor portable mp3 players
08:33:56luminerdLinusN, no?
08:34:07luminerdAhhh! No wonder I'm so confused, never messed with one of those
08:35:17LinusNoxman: there is a cvs change log on the daily build download page
08:35:23 Part luminerd ("Bye, losers. Jk")
08:35:38oxmanok thanks
08:36:12LinusNthis luminerd guy probably has no life whatsoever
08:36:16Slasherifixes are now committed :)
08:36:21oxmanwhere the page ?
08:36:24LinusNSlasheri: goodie
08:37:00LinusNoxman: you can't find the daily build page?
08:37:16oxmanno :(
08:37:29oxmani try
08:37:32oxmanbut i don't see it
08:37:36LinusNon the rockbox home page, there is a menu on the left side
08:37:46oxmanoh sorry
08:37:51LinusNone link is called "daily builds"
08:37:52oxmani find in wiki :D
08:39:49LinusNSlasheri: interesting yielding strategy
08:40:18SlasheriLinusN :) yep, it gives a little better read performance without affecting codec performance too much
08:40:19LinusNhow often do the codecs yield to the gui?
08:40:52Slasheriat least once after every buffer request / new loop
08:40:55amiconnIs it just me or is the iriver line out really much quieter than the archos player line out, or archos recorder set to line level?
08:41:04LinusNSlasheri: and how often is that?
08:41:16LinusNamiconn: haven't thought about it
08:41:17Slasheriit's quite often.. hard to say
08:41:37LinusNit's important to keep the gui fast
08:42:16Slasheriat least vorbis uses quite small buffer sizes (normally below 1 kB) so it will yield always when that much of data are decoded
08:46:23LinusNi guess it's time to separate the codecs from the plugins
08:47:10Slasherigood idea :)
08:47:57Slasheriwould you like to some codec loader stuff should i do something?
08:48:33LinusNi guess i can begin hacking on it
08:48:42Slasherii think it should not be very hard to do because we already have the plugin loader
08:48:46Slasheriok :)
08:57:10amiconnHmm. Does rockbox detect end-of charging on the iriver yet? I know charging is handled in hardware, and I guess the green led is also hardware controlled
08:57:28LinusNamiconn: no it doesn't
08:57:49amiconn...but when charging is complete, the charging animation should stop.
08:58:00amiconnIs there a way to detect this?
08:58:09LinusNthe same way as the fm/v2
08:58:15amiconnA pin?
08:58:36LinusNfm/v2 measures the charging current
08:58:45LinusNiriver has to check the battery voltage
08:59:24amiconnHmm, that won't work too well
08:59:55amiconnCharging is constant-voltage at the end with Li* batteries
09:00:38LinusNi know, do you have a better idea?
09:01:20LinusNi haven't found any other way of doing it
09:01:35LinusNof course, i haven't measured the entire pcb yet
09:03:11amiconnAre all so-far-unknown port pins set to input, and are there more than the 3 adc channels currently shown in the debug menu?
09:04:07amiconnOne method to find out could be adding all adc channels, set all unknown port pins to input (if that does crash we know we missed something ;-) ), and then charge and look at the debug menu
09:04:51amiconnBtw, to add a point for a 2.5 release: There are some important Ondio fixes as well
09:05:27amiconnMultivolume + hotswap, also working on 0308, support for blocksizes != 512 bytes (>1 GB cards)
09:06:00amiconnAnd for fmr / (old) ondiofm: Avoid radio interference from the mas
09:06:14LinusNthere are loads of fixes to be released
09:06:41LinusNi don't think we would have a problem releasing 2.5
09:07:19 Join Chamois [0] (
09:08:20amiconnHmm, iriver q: Did you actually try to enable page mode? If I read the datasheet correctly, this doesn't disable burst mode
09:13:14LinusNmaybe i interpreted the data sheet in a too pessimistic way
09:13:20LinusNlet's try it
09:15:53amiconnI mean , page 100
09:16:31amiconn...the PM flag: 0 = page mode on bursts _only_, 1 = continuous page mode
09:17:41LinusNthat's what i mean too
09:18:50Zoom2what if you just estimated a time
09:18:52Zoom2and set a timer
09:19:08Rickthat's lame
09:19:23Zoom2but an option
09:20:01*LinusN detects a discussion popping up from nowhere
09:20:20***Saving seen data "./dancer.seen"
09:23:43*Rick pops up in front of LinusN and screams bloody murder
09:24:08Zoom2lol you guys are nuts
09:24:32*Rick falls out of the tree onto Zoom2's head.
09:26:32Zoom2So question, I take it the iRiver charging is not good yet?
09:27:00LinusNit works perfectly
09:27:17LinusNin fact, we can't do anything to make it not work perfectly
09:27:33BagerZoom2 the detection is problem ..
09:27:40Bagernot charging
09:27:44Zoom2what are the next big challanges that the rockbox project faces?
09:28:33BagerZoom2 the charging is done in hardware, so rockbox can't do anything to charge battery in bad way or something similar
09:29:14Zoom2ok that makes sense
09:29:49 Nick Lynx_awy is now known as Lynx_ (Lynx@
09:30:09Zoom2just in general terms though, I know the codecs was a big issue for the rockbox project and that seems to be cracked, what are the other big challanges that may be there in the near future?
09:31:38Bagerrecording in format other than .wav ...
09:35:26Zoom2that the only big concern out there?
09:36:05*Bager doesn't see any other
09:36:12Zoom2thats good news
09:36:36Bagerin fact, i don't think we can expect recording in really good quality (such that of lame) in any loseless format
09:37:03Bagerthe CPU just doesn't have enough power
09:37:17Zoom2what is the orginal firmware record in?
09:37:24Bageri mean lossy
09:37:28Bagernot loseless
09:37:30 Join bobTHC [0] (
09:37:42bobTHCmornin' all !
09:38:40t0masmorning :)
09:40:03 Join |maze| [0] (
09:40:30 Quit |maze| (Client Quit)
09:41:25BagerZoom2 the original (as you probably know) records in WAV (loseless) and in mp3 (lossy) format
09:43:20Zoom2so basically the rockbox firmware wouldnt beable to improve on current recordings?
09:43:35Zoom2That doesnt seem too bad for all the new features we would be gaining
09:47:43LinusNthe problem with mp3 recording is that there are no (good) free mp3 real-time encoders out there
09:48:51Zoom2I was aware of this, but is the mp3 codec patented or copywritten? IF it is pattented doesnt it expire in a few years or something?
09:50:53BagerZoom2 the problem is something else
09:51:46Bagersomeone must write opensource GPL compatible good real-time mp3 encoder ....
09:51:52cheriffis it possible to jump to the encoder in the original fw, assuming one is running rockbox off the hdd?
09:52:19Bagerthis is Really Dirty (tm) hack ...
09:52:47Bagercheriff the encoder in the original fw is not so good btw ...
09:52:57Bagerbut it's better than nothing
09:54:03LinusNthat would be very close to a copyright violation
09:54:05Bageron the other side, if we can record in some loseless format (such as WAV (uncompressed), FLAC, etc), you can always reencode the record on the PC with your favourite encoder
09:54:14BagerLinusN i'm not so sure ...
09:54:26LinusNthat's why i said "very close" :-)
09:54:46Bagerin fact, everyone who has bought iriver has paid for all the software inside ...
09:54:50LinusNstill, if a user wants real-time mp3 encoding so bad, let him boot the iriver firmware
09:55:17LinusNjust restart and hold Rec
09:55:40BagerLinusN the built in encoder in the iriver fw is definitely worse than lame as quality
09:55:46Zoom2and what "GOOD" software that was....
09:55:50LinusNof course it is worse
09:56:15cheriffhow much slower than realtime does lame (or similar) actually run?
09:56:19RickI don't think they intended people to be recording live concerts or whatever
09:56:21 Join Aison [0] (
09:56:40Bagercheriff forget about lame on the iriver itself...
09:56:42Rickand I use it for recording class lectures.. and you don't need extreme quality for that :P
09:56:54cheriffthat bad eh?
09:57:04LinusNcheriff: yes
09:57:25BagerRick in fact many people are using their irivers for such purpose .... :)
09:57:33LinusNlame isn't designed for real-time operation, and not for embedded systems either
09:58:02cheriffcan't wait till i get some free time.. finaly can start playing with some of this stuff...
09:58:11LinusNthere is a free mp3 encoder, Shine, but it is very basic and need a lot of work
09:58:25Bagercheriff lame with very hight quality parameters (-q 0) doesn't run realtime even on Athlon 1.5GHz ...
09:58:27RickBagder: yeah! it's really good
09:58:37LinusNsomeone with a clue and a lot of time could start working on it
09:58:38 Nick Bager is now known as Bger (~Bager@
10:00:37Zoom2does realtime recording work well for classes? how close do you have to be to get a good recording?
10:01:09BgerZoom2 you must be above realtime in order to get *anything* usable
10:01:52Zoom2*scratches head* - I am extremly unsavvy in this as you can tell - thanks for being so patient
10:02:49Bgeror your recording will be something such as <1 sec of recording><seconds silence><1 sec of recording (but not the part immediately after the first chunk of recording)>
10:02:53Bger... etc
10:03:48BgerZoom2 don't worry ...
10:04:04Bgerrockbox for sure will record in .wav format well above realtime ;)
10:04:53Zoom2I am just so excited in general somone is picking up our firmware and forgetting all about us
10:04:56Bgermaybe in FLAC also (preglow, do you have anything to way about this one;))
10:05:10Zoom2and not*
10:06:54LinusNwavpack should work too
10:07:33bobTHCfor simple voice recording speex is a good choice imho
10:07:41*Bger looks at MCF5249UM.pdf and after that looks with veneration at Linus for reading all this ...
10:08:02LinusNBger: yeah, it is good bedtime reading :-)
10:08:07Bger :))))
10:08:25LinusNwell, it is a great 1hr commute reading
10:08:45bobTHCa little bit hard to stomach
10:09:04LinusNyou only read the specific sections you need
10:09:30LinusNbut yes, it takes some experience to be able to digest it
10:09:36Bgeryeah, but you must have idea what's it about ....
10:11:15Bgeri don't think that i can read only chapter 7 for example and after that i'll understand everything about the DRAM controller in coldfire...
10:15:06LinusNBger: me neither :-)
10:16:01BgerLinusN you got the idea i had...
10:16:31BgerHCl offt, but what "mrf" does mean ? ;)
10:18:21Zoom2alright guys, I just took 2 sleeping pills alas I am off - thanks again for the quick answers
10:18:31LinusNZoom2: sleep tight
10:18:45 Quit Zoom2 ("CGI:IRC")
10:18:48 Join cYmen [0] (
10:19:06*LinusN doesn't need sleeping pills, he needs amphetamine :-)
10:20:25Bagdersleep is overrated anyway ;-)
10:20:37*Bger hates such pills ...
10:20:57cheriffdoes the CF in the iriver have the 96k scratchpad mentioned in MCF5249UM.pdf?
10:21:11Bagderif scratchpad means iram, then yes
10:21:15Bgerscratchpad ?
10:21:31cheriffthe SRAM, i mean
10:21:54LinusNcheriff: yes
10:22:04Bgertoo late
10:22:13LinusNsection 6 "Static RAM"
10:22:17Bgerit's in 2 banks (iirc)
10:22:35LinusNyes, one DMA capable and one not
10:22:47cheriffso what i have is *exactly* as described here?
10:22:48Bgerjust to say it ...
10:23:02Bgercheriff more or less...
10:23:12cheriffsorry, but i'm used to sony making custom modifications to off the shelf MPIS cores..
10:23:28Bgeras linus said once ... on paper it's better than in reallity ...
10:23:29LinusNcheriff: this is a standard vanilla MCF5249
10:23:43cheriffcool, thanks!
10:24:55BgerLinusN ... "the question" ... do you have any idea when you'll take a look @ your H320 ?
10:26:50 Join Patr3ck [0] (
10:28:26Bgerexpected answer :(
10:30:12 Join Nuxator [0] (
10:30:30Nuxatorhi all
10:31:05LinusNNuxator: shalom
10:31:14Nuxatorast cvs commit was about mp3 file seeking
10:31:18LinusNBger: "the answer" :-)
10:31:21Nuxatorbut it doesn't work on mine
10:31:31LinusNNuxator: oh
10:31:38Bgeryeah, the answer ...
10:31:41Nuxatorif i try to fast forward
10:32:21Nuxatorit restart playing from the begining but the display on the screen is wrong
10:33:35Nuxatorand by the way did you read logs about my player that doesn't leave usb mode?
10:34:18LinusNNuxator: no
10:34:29Bagderwhat tag format does OGG use? Is there a spec URL somewhere?
10:34:33LinusNNuxator: do you build from CVS?
10:34:41Nuxatori got bleeding edge
10:34:45LinusNBagder: apev2 methinks
10:34:58Bagdergoogle does not easily reveal this
10:35:00crwlthat's vorbis comments
10:35:05crwlnothing related to any mp3 tagging format
10:35:05LinusNcrwl: ah
10:35:15crwlFLAC uses the same format, too
10:35:15LinusNcrwl: figures
10:35:42BagderI feel a need to reinvent HCl's wheel
10:35:44crwlvorbis comments are basically unlimited number of key=value pairs (same key can occur multiple times), encoded in utf-8
10:36:15 Join Zagor [0] (
10:36:19Bagderhey Zagor
10:37:29Bagder <= the "Vorbis Comments" spec
10:40:34Bagder... and it doesn't even mention what endian the 32 bit numbers are using
10:41:05Bagderwhat? isn't the whole world Intel?
10:41:30crwlit's in the 2.1.3
10:41:37crwlbitpacking convention, byte order: big endian
10:42:38*Bagder tries to learn
10:42:46 Join niobos [0] (
10:43:41LinusNNuxator: tell me about the stuck-in-usb experience
10:43:49BgerVORBIS_COMMENT: This block is for storing a list of human-readable name/value pairs. Values are encoded using UTF-8. It is an implementation of the Vorbis comment specification. This is the only officially supported tagging mechanism in FLAC. There may be only one VORBIS_COMMENT block in a stream.
10:44:31NuxatorIf i plug my ihp140 it goes to usb mode (in rockbox or in bootloader mode)
10:44:39Bger(got from
10:44:43Nuxatorbut when i unplug it it stay as this
10:45:05Nuxatorjust displays the usb logo or usb bootloader
10:45:17Nuxatorif i press a key backlight swith on
10:45:25Nuxatorbut i can't do anithing else
10:46:22LinusNevery time?
10:46:52Nuxatorand yesterday with mp3 playback usb bug mp3 still plays
10:47:10Nuxatorso it doesn't crash but just stay at usb screen
10:47:59LinusNNuxator: mp3 playback is a different issue
10:48:41Nuxatoryes i know i just did it to see if it's a crash or just if it just doen't detect end of usb mode
10:48:50LinusNNuxator: i can't make it happen on my h140
10:49:31 Join tucoz [0] (
10:49:50Nuxatorand it seems that i'm the only one
10:50:20Nuxatoror maybe an other people but i didn't get if he was speaking for my bug or mp3/usb bug
10:50:41tucozLinusN: now, when I turned on my iriver it started to scream(!), immediatly , as in the bootloader
10:51:05tucozIt's like it loops something quite fast
10:51:48tucozShould the bootloader perhaps reset something, or initialize audio in some way?
10:52:54LinusNtucoz: perhaps
10:53:37Nuxatorby the way: mp3 seeking is working
10:53:55Nuxatori just failed to overwrite my old rockbox.iriver
10:54:07tucozok, It's no worry though, just to unplug the headphones.
10:56:33 Quit tucoz ("CGI:IRC")
10:56:56LinusNPage mode tests: mpa2wav.rock@120MHz catching the digital flow:
10:57:14LinusNWith Page mode: ends at 115.60%
10:57:27LinusNWithout page mode: ends at 115.74%
10:59:31NuxatorIf i let mp3 play plug and unplug: when buffer is empty it retruns to normal mode.... Strange
11:01:15t0masNuxator: you have the newest version?
11:01:22t0masthere was an update this morning...
11:01:39t0mas08:34 - Fixed: Ogg Vorbis crashing, MP3 file seeking, USB mode and a little
11:01:40t0masfaster file buffering.
11:01:49Nuxatori know
11:07:07Nuxatorbut i have bugs with usb
11:07:21Nuxatorand it seems that i'm the only one with them
11:07:30LinusNNuxator: and this never happens with the original firmware?
11:08:06 Quit Nuxator ("CGI:IRC (EOF)")
11:08:49 Join Nuxator [0] (
11:09:41Nuxatorno prob with iriver firmware
11:10:49 Join preglow [0] (
11:10:59NuxatorI use 1.65E patched with fwpatcher.exe
11:11:12Nuxatorso bootloader V2
11:11:25preglowLinusN: and we don't use pagemode right now, no?
11:11:36LinusNno we don't
11:12:50amiconn115%? I thought mp3 was around 300% ??
11:13:06LinusNnot when catching the digital flow
11:13:14preglowwith dithering and wavwriting + 320kbps, i guess
11:13:32LinusNpreglow: which is what mpa2wav does
11:13:41preglowLinusN: not mine ;)
11:14:03LinusNand the current playback code dithers as well
11:14:20Bagdercrwl: I don't see how your URL explains the endian order of the vorbis comments
11:14:44preglowi'll have to do some tests to see if the dithering even matters
11:14:46preglowi suspect not
11:14:53preglowfor me, at least
11:15:43crwlBagder, wouldn't it be odd if the comments used different byte order than everything else
11:16:05Bagdercrwl: but that link doesn't tell what byte order they use, it only explains the concept of byte orders
11:16:40Bagderhm, no
11:16:42BagderI'm wrong
11:16:51LinusNThe Vorbis bitpacking convention specifies storage and bitstream manipulation at the byte, not word, level, thus host word ordering is of a concern only during optimization when writing high performance code that operates on a word of storage at a time rather than by byte.
11:17:09LinusNLogically, bytes are always coded and decoded in order from byte zero through byte n.
11:17:21Bagderand that is...
11:17:29Bagderwhat is byte 0?
11:17:34Bagderin a 32 bit integer
11:17:45LinusNit doesn't use integers
11:17:55LinusN"at the byte level"
11:18:42Bagderso a 32 bit integer is stored how?
11:19:00Bagderyes, I am stupid but I don't understand
11:19:46BagderI _can_ just check an existing file to find out of course
11:20:11LinusNBagder: i agree that the url doesn't tell anything about the vorbiscomments byte order
11:20:21***Saving seen data "./dancer.seen"
11:21:03LinusNwhat the url says is that the vorbis packer is byte-oriented, so endianness is not relevant
11:21:13t0mashm? vorbis is working at the byte level for audio data? so it doesn't care about endiannes there? and in the comments it's host dependant?
11:21:39preglowi expect audio data is coded at the bit level
11:21:52BagderLinusN: yes, but how do they construct "an unsigned integer of 32 bits" from "byte-level" ?
11:21:56LinusNpreglow: yes it is, and the packer packs it into bytes
11:22:06LinusNBagder: does it matter?
11:22:21Bagdersince that's what the comment format uses
11:22:29Bagder 1) [vendor_length] = read an unsigned integer of 32 bits
11:22:35t0masBagder: then I guess big endian...
11:22:41LinusNdoes the vorbis comment format use the bit packer?
11:22:48LinusNi think not
11:23:02Bagderit says very little
11:23:07Bagderone can only assume
11:23:12LinusNtherefore i'd say that the bit packing docs are irrelevant
11:23:14t0masoh wait a moment
11:23:26BagderLinusN: I'd say so too
11:23:28t0masjust found an encoder... with the option to chose... big or little endian
11:23:39t0masso I don't think it's defined by the ogg standard?
11:23:39BagderLinusN: but that leaves us with _no_ hint on byte order for the ocmments ;-)
11:24:06t0mas <−− search the page for endian
11:24:12LinusNBagder: no BOM?
11:24:20BagderLinusN: no
11:24:36 Join stavrosg [0] (
11:24:44Bagderit starts with a 32 bit length
11:24:49Bagdergotta go
11:26:52LinusN"least-significant-bit of the word to be coded into the least significant available bit of the current bitstream octet first"
11:27:00LinusNclear as mud :-)
11:27:41LinusNi interpret this as little endian
11:28:05preglowwow, someone actually using little endian in their format
11:28:58LinusNif we just use oggpack_read(), we should be fine
11:31:31preglowtremor's gotten pretty fast
11:31:59 Quit Nuxator ("CGI:IRC")
11:31:59preglowi wonder why pbvas didn't do the obvious XPROD32 optimisation, works fine here
11:34:36 Join Harpy [0] (
11:35:57 Part stavrosg ("Leaving")
11:38:44 Part LinusN
11:56:15 Join LinusN [0] (
12:17:16 Quit webguest29 ("CGI:IRC")
12:35:02 Join markun [0] (
12:35:32markunpreglow: what was the obvious XPROD32 optimisation he didn't do?
12:38:03preglowlook in asm_mcf5249.h
12:41:37markunDo you know how to rewrite it?
12:41:58preglowi have done so, and it works great
12:42:13preglowhe probably has good reason for not doing it, though, so i've mailed it
12:42:19markunIs there a speed increase?
12:42:29preglowthere's always the chance it overflows, but that chance is there in PROD32 as well, and that's optimised
12:42:34preglowyeah, a small but noticable one
12:42:58markunI noticed my q3 oggs now decode at about 190%
12:43:13preglowyeah, it's gotten pretty fast
12:43:18preglowand there are still things to do with it
12:43:44preglowreplacing the mdct with fft and sorting for one
12:43:51preglowi hope that ends up faster
12:44:06markunIt would be nice if the codecs would use the same fixed point conventions so we could share the mdct code..
12:46:20preglowwell, they might
12:46:22 Join tucoz [0] (
12:46:27preglowfaad uses the same format some places, i know
12:46:39preglowbut i think that might be a bit too much to hope for, heh
12:46:40markunthe same as tremor?
12:46:43preglowlibmad uses its own thing
12:47:03preglowand its own window sizes, so that can just be forgotten about
12:47:09preglowyeah, same as tremor
12:47:10tucozLinusN: Just for the record, I don't think that sound on boot is the bootloader, as it never happens with rec plus play
12:47:17preglowthis is just a wild guess, though, i've just barely looked at it
12:48:35tucozLinusN: But, just now I woke up quite well, as I had my headphones on when this chainsaw-sound emerged through them :)
12:50:41 Join DangerousDan [0] (
12:51:27tucozBut, if noone else has experienced this, then it's possible that it's just my unit that is malfunctioning.
12:51:35ripnetukive had it
12:51:50 Quit DMJC (Connection timed out)
12:51:54ripnetukthe strange buzzing noise on startup. Im guessing its just ebcause the registers on the DAC havent been initialized
12:52:08tucozripnetuk: ok, good to know
12:52:29markunpreglow: can the imdct for faad2 (which uses fft and sorting) be used on tremor?
12:52:41 Join DMJC [0] (
12:52:49tucozripnetuk: buzzing at full volume, right?
12:55:55ripnetukmakes sense
12:56:01ripnetukquite loud yes, maybe not full
12:56:17ripnetukmakes you jump :)
12:56:50tucozyes, if I'm not quite awake, I sure am after that
12:59:23HClits gta san andreas day
12:59:31ripnetukHCl - :)
12:59:36ripnetukxbox yes? or pc?
12:59:40HCli think
12:59:47HCldownloading xbox version now
13:00:11ripnetukHCl - where d'ya get it from? ive not come across it yet? (not asking for URL - just bt, usenet etc)
13:00:17preglowmarkun: depends on the fixed point format
13:00:27HClinternet ;p
13:00:51markunthought you just said faad uses the same format as tremor?
13:00:56preglowmarkun: feel free to have a look, tremors data seems to be sign plus 31 bits of fraction mostly
13:00:58HClmarkun: we mostly gotta do the report, right?
13:01:07 Join zezayer [0] (~chatzilla@
13:01:12preglowmarkun: i also thought i said was a wild guess
13:01:27markunHCl: Yes, think the rest is almost finished
13:01:39HCloh, and we gotta test chars
13:01:44HCland i'd like to add type casting
13:01:47preglowi have seen faad uses that format some places
13:01:52preglownot sure if it's used in imdct
13:01:57*HCl yawns again
13:02:26markunpreglow: I'm looking at it now.
13:02:33 Quit edx (Read error: 60 (Operation timed out))
13:04:32ripnetukHCl - yay for bittorrent :)
13:04:45HCli don't think the pal version is out yet :/
13:04:58HCland our cruddy old tv can't do ntsc
13:04:58ripnetukno, thats the 10th
13:05:06HCloh :(
13:05:08ripnetukdang :(
13:05:09HCli didn't know that.
13:05:49preglowHCl: weren't you download this ten days ago or something? :PP
13:05:52ripnetukcant you switch the xbox from pal to ntsc using the enigma boot loader or somethign?
13:06:06HClpreglow: partly, but it was a fake :)
13:06:15ripnetukHCl - i donwloaded that as well...
13:06:18HClripnetuk: yes, but then the output will be in ntsc
13:06:25HCland our tv can't handle that
13:06:38ripnetuki mean leave it in pal, then dont all games kick out a pal signal?
13:06:50ripnetuknot sure if a specific game can override tho
13:07:04HClsome games refuse to run in pal mode
13:07:12HCleg, super monkey ball deluxe
13:07:40ripnetukoh well... only 3 days to wait
13:08:04ripnetuki probly wont play it until next week anyway, as im off to download festival very soon :)
13:09:37HClits going dissappointingly slow... 23k/s :/
13:09:50HCli'm dumb :)
13:09:55*HCl grabs it off their student net
13:10:11HClwhy use the internet when its on our 100mbit network :)
13:12:39markunpreglow: I think FRAC_CONST does the same as tremor, can you check?
13:14:51preglowfracs are the same, yes
13:15:20preglowthat's the format the emac unit uses
13:15:39preglowand the format tremors's PROD31 functions use
13:18:31 Join XavierGr [0] (
13:18:56XavierGrHi All!
13:18:58markunI like the way the floats are converted to fixed point at compile time.
13:20:23***Saving seen data "./dancer.seen"
13:20:34preglowassuming you've got a compiler that's decent, heh
13:22:27markunof course. I guess gcc doesn't have a problem it.
13:22:45Bagderseems suitable for songdb
13:23:23markunAh, hi XavierGr btw :)
13:24:56 Quit lostlogic_ (Client Quit)
13:25:35*HCl looks at his san andreas download
13:25:38*HCl looks at the time :/
13:26:30XavierGris that for PC?
13:29:57 Part LinusN
13:46:25 Join ashridah [0] (
13:50:27 Part tucoz
13:54:52 Join spiralout [0] (
13:55:00spiralouthi there
13:55:34tvelocityt0mas, *ping*, are you there?
13:55:55XavierGrhi spiralout
13:56:40spiraloutjust one question is the sound-playback support already implemented in the daily build or only in the *zip from slasheri?
13:57:30 Nick zezayer is now known as zezayer_away (~chatzilla@
13:57:40XavierGrit's commited in CVS so every daily build has it
13:57:54 Join muesli- [0] (
13:57:56spiraloutok thank you:)
13:59:09XavierGrhi muesli-
14:00:29 Join Shagnar [0] (
14:00:40muesli-hi XavierGr
14:04:11spiraloutand how can a normaly windoze user without coding skills integrate a patch?
14:05:05Bagder1) don't 2) install the dev environment and learn how to
14:06:30spiraloutah ok
14:08:24XavierGrOR use the Rockbox Devkit!
14:08:39Bagderthat _is_ a dev environment
14:08:41XavierGrI did and I am quite
14:08:54XavierGrAh thought you said cygwin
14:09:07spiraloutok thank you guys i take a look
14:09:30BagderI recommand using linux any day ;-)
14:09:43muesli-i dont :D
14:10:04Bagderthat only shows you don't build rockbox very often :-)
14:10:05muesli-pushing my mouse arround rulez :D to stupid for coding...
14:10:40 Join splat [0] (
14:13:02*Zagor just bought a new digicam
14:13:18preglowwhich one?
14:13:37Zagorthe Fuji F10
14:14:05muesli-i am keen on canons 350d..
14:14:21Zagori was looking at the canons, sonys and casios at first but then found this "dark horse".
14:14:26preglowmuesli-: why, you're not alone in that...
14:14:41preglowi'm gonna get myself a new digicam once i get cash
14:15:37Zagorpersonally I can't see myself lugging around cameras weighing as much as the 350d.
14:15:55preglowhaving a smaller one as well is pretty clever
14:17:51preglowbut i'm very interested in photography as well, so i'd like an slr too
14:17:54muesli-i had a small compact cam like a canon ixus 40. its excellent for fun pics while parties but for landscapes it suxxx
14:18:10muesli-preglow same for me
14:18:32preglowbeing able to switch lenses is pretty neat
14:18:55*Bagder has ogg support in songdb now
14:19:10preglowand so forth
14:19:15*preglow denounces java
14:19:56thegeekhey, an idea: when the iriver is charging and numerical display of the battery is on, why not display just a tiny "charging icon" and show the "real" percentage of the battery instad of that "animation/numerical-loop"
14:20:24 Join dapureplaya [0] (
14:20:26bobTHCthe best "compact camera" is the Panasonic Lumix DMC-FX7
14:20:29*muesli- denounces the costs for lenses/objectives..
14:20:46preglowmuesli-: amen
14:21:36muesli-(same word like in german) thats why i will have one more expensive one
14:21:46Shagnarcanon a80 :rock: ;D
14:21:49Bger"the best camera" is one with open protocol for me :P
14:22:07bobTHCopen protocol for what ?
14:22:10Bgerlike PTP or UMS
14:22:34ZagorBger: most cameras have ums these days
14:22:38bobTHCi dont care about plug my camera
14:22:48BgerZagor: yes, if you exclude Canon ones ...
14:22:52 Join LinusN [0] (
14:23:03Bagderfeel free to test my commit
14:23:05bobTHCi prefere use a cardreader
14:23:19BagderI'm off for a while
14:24:28preglowi will
14:24:33preglowi've got more or less only vorbis files
14:25:40Bgerdefinitely legal ones... ;)
14:26:04preglowBagder: but i thought people vehemently opposed having perl modules external to
14:27:26Zagornot vehemently oppose. but it makes running it from the usb disk more difficult in windows.
14:28:04muesli-l8er girlz
14:30:35XavierGroof snake2 porting finished just now!
14:30:42XavierGrfancy a pic?
14:30:57preglowwhere does rockbox save its settings?=
14:31:15Zagorin a sector on the disk
14:32:29preglowisn't that a bit nasty?
14:33:12Zagorno, it's a sector that is never used.
14:34:21Zagorhowever we have discussed for a long time to dump the sector and use our config file system for all settings. makes things more visible and transparent.
14:34:32Zagorit also make it slower though.
14:35:35preglowplayback has some problems with browsing, yes
14:35:59preglowjust completely locked up with the disk in permanent activity here
14:36:20preglowno shit
14:36:20dapureplayaHi everybody
14:37:04 Nick XavierGr is now known as Xavier|away (
14:37:12zezayer_awayXavier|away: thats amazing!! Well done
14:37:38Xavier|awaythanks to your images it is good looking
14:38:03zezayer_awayu did the hard bit, making teh images was the easy bit
14:38:06preglowit's never been this buggy before
14:52:44 Quit muesli- (Read error: 110 (Connection timed out))
14:52:59 Join edx [0] (
14:54:27 Part dapureplaya
14:57:55 Quit mirak (Read error: 110 (Connection timed out))
14:58:48 Quit splat ()
15:00:14 Join mirak [0] (
15:13:59 Join ghostiger [0] (~ghostiger@9f6383d7b5a46279.session.tor)
15:14:26 Join Tomhab [0] (~tom@
15:16:28 Join bipak [0] (
15:16:31 Join pabs [0] (
15:20:27***Saving seen data "./dancer.seen"
15:29:07Xavier|awayhave to go see you all later!
15:29:14 Quit Xavier|away ()
15:36:31Chamoiswps files are still limited to 13 char
15:36:45Chamoisbut in the wiki i read 26 lines
15:36:50Chamoisoops char means lines
15:45:39 Part Tomhab
15:46:45 Join bg_ [0] (
15:46:54 Join Tomhab [0] (~tom@
15:46:58LinusNChamois: fixed
15:47:41Bgerin code or in the wiki ;)
15:47:46bg_*something to add to the wiki and be aware of... rockbox may not boot if formatted with 3rd party tools such as norton partition magic
15:48:02 Join Musicmad [0] (
15:48:06Bgerbg_ that's not problem with bootloader v2 ;)
15:48:11bg_yes it is
15:48:16bg_i had bootloader v2
15:48:20bg_and it would not boot
15:48:36Bgeri mean that you can always repair it with boot time USB
15:49:09Bgerbtw, isn't this problem with iriver fw also ?
15:49:29bg_worked fine under iriver fw
15:49:42bg_i ended up formatting using irivers built in method
15:49:47bg_and rockbox boots fine now
15:52:58Musicmadis it sufficient to just copy the daily build over and reboot?
15:53:27MusicmadI have Rolo working etc. already
15:53:27Bgerit should be
15:53:34bipakyes works :)
15:53:47Musicmadok - I wonder why I don't have wps then.
15:54:11ChamoisLinus : thx
15:54:28LinusNbg_: "formatted with 3rd party tools"
15:54:44LinusNi seriously doubt it has anything to do with that
15:54:50bg_i bet it does
15:54:56LinusNhow did you partition it?
15:55:15bg_i formatted with partition magic, it set up the partitions itself
15:55:20Bgerbg_ in fact, did you made any partition
15:55:30Bgerthat's the problem
15:55:40bg_when i mounted it under linux, i had to mount it as sdc5
15:55:47bg_after formatting with iriver, its sdc1
15:55:50Bgeryou must make
15:55:54LinusNbg_: then you had no partition table at all
15:56:06LinusNso called "super floppy mode"
15:56:44Bgers/did you made/did you make .....
16:01:30LinusNrockbox is supposed to handle superfloppy mode, but we haven't tested it much with the iriver
16:01:31 Quit Tomhab (Read error: 60 (Operation timed out))
16:02:17CoCoLUSwhy would anyone want to use this mode instead of just using an usual partition table?
16:02:56LinusNbg_: so it has nothing to do with the tool itself
16:03:58LinusNCoCoLUS: it's handy for compact flash etc
16:04:08LinusNwhere you normally only have one partition
16:04:27LinusNand it *should* work with rockbox as well
16:05:05LinusNbut for now, rockbox still requires a partition table to be able to allocate a sector for saving the settings
16:05:24LinusNgotta go now
16:05:26LinusNcu all
16:05:29 Part LinusN
16:05:29Bgercu, LinusN
16:05:34Bgertoo late
16:05:45Slasherinice this nokia 9500, ircing with ssh and going to town by bus :)
16:06:21 Quit bg_ ("Chatzilla [Firefox 1.0.4/20050515]")
16:06:24*Bger hates nokia ...
16:07:13Slasheriand of course rockbox is playing music ;)
16:08:13preglowSlasheri: current cvs really has problems with browsing and music at once
16:08:21Bgerhaving the fact that nokia is finland corp ... ;)
16:08:23preglowSlasheri: i'm hacking up a quick wavpack codec here, btw
16:09:02Bgerpreglow optimising it or ?
16:09:09SlasheriHmm, great. i will try to solve biggest problems soon
16:09:09preglowno, making a codec plugin for it
16:09:21preglowdavid bryant's doing the optimising
16:15:05[solid]preglow: how bout musepack and flac?:)
16:15:59preglow[solid]: no time
16:16:04preglow[solid]: and musepack isn't even realtime
16:16:38[solid]suppose 'hacking a quick' means no support for lossy dualstream playback?
16:18:03preglowour wavpack lib doesn't support that
16:18:11preglowdon't know if it ever will
16:18:18preglowyou'll have to ask david about that
16:18:30[solid]methinks wavpack and musepack support will be the real marketing shiznit for rockbox... everyone will start wanting to have a h1xx then :D
16:18:57preglow"everyone" being people that actually knows musepack and wavpack exists, yes? :P
16:19:02[solid]but i shall not bother you anymore... have fun and good luck
16:19:20[solid]yeah, musepack has a lot of fans unhappy that it has no portable support
16:19:46[solid]and wavpack is slowly growing to be the lossless codec of choice imo
16:20:05[solid](still using flac tho, it's nice to have oggenc encode straight from flac's on linux)
16:20:18preglowyeah, i think wavpack looks nicer myself
16:20:30preglowi also think musepack looks sweet
16:20:37preglowit's a nice and simple codec
16:21:09[solid]and has replaygain supported in the headers, independent of tags, i like that
16:21:21tvelocityer... who can i ask about dict2rdf?
16:24:29*[solid] floats away with portable gapless vorbis in tha hedphones
16:25:21 Join Tomhab [0] (~tom@
16:30:12 Quit Musicmad ("Trillian (")
16:31:59preglowSlasheri: why are the AFMT_ #defines formatted like bitfields, btw? do you predict we'll ever need to combine these in a bitfield?
16:33:00preglowdoing it like this does give us an artificial limit of 32 codecs
16:35:17preglowand why the hell does ".rm" files return CODEC_OGG_VORBIS? :PPP
16:35:23preglowAFMT, i mean
16:36:48 Join crashd [0] (
16:38:45 Quit Tomhab (Read error: 60 (Operation timed out))
16:44:11 Quit ashridah ("Leaving")
16:46:07 Quit DMJC (Read error: 104 (Connection reset by peer))
16:49:06tvelocityt0mas, are you there?
16:49:58t0masI am
16:50:45tvelocityok, some questions about the rdf
16:50:48 Join MO-Pantsu [0] (
16:51:13MO-Pantsuyo I have a weird prob on latest iRiver build
16:51:53tvelocityi keep entries in lowercase and sorted... but the dict format allows spaces and other characters to be used. should i keep them or filter them out?
16:52:12MO-Pantsuif I play an mp3 file 01 in a folder it skips past 02 to 03. But if I put it in a playlist it does not skip the track
16:52:30MO-Pantsuanyone know why? :)
16:52:42preglowplayback is far from finished
16:52:50preglowi'm sure slasheri will look into your problem when he's got time
16:53:45 Join Allex [0] (
16:53:48MO-Pantsuwill there be work on making gapless playback on mp3 flow better too do you know?
16:53:59preglowthere will, as i said, it's far from finished
16:54:12preglowmp3 files that have lame info header will be proper gapless
16:54:14MO-PantsuOk just checking. Amazing work thus far
16:54:32preglowall other mp3 files will have a short gap, there's nothing to be done about that
16:54:51MO-Pantsuoh I thought you might use the trick Winamp does
16:55:03t0mastvelocity: ascii isn't a problem
16:55:07t0masas long as you keep them sorted...
16:55:08preglowMO-Pantsu: which is?
16:55:08MO-PantsuI can get nearly any mp3 gapless using Winamp's buffering method
16:55:11t0masand under the maxwordlen
16:55:15t0mas(32 chars atm)
16:55:21preglowbuffering method?
16:55:22MO-Pantsuas in does not skip or jump forward
16:55:28tvelocity32 chars? ok
16:55:38preglowwell, i'm sure winamp too has short gaps in mp3s that don't have the lame info header
16:55:40t0mastvelocity: make that definable please?
16:55:53MO-Pantsupreglow I don't hear them
16:56:00tvelocitywith a parameter?
16:56:20 Quit Allex (Client Quit)
16:56:24preglowMO-Pantsu: well, are you sure the files haven't got the info header?
16:56:24MO-PantsuBuffer length 2000 ms Prebuffer 500 ms Buffer ahead on track change 500 ms
16:56:36preglowMO-Pantsu: that's a winamp hack and has nothing to do with us
16:56:38t0mastvelocity: no in code is ok
16:56:39preglowMO-Pantsu: we always prebuffer
16:56:40tvelocitydict also has newlines in the definitions. should I just replace them with spaces?
16:56:47t0mastvelocity: yes
16:56:53tvelocityok then
16:56:58tvelocityi'm almost done
16:57:01MO-Pantsujust testing a mixset have and it jumps ahead too far on track change
16:57:10MO-Pantsudoes not do that in winamp
16:57:24MO-Pantsulike it cuts off too much in rockbox
16:57:26preglowwell, it will be fixed
16:57:38preglowthe gapless mp3 thing is just a hack now, since slasheri doens't know how it's supposed to work
16:57:42preglowi gotta go
16:57:45MO-Pantsuoh ok
16:58:07MO-Pantsuif it;s not done it's not done I didn't think it was finalised yet
17:07:15 Join DMJC [0] (
17:08:02 Part DMJC ("Leaving")
17:10:10 Quit thegeek (Read error: 54 (Connection reset by peer))
17:13:21 Join Stryke` [0] (
17:20:28***Saving seen data "./dancer.seen"
17:23:12HClhmm, too bad
17:23:23HClgta san andreas insists on ntsc mode :/
17:23:34preglowwell, it's an ntsc game...
17:23:39preglowmost newer televisions do ntsc anyway
17:23:46HCli know
17:23:49HClwe have a crappy old one
17:23:53preglowmine isn't one of them
17:23:54HCli'm saving up for an lcd tv
17:24:15preglowi'll just use my monitor :P
17:24:21HClsame thing
17:24:28HClmy lcd tv will double as my tft monitor.
17:24:32HClreally nice, actually
17:24:41HCl26" screen
17:24:42HCl :)
17:24:50preglowsounds cheap
17:24:54HCl :P
17:25:00HClit is, relatively :)
17:25:16HClcause tft tvs don't sell for as much as tft monitors
17:25:23HClyou can get a 26" for +- 800 euro
17:26:14 Join thegeek [0] (
17:27:00preglowcrappy resolution, then
17:27:06preglowwouldn't exactly use that for a monitor
17:27:07HClnot too bad
17:27:22preglowwell, yeah, a bit strange for a monitor ;)
17:27:31HClmy laptop is 1280x800 as well.
17:27:32Zagorpixels the size of flies :-)
17:27:41ZagorHCl: yeah, but your laptop isn't 26"...
17:27:44HClstick it on your wall, wireless keyboard, wireless mouse :)
17:28:20HCli only use my computer for games and movies
17:28:52 Join tedboer [0] (
17:29:58preglowwhy doesn't rolo work consistently?
17:32:42tedboeris there any news on the H3X0 lcd display controller?
17:38:42niobosHCl: I have a question for the album thing
17:38:52niobosI don't get what is defined as "album"
17:39:04niobosa directory? all files with the same "album" field in the tag?
17:39:13HClall songs in a directory with the same album field
17:39:21HCloptionally, you can change that to all songs in a directory
17:39:40niobosso songs in a different dir are ALWAYS different albums?
17:40:05nioboswhat do I use for the "album" field in the database then?
17:40:18niobosthe directory
17:40:21HCllook at the java version, the java version has an option
17:40:24HClof directory
17:40:28HClalbumname + directory
17:40:32niobosI know, but I didn't understand it
17:40:43HClits just the key to store it..
17:41:10niobosbecause you do the album_name ___ directory thing
17:41:32HClalbum name + directory to identify an album
17:41:38niobosso in the database, the stored "album" field should always equal the field in the file's tag
17:41:52HClunless its an multi artist album
17:42:02niobosbut 2 same "album" fields are only considered to be equal if the files are also in the same diretory
17:42:35tedboerdid anyone try if maybe the H3x0 display controller is compatible with the H1x0 one?
17:42:47 Nick Lynx_ is now known as Lynx_awy (Lynx@
17:43:04niobosif the directory's are named "artist - album_name"
17:43:20niobosthe dir=album will work
17:43:40niobosbut the dir=album=name will give you album-names including the artist, correct?
17:44:14niobosok, I think I get it... thx
17:49:56 Join F1^Aison [0] (
17:50:49HCloh, also, when dirisalbum is on, file fields needn't be the album anymore..
17:52:13*niobos is confused
17:52:15HClcause albums are only associated with directories
17:52:17 Part oxman
17:52:43niobosisn't that with dirisalbumname?
17:53:20HCldirisalbumname makes albumnames default to directory name when no album tag is present
17:53:49nioboshmm... seems like I'm lost again...
17:53:56HClsorry :/
17:54:29niobosin the perl-tool you can set dirisalbum and dirisalbumname seperately
17:54:49niobosso situation1: none set, what do I do?
17:54:51HClthey are seperate options.
17:55:05HClalbums are identified by album name and directory
17:55:22HClwhen no album tag is present, "<no album tag>" is used for the album tag
17:56:15niobosand all files without album tag are seen as 1 album with the name "<no album tag>"?
17:56:29HClalbums are still associated to album and directory
17:56:38HClso you can get multiple no album tags
17:57:00niobosok, so all files without album tag in the same directory are seen as....
17:57:07 Quit Aison (Operation timed out)
17:57:14HClan album without an album tag
17:59:04niobosso I have an album entry with album="", artist=0x0000000 and songs[] all the songs without album tag
18:00:34HClalbum="<no album tag>"
18:00:36nioboswell, drop it, I'll study the perl code
18:00:38HClartist = whatever artist those songs have
18:00:49HCland songs[] all the sonngs in that directory.
18:01:15niobosI'll come back when I'm educated...
18:03:23 Nick F1^Aison is now known as Aison (
18:04:14 Quit Patr3ck (Read error: 104 (Connection reset by peer))
18:16:34bobTHC61, god it's a peak !
18:16:47 Join linuxstb [0] (
18:17:53amiconnCoCoLUS (and LinusN, for the log): The problem with rockbox after formatting with 3rd party tool wasn't the 3rd party tool, and it wasn't superfloppy mode
18:18:01preglowlinuxstb: wow, haven't seen you around for a while
18:18:41amiconnThe real problem was (derived from mentioning sdc5) was that the drive was partitioned as a logical drive within an extended partition
18:18:52linuxstbpreglow: No, Real Life has took over. Anything new? :-)
18:18:57amiconn...and logical drives are (still) unsupported in rockbox
18:19:35preglowlinuxstb: oh, this and that
18:19:45*t0mas np: Ashlee Simpson - La La (0:23) []
18:19:52t0masfond an ogg :D
18:19:57bobTHCcya fellows!
18:20:01 Quit bobTHC ("Smoke Weed Every Day !")
18:20:35linuxstbI noticed that only MPA and OGG have codecXXX files - is there a reason for that, or is it just that no-one has got around to writing them for the other codecs?
18:21:06preglowlinuxstb: i'm sitting on a codecwavpack as we speak, but can't get it to work
18:21:25preglowand i'm really short on time, so i'll let it be until slasheri appears again
18:21:45linuxstbI was thinking of trying to get a codecflac written this evening.
18:21:55preglowsounds good
18:22:30linuxstbThe other thing I noticed is that the iRiver only appears to support 11.025, 22.5 and 44.1KHz samplerates - i.e. no 48KHz :-(. Is that a big problem?
18:22:44preglowwe'll have to resample 48khz
18:23:11linuxstb:-( Are you saying the iRiver can't do 48KHz? Almost all my files are 48KHz....
18:23:29preglowit can do it, but then it can't do 44.1khz and those guys
18:23:33preglowit can't do both
18:23:42preglowthis is how iriver does it, even
18:23:53pregloweven optical out is resampled
18:24:15linuxstbSo what's stopping us switching between the two modes, depending on what's playing?
18:24:22preglowyou'd have to ask linus
18:25:04preglownow that i think of it, i'm not even sure it can support 48khz
18:25:07linuxstbOK - hopefully he'll read the logs.
18:25:34preglowthere is an explanation in the logs already, though
18:26:15linuxstbI've searched the logs, but only the last couple of weeks. Any idea when it was?
18:28:35preglowmonths ago
18:29:27linuxstbI've just found something on 20050426 - I'll keep browsing...
18:29:27preglow07.40.51 # <LinusN> damn! it looks like we will have to resample 48000hz to 44100hz, since the hardware can't handle 48000 :-(
18:29:52preglow20050317 <-
18:30:12preglowresampling from 48khz to 44.1khz properly is a pain...
18:31:06preglowwill need very large filters, so we'll probably go for a low quality thing
18:31:12preglowlike ordinary interpolation
18:31:47preglowi'm pretty sure i won't be able to hear the aliasing frequencies anyway
18:31:52CoCoLUSthat means no direct ac3 output?
18:32:07preglowCoCoLUS: sure, it'll just be resampled
18:32:13preglowCoCoLUS: poorly
18:32:22linuxstbSo it sounds like it was a hardware design choice to limit it to the 44.1KHz family. I suppose that makes sense in a way, but it's very annoying.
18:32:24 Quit Aison (Read error: 145 (Connection timed out))
18:32:34preglownot so much a choice as a constraint
18:32:42preglowit's a coldfire issue, probably
18:32:53preglowdoesn't have enough clock divider options or something
18:34:18 Quit Stryke` (Read error: 110 (Connection timed out))
18:35:49preglowCoCoLUS: i don't think i'll be able to hear it, no, the distortion will be in the frequences 20khz
18:35:58preglowfrequencies above 20khz
18:36:39 Join Aison [0] (
18:45:05markunpreglow: I noticed something about the DC offset problem. When I press stop when there is silence in the song I hear no 'plop' when I plug in my headphones
18:45:20markunbut when I press stop during some sound I can hear the plop.
18:46:40markunDoes that sound strange?
18:47:10markunI'll test some more.
18:47:12preglowthe might very well be a plop during sound
18:47:14preglowtry it for iriver fw
18:47:19preglowi suspect it's there as well
18:47:26preglowthere should be no dc offset while audio is playing
18:47:37preglowafter stop it's more likely
18:50:07markunI'm quite positive, can you try it? I think we have to reset the dac after stop or something.
18:50:46amiconnMaybe it's just that the dac isn't silenced at stop
18:50:48preglowyes, sure, i'm referring to while playing
18:51:01preglowdc offset after stopping isn't that surprising
18:51:11preglowdac is probably stuck at its last value
18:52:09markunIn the UDA1380 datasheet there are some tips on how to avoid plops
18:55:10 Join F1^Aison [0] (
18:57:04 Join TCK- [0] (
19:01:02 Quit Aison (Read error: 145 (Connection timed out))
19:01:13 Nick F1^Aison is now known as Aison (
19:08:18 Join bobTHC [0] (
19:09:02 Nick cheriff is now known as cheriff_AWAY (davem666@wagner.orchestra.cse.unsw.EDU.AU)
19:14:22 Quit TCK (Read error: 110 (Connection timed out))
19:15:54 Join DMJC [0] (
19:16:21 Join simple_sample [0] (
19:16:28DMJCI'll be giving you money for this
19:16:36DMJCfinalyl working audio..
19:17:44HCl :p
19:17:49DMJClistening to some rammstein .oggs now
19:17:54HCli need to get on runtime database.
19:18:04HClSlasheri: awake?
19:18:09linuxstbQuick question - does the ci->audiobuffer_insert function take big-endian or little-endian samples?
19:18:17HClor maybe preglow knows what code gets executed when you select to play a file?
19:18:20t0maslinuxstb: ogg?
19:18:28simple_sampleHello! Is it possible to replace the harddisk of the "Archos Jukebox Recorder 20"?
19:18:48linuxstbt0mas: I'm working on adding FLAC support to the new playback system.
19:19:10t0masah ok... then I don't know... but I guess big endian as rockbox is bigendian?
19:19:41DMJCit's like christmas all at once
19:19:54DMJCosx on x86... rockbox..
19:20:30***Saving seen data "./dancer.seen"
19:20:52SlasheriHCl: yes, just came to home. now eating something :) ->
19:20:58preglowlinuxstb: big endian
19:21:12HCli was just wanting to know where i can find the code of when a song ends
19:21:25HClsince its easier to ask than to dig through the code
19:21:34linuxstbpreglow: thanks.
19:21:51preglowSlasheri: yo, i've added wavpack codec, but rockbox still just seems to call wv2wav when i try to play it
19:22:21SlasheriHmm, wv2wav?
19:22:27SlasheriThat's really strange
19:22:40Slashericheck if viewers.conf has some defines about it and remove those
19:23:10Slasheribtw, have you added correct information to playback.c?
19:23:12 Join Hansmaulwurf [0] (
19:23:43SlasheriYou should also define wavpack file extension to probe_file_forma
19:23:56simple_sampleDoes nobody know it? (Is it possible to replace the harddisk of "Archos Jukebox Recorder 20"?)
19:24:43bobTHCfor sure simple_sample
19:25:01SlasheriHCl: Hmm, that real track end event is not still added to playback.c
19:25:15HClso i should wait till thats implemented
19:25:18SlasheriIt detects song changing but timing is not exact
19:25:25Slasherimaybe, depends what you need
19:25:33HCli need something that can tell me the song has ended
19:25:38preglowSlasheri: bah, didn't think of that ;)
19:25:39HCland the absolute path
19:25:41HClof what was played
19:25:57HClthats all i need for the runtime database
19:25:58SlasheriHCl: Hmm, that absolute path stuff needs to be added
19:26:02Slasheribut back soon ->
19:26:04HClyea, i assumed as much
19:26:07HClsame, i'm gonna make dinner
19:26:08Slasheripreglow :D
19:26:41preglowSlasheri: well, ogg and mp3 is still listed there, yet they play fine
19:26:48preglowwhy shouldn't .wv as well?
19:32:24 Nick DMJC is now known as DMJC-sleep (
19:33:06HansmaulwurfQ: is there maybe a bug that rockbox cant play a MP3 file if the file-name is too long?
19:33:45bobTHCeverything is possible it's a CVS version
19:34:56bobTHCi dunno if it's already reported but now it is :)
19:35:50amiconnpreglow: Aas you did quite some asm functions, what is the m68k convention for return values? A register? Using the stack like the parameters?
19:36:19amiconnAlso, which registers are considered scratch?
19:39:18simple_sample@bobTHC: Thanks.
19:39:21preglowamiconn: d0-d1, a0-a1 is considered scratch
19:39:27preglowi've never returned a value from a function before ;)
19:39:29 Part simple_sample ("Leaving")
19:39:51preglowbut i think d0 or a0 are the return value regs
19:40:13amiconnpreglow: 4 registers isn't much...
19:40:33preglowamiconn: indeed
19:40:33amiconn...need to push some then.
19:40:41preglowamiconn: you need to push quite a lot, yes
19:40:51preglowin my biggest functions, i need to push everything
19:41:10amiconnYou don't need to return anything in all your functions?
19:41:48preglowi haven't done _that_ many
19:41:53Slasheripreglow: i don't know. What happens if you remove wv2wav.rock?
19:42:51preglowcomplains that it cant find wv2wav.rock
19:43:05SlasheriHmm, what does it exactly say?
19:43:34preglowgimme a sec, i'm removing the entry
19:44:15preglowdoesn't do anything at all if i remove the entry
19:44:33preglowbut yeah, i've added .wv to the extension checked
19:45:00linuxstbpreglow: Have you added .wv to apps/tree.c ?
19:45:12preglowi tried, but that didn't work either
19:45:19preglowi think the extension list isn't big enough
19:45:23preglowsaw a complaint about that
19:46:03Slasheriah, that might be the problem. try increasing the size?
19:47:59preglowwhere do i do that? can't find any obvious defines
19:48:59Slasherii don't know.. :D
19:49:02 Join Yokalosh [0] (
19:49:19Slasheritry to find what function says the warning message
19:49:46YokaloshSorry in advance for being an idiot again, but i don't really get the conditional tags in wps
19:50:25preglowMAX_EXTTYPES is 32
19:50:33preglowand i sure as hell haven't got 32 yet
19:50:34Yokaloshcan anyone enlighten me?
19:50:41preglowunless viewers.conf also populates the list, of course
19:50:48preglowi'll bump it up to 64
19:51:25preglowSlasheri: probably should add a codecs.conf or something
19:51:44Slasheripreglow: yep, i have also though about that
19:52:43HClHansmaulwurf: the absolute path limit is 224, iirc
19:53:10Hansmaulwurfhm, i think then the problem is something other
19:53:12preglowSlasheri: seems to work better now, but my codec bugs, heh
19:53:14Hansmaulwurfmaybe ID-Tag
19:53:17Hansmaulwurfi will test it
19:54:43 Quit Yokalosh ("CGI:IRC")
19:58:03amiconnpreglow: Found it, d0 is return value
19:58:52preglowamiconn: a0 if you return a pointer, i think
20:01:08 Join webguest40 [0] (
20:03:55amiconnIt's always d0
20:04:03amiconnI checked disassembled rockbox
20:05:00 Join Stryke` [0] (
20:05:09webguest40Hello there. I have a quick question about building rockbox for iriver. Should I be using gcc 3.4.x or 3.3.4 The wiki page mentions both!
20:05:54MO-Pantsuis there any chance of having a user conf for gapless playback of MP3 files?
20:06:50webguest40Thanks. 3.3.4 seems to crash anyway ;-)
20:07:13preglowMO-Pantsu: what do you mean by user conf?
20:07:28 Join muesli- [0] (
20:07:37MO-Pantsuto adjust buffer/prebuffer etc
20:07:52preglowwhy do you need that?
20:07:54preglowgapless should just work
20:08:11preglowdon't use winamp as a reference for gapless playback, winamp is shit at gapless playback
20:08:15MO-Pantsubtw I have some Lame file mixsets that it skips too far ahead on and is not a seamless transition
20:08:30MO-PantsuI find it the best actually
20:08:43preglowhaven't seen anything that beats foobar
20:08:50markunme neither
20:08:55MO-PantsuI have tried all the gapless plugins and foobar2k also and winamp's still has the most seamless transitions on dance mixsets
20:09:14MO-Pantsufoobar2k's transitions can sometimes glitch I noticed
20:09:15HClisn't that cause winamp has crossfading?
20:09:24MO-PantsuI don't crossfade
20:09:26preglowcrossfading is most definitely cheating
20:09:41tvelocitywhy don't people just use ogg?:P
20:09:54markunSome people do
20:10:19MO-PantsuI set winamp's DirectSound out to remove silence and buffer 2000 ms, prebuffer 500 ms and track ahead 500 ms. ALWAYS get perfect transitions
20:10:30amiconnThere are only 2 ways how to get true gapless mp3 playback
20:10:33MO-Pantsubecuase most people have MP3 they download roffle
20:10:34muesli-they are not that popular in file sharing tools..
20:11:13amiconn(1) The files need to be encoded as a consecutive set, and the split points adjusted to mp3 frame boundaries. lame −−nogap does this
20:11:45muesli-i have tried −−nogap in eac..but didnt work
20:11:50MO-Pantsuwell this mixset was encoded using 3.87 beta 1 and it glitches on rockbox because rockbox jumps too far ahead into the next track
20:11:54amiconn(2) When the files are encoded with lame, the lame tag can tell how many samples to cut from the last frame
20:12:17MO-PantsuI will test it on fb2k if you insist though to make sure it works there as a nogap fileset
20:12:42amiconnmuesli-: −−nogap does work only if you give lame _all_ .wav files on the same command line
20:12:51preglowMO-Pantsu: rockbox doesn't have gapless mp3, stop mentioning it
20:12:57amiconn...for obvious reasons, and that cannot work with eac
20:13:15MO-PantsuI thought it did?
20:13:23muesli-−−preset standard −−id3v2-only −−pad-id3v2 −−tt "%t" −−ta "%a" −−tl "%g" −−ty "%y" −−tn "%n" −−tg "%m" %s %d
20:13:31muesli-used this setting
20:13:46preglowMO-Pantsu: it tries, but like you hear, it fails
20:13:59preglowMO-Pantsu: slasheri still doesn't know how to use the info in the lame header to make gapless playback work like it should
20:14:06preglowand no one has cared to update cvs to fix it
20:14:09amiconnIf you want to use −−nogap, you need to extract to .wav with eac first, and encode all .wavs with one 'lame' command line
20:14:26preglowyou don't want to use −−nogap unless you're on a hardware codec box anyway
20:14:33t0mashow "long" is one mp3 frame? (in miliseconds)
20:14:53MO-Pantsupreglow that is all I needed to know
20:15:13MO-Pantsuwhen you mentioned it earlier today I thought you meant ALL MP3. Not just Lame ones
20:15:16amiconnt0mas: Depends on the sample rate
20:15:16preglowt0mas: depends on the sample rate, for 44100, it's 25ms or something
20:15:34preglowMO-Pantsu: there is no possible way to make non-lame mp3s gapless
20:15:45MO-Pantsuwhy not?
20:15:45amiconn48/44.1/32 and 24/22.05/16: 24 ms / 26.12... ms / 36 ms
20:15:57amiconnDouble the times for 12/11.025/8
20:15:59t0mas25 ms... so the gap < 25 ms if I do it without −−nogap ?
20:16:15preglowMO-Pantsu: because the accurate song length isn't saved, you don't know which parts of the last frame in an mp3 contains sound that shouldn't be there
20:16:35Slasheribtw, if anybody has 48, 32, 22 kHz etc. mp3 or ogg files, please send
20:16:41MO-PantsuI'd accept a good compromise that cheats a bit like Nullsoft's output does
20:16:45t0masI can create one...
20:16:59Slasherit0mas: hmm, great :)
20:17:07t0masanybody using windows here?
20:17:12oddi am t0mas
20:17:18t0masinstall fruityloops... and click some beats together
20:17:25t0masthen output as mp3 and chose 48 khz...
20:17:27preglowSlasheri: are you on the mailing list?
20:17:35oddand if i can just input on this convo, i would like to second MO-Pantsu's proposal for cheating like nullsoft
20:17:42muesli-there is a new beta
20:17:42amiconnSlasheri: I have all sorts of mp3 and mp2, a whole test set
20:17:44tvelocityfruity loops? yuk!:P
20:17:45Slasheripreglow: no i'am not. How can i join in?
20:17:51muesli-maybe it supports −−nogap
20:17:57Slasheriamiconn: cool
20:18:03MO-PantsuBTW it works gapless in FB2K I just tested but FB2K still glitches slightly even on -nogap Lame MP3's
20:18:05Slasherimarkun: thanks :)
20:18:10preglowSlasheri: just go and click mailing list, daniel commented on your latest patch now, check the archive
20:18:12MO-Pantsuwhich is why I switched back to Winamp
20:18:24amiconnSlasheri: All possiblke sample frequencies encoded with all possible bitrates. Everything the same song; a total of 600 MB
20:18:25preglowMO-Pantsu: that's impossible
20:18:25markunSlasheri: no problem.
20:18:31MO-Pantsuyes possible
20:18:35Slasheripreglow: ah, ok :)
20:18:45MO-PantsuUnless FB2K needs something additional done in settings
20:18:48preglowMO-Pantsu: no, it they're −−nogap encoded, the entire last frame is filled, so there is no gap
20:19:15MO-PantsuHow do I know if the files are correctly -nogap?
20:19:19preglowyou don't
20:19:27oddyou encode them yourself ;p
20:19:30preglowif you don't know, why do you say you tried -nogap mp3s
20:19:41preglowvery few people encode with -nogap
20:19:52MO-PantsuI think I tested this myself a while back. I encoded myself with -nogap and stil had the same issue
20:19:54markunSlasheri: I also have some Autechre music in 48kHz for you if you want.
20:20:00t0maspreglow: loads do it... but encode just 1 file with it... and that doesn't work
20:20:00MO-PantsuFB2K just does not cut it
20:20:10preglowMO-Pantsu: encoding correctly with nogap is a bit tricky, you need to pass lame ALL your wav files
20:20:24MO-PantsuI do that in EAC
20:20:35amiconnSlasheri: I used it for analysing the MAS limitations described here:
20:20:39Slasherimarkun: everything is welcome. I just want to test different sample rates switching with as many sample rates as possible
20:20:41preglowt0mas: well, they should stop, then ;) using the ordininary info header is better than splicing wav data around in the mp3s
20:20:47MO-PantsuI am gonna do one now just to get it off my chest
20:21:05preglowSlasheri: rockbox doesn't support 48/24/etc khz sample rates
20:21:11preglowSlasheri: they need resampling
20:21:14odddoes .ogg not use frames like mp3 does?
20:21:17preglowodd: yes
20:21:23oddyay for ogg :D
20:21:24preglowodd: but vorbis saves the exact song length
20:21:31oddah, i see
20:21:39tvelocityvorbis rulez
20:21:43preglowMO-Pantsu: fb2k does very much cut it
20:21:47Slasheripreglow: is that sure? rockbox or uda doesn't support them?
20:21:57tvelocitythe way you can have multiple streams in one file... it's so elegant!
20:21:57preglowMO-Pantsu: i suspect you're doing something wrong, the people who use fb2k are _very_ demanding when it comes to this
20:22:04preglowalmost outright anal
20:22:12oddi ought to make some stickers that advertise ogg vorbis... mp3 must be destroyed
20:22:18preglowSlasheri: uda supports them, but there's a problem somewhere
20:22:27MO-PantsuI am doing an EAC rip of the same mixset with the -nogap added in EAC
20:22:27preglowSlasheri: even irivers firmware resamples 48khz files to 44.1
20:22:31Slasheripreglow: Hmm, that needs to be fixed
20:22:41preglowSlasheri: actually, it resamples _everything_ to 44.1, i think
20:22:46preglowSlasheri: can't be fixed, it's a hardware limitation
20:22:53Slasheriok :/
20:22:57preglowwe need to resample
20:23:03preglowdo some cheap linear interpolation, or something
20:23:09preglowreal resampling is too heavy
20:23:36amiconnMO-Pantsu: −−nogap won't work directly within eac
20:23:45 Join ehntoo [0] (
20:24:01preglowMO-Pantsu: using −−nogap is a waste of time, i can guarantee you fb2k wont glitch with −−nogap
20:24:03amiconnIt's described in the lame docs
20:24:07preglow_nothing_ should glitch with nogap
20:24:49markunSlasheri: I think the audio should be muted after you press stop. Should it be done in dma_stop?
20:24:54MO-Pantsuit glitches for me
20:25:20amiconnMO-Pantsu: [20:25:39] <amiconn> MO-Pantsu: −−nogap won't work directly within eac
20:25:29preglowwhere does DEBUGF messages end up?
20:25:30MO-Pantsuamiconn I just noticed
20:25:35MO-Pantsuthat sucks
20:25:52MO-Pantsuit errors when you add it as an option to the options line of Lame
20:25:56preglowMO-Pantsu: not really, you don't need −−nogap unless you use mp3 players with hardware decoders
20:26:03Slasherimarkun: Hmm, maybe. Just remember to add unmute too :)
20:26:19t0maspreglow: debugf is nothing in a device build
20:26:26t0masin simulator you get the info in console
20:26:35MO-Pantsuso why is fb2k glitching?
20:26:36t0masthat's why Bagder created logf
20:26:38preglowthen someone needs to add logf to the plugin api
20:26:46amiconnpreglow: DEBUGF() is nonvoid only for debug builds, or in the simulator
20:26:47t0maspreglow: no
20:26:59*tvelocity goodnightz everyone
20:27:04preglowt0mas: why not?
20:27:05t0maspreglow: then it will give errors when you compile without it
20:27:14preglowso no debugging in plugins?
20:27:14t0mascheck IRC logs... or ask Bagder
20:27:16t0masI wanted to add it to
20:27:33markunSlasheri: Maybe we could even power-down the dac when it's stopped to save some power
20:27:36t0maspreglow: Bagder is looking for some good solution to it...
20:27:58preglowit's should be very easy, just add a wrapper that's always compiled in, but does nothing if logf support isn't added
20:28:08preglowplugin_logf() or whatever
20:28:09t0masI proposed that too
20:28:11MO-Pantsubtw my soundcard has no DirectSound hardware mixing
20:28:20t0masbut that'll take resources
20:28:21MO-Pantsuprobably why it glitches
20:28:29preglowMO-Pantsu: i sincerely doubt it
20:28:37Slasherimarkun: that's good idea. At least it would make sure there remains no dc offset after audio stop
20:28:45*t0mas just has crossfade enabled in amarok...
20:28:50t0masfor 1 sec...
20:30:00MO-Pantsubut anyhow it's not FB2K I am interested in so I don't know why I am bothering testing that
20:30:05markunSlasheri: Digital mute should also take care of that. Don't know if we should mute during pause..
20:33:05MO-Pantsuoh well it was either the ripping or that beta version of Lame but you are correct. Just ripped the same 2 tracks using release version of Lame and no gap in FB2K
20:33:33MO-PantsuBut I still have a lot of Non-Lame tracks that could do with some cheating gapless method
20:34:00HCldon't forget an option to insert gaps on demand..
20:34:07HCl :/
20:34:10MO-PantsuGapless is all I am interested in on Rockbox. Not even bothered about on the fly playlists or games or anything else.
20:34:28*HCl mostly in on the fly playlists
20:34:45MO-PantsuI listen to a lot of dance and live stuff
20:35:02HCli listen to so much different stuff that i almost never know what to listen to
20:35:14HClsearchengine + runtimedatabase will be able to generate me playlists of stuff to listen to
20:36:53MO-PantsuGonna rip that entire album again just to see how it sound on Rockbox anyhow with the correct Lame headers
20:37:21markunMO-Pantsu: You can try ripping to ogg if you want to have gapless now
20:37:36MO-PantsuI know but most of my stuff is MP3
20:37:47MO-PantsuI don't 'own' most of my collection ;)
20:38:10MO-Pantsuand no way am I transcoding it to Ogg and losing quality
20:38:21markunMO-Pantsu: I meant just for now because gapless mp3 is not supported.
20:38:25MO-Pantsubesides MP3 to Ogg won't work for gapless
20:38:38markunMO-Pantsu: I thought you wanted to re-rip?
20:38:43amiconnSlasheri: Could you add .mp2 extension as well?
20:39:00MO-Pantsuonly re-ripping this particular mixset to test
20:39:05preglowi think we need a codecs.conf soon :]
20:39:23MO-PantsuI don't have to re-rip anything to get seamless playback of any filetype in winamp
20:39:27Slasheriamiconn: ah, i will do that :)
20:39:34 Quit noclue2 (Read error: 110 (Connection timed out))
20:39:39markunMO-Pantsu: Than just be patient.
20:39:49MO-Pantsuof course
20:39:51amiconnIs there such thing as an .mp1 extension?
20:40:09preglowyou dont see it often, though
20:40:16MO-PantsuI was just pointing stuff out and preglow thought he was steering me right. which he was but only for Lame :)
20:41:04pregloweverybody with at least half a brain uses lame
20:41:19MO-Pantsuyes but I have a lot of stuff from before Lame
20:42:05*preglow huggles vorbis
20:42:15MO-Pantsuand that needs some magic playback trickery to get gapless. The Archos code. What does that do to get gapless? Does it only support Lame gapless?
20:42:24preglowit doesn't do gapless
20:42:30preglowthe only gapless archos can do is −−nogap encodes
20:42:38preglowit doesn't support lame gapless
20:42:42MO-Pantsuoh I was mislead by others then
20:42:58preglowbut yes
20:43:06preglowthe _only_ proper gapless is lame info header gapless
20:43:22pregloweverything else is a hack
20:43:23MO-PantsuI am going to type this all up on misticriver so people don't get confused. a lot of fud going around
20:43:47MO-PantsuEVERYONE there thinks gapless is gapless for everything MP3
20:44:25HCli must be one of the few people completely uninterested in gapless
20:44:28HCli *want* gaps :P
20:44:55thegeekit's only needed when you have albums that are meant to be played continously
20:45:05thegeeksay, a recording of a live concert
20:45:05HClwhich i barely have
20:45:09HCland even then
20:45:11HCli couldn't care less
20:45:13HClabout gaps
20:45:13thegeekI only have a few
20:45:17thegeeksame here
20:45:18HCli have maybe 2
20:45:22thegeekI don't really care all that much
20:45:32thegeekcertainly get what the _huge_ fuss is about;)
20:45:37HCl :p
20:45:41preglowhalf of my music is supposed to be played gapless
20:45:43linuxstbI've sort-of got a codecflac.c working, but the WPS screen isn't coming up when the track is playing...
20:45:54preglowlinuxstb: wps doesn't work for anything but mp3
20:46:09linuxstbOK. Should I commit it to CVS then.
20:46:15preglowif it works, then hell yes
20:47:36linuxstbI think it's working (I'm listening to a track now). I'll commit it.
20:48:23zezayer_awaylinuxstb: am i hearing u rite :D
20:48:42 Quit webguest40 ("CGI:IRC")
20:49:03MO-PantsuA big portion of my albums are DJ mixes
20:49:16preglowmost of my mixes are just one mp3
20:49:39MO-PantsuI like them that way but only for gapless. I would prefer CUE support for those in Rockbox ;)
20:49:46MO-Pantsuso you can skip tracks
20:50:15preglowi've got a couple of mixes on cd, though, but i rip to vorbis, so...
20:50:31linuxstbzezayer_away: it's even more buggy than MP3/OGG, but it's sort-of working. Congratulations go to whoever optimised it.
20:51:01preglownot much optimising in libflac apart from the coldfire.S i did
20:51:08preglowit doesn't even use iram
20:51:20preglowit's coded so embedded unfriendly it's almost a pain
20:54:54preglowwhy does codecflac.c credit zagor? :PP
20:55:07linuxstbOK, it's in CVS now. But I think it's stopping decoding when the memory buffer is initially depleted - is that my bug or a limitation of the playback system?
20:56:09Slasherilinuxstb: Hmm, the playback system should keep buffer always filled
20:56:22SlasheriJust remember not to request too large chunk size
20:56:32linuxstbSlasheri: Can you have a quick look at my code?
20:56:37Slasherilinuxstb: I will
20:57:21linuxstbThe problem will be in the read callback - libFLAC asks for "bytes" bytes of data, and my read callback tries to give it, and if not, returns how many bytes were actually available.
20:58:27SlasheriHmm, that might not be good idea to use seekable flac because if the whole song can't fit into memory, then seeking backwards is not possible when buffer is being refilled
20:59:14SlasheriI think that might be the problem
20:59:24SlasheriYou should try streaming flac if possible
20:59:51SlasheriAh, hmm
21:00:07linuxstbI don't think the seek callbacks are ever actually called, so hopefully it's not a problem.
21:00:17Slasherii hope too :)
21:00:37Hansmaulwurfwhoops, "repeat one song" crashed my H120
21:00:42linuxstbBut we do need to support seeking in the future.
21:00:56SlasheriHmm, i have to test the codec also. At least the blocksize 4608 is not too large
21:01:46SlasheriHansmaulwurf: If you find a bug, please tell what version you was using and how the problem can be reproduced
21:03:16 Join webguest40 [0] (
21:03:32webguest40hi everyone.. just curious, can the iriver support crossfading?
21:03:52webguest40I know it says it can't for archos, but that has its own MAS chip for decoding
21:04:13webguest40is it possible to load two codecs in iriver and crossfade between the two?
21:04:55preglowif we try hard enoug
21:05:27webguest40lol... excellent! keep on trying! :)
21:05:30 Quit muesli- (Read error: 54 (Connection reset by peer))
21:05:33 Join wake [0] (
21:05:49Hansmaulwurfah ein deutscher
21:06:06preglowplease don't stop speaking english
21:06:56 Quit wake (Client Quit)
21:07:12webguest40Ich bin im schuele ein bissen Deutsch lernenen :)
21:07:13HClthe xbox pal version of san andreas is out :)
21:07:15webguest40I think I butchered that...
21:07:56niobosHCl: I read the perl code, and still have a problem:
21:08:16niobosWhen I have 2 files, 1 with artist "" and 1 with artist "something"
21:08:37niobosthey'll both map to the first artist-entry = "something"
21:08:45HClhow do you manage that?
21:08:48niobosso you'll confuse "" with "something"
21:08:54HClartist "" needs to be "<no artist tag>"
21:09:05 Quit webguest40 ("CGI:IRC")
21:09:08niobosin the perl-implementation that is currently in CVS, it's ""
21:09:14 Join wake [0] (
21:09:18*HCl checks
21:09:41HCl# fallback names
21:09:41HCl $$id3{'ARTIST'} = "<no artist tag>" if ($$id3{'ARTIST'} eq "");
21:10:21 Quit wake (Client Quit)
21:10:23niobosI know, read that too
21:10:28niobosbut he doesn't do it...
21:10:32linuxstbWhy does the WPS screen only work for MP3? Do I need to fill some structure with info about the FLAC file or is it not that simple?
21:10:46preglowlinuxstb: it hasn't been adapted yet
21:11:30markunlinuxstb: after your latest commit I get "Extension array full" and rockbox doesn't respond anymore.
21:12:17MO-Pantsuanyone know anything about EAC? I don't have a drop down option to compress to Ogg Vorbis
21:12:24preglowyou didn't increase the extension array size?
21:12:36linuxstbIs that #MAX_EXTTYPES ?
21:12:49preglowbut yo
21:12:54preglowwrap an ifdef around it
21:12:58preglowdont increasie it for achos
21:13:12linuxstbDo you have a version you can commit?
21:15:10linuxstbOK, so I'll add #ifdef CONFIG_HWCODEC == MASNONE and increase it to 64?
21:15:46preglowsounds good
21:17:04amiconnThe mas has nothing to do with the # of extensions
21:17:36preglowplayers with hardware codecs wont need so large an extension array
21:17:39linuxstbamiconn: That's how it's done in tree.c
21:17:41preglowsince they can't load very many files
21:17:55linuxstbSoftware codecs -> lots more extensions
21:18:53linuxstbI'm happy to do something else though...
21:18:58[solid]oh boy oh boy oh boy flac is sooooo working xD
21:19:11[solid]i'm senfing all my love to you guys :D
21:19:25*linuxstb ducks
21:20:08SlasheriHmm, it's even not necessary to load to codecs same time to get crossfading. Only both codecs should run fast enough and even with the current implementatio, i think that crossfading is possible
21:20:14SlasheriMaybe i should try that some day :)
21:20:33***Saving seen data "./dancer.seen"
21:20:35preglowthe plan isn't to run them at the same time, no
21:21:07linuxstbamiconn: any better ideas for deciding when to increas MAX_EXTTYPES ?
21:21:34Slasheriin fact i think that i will try crossfading right now. But some software mixer code should be written before that
21:21:45amiconnLoading one after the other will limit the cross-fade time, but other than that, would be the cleaner solution
21:22:09preglowloading them at the same time will require codec slots
21:22:34amiconnApart from that, we should probably be able to run 2 codecs at once (thinking voice ui here)
21:22:54Slasheriamiconn: Still the cross-fade time will be around 6s with 1 MiB audio buffer and 44.1 kHz samplerate
21:23:28amiconnAnyway, I don't care about crossfade at all
21:24:01Slasheriamiconn: Hmm.. Unless voice ui is done in wav format
21:24:11amiconnForget .wav here
21:24:15Slasheriok =)
21:24:37amiconnThe voice file is ~1.5 MB on archos, and that's with 12 kHz sample rate and tight vbr compressin
21:25:27Slasheriah yes, the whole while should be loaded into memory?
21:25:51Slasheriok, then compression will be necessary
21:26:23amiconnPreferably we would set the buffer mem aside on iriver, in order to avoid reloads
21:26:56amiconnOf course this would then be done only if voice ui is enabled
21:27:44t0masoh... about mem...
21:27:57t0masamiconn? where should I allocate a buffer for graphics on the wps?
21:28:09t0mas(2560 bytes, so I guess not on stack?)
21:28:20t0masoutside function? as a static?
21:28:50amiconnThink so (but I still don't like that idea much...)
21:28:52Slasherit0mas: btw, have you thinked about implementing wps to remote control? That would be really nice :)
21:29:05niobosHCl: I tried it again: the perl version won't write <no artist tag> into the file... I'm no Perl specialist, so I haven't found the problem yet
21:29:16HClthe perl version is crappy anyways.
21:29:25 Join belgarath_the_so [0] (
21:29:29niobosuse your java as definition?
21:29:38HClyea, heh.
21:29:48HCli can't even read the perl version properly
21:30:50*niobos is releived that it isn't only him...
21:31:15nioboshow do I have to run your java-rhing again?
21:31:35belgarath_the_sohas anybody had difficulty using the id3 database with a newer version of the rockbox firmware?
21:31:36HCljava -jar SongDB.jar
21:31:52t0masSlasheri: erm
21:31:56HClbelgarath_the_so: format was updated. why?
21:32:00t0masno... I haven't
21:32:09t0masbut it shouldn't be verry difficult... I guess?
21:32:30niobosException in thread "main" java.lang.UnsupportedClassVersionError: SongDB (Unsupported major.minor version 49.0)
21:32:37belgarath_the_soit gives me an error message about it when I start up
21:32:38Slasherit0mas: ok, it would be nice to have at least remote buttons (volume and skip track) working :)
21:32:43HClthen its not recompiled for 1.4 properly, as i suspected
21:32:56t0masSlasheri: the buttons are your problem :)
21:32:58nioboshow do I do that?
21:32:59belgarath_the_sosomething like "version 2 db not supported"
21:33:00*amiconn wonders why the cpu is boosted right after boot
21:33:05HClunpack the .jar
21:33:18HClthen javac -cp . `find -name '*.java'`
21:33:22t0masSlasheri: or can I edit the patch?
21:33:26t0masin cvs to add the remote buttons?
21:33:30HClyou'll most likely get some assert errors
21:33:53nioboswhat's the -cp? he doesn't recognize it
21:33:56Slasherit0mas: of course you can, just edit and commit :)
21:34:11HClwhat odd java compiler do you have that it doesn't recognise that o_o
21:34:45 Quit ehntoo ("Leaving")
21:34:52niobos-classpath <path> Specify where to find user class files
21:34:59HCluse that
21:35:12amiconnIs there a reason for immediately boosting the cpu?
21:35:23niobosHe complains about assert() things.. I'll remove them
21:35:34Slasheriamiconn: where?
21:35:49amiconnDirectly after boot the boost count is 1
21:35:59 Join TCK [0] (
21:36:36SlasheriThat's not good
21:37:02belgarath_the_soi extracted the song db.jar and have downloaded a java compiler, what do I need to do now?
21:37:14[solid]hm... i might have found a bug in flac playback, it seems it won't play the next track at all... doing some more testing now
21:38:45 Join Zoom2 [0] (
21:38:55linuxstb[solid]: Yes, I'm aware of that - I'm looking at it now. You just need to move the joystick right to skip to the next track.
21:39:24[solid]oh, ok
21:39:36HClbelgarath_the_so: you neither need to extract it, nor get a java compiler.
21:39:43HCljust do java -jar SongDB.jar
21:40:02niobosHCl: got it working, but still have (yet another) question:
21:40:03HClat the moment you need java 1.5
21:40:03belgarath_the_sowhat program would I do that with?
21:40:09HClniobos: can you put the recompiled version on wiki/
21:40:17HCljust the java runtime environment
21:40:29niobosshouldn't <no artist tag> be sorted before all the rest?
21:40:37 Quit zezayer_away ("Chatzilla [SUSE 1.0.4-1.1/20050511]")
21:40:38HClyea, its one of the issues with the java version so far.
21:40:41 Join ehntoo [0] (
21:40:51HCli'll fix it soonish
21:40:54HClits an very easy fix.
21:40:55niobosnp, it's just that I implement it the good way
21:41:15nioboswell, do the sorting with <notag> first
21:41:28niobosisn't hard, but it's easyer when you know what to do
21:41:30HClwhat do you mean?
21:41:44HCli'm just gonna adjust the compareTo()'s to make < go before any other char
21:42:07nioboswell, I could fill in "<no tag>" first, and sort; or I could fill in "" first and sort and fill "<notag>" later
21:42:08t0masSlasheri: you have some main loop where you do all messages and keys right?
21:42:34 Quit Hansmaulwurf ("( :: NoNameScript 3.81 :: )")
21:42:37HClwell, you can do it whatever the way you want it, i prefer to just change the sorting criteria in my java version
21:42:47niobossince if someone is foolish enough to tag his file with the artist "<no artist tag>" that should be sorted in between...
21:42:59niobosk, thx
21:43:00 Join zezayer_away [0] (~chatzilla@
21:43:59 Quit TCK- (Read error: 145 (Connection timed out))
21:45:41*amiconn needs to get familiar with coldfire asm
21:46:03preglowamiconn: shoot if you've got any questions, i've picked up one or two things
21:46:05amiconnMy first try on memset() already runs faster than the C version
21:46:21[solid]nice one that
21:46:44amiconn...even though I was lazy and do write unaligned longs
21:47:02amiconn(wanted to see whether this really does perform like the datasheet says - it does)
21:47:31niobosHCl: am I right about the dir-is-album-... :
21:47:45HClprobably not, cause you say
21:47:49HCland they're *not* the same
21:47:58amiconnpreglow: It's just that I need to learn which instructions are there, what their restrictions and execution times are etc
21:48:02nioboscan I complete my senteces?
21:48:07HCl :x
21:48:38niobosdia: when set, all files in a dir are considered to be the same album; take the first tag as album name
21:48:45mirakisn't it frustrating to do constrained programming ? :D
21:48:51mirakwe tend to be used to have a lot of power
21:48:53HClpretty much, yes.
21:48:56niobosdian: when no album tag is found, use the directory as it
21:48:58[solid]btw, is anyone working on wps currently? (ie support for formats other than mp3, esp since flac and ogg vorbis you basically the same system)
21:49:04HCluse the directory name as album tag
21:49:25*niobos is pround of himself to finaly, after 2 hours, get the concept
21:49:50preglowamiconn: it's pretty easy, the timing almost consistenly depends on the addressing mode you use
21:49:54amiconnmirak: Wasting resources is never desirable, even with the fastest processor
21:50:18HClmicrosoft could learn from taht.
21:50:30 Quit zezayer_away (Remote closed the connection)
21:50:52miraksun also
21:51:03mirakthough java is not that bad now
21:51:21mirakbut that's heavy
21:51:27t0mas<hilight> Slasheri :)
21:51:27mirakand modularity is wanted first
21:52:04Slasherit0mas :D
21:52:10t0massaw my question?
21:52:14amiconnpreglow: There is a table of execution times in one of the datasheets, but with coldfire there are a lot more factors to take into account that aren't there for SH
21:52:30 Quit niobos ("off")
21:52:53t0masSlasheri: where do you handle the track change? you handle the buttons yourself? or does rockbox do that somewhere else?
21:52:59Slasherit0mas: Hmm, i don't control the keys at all (keys are controlled by wps itself or something like that)
21:53:06t0mashmz... a lot more addressing modes, variable instruction length, instruction cache, more situations influencing the pipeline etc
21:53:10t0masnever saw that code...
21:53:12Slasherirockbox does that
21:53:21t0massearching time then :)
21:54:20preglowamiconn: the addressing mode used is also consistent with regards to instruction size
21:54:57amiconnYes... but it's a thing that simply doesn't exist in SH asm
21:55:11amiconnAll sh instructions are 16 bits
21:55:12t0masSlasheri: what function is called in your playback.c to change track?
21:55:21t0mas(so I can search from where that's done)
21:55:39 Quit belgarath_the_so ("CGI:IRC")
21:55:44 Join belgarath_the_so [0] (
21:55:50Slasherit0mas: it depends, normally codec_request_next_track_callback in playback.c
21:56:07preglowamiconn: use postincrement or predecrement addressing as often as possible
21:56:13preglowall other addressing modes add sixteen bits
21:56:14amiconnpreglow: I already had 1 encounter with cache and pieline effect:
21:56:16t0masSlasheri: ok, tnx
21:56:24Slasherit0mas: you really want to enable logf while listening to songs
21:56:37belgarath_the_sosorry about this but I am in command prompt and as HCl suggested I typed java -jar SongDB.jar - what do I need to do now? memset() execution times are _exactly_ identical for filling 0 and 1 byte
21:56:59linuxstb Well, FLAC playback is nice and gapless :-).
21:57:26[solid]linuxstb: can't wait to test the hell out of it :o)
21:57:55linuxstbSlasheri: How should I detect end of stream when calling ci->read_filebuf ?
21:58:09HClit gives a clear usage.
21:58:12preglowamiconn: coldfire doesn't have much of a pipeline
21:58:34Slasherilinuxstb: it will return 0 bytes as size
21:58:35t0masSlasheri: found it
21:58:38t0masit's done in wps code
21:58:44Slasherit0mas: ah :)
21:58:59amiconnpreglow: I think there must be a pipeline. Even SH1 has a 5-staged pipeline
21:59:15preglowamiconn: like i said, not much of a pipeline
21:59:15linuxstbSlasheri: OK, that's what I am testing at the moment. I'll commit it.
21:59:23Slasherigreat :)
21:59:23belgarath_the_soi have a list of options such as [−−<strip directory>]
21:59:30preglowamiconn: i think they would have mentioned a pipeline in their docs
21:59:38preglowamiconn: and i can't remember reading about any pipeline
21:59:40t0mashm... general question... when I switch on a bitmask (like 1000) does that match case 1011 too?
21:59:50HClyes. they're all in standard commandline options format...
22:00:01preglowamiconn: probably has a pipeline, yes, but i don't think you need to think much about it
22:00:01HClin case you're unfamiliar with it. [] is optional, < > is mandatory
22:00:03t0masoh stupid... ofcourse it doesn't
22:00:24belgarath_the_soi'm really sorry but could you tell me what to type because it keeps saying file not recognised
22:00:36HClfile not recognised?
22:00:40belgarath_the_soor cannot find specified file
22:00:50HCljava -jar SongDB.jar <whatever your path to your music files is>
22:01:23amiconnpreglow: page 34: "The ColdFire processor Version 2 core consists of two independent, decoupled pipeline structures..."
22:01:37belgarath_the_sook thx a lot its working now...
22:03:28preglowamiconn: don't think it's very deep, though
22:03:40Slasherilinuxstb: btw, you should place next_track: before codec_init so all dynamically allocated memory will be freed
22:04:07SlasheriThat was a problem with vorbis and caused eventually memory overflow to the file buffer
22:05:30belgarath_the_sowhat do I do once all of the albums have been shown on the screen (do you have to save it or something?)
22:06:17belgarath_the_sobecause I restarted rockbox and the id3 database didn't show up
22:06:26HClthere's a howto on the wiki..
22:06:46HClits okay
22:07:12linuxstbSlasheri: I've already changed that. It will be in my next commit.
22:07:50Slasheriok :)
22:09:53 Join muesli- [0] (
22:10:15linuxstbI've just committed some fixes for codecflac - it should be working much better now.
22:10:30linuxstbAny comments/fixes welcome :-)
22:11:23SlasheriI think i should start encoding more albums to flac :)
22:11:44ehntooSlasheri, I think I need to do the same
22:12:25linuxstbFLAC probably uses more battery though.
22:12:38[solid]i already have been using my iriver as a storage place for some of my flac's so i'll be testing as soon as it builds..
22:12:45preglowthere's room for much, much optimisation in flac
22:13:05[solid]it's great that they are no longer useless sitting there
22:13:27[solid]preglow: but it will always use a lot more battery power because of the size, no?
22:14:00linuxstbEverything I've tried seems to play in realtime though. That's the biggest hurdle overcome.
22:14:05ehntooI don't worry about battery life... I'm never away from the house for more than, say 10 hours
22:14:55[solid]linuxstb: i have only max compression (-8) flac's and it's all great:D
22:15:05belgarath_the_soHCl: I'm really sorry and really don't want to waste your time but I looked on and couldn't find anything to solve my problem
22:15:19linuxstb[solid]: I don't think the compression rate affects the decoding time, only the encoding time.
22:15:21belgarath_the_sothat too
22:15:27belgarath_the_soi have looked
22:15:42HClyea, i guess you're right, it needs updating
22:15:51HClanyways, move the rockbox.id3db to .rockbox folder
22:15:54HClafk now
22:15:56 Quit bipak (Read error: 145 (Connection timed out))
22:16:09belgarath_the_sook thx
22:16:40[solid]linuxstb: but at lower compression levels it uses less complicated maths, doesn't this affect it?
22:17:23t0masaccording to cvs annotate...
22:17:23t0masyou added the #define WPS_RC_NEXT lines to wps.c
22:17:23 Quit belgarath_the_so ("CGI:IRC (EOF)")
22:17:24t0masRC = remote controller here?
22:17:34t0masbut they're never used in the file... just defined...
22:18:24t0masoh wait
22:18:29linuxstb[solid]: I think you're right - it's probably the "-l" option that increases the decoding time. -8 implies "-l 12". But anyway, Rockbox seems to cope :-).
22:18:30t0masnever mind... found them :)
22:19:04preglowcompression rate affects decoding time
22:19:25preglowit uses higher order lpc, and i've just coded a generic decoder loop for that
22:19:29preglowthough it should be pretty fast
22:20:22[solid]linuxstb: yeah there was this issue with some maudio players that couldn't cope with levels over -3 or something like that. i'm listening to flac on my h120 as we're speaking...:)
22:20:30[solid]it's just so cool :D
22:20:58[solid]wee and the build is done
22:22:09 Join mborus [0] (
22:23:27 Quit Shagnar (Remote closed the connection)
22:24:58 Quit mborus (Client Quit)
22:25:05[solid]hmmm... it told me 'extension array full' on startup now
22:25:08[solid]but it's working
22:25:40linuxstb[solid]: You need to update the apps/filetypes.c file from CVS.
22:25:56[solid]i just fetched a daily build o_O
22:26:15[solid]i thought it's supposed to have that patch already
22:26:55HClyay gta san andreas works
22:27:17Zoom2on iriver?
22:27:28HClsure, on iriver *rolls his eyes*
22:27:49Zoom2a man can hope cant he not?
22:28:20linuxstb[solid]: Yes, it should be. Not sure why you are getting that message.
22:28:55[solid]linuxstb: waiting for the gap now... or rather, waiting for the gap not too happen:)
22:29:24[solid]maybe i should delete the .rockbox directory and do a fresh copy?
22:29:42HCldaily builds are always lagging behind cv
22:29:46HClbecause they need to be compiled
22:29:53[solid]HCl: yeah, about 20 minutes
22:30:11[solid](from my experience)
22:30:51t0mashm? they're not really daily?
22:32:13amiconnpreglow: Spooky... using movem to make use of burst mode (as we thought) doesn't help a tiny bit to increase speed... :(
22:32:38[solid]oh yeah, somethings wrong... it stopped after the track again... i'll delete the .rockbox dir and try again
22:32:38 Join firmwarewontload [0] (
22:32:49t0masnice nick
22:33:06amiconn[solid]: You're talking about bleeding edge. The dailies are compiled at 06:00 ce(s)t
22:33:11t0masfirmwarewontload: what's your problem? Doesn't the rockbox firmware load? or the iriver?
22:33:28t0mas<t0mas> hm? they're not really daily? <−− I was right :P
22:33:28firmwarewontloadhey. great workwith rockbox
22:33:33t0masthey are daily
22:33:39t0masand not 20 minutes behind...
22:33:39firmwarewontloadtrue the iriver firmware wont load
22:33:47t0masfirmwarewontload: charge batterys
22:33:51t0masand try again in 10 minutes
22:33:56amiconnt0mas: The dailies are daily, bleeding edge is... bleeding edge
22:34:03firmwarewontloadok. thanks a lot
22:34:19[solid]oh... i've been using bleeding edge since... since... since i flashed my player
22:34:41preglowamiconn: not a thing?
22:34:48preglowamiconn: what did the old code look like?
22:34:51 Quit firmwarewontload (Client Quit)
22:35:19amiconnpreglow: I compared a simple loop storing longs with one that stores 4 longs at once using movem
22:35:40t0maswho did the iriver remote handling code?
22:35:41 Join mborus [0] (
22:35:52amiconnThe version using movem shows exactly the same speed as the simple long loop for very large chunks
22:36:04preglowamiconn: well, that's pretty strange
22:36:09amiconnIt's even a bit slower for smaller chunks due to more precalculation
22:36:11t0masI guess he did something wrong... the volume buttons aren't working...
22:36:18HClmaybe some stuff wasn't initialized properly?
22:37:14preglowamiconn: but anywho, how much faster is it as it is?
22:37:15t0maswell... the backlight doesn't activate when using the volume or trackchange buttons
22:38:06amiconnpreglow: If the destination is longword aligned to begin with, my asm version is equally fast for large chunks
22:38:26preglowamiconn: i mean compare to the c version
22:38:40preglowagain, strange
22:38:48amiconnFor non-aligned, I can get it up to 4x (same speed as aligned, doing start and trail bytes separately)
22:39:16amiconnFor small chunks it's a bit faster as well due to optimised calculations
22:39:51Slasherilinuxstb: the flac playback seems to work quite well and sound quality is great :)
22:40:19linuxstbWhat's needed to get the WPS working?
22:40:33Slasheritag info has to be read at least
22:40:53Slasherione temporary solution would be writing tags to proper id3 fields
22:41:01BagderSlasheri: you do read the mailing list, right?
22:41:13SlasheriBagder: i just subscribed :D
22:41:23HCltag info can be gotten from the tag database
22:41:31HClif the tag database is present
22:41:40BagderSlasheri: so I guess you read my mail?
22:41:45t0masSlasheri: remote volume, pause/play and stop are working
22:41:46BagderHCl: it should not
22:41:54HClwhy not?
22:41:59Bagderplaying songs should get tags from the song
22:42:00HCli said it could, not that it should
22:42:02preglowBagder: he's already commited a fix :P
22:42:04[solid]linuxstb: oh indeed, deleting the .rockbox dir helped... still waiting for the gap tho, damn long song :)
22:42:05t0masSlasheri: next and prev. song don't work... the buttons aren't handled as I want them...
22:42:06SlasheriHCl: Hmm, that database is something i really should study better
22:42:21HCleither way, we're going to have to look up the song in the database anyways
22:42:21SlasheriBagder: yes, thanks :)
22:42:25HClfor the runtime information
22:42:31mborusHi. Just installed RB on my H140 today - do you want feedback in IRC or on the mailing list?
22:42:33 Join MoosCamaro [0] (
22:42:37HClso you might as well fetch the info from the database
22:42:40SlasheriBagder: i removed the rindex right away
22:42:41MoosCamaronight guys
22:42:42Bagderthe db is secondary, playing is primary
22:43:06Slasherit0mas: oh, really cool :)
22:43:26 Quit TCK (Read error: 145 (Connection timed out))
22:43:27 Quit muesli- (Read error: 110 (Connection timed out))
22:43:37HCli still don't see why we wouldn't use info thats available anywho
22:43:46HClbut whatever
22:43:47Bagderwe should
22:43:52Bagderbut thats _secondary_
22:44:04Zoom2anyone know if rockbox is janes compatible?
22:44:11 Quit preglow (Read error: 104 (Connection reset by peer))
22:44:11HCli dunno how much cpu it costs to parse tags in a file
22:44:13linuxstbDo I just need to fill in the ci->mp3entry structure with the tags from the FLAC file, and then set ci->taginfo_ready to true?
22:44:27HClbut for vorbis, i think it would be more than to look it up in the database?
22:44:29ZagorZoom2: you mean janus?
22:44:35Zoom2yeah =)
22:44:42Slasherilinuxstb: that should work
22:44:49 Join preglow [0] (
22:45:02Zagorno it's not. janus is Microsoft stuff, which means it can't be done on open source.
22:45:05SlasheriBut it really is not the best solution..
22:45:18Zoom2Zagor: k, thanks
22:45:29[solid]and so the next song came, and the gap did not happen... great work linuxstb
22:46:19MO-PantsuIs anyone else having the problem with tags not displaying when playing back Ogg files?
22:46:27preglowit's not a problem
22:46:33HCllike preglow said.
22:46:37HClits in development
22:46:37linuxstbSlasheri: What is the best solution? (I'm assuming you're talking to me...)
22:46:40HClits not meant for end users.
22:46:40MO-PantsuOK it's not coded I guess
22:46:56MO-Pantsuwas thinking I was going nuts ;)
22:47:12Slasherilinuxstb: currently there is no better solutions.. it has to be implemented =)
22:47:31*MO-Pantsu smacks EAC for giving an error when using OggEng
22:47:37Slasherii think it's best at the moment to use the id3 fields to get tags to wps..
22:48:06MO-PantsuIs there a better ripper than EAC for Ogg?
22:48:43preglowhas nothing to do with the ripper
22:48:48MO-PantsuI get a cryptic error message about not adding UTF-8
22:48:51preglowhas everything to do with the encoder the ripper uses
22:49:16MO-Pantsugonna try another encoder methinks
22:50:11 Join theebag [0] (
22:52:49 Quit ghostiger (Remote closed the connection)
22:53:40 Join Xavier|away [0] (
22:53:56Xavier|awayHello everybody!
22:55:38 Nick Xavier|away is now known as XavierGr (
22:55:59 Join leftright [0] (
22:56:38t0maswoohoo :D
22:56:51amiconnHmm, no Linus :(
22:56:54*t0mas is enjoying his iriver... with a graphical wps
22:57:11t0masnicely formatted.. all info I need
22:57:15amiconnI spot some weirdness in crt0.S
22:57:16t0masand control from the remote works too :D
22:57:27[solid]t0mas: sounds nice
22:57:35XavierGrwow t0mas where!
22:57:45XavierGrdid you patch it?
22:57:55[solid]t0mas: control from the remote is the only thing that keeps me using the original firmware now :D
22:58:01XavierGrme too!
22:58:13t0masok, I'll reformat it... and test for archos tomorrow
22:58:16t0masand then commit ok?
22:58:17XavierGrdo you have wps in the remote too?
22:58:24t0masthat's still left todo
22:58:35[solid]who cares, the buttons are working \o/
22:58:35XavierGrgotta do it man its perfect!
22:58:35t0masbut I'll try to get that working :)
22:58:50t0masdamn... I'm _really_ tempted to do it right now
22:59:02t0masbut I have to go to sleep... have to get up at 6:30 tomorrow
22:59:02Zoom2got screenshots?>
22:59:06Bagderthat takes some thinking
22:59:10t0masZoom2: from simulator?
22:59:17XavierGrah then go to sleep But aupload some pics first
22:59:27Bagderlike what to do on the remote/main screen
22:59:29Zoom2doesnt matter where
22:59:33Bagderfor WPS/file browser
22:59:36Zoom2I just wanna see what it looks like
22:59:56t0maslike that now :D
23:00:28t0masBagder: shall I just try to make it use wps in the remote? from a separate wps config file?
23:00:35t0masor are there better suggestions?
23:00:50Bagderthat's what I mean with the required thinking
23:01:06Bagderit could for example remain in WPS in the remote even if you browse files on the main screen
23:01:30t0masyes, but writing wps code for the remote is needed anyway?
23:01:34Bagderyou want another font for the WPS on the remote
23:01:40t0masso I can do that when I have some time tomorrow?
23:01:41Bagdert0mas: true
23:02:28Zoom2good work t0mas
23:02:46[solid]oh man, can;t wait to test that
23:03:01XavierGrwhat files did you reprogrammed t0mas to work that out?
23:03:44XavierGronly that?
23:03:59t0masI already knew a lot of that code... because I did the graphics thing before
23:04:09t0masjust added the remote buttons there
23:04:19t0masand re-added my graphics stuff
23:04:32XavierGrhmm maybe a wps with graphics thing (customizable) would be perfect!
23:05:19XavierGrwhat about if you open a wps file? Does your patch crash or it can choose
23:09:50 Quit Aison (Read error: 54 (Connection reset by peer))
23:10:25t0masit doesn't crash...
23:10:47t0masbut if your wps doesn't fit in my image it doesn't look really good
23:10:58t0masthat's why I'm going to change that... to read the image from a bmp file
23:11:06t0masas that's not really difficult
23:11:48Zoom2that is really good stuff t0mas, if we could get custome graphics on wps that would be awesome!
23:12:52 Quit thegeek (Read error: 54 (Connection reset by peer))
23:12:56t0mas:) maybe it will just be a separate patch...
23:13:03t0masas it uses some memory... wich amiconn doesn't like
23:13:56Bagderhe likes it if you do it effectively
23:14:56t0masok, so don't use more buffer than we have lcd space...
23:14:56 Quit mborus ("CGI:IRC")
23:14:58t0masand what else?
23:15:27t0masoh btw... a reallife "screenshot" ->
23:16:37[solid]t0mas: can't wait :D
23:16:40Zoom2looks good, im trying to find some post for you on the forums
23:16:47Zoom2there was an artist who did a really good rendering
23:16:54Zoom2I was wondering if it were possible
23:17:24[solid]that's a plug from koss'? ;)
23:17:38Bagderthe WPS should use whatever pic you want, so t0mas won't need that artist
23:18:54Zoom2would that tabbed page interface be possible?
23:19:21Bagdereverything is possible, if you write the code for it
23:20:35***Saving seen data "./dancer.seen"
23:21:02t0masZoom2: I won't write that
23:21:20t0masand I'm going to bed now
23:21:23t0masgood night :)
23:21:28Bagdernight t0mas
23:21:34XavierGrgood night
23:22:30t0mas(and the reason not doing it is that it's way to complicated to integrate into rockbox... don't think many people can make something like that)
23:22:55Bagderand those who can don't see the point of doing it ;-)
23:23:06t0masyes, I guess Bagder can do it :P
23:23:29t0masgnight :) this time for real... ghehe...
23:23:47Zoom2nite t0mas, thanks for the input
23:24:09 Join thegeek [0] (
23:25:48 Quit theebag (Read error: 104 (Connection reset by peer))
23:27:03Zoom2would this interface be easier to code?:
23:27:47Bagderits not about how easy or hard it is to write
23:27:49Zagorit's not about how easy it is to code. we already have an interfac.
23:28:26Zoom2well I dont know what I am talking about then =(
23:28:42*amiconn spotted some weirdness in system.c too :(
23:31:12Zagorbtw, i'm ready to press the release button. but I'm pretty out of touch about the current status. do we have any outstanding issues for the archos that warrants a delayed release snapshot?
23:31:48BagderZagor: I think amiconn is most on top of these things atm
23:32:01Zagoras in "has fixed" or "is fixing"?
23:32:13Zagoror "knows best" :)
23:32:35 Join Aison [0] (
23:32:47XavierGrrelease what release?
23:33:01BagderI'm not personally aware of any remaining major flaws
23:33:01amiconnI think I have fixed the major issues, but e.g. for the multivolume probs on 0308 Ondios, I did not yet receive confirmation from Trevor
23:33:16amiconnHe told me that he can't find his other 0308 Ondio :(
23:34:06amiconnAnd of course, the recording fixes aren't tested much on other devices than my own recorder
23:34:23amiconn(but they probably never will be if we don't release)
23:34:44Zagorwell, i just made a snapshot and tagged cvs. it'll take a little time to write the release notes (bagder, will you do the usual honors?)
23:34:56amiconnPeople seem to stick to releases, even if the dailies are way better
23:34:57Bagdercertainly, hang on
23:36:07 Join LinusN [0] (
23:36:13Bagderaha, dr L!
23:36:24amiconnUuuhh, LinusN himself! :)
23:36:46*LinusN is writing a linux driver for his logic analyzer
23:37:05amiconnLinusN: There are some things in crt0.s and system.c that look a bit suspicious to me
23:37:14LinusNamiconn: shoot
23:38:11amiconn(1) Am I right that the SDRAM setup is done in the bootloader only? If so, did you write a modified bootloader to test page mode?
23:38:41LinusN1) the setup is done in the boot loader, but page mode can be changed on the fly
23:38:53amiconnAH, ok
23:39:28amiconn(2) DCR setup (lines 193,194). It sets RC to 5, but claims to set 80 cycle refresh
23:39:54amiconnHowever, refresh cycle count is (RC+1)*16 so it sets 96 cycles...
23:40:41amiconnMore oops'es to come... system.c
23:41:16amiconnIn set_cpu_frequency:
23:41:50amiconn(1) pll bypass means 11 MHz CPU frequency again, and 5.5 MHz sysclk, as in the bootloader
23:42:29amiconnWhy do you set RC=1 then. RC=4 (80 clks as intended in the bootloader) should sufficew
23:43:07LinusNi remember i had a reason for that...
23:43:40preglowand didn't comment it? ;)
23:43:43LinusNbut i can't remember now
23:44:25LinusNi guess it could be 4
23:44:29amiconn(2) CPU_FREQ_MAX:
23:44:49CoCoLUSif i want to build rb myself, do i have to check out anything else except the base package and the fonts?
23:45:26ZagorCoCoLUS: tools
23:45:26amiconnYou set the refresh timer to 28. If my calculation is correct, it should be 57
23:46:14amiconnDitto 22 instead of 10 for CPU_FREQ_NORMAL
23:46:26preglowBagder: shouldn't make be run in tools/ if they aren't built? that used to be the case before, and now we get people asking why it fails all the time
23:46:55LinusNamiconn: you mean i have mixed up pll freq and sysfreq?
23:47:05amiconnLinusN: (3) default: The comment for i2c frequency is obviously wrong, as it says 48 MHz.
23:47:24Bagderpreglow: it should, but I didn't like the way it was done before. I should re-add it in a better way.
23:47:43amiconnThe clock divider might be wrong too. (Didn't check yet, but it has the same values as for 48 MHz)
23:48:10LinusNamiconn: i must have forgot to set the correct values for the default case
23:48:38amiconnIiuc the formula for calculating refresh is
23:48:41Bagderall commits since 2.4
23:49:01LinusNamiconn: the refresh rate is based on the bus frequency
23:49:16amiconnsysclk = cpuclk / 2; RC = (sysclk * 15.625 us) / 16 - 1
23:49:29LinusNlooks correct
23:49:58amiconn...and if I calculate that, I get 57 / 22 / 4 for ~120 / ~48 / ~11 MHz
23:50:32ZagorBagder: yikes! yeah, this'll take a while...
23:50:44LinusNamiconn: looks ok
23:50:44Bagderyes, its... BIG
23:50:46amiconn...and it's obvious that the value for 120 MHz must be >10 times the value for 11 MHz
23:51:06LinusNi made a nice excel sheet for all those calculations, i wonder where i goofed
23:51:21amiconnDo you still have that sheet available?
23:51:27LinusNyes, at work
23:51:32amiconnHmm :/
23:51:49amiconnOpenoffice should be able to open excel files
23:52:53preglowwill this affect performance much?
23:52:55Bagdernumber of commited files since 2.4:
23:53:00Bagdertop 5
23:53:04Bagder 574 amiconn
23:53:05Bagder 482 bagder
23:53:09Bagder 404 linus
23:53:09Bagder 269 hcl
23:53:09Bagder 221 dave
23:53:28Bagderhm, how did I make it there?
23:53:28LinusNpreglow: it will affect performance somewhat
23:53:38amiconn574. wow!!!
23:53:54Bagderout of 2703
23:54:30preglowwhere am i :/
23:54:44Bagder 81 preglow
23:54:50linuxstbCommitting all those codecs was an easy way for me to get on that list...
23:55:03Bagder 167 jyp
23:55:03Bagder 119 hohensoh
23:55:03Bagder 94 zagor
23:55:03DBUGEnqueued KICK Bagder
23:55:03Bagder 81 preglow
23:55:03Bagder 71 christian
23:55:44linuxstbLinusN: Is there anything we can do to get 48KHz playback on the iRiver? Or was your statement a few months ago that we have to do software resampling the final word on the subject?
23:56:03t0masBagder: and me? 3 or 4 ? :)
23:56:12LinusNlinuxstb: resampling is the only way to go
23:56:13Bagder 15 tomas
23:56:19t0masoh ok :)
23:56:31Bagdert0mas: aren't you sleeping? B-]
23:56:50linuxstbHas anyone looked for suitable resampling algorithms yet?
23:56:54t0masyes I am
23:56:58preglowinterpolation is mroe or less our only choice
23:57:04t0maswas just returning to scream: "I love gapless"
23:57:19t0masas I was listening some music while cleaning up before I was going to sleep
23:57:29CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
23:57:29*LinusN screams "i've had gapless for years"
23:57:31linuxstbpreglow: I mean do we have any code ready to be slotted in, when we have a place to slot it?
23:58:17preglowmaking a quick linear interpolator shouldn't take too much time
23:58:35preglowbut i'm concerned linear interpolation might kill too many high frequencies
23:58:46linuxstbAre we keeping the DAC at a constant 44.1KHz and resampling everything to that (even 22.05KHz etc)?
23:58:49preglowjust one way to find out, i'll try a few experiment
23:58:58preglowi can't seem to get any work done anyway

Previous day | Next day