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

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

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

#rockbox log for 2005-04-12

00:00:40preglowwell, yeah, but that's pretty much a special case, being a part of a bigger work
00:00:45preglownamely libavcodec
00:00:56preglowor ffmpeg or whatever
00:01:16preglowno, i'm plain wrong
00:05:22rasher//#define MPC_FIXED_POINT
00:05:28rasherexcellent :)
00:05:41rasheralthough somewhat malplaced
00:06:14preglowyeah, it's fixed point
00:06:34rashernot by default though
00:06:48pregloweasily done
00:06:56rashervery :)
00:07:00preglowi'm looking forward to seeing performance numbers
00:07:03preglowit should be pretty fast
00:07:10preglowsince it's just a simple subband coder
00:07:29rasherWasn't it a musepack developer who bought a h120?
00:07:37rasheror 140
00:08:00preglowyeah, the wavpack guy mailed me and said he was very interested
00:08:20preglowhe has a h140, embedded experience and interest
00:09:42 Quit matsl ("Leaving")
00:09:59rasherGuess I could redo this without touching as much
00:10:14rashershould I?
00:10:59preglowlibmusepack is pretty small, doing a merge on a per file basis won't be much hassle
00:11:05preglowif it's okay by the big boys
00:12:26 Quit Harpy (Read error: 60 (Operation timed out))
00:15:04 Quit lolo-laptop ("Client exiting")
00:17:30preglowi'm having trouble applying the patch, thoguh
00:19:15rasheroh ah
00:19:27rashershouldn't be too hard manually
00:21:59preglowwell, no, but i don't want to do it manually :PP
00:23:12preglowdid you do -u ?
00:23:25rasherno, just cvs diff
00:23:40rasherdidn't really know what I could do with it
00:23:44preglowdoesn't matter, i guess
00:23:59rasherit shouldn't
00:24:49preglowwell, when i try to patch, it just asks me what files do patch
00:25:07Seedrasher/preglow: Shy and I are the Musepack people. if you need any hints...
00:25:36preglowSeed: will let you know
00:25:51preglowi'll try to make it output some sound now
00:25:55Seedgoing to sleep.. but the programmers of libmusepack will be around tomorrow
00:26:08preglowaight, good night
00:26:34Seedthey are on freenode. but if there's something urgent, #MPC is where you'll get answers
00:32:27preglowrasher: could you try making a patch with -u as well, just for shits and giggles?
00:33:22 Quit DangerousDan (Read error: 60 (Operation timed out))
00:34:20rashercvs diff -u ?
00:35:24rasherweird.. what are the "RCS file:" lines doing there?
00:35:47preglowno idea">
00:37:27preglowno dice
00:37:33preglowi'll just do them by hand
00:39:27preglowparts of it have been patched several times
00:39:31preglowwhat the hell is this
00:40:43preglowignore all this
00:40:56preglowthe laugh is on me, as per usual
00:41:30preglowthe libmusepack tarball i got from didn't have a root folder, so i managed to unpack it straight into apps/codecs, and had to clean it up afterwards
00:41:34preglowand i cleaned away the Makefile as well
00:41:42preglowsmall wonder patch didn't get shit
00:43:08preglow-u is always better
00:43:17preglowthat way things can't be patched several times, afaik
00:44:51rasherJust didn't know that cvs diff would accept options in that way
00:46:07preglowmy musepack compile sure vomits
00:46:21rasherwhere did it vomit?
00:46:34preglowfirst file
00:46:34preglowreader.h:68: error: parse error before "FILE"
00:46:34preglowreader.h:68: warning: no semicolon at end of struct or union
00:46:37preglowand tons more
00:47:12preglowrockbox doesn't have file
00:47:20preglowso small wonder
00:47:27preglowyou actually did compile it?
00:47:27rasherthat's where mine starts as well
00:47:34rashernot at all
00:47:36preglowahh, i thought you compiled it
00:47:43rasherdon't think I claimed to
00:48:10preglowthen my claim to have it producing sound tonight might very well be a bit optimistic :P
00:51:38 Join lostlogic [0] (
00:52:06rasherI just don't know which modifications to make
00:52:11preglowthere, compiled
00:52:20rasherthat was fast
00:52:37preglownot much work needed, just ripped out the default file reader
00:52:42preglowwhich we won't be needing
00:52:56preglowi'll get to work on mpc2wav
00:53:18 Join XShocK [0] (
00:55:26rashernow if I had any musepack files..
00:56:15ShyKrasher: that's no problem :)
00:56:25preglowmusepack files i'll make
00:57:06ShyKmpc inside (tm)
00:58:28preglowbut yes
00:58:43preglowdave chapman most certainly seems to be right in saying most codecs work with a file io callback system
01:06:53rasherthink I'll make flack, musepack, wavpack and vorbis versions of iriver.mp3
01:09:32preglowmake note of the options you use to encode
01:09:33preglowit does matter
01:13:22preglowyou might even want to make multiple bitrate version of things if you want to use these things as benchmarks
01:14:17rashergood idea
01:14:36rashersource is a 320kbps mp3, so it should be fairly alright
01:14:40rasherand so much fun!
01:15:36ShyKwhat's this mp3? intro?
01:16:17rasherit's the iRiver song
01:16:19rasherit's so cheesy
01:16:27rasher30 seconds or something
01:16:39 Part MoosCamaro
01:16:54ShyKofficial or made by someone unrelated? :)
01:18:19preglowand it makes me want to cry every time i hear it
01:18:27preglowso what better track to make multiple versions of!
01:18:49ShyKi'd like a copy. i might do a remix that makes it listenable :)
01:20:08rasherhaha.. I'll put it on my webspace after I've made all the versions
01:21:56preglowmake a terrorcore remix of it
01:22:11rasherflac −−help ... holy crap...
01:22:13preglowi'll consider it suitably defiled then
01:22:19preglowrasher: −−best is a useful setting
01:22:24preglowrasher: but make one without as well
01:22:43ShyKfor wavpack use -h -x
01:22:45 Quit Aison ("( :: NoNameScript 3.72 :: )")
01:23:16ShyKpreglow: flac settings don't affect decoding
01:23:52rasheryou'd hope not
01:23:54ShyKit's different in that all encoding settings make files that decode at same speed
01:24:40rasheroh well
01:24:49ShyKwavpack's -h settings results in ~half decoding speed than normal
01:25:11ShyKwhich is still very fast compared to ape :)
01:25:40preglowShyK: why, yes they do, internally
01:25:49preglowShyK: it sets lpc order above eight
01:26:08preglowShyK: and i've written an assembler core that only does lpc equal to or lower than eight
01:28:10preglowquality wise it doesn't matter, of course, but it gives a somewhat smaller file that will decode more slowly
01:29:27rasherwhy are none of these tools in debian..
01:29:38preglowbecause debian is an outdated piece of shit? :-)
01:29:46rasherit so is not
01:29:51rasherat least, unstable isn't
01:30:06preglowwell, at least not when it's not frozed
01:30:08preglowlike it is now
01:31:13rasherunstable isn't frozen
01:31:19rashertesting is
01:34:34***Saving seen data "./dancer.seen"
01:36:22 Join DMJC [0] (
01:36:58preglownot frozen, but it's on some kind of hold
01:37:02preglowtons of packages are queued
01:37:15preglowit just recently got gnome 2.10
01:38:18rasherah, like that.. well the gnome people are working on ironing out 2.8
01:39:12 Join Aison [0] (
01:39:18preglowmeanwhile, i have thoroughly enjoyed 2.10 for a month now :P
01:40:01rasherI have for even longer!
01:40:21preglowyes, i'm on ubuntu as well
01:40:31rasherI thought you were on gentoo
01:40:33rasherthen I haven't
01:40:35preglowmy server's gentoo
01:40:40preglowmy desktop not
01:40:47rasher$ ./mppenc
01:40:51rasherbash: ./mppenc: Permission denied
01:41:37preglowyou didn't switch on fixed point in musepack?
01:43:04rasheroops, no
01:43:10rasherlibmusepack that is
01:43:27rashernow what the hell am I doing wrong here
01:45:54preglowarghh, why haven't anyone just written xxx2aiff
01:48:21rasheroh jesus
01:48:26rasherI'm an idiot
01:49:23rasherI was working on my iriver.. apparently the permissions I set up for this mount doesn't allow execution
01:49:31rasherand my chmod obviously wasn't doing any good
01:49:55rasherthat's better...
01:51:58 Quit Aison ("( :: NoNameScript 3.72 :: )")
01:53:28rasherthat's musepack files done
01:53:46 Quit Sucka ("a bird in the bush is worth two in your house")
01:55:13preglownothing beats listening to psytrance when coding
01:55:42preglowright leg going 180 bpm, fingers typing like mad
01:58:49*rasher does wavpack
02:01:09preglowhmm, why the hell isn't config_types.h included anywhere?
02:01:20Rob-what is the status of the iriver port?
02:01:34rasherpreglow: that's a very good question
02:01:40preglowwhy, it is included
02:01:42preglowin musepach.h
02:01:48rasherah :)
02:01:50preglowRob-: all's well, lacking codecs
02:02:05preglowthen why the hell does this error occur
02:02:44Rob-does it play and record wav at the moment?
02:03:05rasherno.. it doesn't really play, certainly not record anything
02:03:56rasherinternal architecture changes are needed to allow software codecs
02:04:01rasherthose haven't been done yet
02:04:39rasherwhat the fuck.. this error looks like a real code error
02:04:39preglowwell, the fixed point broke it nicely
02:04:57rasherprobably not well-tested?
02:05:10rasheror maybe just not on "limited" hardware
02:05:21preglowno, not that
02:05:24preglowsome compiler error
02:05:49*rasher is confused
02:05:53preglowalmost seems like mpc_int32_t etc aren't defined
02:06:17rasherand you checked that the values in config_types.h are acceptable?
02:06:45rasherbecause those are set by ./configure .. I just did a configure −−build m68k-elf and copied it from there
02:06:49preglowoh, indeed
02:06:53preglowthey're dandy
02:07:21preglowthe file isn't included early enough
02:07:33rasherthat's excellent
02:07:42rashercould be my fault I guess
02:07:44preglowan "#error EHEHEHEHEH" inserted in strategic places can reveal many a thing
02:08:16preglow#include <math.h>
02:08:24preglowi'm willing to bet that's the culprit
02:08:34rasheroh well, guess I'll do wavpack on win32
02:08:50preglowit was
02:10:26preglowwhat the _HELL_ ?
02:10:26preglow#ifdef MPC_FIXED_POINT
02:10:26preglowstatic mpc_uint32_t find_shift(double fval)
02:11:03preglowahh, lookie here
02:11:22preglowi'm more than willing to call this a flaming hack
02:12:03rasherwhat's the problem there?
02:12:50preglowthere's a floating point value in the parameter list there, you see
02:12:55preglowafter a flaming fixed point ifdef
02:14:07preglowwhy, yes
02:14:19preglowthey then go on to interpret the floating point number with some crazy stuff
02:14:53preglowbut i've bumped into a more serious issue here
02:16:02rasherwow, wavpack -h -x sure is slow
02:16:07preglowyou very nearly might have a point with it not being tested for fixed point
02:16:19preglowit just fixed a bug that would prevent it from compiling anywhere
02:16:29preglowit = i
02:17:11rasherSounds silly
02:17:15ShyKit has been tested for fixed point and pocket pc support exist since a long time
02:17:43preglowShyK: well, i just had to fix the source to make it compile here
02:17:58preglowit lacked a pointer dereference
02:18:08preglowi find it hard to believe that source ever compiled
02:18:21ShyKpreglow: feel free to post this on the forum. i'm sure they could apply it
02:18:31preglowi will, i'll just try on another box first
02:18:35ShyKpreglow: it most definitely did compile
02:18:46rasherShyK: any other interesting options for wavpack?
02:18:56preglowwell, this musepack 1.1 surely didn't
02:19:07rasherI have one done with defaults and -h -x
02:19:25ShyKrasher: nope. the interesting feature is -x that adds extra compression with no decoding speed influence
02:19:41rasherpreglow: for great justice
02:20:44*rasher creates a -c
02:21:19rasherI meant b
02:22:29 Quit DMJC ("Leaving")
02:28:30 Join DMJC [0] (
02:28:33preglowBOOL was a poor choice of name for an enum on rockbox' part
02:29:03rasherhm, what am I missing..
02:29:21rashermp3, wav, flac, ogg, wavpack, musepack
02:30:14ShyKmodules :)
02:30:34rasherthat'd be so pointless
02:30:45preglownot to mention impossible
02:30:56preglowcould of course make a mod with one pattern and one sample
02:30:57rasherwell you could chop it up in little bits
02:30:58preglowand one note...
02:31:28preglowthis flaming BOOL is ticking me off
02:31:31preglowit's all that's standing in the way
02:32:11preglowrasher: could you send me the mpc?
02:32:31rasheryup, hang on"> .. not up yet
02:33:55rasherthere, first version uploadet
02:34:30rashershould be
02:34:31preglow404 still
02:36:17preglowno 404, but no permission
02:36:23preglowdo the chmod
02:36:40rasherweird, the dir wasn't created
02:36:53rasherno dirlistings
02:37:25rasherI wonder if I can turn on dirlistings
02:37:49preglowyou can
02:38:15rasherassuming apache is set up to allow me
02:38:28preglowis it your own server?
02:39:34preglowanyone know how i can do a search and replace in all files?
02:39:51preglowsed s/BOOL/bool/ * spits everything to stdout
02:39:55preglowwhich i'd rather it didn't
02:40:22rashersed s/BOOL/bool/ > tempfile; mv tempfile $filename
02:40:34preglowi'd rather it did so automatically
02:40:51preglowcatch the digital flow.....
02:41:23*rasher uploads flacs
02:41:25preglowi hope this "artist" has been drawn and quartered by the people who used to be closest to him
02:41:49ShyKoh god, like the mp3 pre echo is not enough, this is indeed very spoony
02:42:04preglowit's a 320kbps file...
02:42:13ShyKdoesn't matter
02:42:22ShyKseems like they used xing or something
02:42:34ShyKwhich would have horrible pre echo at 3,000kbps :)
02:42:41preglownot impossible
02:42:55preglowwe are talking about iriver here
02:42:58rasherthere's actually a comment in the file of which encoder iirc
02:43:10preglowrasher: then do indulge us with that information, please
02:43:43rasherTSSE (Software/Hardware and settings used for encoding): N2MP3
02:43:51ShyKholy shit
02:43:57rasherCOMM (Comments): ()[]: easier than ever. Mobility. Simplicity. Superiority. The most discreet way to fiercely express yourself. Can you Flash?
02:44:29preglownever heard of that
02:44:39preglowseems to be some mac piece of shit
02:44:56preglowcan you flash?
02:45:13rasherI can.
02:45:24preglowdepends on what he means
02:45:29 Join DJ_Dooms_Day [0] (
02:46:04rasherif it's some bizarre sexual practice, then I'm not so sure
02:48:06preglowi'm perfectly capable of exposing my genitals to innocent bypassers
02:48:11preglowhaven't tried, but think i can do it
02:48:28*rasher smacks forehead
02:49:41rasherThat's not so bizarre I guess
02:49:53rasherat least.. nothing I haven't heard of beforee
02:50:34preglowlo, he awakens!
02:50:38HClno no.
02:50:40HCli'm going to sleep.
02:50:57HCljust wondering what i dropped into >.>
02:51:00preglowi can't announce that dramatically :/
02:52:16preglowi'm trying to make libmusepack run
02:52:33preglowthen ran into the well-named BOOL enum in rockbox
02:52:55 Quit DMJC ("Leaving")
02:53:33preglowfuck this, says preglow
02:53:39preglowi'll continue tomorrow morning
02:53:44preglownow, sleep
03:01:19 Quit preglow ("fop")
03:13:23rasherme too..
03:13:26 Quit rasher ("CGI:IRC 0.5.4 (2004/01/29)")
03:23:38 Nick tvelocity[away] is now known as tvelocity (
03:34:35***Saving seen data "./dancer.seen"
03:58:44 Quit cYmen ("zZz")
04:05:58 Join QT_ [0] (as@area51.users.madwifi)
04:08:44 Quit QT (Read error: 60 (Operation timed out))
04:24:09 Join elinenbe [0] (
04:24:17 Nick tvelocity is now known as tvelocity[away] (
04:43:08 Quit asdsd_ ("Trillian (")
04:55:02 Quit XShocK (" HydraIRC -> <- Go on, try it!")
05:34:37***Saving seen data "./dancer.seen"
06:02:54 Join bobdbob [0] (
07:07:00 Join LinusN [0] (
07:34:42***No seen item changed, no save performed.
08:15:30 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
08:25:58dwihnoyay! driver download finished!
08:26:04dwihnotime to setup the wavelan card.
08:35:50 Join Harpy [0] (
08:36:56*LinusN needs lots of coffee
08:58:49 Nick Lynx_awy is now known as Lynx_ (HydraIRC@
09:01:24dwihnoYay! I got another wireless network card to fungieren!
09:14:48 Join bobTHC [0] (
09:16:07bobTHCmorning all!
09:16:30Strathmorning bobTHC
09:34:45***Saving seen data "./dancer.seen"
10:00:15 Join ashridah [0] (
10:29:46 Join preglow [0] (
10:30:59preglowyou guys know there is a penalty in the afterlife for using names like BOOL for enums, right?
10:32:08ashridahhah. may as well. like i'm going to anyone's version of heaven anyway
10:32:59*ashridah resumes his sleep-deprived state of delirium
10:36:53bobTHCpreglow > have u found a "decent" digital audio editor for nux ?
10:37:27preglowbobTHC: nah, forgot all about it yesterday
10:45:30 Nick QT_ is now known as QT (as@area51.users.madwifi)
10:49:12bobTHCa very good news on the forum for Archos owners :" Both Archos and Micronas decisionmakers have agreed to publish the MAS software to us"
10:49:58bobTHCnothing is publish for the moment but, i hope it will be soon
10:55:04preglowwow, cool
10:55:20preglowthat definitely rocks
10:55:28preglowbut is there a mas compiler?
10:55:36Bagderbobdbob: what forum?
10:56:00Bagderbesides, the docs is already "out"
10:56:31preglowbut compiler? :V
10:56:53BagderI don't remember exactly what it contained
10:57:03Bagderthere was more than just docs, but I remember some "gaps"
10:57:18Bagderand of course, it wasn't found "the right way"
10:58:18Bagdersome eastern european university web site published it
10:58:29bobTHCthe other problem is the mas connction with other parts who have some bottlenecks, isn't it ?
10:59:01BagderbobTHC: yes
10:59:24Bagderbut I would say that the biggest problem is that someone would actually need to go and do a lot of programming
10:59:29Bagderon a weird DSP
10:59:36Bagderwith very little debugging aids (I guess)
11:00:48preglowwhy, yes, codecs would pretty much need to be re-implemented
11:01:01bobTHCit's a job for a coding wizard, but it's not the first ;)
11:01:42LinusNthe MAS software in question is a WAV codec
11:02:01preglowthen i sure as hell hope they have docs as well
11:02:42LinusNthe docs we have only cover the DSP core, but not the onboard peripherals
11:03:00LinusNlike the I2S serial controllers etc
11:03:36LinusNso we can develop and simulate pure MAS DSP code, but we can't make the MAS communicate with the SH1 CPU
11:03:40preglowso, what are the coarse specs of the mas?
11:03:51bobTHCoki, so the dsp code for wav codec is already in our hands ?
11:04:02LinusNbobTHC: not yet
11:11:57preglowdoes anyone care if i alter the directory structure of codecs before commiting?
11:12:11preglowlike new codecs
11:12:35preglowwe're supposed to change as little of them as possible
11:12:42preglowdoes that apply to directory structure as well?
11:12:42LinusNah yes
11:13:01LinusNif it serves a purpose, i don't mind if you change the structure
11:13:10BagderI agree
11:13:40preglowi don't really care, rasher sent me his libmusepack for me to finish yesterday, and he just moved things out of src/ and include/
11:14:17LinusNso all .h and .c files are in the same dir?
11:14:30LinusNthat's ok
11:15:27preglowbut yes, i'll try to make mpc2wav work now, and commit it if it does
11:16:08preglowwhy does everyone have to define their own BOOL ://
11:16:33Bagderpreglow: the pains of pre-C99
11:16:59preglowwhat should i change it to? MPC_BOOL? bool?
11:18:09Bagderperhaps make it use 'bool'
11:18:18Bagderand include <stdbool.h>
11:18:24preglowit also has TRUE and FALSE, heh
11:18:28*preglow kickstarts sed
11:18:35 Join Aison [0] (
11:18:50LinusNi just love when they compare to TRUE
11:18:58LinusNif(mybool == TRUE)
11:19:39ashridahLinusN: yeah, nevermind that TRUE can only be defined as !0 iirc :)
11:21:55preglowi also assume removing unneeded files is not a crime?
11:22:12preglowdumb needs some deletion, then
11:22:21preglowthere are files commited that will never be needed
11:23:01preglowlike allegro interface files
11:23:52 Join Mirfle [0] (
11:27:41LinusNlunch time
11:33:15ashridahdamnit. C89/90 didn't include any kind of regex processing did it?
11:33:28ashridahi really hate having to do input validation by hand :(
11:34:32Bagderregex shouldn't be in the C standard, it belongs in posix
11:34:47***No seen item changed, no save performed.
11:35:11ashridahyeah, i know, but i was hoping they were tools and put one in
11:35:22ashridahsince i can only use functions that come with C89's libc.
11:35:36*ashridah mutters and goes back to writing a state machine.
11:45:39*preglow sighs and gets his reset button implement
11:53:25 Quit Strath ("Client closed")
12:01:18preglowi seem to have made a plugin that doesn't even load...
12:01:46preglowi assume this means i'm skilled
12:13:32preglowany known stupid blunders that would result in my mpc2wav just doing nothing? the first thing i do is print a splash, and even that doesn't display
12:16:38 Join MoosCamaro [0] (
12:16:50MoosCamaroHey all
12:36:50 Quit Mirfle ("Chatzilla 0.9.67 [Firefox 1.0.2/20050317]")
12:45:47 Join rasher [0] (
12:46:48*rasher prods Bagder
12:50:06preglowrasher: compiles fine and all, but my plugin refuses to work
12:50:13preglowwill soon have to start doing other things
12:51:03ashridahare there any macros you need to use to define standard entry points into a module or whatnot?
12:51:46preglowwhat modules?
12:52:23preglowwell, apart from using plugin_start(), nothing i can think of
12:53:00ashridahare you accidentally stripping the module? :)
12:53:44preglowrocks are stripped, afaik
12:53:48preglowthey're just a plain binary image
12:55:51ashridahah. i just remember stripping a kernel module once, and getting nothing useful in return :)
12:56:12LinusNpreglow: the .map file could give some clues
12:56:32*Bagder awakes and looks around
12:57:41preglowwell, can't see anything obvious
12:57:53preglowentry point is correct
12:58:19pregloware there any limits imposed on static content?
12:58:37preglowbah, i don't use nearly enough anyway
12:58:43rasherBager: About the digest thingy.. do I just commit it and hope for the best, because I can't make the scripts work
12:59:00Bagderrasher: sure, commit and we shall see what happens! ;-)
13:01:06rasherHaha... I guess I should have saved the password I told you to set... do you log?
13:01:41 Quit ashridah ("Leaving")
13:01:49BagderI do
13:06:33LinusNpreglow: still problems?
13:07:07LinusNemail me the plugin src
13:07:21LinusNlinus at haxx dot se
13:07:36LinusNor put it up on a web server
13:10:20preglowthe relevant info would be that the topmost splash doesn't even display
13:10:22preglownor anything else
13:10:32preglowit just hangs with the backlight on, and nothing on display
13:10:41*preglow braces for a stupid bug and goes to make coffee
13:11:50 Join R3nTiL [0] (~zorroz@
13:15:26 Join cYmen [0] (
13:17:42 Join Sucka [0] (
13:19:48LinusNpreglow: can i see musepack.h?
13:20:00LinusNand your .map file
13:21:06preglowwhy, yes
13:21:40preglow and
13:22:27preglowentire libmusepack dir is there as well if you need more
13:23:01LinusNyour web server doesn't like the .map extension
13:26:09preglowahahah, what the hell?
13:26:30preglowi shall have to remedy that
13:27:00preglowin the meantime, it's there as txt
13:27:31LinusNcan you give me a tar file with the musepack src?
13:34:51***Saving seen data "./dancer.seen"
13:45:08preglowhaha, apache had the .map extension used for pixmaps
13:46:38preglowi never ever used those, so disabled the whole thing
13:48:41LinusNpreglow: the mpc_decoder type is 122468 bytes, you don't want that on the stack... :-)
13:49:00preglowholy friholies
13:49:15preglowi didn't think it was THAT large
13:49:18preglowand i considered using iram for that
13:50:10LinusNcase closed
13:50:26preglowi expect it'll work now, yes
13:50:28preglowthanks a bunch
13:50:31LinusNyou're welcome
13:54:06preglowhaha, well, it makes sounds now
13:54:16preglowat least party reminiscent of iriver.mp3
13:54:45preglowthe efficiency is not at all what i had hoped
13:54:49preglowbut i think something is wrong
13:57:53rasherare you usingn a very low quality file?
13:58:21preglowi know what coding errors sound like
13:59:53preglowarghh, i really, really wonder how i'll handle iram with these large structs
14:00:17preglowwill probably have to modify everything to use two structs
14:00:27preglowwhich will work wonders for later patchability
14:00:47Bagdergo #MPC on :-)
14:01:13*Bagder points out that Seed is a musepack dev
14:03:28rasherhaha, that disco-floor on slashdot is excellent
14:04:05preglowmaking my apartment into the popular demo
14:11:16 Quit R3nTiL (Read error: 110 (Connection timed out))
14:20:22*LinusN wants a dance floor plugin for the H300 USBOTG port
14:20:45preglowmy god, i surely would need that sometimes
14:25:20rasherAny idea if the usbotg is a generic usb host? or just for usb-storage?
14:25:43 Quit ShyK (Read error: 145 (Connection timed out))
14:27:01LinusNit's whatever we make it, as far as i understand
14:27:43rasherSo if anyone wanted, they could plug in a usb wifi nic and knock themselfes out?
14:28:16LinusNas far as i can see, yes, but i haven't studied the data sheets closely
14:28:34LinusNi might very well be wrong
14:29:50Bagder <= several new archos players
14:30:27 Part lostlogic ("Kernels are for popping")
14:41:44rasherAnother thing Bagder...
14:42:05rasherI posted a patch to and discovered another "problem"
14:42:36rashermulti-artist cds get an album entry for each artist :-\
14:42:54Bagderyeps, I know Zagor had some idea on how to deal with it
14:43:25rashersomething about considering the dir they're in as well?
14:43:32Bagderyes, I believe so
14:44:00BagderI think that would be a fair way to guess slightly better
14:44:36rasherso compose the albumid of path__albumname instead of artist___albumname?
14:44:53rasherwouldn't that work 90% of the time?
14:44:59rasherif not more
14:45:02BagderI think it would
14:45:14Bagderexcept for those who have all songs in the same dir
14:45:21Bagderbut we can't work for those
14:45:49rasherI think this way of doing it works for more people
14:45:54rasherand in more cases
14:46:12Bagderyou do the patch?
14:46:25Bagderbtw, you can even commit your previous patch, since you have access now
14:46:38rasherwell I'd like someone to have a look at it first :)
14:46:55preglowwhat percentage realtime should we aim for?
14:47:08LinusNi think 300 is fair
14:47:38LinusNpreglow: did we perform some kind of profiling for libmad?
14:47:48preglowno, we did not, but some other people did
14:47:59preglowthat won't matter for us, of course
14:48:02preglowbut no
14:48:09preglowwe have no profiling possibilites
14:48:17preglowthat don't require us hacking the lib to pieces, that is
14:50:03preglowlibmad is almost 80% realtime for iriver.mp3
14:50:05LinusNwell, didn't we add that profiling timer?
14:50:08preglowat 40mhz, that is
14:50:20preglowyeah, i think we did
14:50:46LinusNsynth_full takes a lot...
14:50:49preglowit does
14:50:55preglowbut i don't think there's much more to be done about that
14:51:03LinusNIII_huffdecode is next
14:51:05preglowi'm going to rewrite it in pure assembler
14:51:06preglowbut that's that
14:51:28preglownot much to be done about huffman decode, it's strictly table based, afaik
14:52:06preglowi dare not touch dct32
14:52:41preglowour best shot for dct32 is replacing it completely, with something more iterative in structure, and that's probably not going to be faster anyhow
14:53:07preglowLinusN: i've lost the link the the libmad profiling page, care to share?
14:54:03preglowand see all the other dcts?
14:54:07preglowthey're called by dct32
14:54:12preglowso they're part of the dct32 algorithm
14:54:16DJ_Dooms_Daysup Linus, how goes PCM playback?
14:54:26LinusNDJ_Dooms_Day: slow :-)
14:54:30preglowno, now i'm speaking bullshit
14:54:34preglowthey're used by imdct_l
14:54:43preglowso imdct_l is also pretty intensive
14:54:47DJ_Dooms_Daylol, grand
14:55:52LinusNhehe, audio_linear_dither is about 13% (including prng)
14:56:24preglowi refuse to believe that can't be made faster
14:56:42preglowdone in asm, using all registers, with sequentail access to the samples
14:56:43LinusNit probably can
14:57:21LinusNhave you done any more optimizations not in cvs?
14:57:28preglowyes, let's see
14:57:43preglowi've started on making an imdct_mcf5249.s
14:58:06preglowimdct_s is done, and i've started porting the arm imdct_l to coldfire asm
14:58:14preglowbut it's a pretty big job
14:58:34preglowi've also got the old imdct_l i did, which is shit ugly, but nevertheless helps some
14:59:31LinusNyou chould consider committing some, if not only for backup purposes
14:59:56preglowhmm, i will
15:00:05preglowi'll just try out some other minor opts i've got going as well
15:01:02preglowmaybe tomorrow, i'll have to start working on my masters stuff again now
15:01:37preglowit would also be a sin to not go outside in this weather :/
15:02:43 Join lolo-laptop [0] (~lostlogic@
15:04:10rasherAny way to "fix" a bdf font?
15:04:24rasherError: Invalid BDF file, requires FONT_ASCENT/FONT_DESCENT/ENCODING
15:04:28rashersays bdfconf
15:05:10LinusNwell, you could load it in a font editor and save it again
15:05:33rasherThat's how I made it.. I blame the editor
15:05:52LinusNwhich editor?
15:06:12LinusNyou made the font from scratch?
15:06:20rasherno, converted from a .fon
15:07:36LinusNgotta go
15:07:47LinusNcu around
15:07:50 Part LinusN
15:08:23preglowwhy the hell doesn't libmusepack write the entire file
15:11:05rasherweird, xmbdfed claims that FONT_ASCENT is set
15:12:03preglowwell, have a look at the file, then
15:12:05preglowit's plaintext
15:13:08rasheroh, no ENCODING
15:13:15preglowthat is a showstopper
15:13:20preglowthe parser is pretty strange
15:13:45preglowi started to redo it in perl, but i accidentally deleted the source yesterday, heh
15:14:08rasherwait.. I don't understand this
15:14:59rasherat all..
15:15:12rasherFONT_ASCENT and _DESCENT are set in the header
15:15:19rasherENCODING is set for each char
15:15:26rasherI thought it was charset
15:16:10rasherI can't see anything wrong with this file
15:16:39Bagderconvbdf cannot have any bugs ;-)
15:17:01rashersurely not
15:20:00preglowoh well, i'm out of here for a while
15:27:20rasheris convbdf expecting properties in a certain order or something?
15:31:49 Join t0mas [0] (
15:34:53***Saving seen data "./dancer.seen"
15:51:05 Join asdsd_ [0] (
15:52:44 Part asdsd_
15:52:56 Join DangerousDan [0] (
16:13:21 Quit tvelocity[away] ("Leaving")
16:26:38 Join elinenbe_ [0] (~elinenbe_@
16:33:52 Join muesli- [0] (
16:36:06rasher lo
16:37:27t0massomething wrong?
16:37:56muesli-nothin with me :D
16:44:10*rasher fixes flipit for h100
16:45:31muesli-whats that good for?
16:45:47rashernot much
16:47:00preglowi've bounced into problems with libmusepack, yes
16:48:32rashererp, what's up?
16:56:30preglowi get a corrupt frame error
16:56:38preglowand i don't know why ./
16:56:40preglow:/ <-
16:56:55Bagderpreglow: so all builds fine, but the output is bad?
16:57:22preglowall builds fine, but output is semi-ok, and way too small, seeing as how it exits due to a faulty frame
16:57:25preglowand the frame isn't faulty
16:57:28*rasher tests songdb
16:57:29preglowso it's probably my io
16:57:59preglowdebugging this stuff isn't exactly simple either
16:58:16*Bagder disappears
16:58:27rasherdammit, my sync script/rsync is too smart
16:58:35rasherdidn't run
16:58:41rasherbecause no files were transferred
17:00:14rasherIn other news, I finally found the right object to push reset
17:00:36rashera piece of wire will easily miss the switch
17:00:54rasherbut this small nail is just the right size :)
17:01:12preglowi use my smallest torx driver
17:04:17rasherMax songs per album: 316 (<no album tag>___/mnt/iRiver/Music/New/Others)
17:04:19rasheroh well...
17:06:02rasherfunky stuff is happening here
17:06:31rasherI think the id3db reader is off its rocker
17:08:53preglowplease don't tell it suddenly started working after i added debugging measures
17:09:23preglowclever, that was
17:09:47rasherObviously it's not something I made up
17:12:23rasherbut I like the term so much
17:12:40preglowIT SUDDENLY WORKS NOW
17:13:35rasherdon't look back
17:13:36rasherjust accept it
17:13:48preglowwell, it ALMOST works
17:13:50preglowit's double speed
17:14:04rasherlet me guess.. you added debug measures, then removed them again?
17:14:04rasherand now it works?
17:14:47preglowhaven't tried removing them
17:15:46rasherI'm sure *someone* will be glad to have musepack support
17:16:50preglowwell, i would
17:16:53preglowi think it's a cool format
17:17:01CoCoLUSi would if i could actually -listen- to it :)
17:17:14HClwhat kind of format is it?
17:17:25preglowit's basically mpeg2 on acid/speed
17:17:30preglowmpeg1 layer2
17:17:49HCleh, ok o.o
17:18:02*HCl smiles and nods
17:18:06 Join tvelocity [0] (
17:18:09preglowit's good for higher bitrate stuff
17:18:12preglowlike 160-192 ish
17:18:16HClkay o.o.
17:18:20preglowand also patent free, they say
17:18:33preglowand, it SHOULD be a pretty efficient codec
17:18:37preglowbut right now, it sure as hell isn't
17:18:49preglowruns at about 35% realtime at 120 mhz
17:18:58preglowbut i guess i can massage it into submission
17:19:43rasherogg first! ogg first!
17:20:07HCli'm interested in taking the original firmware apart
17:20:11HCland snatching their ogg codec
17:20:30rasherhm.. you probably shouldn't
17:20:49preglowi quite simply haven't got time to touch vorbis for a long time
17:20:56preglowmusepack seems to be such a cute little codec
17:21:00preglowso i thought i'd try that
17:21:08preglowi also have libmad and libflac on my hands
17:21:17HClrasher, why not?
17:21:24rasherheh, of course.. do whatever you're most interested in
17:21:25HClreverse engineering is allowed for portability
17:21:31HClin america anyways
17:21:31preglowHCl: i sincerely doubt it's worth stealing
17:21:40preglowHCl: and reverse engineering that is going to be a nightmare
17:21:43HClin europe, i think its allowed, heh.
17:21:47HClpreglow: is it?
17:21:51preglowHCl: yes
17:22:01preglowbecause it's a pretty large codec
17:22:08preglowand dsp code isn't the easiest to read
17:22:18 Quit muesli- (Read error: 60 (Operation timed out))
17:22:20HClwell we don't have to write it back into code.
17:22:21preglowbut hell
17:22:24preglowgo for it!
17:22:24HCljust the routine that loads an ogg
17:22:30HClheh, nah, not yet.
17:22:33preglowi wouldn't want that anyway
17:22:37preglowi like clean stuff
17:22:45preglow.h file in binary != clean
17:22:55HCli was sort of thinking of just calling the firmware code.
17:23:04preglowwell, if you can figure out the addresses, go for it
17:23:06HClits just gonna be a subroutine.
17:23:09t0masHCl: cheater :P
17:26:40rasherokay, the tag-db is definately buggy
17:26:43rasherand it's not my fault
17:29:36 Join rjamorim [0] (~blah@
17:29:46 Join XShocK [0] (
17:30:17rasherhi, XShocK
17:33:08preglowTHERE i got the blasted turd working
17:33:15preglowslow as hell, but it makes coherent sound
17:33:34rasherwhich speed?
17:33:43preglowat 120 bleeding megahertz
17:33:48CoCoLUSnow if only... we could actually test that... with headphones :P
17:33:55preglowCoCoLUS: code away, ples
17:34:11rasherwe can.. if we rename to sample.wav and go info>debug>audio test
17:34:18preglowindeed, good point
17:34:23preglowi don't see what they're nagging about anymore
17:34:26preglowwe have audio out!
17:34:36CoCoLUSyeah the first 8 mb ;)
17:34:43preglowso, you want more???
17:34:45rasherthat's all you need for iriver.wav
17:34:53preglowwhich is all i listen to
17:34:55***Saving seen data "./dancer.seen"
17:34:55CoCoLUSwhich is, after all, the best. song. ever.
17:35:23rasheroh, maybe it *IS* my fault that the songdb was confused
17:35:26t0masrasher: I start to like that song ;)
17:35:38rashert0mas: kill yourself at once
17:35:51*preglow gently removes the libmusepack debug code
17:35:58 Quit t0mas ("Connection reset by jump off the roof :P")
17:36:17rashercareful now
17:36:22preglowi'm using pliers
17:36:51 Join El_Gringo_ [0] (
17:36:59*rasher rebuilds songdb
17:37:04 Join t0mas [0] (
17:37:16t0mashm... need to get a higher roof :P
17:37:28rasherThe thing gets confused if you change the db underneath its feet
17:37:53El_Gringo_I've some informations witch could be usefull for the audio API
17:37:59preglowEl_Gringo_: spit
17:38:47rashermy change is actually working
17:39:08 Quit XShocK ("CGI:IRC (EOF)")
17:39:13preglowrasher: then starting coding a gui tool which embeds perl so we don't require people to install perl!
17:39:30El_Gringo_preglow, Tang and Gabriel Bouvign have a message for you
17:39:35Seedpreglow: feel free to abuse me verbally if things go wrong :)
17:39:37rasherA C version would probably be a better bet
17:39:44preglowSeed: things are working nicely now
17:39:51SeedI'm glad
17:39:53preglowrasher: yes, but take ten times the time
17:40:02preglowrasher: embedding perl is very easy
17:40:26rasherI ain't doing it!
17:40:32rashernor am I writing a c version
17:40:35El_Gringo_Gabriel Bouvigne Something to add about Vorbis low bitrate limitation for iFP/iMP:
17:40:40El_Gringo_In oldest iMP there is some memory in the processor (ARM) and also main memory.
17:40:47El_Gringo_The main memory is large but the access is very slow. On the contrary the processor memory is very limited but access is very quick for the processor (quite obviously ;))
17:40:53El_Gringo_Unfortunately the big windows of Vorbis don't fit in the "on chip" memory, so the processeur need to keep this data in the main memory which is too slow to allow the processor to make real-time decoding...
17:41:00El_Gringo_It's partially an iMP limitation (hardware undercalculated to get low cost) but it's also a critical flaw of Vorbis design.
17:41:08El_Gringo_Indeed Vorbis should have:
17:41:13El_Gringo_- not allowed so big windows
17:41:17El_Gringo_- created profiles has MPEG ones...
17:41:21 Nick Sucka is now known as Sucka`away (
17:41:23*rasher ponders committing this change
17:41:40 Quit Harpy (Read error: 60 (Operation timed out))
17:41:48preglowEl_Gringo_: yep, i pretty much figured all that, and agree as well
17:42:07El_Gringo_And something else
17:42:34El_Gringo_here is a MP3 decoder full integer : [...] 969&st=0&#
17:42:49 Quit bobTHC ("Smoke Weed Every Day !")
17:42:53El_Gringo_here :
17:43:15rjamorimRead my replies there
17:43:22rjamorimI'm pretty confident it's not GPL-compatible
17:44:03El_Gringo_rjamorim, you speak about the decoder ?
17:44:04preglowyeah, i know
17:44:16preglowand: i have already invested tons of time in libmad, and i like it
17:44:17rjamorimOh, sorry. I thought you meant the IPP encoder
17:44:18preglowit's a good library
17:45:09preglowintel libraries probably have tons of x86 assembler in them as well
17:45:24rjamorimYes. And Xscale optimizations too
17:45:46 Nick tvelocity is now known as tvelocity[away] (
17:45:54rjamorimEl_Gringo_: that thread is not mentioning an MP3 DEcoder :B
17:46:29preglowencoders i won't touch with a five foot stick
17:46:33preglowtoo much work
17:46:34rjamorimThat's why I got confused
17:46:35El_Gringo_rjamorim: you're right ! I'm only a "messenger"
17:46:40rjamorimheh, right
17:46:54preglowEl_Gringo_: why can't they come online and speak themselves?
17:47:11rjamorimTang used to be around here often...
17:47:14preglowi don't feel courier based communication is the best, heh
17:47:17rjamorimMakes me wonder where is he
17:47:20preglowyes, he comes on from time to time still
17:47:25preglowspotted him yesterday
17:47:46El_Gringo_preglow: simply because of an Internet connection problem
17:47:53rjamorimAnyway, I'm pretty confident there's no open source psychoacoustical integer encoder.
17:48:06rjamorimThe closest you can get is probably WavPack lossy
17:48:16El_Gringo_It's a message from Tang
17:48:44*rasher boggles
17:48:45rjamorimTang sent you the wrong link :P
17:48:49preglowrjamorim: yes, i'm pretty confident in that as well
17:48:55rjamorimOr maybe he too confused En/Decoder
17:49:04preglowand i also happen to not care
17:49:10preglowi will do my recording losslessly
17:49:22preglowand do the encoding on something that doesn't have a realtime requirement
17:49:22rjamorimI see
17:49:41El_Gringo_pregolw, you are the guy who work on the MP3 decoding plugin ?
17:49:46preglowEl_Gringo_: yes
17:49:59preglowmusepack right now, heh
17:50:45*rasher tries to commit and fail
17:51:03El_Gringo_preglow: And now, you have still some optimisations the make, or you're juste waiting for the Audio API ?
17:51:16preglowEl_Gringo_: oh, tons of optimizations left
17:51:27preglowEl_Gringo_: but it is fast enough as it is
17:51:29 Join Harpy [0] (
17:52:06El_Gringo_I thought MP3 decoding was 200% of rel time ! Is it insufficient ?
17:52:24preglowthat's for vbr files
17:52:33preglowwe'll need to have it running at at least 300% for 320kbps
17:52:38preglowthen i'll be more or less satisfied
17:52:38rjamorimI suppose the faster you decode, the less battery it consumes...?
17:52:46rjamorimhaha, wow
17:52:54preglowand we need fast codecs for stuff like crossfading
17:53:00El_Gringo_your right
17:53:07preglowwhere we'll pretty much need to codecs running in parallel
17:53:12rasherdo we really?
17:53:22preglowrasher: well, how do we do it otherwise?
17:53:28rasherwon't the file be fully decoded a good while before it's done playing?
17:53:43rasherdepending on the size of the pcm buffer of course
17:53:45preglowrasher: well, yes, if not in actuall parallel, they'll need to use the same amount of cpu time
17:54:04preglowrasher: the first codec will have to work harder at the end of the file, to prebuffer the fade area
17:54:17El_Gringo_I've to go, bye
17:54:18preglowrasher: then the other codec needs to throttle up like hell when it's started, to fill the fade area as well
17:54:24preglowthen normal operation continues
17:54:26preglowEl_Gringo_: bye
17:55:56 Quit El_Gringo_ ("Chatzilla 0.9.67 [Firefox 1.0.2/20050318]")
17:56:47preglowis it just me, or does play NOT reboot the device as rolo says?
17:56:58rasherhit select
17:57:03rashertext is wrong
17:57:23preglowselect is joystick?
17:57:27rashersame deal with delete, rename and a couple of other things
17:57:39preglowi plied away the debug code, and it still works
17:57:45preglowa round of applause for yours truly
17:58:03*rasher applauds
17:58:16rashercvs commit: authorization failed: server rejected access to /cvsroot/rockbox for user rasher
17:58:28rasherkindly stop that!
17:58:40preglowwell, sorry, can't help you there
17:59:37rasherI'm probably not properly logged in
18:01:57rasherI don't get it...
18:02:13amiconnDid you do the checkout of your working copy while being logged in?
18:02:33rjamorimCVS is such a delight
18:02:41amiconnIf not, you need to do so. You can't turn an anonymous readonly working copy into a read-write working copy
18:03:09rashercvs commit: used empty password; try "cvs login" with a real password
18:03:13rasherI so didn't!
18:04:24rasheramiconn: actually you can.. just edit CVS/Root :)
18:04:26rasherbut yes, I did
18:05:11rasheror did I...
18:05:13*rasher goes insane
18:06:27rasherwell there we go
18:06:37rasherGuess I didn't.
18:07:30rashercvs commit: Using deprecated info format strings. Convert your scripts to use the new argument format and remove '1's from your info file format strings.
18:07:32rasherwhat's this then?
18:08:10preglowso alac actually compresses very badly?
18:08:42preglowi would have thought it to be better than flac, what with adaptive coefs and all
18:08:55rjamorimWell, not very badly. Matter of fact, at the same efficiency as FLAC
18:09:12 Quit Nibbler (Read error: 113 (No route to host))
18:09:16rjamorimBut it's worse than WavPack, Monkey's, WMA Lossless...
18:09:26preglowseems like you rate wavpack pretty highly :)
18:09:36rasherWMA ... ><
18:09:50preglowwould be cool if the wavpack guy showed up some day
18:10:19rjamorimpreglow: I know. I'm an enthusiast of this codec :)
18:10:28rjamorimI'm afraid he's not much into IRC...
18:10:41preglowdavid bryant mailed me and said he was very interested in helping
18:10:45preglowhe also said he'd never been on irc, heh
18:11:03rjamorimHe said he would now start working on ASM optimizations for M68K (RockBox) and TI DSP (Neuros)
18:11:08rasherwhat better timt to start than now
18:11:13rjamorimhehe, yeah. He doesn't use instant messaging either
18:11:26preglowohh, i'd like to do that :P
18:11:27rjamorimGod knows I try to convince him...
18:11:33 Join Nibbler [0] (
18:11:34preglowbut i like commmunicating too much
18:11:56preglowit is distracting, though
18:11:57preglowbut anyway
18:12:05preglowit's nice to have the codec dev himself interested in helping
18:12:12rjamorimI agree
18:12:31preglowespecially when he has embedded development experience :P
18:12:31rjamorimSad thing is, he is a quite slow coder :(
18:12:32amiconnrasher: The 'Using deprecated info format strings' is harmless. All committers get is. Someone needs to update the server config...
18:12:40rjamorimSo, it might take months until teh optimization is ready
18:13:00preglowrjamorim: well, yeah, but we're not really in any hurry, heh, things aren't exactly swift here at the moment
18:13:13rjamorimHehe, good
18:13:18preglowbut ok, i think i'll commit musepack support, then
18:13:36*rasher updates SoundCodecs
18:24:11preglowrasher: this is libmusepack 1.1 vanilla tar.gz you found on the site, hyes?
18:25:44rasherI only changed the #includes and directory structure
18:29:53 Quit DJ_Dooms_Day ()
18:32:43 Join Stryke` [0] (
18:36:23 Quit rjamorim ("On the turning away...")
18:37:38 Nick Lynx_ is now known as Lynx_awy (HydraIRC@
18:53:36rasherpreglow: no commit?
18:54:38 Nick Sucka`away is now known as Sucka (
18:58:28preglowtied up
19:08:11preglowi'm working on it
19:08:18preglowextremely boring, so i'm easy to distract
19:08:55preglowrasher: give me a quick rundown on what you removed, if anything
19:08:58preglowsample.cpp is all i can see
19:10:31rasherall the autoconf/make stuff
19:10:52rasherbut from include/ and src/ yes, only sample.cpp
19:25:41preglowany ideas on what i should do with these sprintfs?
19:33:20rashercomment them out?
19:33:25rasheroh wait
19:33:42rasherignore me.
19:34:58***Saving seen data "./dancer.seen"
19:35:02preglowi have commented them out
19:35:11preglowand that's the way i'll leave them for now
19:35:31preglowstreetfighter 3 and a pair of arcade controllers are calling for me from about 50 metres away
19:36:50 Quit Nibbler ("blubber")
19:38:22rasherso hurry up and finish :D
19:46:12 Quit preglow (Read error: 104 (Connection reset by peer))
19:48:48rasherI guess streetfighter won
19:49:37 Join preglow [0] (
20:00:49preglowi've gotta run for now, i'll commit musepack in a couple of hour, just so i'll be here in case i break any builds
20:01:53 Quit DangerousDan (Read error: 54 (Connection reset by peer))
20:05:09rashergood plan
20:05:35rasherit's not like I'm in a hurry or anything
20:05:37rasherjust bored
20:20:30 Join webguest43 [0] (
20:21:58webguest43Hi ! I know this is not a iriver support forum, but I seem to remember that there was a way to sort the folder structure on the player so it would play one folder after the other in alphabetical order, and not jump by order in which they have been added. Kinda sucks to listen to Bruckner and then jump into metallica...
20:25:01HCllet me guess, you have different capital letters for the first letter.
20:25:07rasherno idea
20:25:39*HCl wonders how one could do that anyways
20:25:49HClif i'd do that, i'd be listening to music from the same artist for hours :P
20:28:27rasherI tend to either listen to one album, then switch to another, or shuffle entirely
20:29:21webguest43ah, sorry I dont understand, I have Genre-Artist-Album-Song folder structure, it jumps from somewhere in the tree to a different genre when an album is finished. I think its doing them in order of when they where added, but am not sure.
20:29:40HClwebguest43: its most likely looking at the first character of your directory name
20:29:49HCland its most likely having different cases.
20:29:51HClsay you have
20:29:57webguest43could it be that its just going to the next entry in the fat ?
20:30:06HCl"Abra" "Balony" "Zoxar" and "balla"
20:30:18HClthen it plays Abra Balony Zoxar balla
20:30:28HClrather than Abra balla Balony Zoxar
20:30:32HClbecause of the different case.
20:30:41HCland thats as far as i'm gonna go with explaining, sorry
20:33:02rasherdammit, I wish I could make convbdf read this file
20:37:12 Join Tang [0] (
20:45:07 Quit webguest43 ("CGI:IRC (Ping timeout)")
21:10:06MoosCamaroHi Tang
21:10:17TangHello Seed and Moos
21:10:34Tangi'm upset
21:10:42Tanglost 2 version of a mail
21:10:49Tanghave to rewrite again
21:12:14Tangdont know if Linus is here
21:12:29Tangbut i want to thanks for the noobs edition
21:12:39Tangvery best organisation than mine indeed
21:13:22TangWhat about MPC moos
21:13:35Tangi hadn't read all the logs
21:14:04Tangjust saw it was in discussion since yesterday
21:14:32MoosCamaroanswer Rasher & Preglow
21:15:12MoosCamarothey had make good works
21:16:07TangOkay but yuo can't make a short resume for me?
21:17:48MoosCamaromy english's very poor
21:17:56MoosCamaroprivate for french version
21:18:33Tangoki i on
21:23:10rasherBasically, preglow is changing whatever's needed to make it actually compile
21:23:14rasherthat's done
21:23:27rasheralso a mpc2wav plugin
21:23:36rasherjust not committed yet
21:24:07rasherAll I did was clean up the libmusepack sources and make it fit into the rockbox build process
21:24:18TangAh thanks Rasher
21:24:24Tangvery sweet
21:24:28rasher"clean up" meaning that I removed files that wouldn't be needed
21:24:49rasherdidn't touch the source at all
21:25:45TangOkay so no real optimisation if i understand well
21:26:02rasherNo.. just the absolutely needed changes
21:26:27rasherbut it may be the next codec preglow starts optimizing
21:26:31rasherif I understood him right
21:26:40rasherbecause it's fairly simple
21:30:34TangSo very good news for many "audiophiles"
21:32:17 Join muz [0] (
21:32:44muzhey guys, i know you guys hate people reporting bugs but i think i have one with the tag database
21:33:40muzwhenever i try to rename an album the one it selects for renaming isnt the one i selected, if that makes any sense
21:34:08Tangbe careful Muz
21:34:17muzand also it wud be cool to be able to make playlists using the database
21:34:25TangRockbox is known to kill any bug reporters
21:34:35*muz runs for life
21:35:00***Saving seen data "./dancer.seen"
21:35:14muzbtw is it bad to delete the system volume information folder on my iriver
21:35:18 Nick Sucka is now known as Sucka`away (
21:36:21muzi got my iriver to play an mp3 which i converted to a wave then renamed to sample.wav the other day
21:36:27muzi was really proud
21:36:40rashernot at all.. I should figure out how to make windows stop creating that folder
21:36:59muzstupid windows
21:38:40rasherbut tell more about the rename thing
21:39:19muzbasically i hold the joystick in to get to the rename, open with, delete etc menu
21:39:29muzafter selecting an album
21:39:56rasherWorks for any folder I guess?
21:39:58muzi go to rename to hopefully rename that album tag field for all those songs in that album
21:40:07muzno this is in the tag database
21:40:34rasherVery important piece of information there :)
21:40:38muzbut the one which i selected isnt the one it renames
21:40:46muzyeah i guess cos the directories are virtual right?
21:41:53rasherI'm not quite sure how it's supposed to handle the rename function in tag-db mode
21:42:24rasherBut I see what you mean.. it shows a different value than what you opened
21:43:03rasherI think it's more likely that "rename" shouldn't be accessible at all
21:43:07muzwhen will genre/year be added just out of interest
21:43:22rasherbecause I don't think retagging is available
21:43:38muzit wud be cool to rename the tags
21:43:59rasherI agree..
21:44:01muzanyway isnt it easy to add playlist support to the database
21:45:29rasherno idea
21:45:54muzok cool, btw i saw u were doin some stuff on mpc hows that goin?
21:45:54rasherI don't see what the obstacle would be.. but then, I'm not much of a codeer
21:46:31rasherpreglow did nearly all of the actual work.. I think it's in a working state now and just waiting for him to commit
21:46:38rasherno optimizations yet though
21:46:47muzwhat speed are we looking at
21:46:58rasherpretty slow
21:47:06rasher30% at 120MHz
21:47:35muzyeah, but then again i dont think there are any mp3 players that can decode musepack
21:47:48rasherThere is now
21:48:01rasherif not fast enough for realtime playback (yet)
21:48:32muzumm i was wonderin, if rockbox can decode an mp3 to wav and then play that wav
21:48:50 Quit Tang ("Chatzilla 0.9.66 [Mozilla rv:1.7.5/20041108]")
21:48:55muzshudnt it be easy just to combine the two right now
21:49:10muzto get mp3 playback
21:49:48rasherI don't think anyone's going to bother doing that work.. because it'd be a bunch of work for a temporary gain that'll be useless when the api gets here
21:50:46muzur right, just out of interest does wav playback use up lots of cpu power?
21:51:15rasherIt uses practically nothing
21:51:31muzreally thats cool
21:51:56rasherI.. think.. it's more or less in the right format for the hardware to accept
21:52:36muzcool well thanks for fillin me in
21:53:35 Quit muz ("CGI:IRC")
21:54:26rasherlet's hope I didn't get any of that wrong :)
21:57:50*t0mas is away: tv
22:00:20 Join amiconn_ [0] (
22:01:28rasherevening amiconn_
22:01:34HCllinus was working on adjusting the code to add mp3 playback..
22:02:20rasherwell isn't this more general than what muz was asking?
22:02:53rasheroh well
22:04:09rasheramiconn_: tried looking at the viewer patch by pillo?
22:10:56HClwhat does it do?
22:14:49 Join asdsd_ [0] (
22:17:20rasheradd support for proportional fonts
22:17:29rasherand a new flow mode
22:17:32rasherit's very good
22:17:48rasherapart from a bug I discovered
22:18:13rashersupport as in previously proportional fonts were treated as fixed
22:18:24rasherwhen calculating how many characters would fit on a line
22:18:53 Quit amiconn (Read error: 110 (Connection timed out))
22:18:53 Nick amiconn_ is now known as amiconn (
22:34:59rashererp, digest exploded
22:36:34 Join LinusN [0] (
22:36:38Bagderme check
22:36:42rasherevening Linus
22:38:46elinenbe_LinusN: what
22:40:36elinenbe_LinusN: what's the truth to the statement that you are close to getting mp3s up and running on the iriver?
22:42:59LinusNwell, i'm working on the infrastructure to be able to decode and play mp3
22:43:23elinenbe_I have a nice shiny nickle for you if you get it working soon! ;)
22:43:24LinusNand there isn't much left to do before i can have it running
22:43:49elinenbe_oh awesome!
22:45:30elinenbe_LinusN: have you finalized the api, or are you just doing mp3s?
22:46:35LinusNi'll start with mp3
22:47:00LinusNthen the api will grow when i do the next format
22:47:29elinenbe_any sort of eta? or more like a when it's done it's done... ;-)
22:48:47LinusNas always
22:50:21rasherI'm currently wondering if my patch to broke things
22:52:33rasher something's all wrong...
22:55:30rasheroh darn
22:58:58 Quit Harpy (Read error: 60 (Operation timed out))
23:04:29CoCoLUSuh sounds bad
23:05:14preglowstreetfighter 2 rocks pretty badly
23:05:20preglowas do arcade controllers
23:06:10preglowLinusN: should i commit the giant, huge, ugly imdct_l thing i did while waiting for the more proper version to be done?
23:06:52 Nick Sucka`away is now known as Sucka (
23:09:23*rasher votes for yes
23:09:28rashercan't do any harm, surely
23:09:57CoCoLUSfamous last words.
23:09:57thegeekwhy not?
23:10:19thegeekit's not like it would break anything is it?, it's just not perfect?
23:11:21rashereven if, it can be reversed easily
23:13:09preglowthegeek: break anything, no, pain me everytime i see it, yes
23:13:09preglowbut yes
23:13:12preglowi'll put it in
23:13:15preglowit's better than nothing
23:13:22preglowand i'm the only one looking at it anyway
23:13:58preglowi'll see to putting in musepack now
23:19:12 Join edx [0] (
23:20:39LinusNpreglow: i think you should commit it
23:22:37preglowLinusN: yeah, i'll commit everything i can tomorrow
23:22:41preglowneed to split some files first
23:23:24thegeekI think I'm going to get the dell 2005fpw
23:23:32thegeekjust like the 2001fp only widescreen
23:24:25preglowi can't get this shit warning free, i'm afraid
23:24:37preglowso i'll just commit it and let people who now more about that kind of thing deal with it
23:25:05LinusNthat's the spirit
23:25:37thegeekyou guys with the 2001fp, is it good?
23:25:55preglowhmm, there was an info format string error here
23:26:10preglowbut i don't think libmusepack uses those :/
23:26:22rashermaybe that's the error?
23:27:00LinusNthegeek: i have the 2000fp, it's wonderful
23:27:18thegeekthe 25% coupon expired last night
23:27:49thegeeknow I'll either have to wait for another coupon to come along, or go with the 30% coupon for the 2001fp
23:29:55preglowok, here goes
23:30:03*preglow waits for stuff to break
23:30:44*CoCoLUS breaks stuff.
23:31:07 Join Strath [0] (
23:31:12*preglow makes a mental note to add more things with on cvs command line
23:31:39*t0mas sees commit :)
23:31:50t0masDirectory /cvsroot/rockbox/apps/codecs/libmusepack added to the repository
23:31:58t0masand wait's for the rest of it :P
23:32:15Bagderreminder: edit FILES as well
23:33:13preglowBagder: i have
23:33:24preglowlast thing i did :)
23:33:54*Bagder pads preglow on his shoulder. "good work my son"
23:34:12*rasher stares at
23:35:00SeedBagder: I don't think you're old enough to be preglow's father
23:35:01***Saving seen data "./dancer.seen"
23:35:22Bagderdarn, you got me
23:36:05preglowwell, it's not impossible, if i don't remember incorrectly
23:36:15preglowhe would have had to started out young :P
23:38:30*t0mas is going to sleep :)
23:38:36 Quit t0mas ("good night")
23:39:58 Quit Hadaka (Read error: 60 (Operation timed out))
23:42:28 Join muesli- [0] (
23:43:30muesli-any exciting news, filesize of increased by 20kbyte!
23:43:50preglowhah, nothing broken
23:44:23rasherbunch of new warnings
23:44:24muesli-just tell me, whats neeew!? :D
23:44:35muesli-i love warnings ;)
23:45:44preglowi knew of the warnings
23:46:57 Join Biptoria [0] (
23:47:26LinusNhmmm, maybe the configure script should enumerate the codecs instead of hardcoding the list of targets
23:47:34BiptoriaHello my freinds the time has come, to raise the roof and have some fun
23:47:41Biptoriayour kidding me
23:48:19Biptoriaforget everything i just said, thought IE had crashed
23:48:34BagderLinusN: the list is there like that to enable the gmini build to pick specific ones
23:48:55Bagderit used to be "yes"
23:49:22preglowBagder: why doesn't the daily build gen list any changes?
23:49:34preglowi've asked this before, but i can't remember things
23:49:34Bagderbecause it bugs somehow
23:49:44rasheryeah, the changes bit seems a bit .. dim
23:50:19BagderI don't really understand how I can fail, but it does ;-)
23:50:25Bagderhow it can
23:50:41rasherweird.. every artist lists the same albumname for every album
23:51:01rasher(the first one in my list)
23:51:15LinusNpreglow, Bagder: you mean the Changelog?
23:51:45Bagderyes, the change file the time stamp links to
23:51:47rasherso if I go into say, Metallica and I have 3 albums by them, I get 3 times "−−- blablbalb −−" (some bogus tag..)
23:51:52 Quit bobdbob (Read error: 110 (Connection timed out))
23:52:24LinusNBagder: i don't get it
23:52:30BiptoriaRasher: its probably broken
23:52:35preglowcan anyone remember libmad performance with no opts?
23:52:50BagderLinusN: the daily build table's first line has a time stamp, click the link
23:52:50LinusNoh, like 8% or sth
23:53:02preglowahh, ok
23:53:11preglowlibmusepack isn't exactly blazingly swift either
23:53:15rasherBiptoria: no kidding
23:53:29BiptoriaRasher: im not mate
23:53:40preglowoh, how i wish mac did not have that parallell load version
23:53:57LinusNBagder: interesting URL, with spaces and colons
23:54:34BagderLinusN: colons are OK, and the whitespace is actually %20 in the url
23:55:34 Join muesli_ [0] (
23:55:45 Quit lolo-laptop ("Client exiting")
23:57:55rasherthis is just dandy.. I can't for the life of me figure out why the tagdb is behaving like this
23:58:02BiptoriaSterlling job old chaps
23:58:21BiptoriaRasher: its broken
23:58:41Bagderrasher: did your patch introduce this?
23:58:56rasherI'm not even sure of that.. testing now

Previous day | Next day