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 2005-06-21

00:00:19linuxstbBagder: I04 at 3CF04000 (FLAC)
00:00:23preglowsame address regardless of codec
00:00:54preglowlinuxstb: strange..
00:01:30preglowman, rockbox kicks iriver in the ass with great efficiency
00:01:38preglowthis fading backlight was the icing on the cake
00:01:43linuxstbSame message for WAV files....
00:01:56Bagderyes, all codecs should have the same prob
00:02:08n0bbyi wouldnt be surprised if the battery life is anywhere near double at this rate
00:02:49ep0chpreglow: thanks for the non dithered codecmpa, am gonna stick with it :)
00:03:05ripnetukdoes the iriver have a driver for the radio yet?
00:03:07n0bbycould one of you email me that?
00:03:18preglowep0ch: no problem, i'm actually contemplating ripping out the dithering code
00:03:30ep0chnobby: what do you want?
00:03:39HClripnetuk: austriancoder worked on that...
00:03:41n0bbythe non dither codecmpa
00:03:49ep0chscroll up somewhere
00:04:16hubbelpreglow: yeah.. I had a problem with my 4.5h recording.. when I put the charger in and turned it on, the file had not been properly saved (old record-1.wav was still there), but after running scandisk windows found a 512 MB file (should have been > 2.5 GB)
00:04:38HClis the 2gb limit implemented?
00:04:39ep0chsee if you can tell a difference with sound, and check the cpu boost
00:04:46Bagderthere's a 2gb limit in fat
00:04:47HClcause fat can't have files bigger than 2gb
00:04:53hubbelmight be a fat-flush issue or how is the file-system limits handled?
00:05:06*n0bby is going to bed
00:05:10n0bbynight all
00:05:11 Quit n0bby ()
00:05:16preglowi assume you're not actually going to listen to the file contents to see if the recording glitches? :>
00:05:32Coldtoasthaha! the flickerless backlight fade is annoying
00:05:48Coldtoastcos I can't decide which length I like most
00:06:03preglowi use 500ms for both of them
00:06:06preglowi like subtle fade
00:06:21ep0cham off to bed too
00:06:38ep0chits just a coincidence me and n0bby are off at the same time i swear
00:06:49Coldtoastthe improved crossfading seems to be working 100% of the time too
00:06:59hubbelpreglow: I will eventually..
00:07:10 Quit ep0ch (" HydraIRC -> <- Leading Edge IRC")
00:07:23preglowhubbel: there are shitload of misticriver forum people waiting to have your babies if you fixed that
00:07:33Coldtoastwill crossfade length be configurable?
00:07:39hubbelpreglow: hehe
00:08:43preglowcan't wait until we have wavpack and speex recording
00:08:45preglowthat'll be friggin leet
00:08:55preglowi sure hope speex is fast enough
00:09:10Coldtoasthey. how about animated peak meters for playback?
00:09:32preglowColdtoast: it'll come
00:09:46Coldtoastlow on the list tho I guess?
00:10:07Coldtoastactualy, I rarely look at the main lcd anyway
00:11:19Coldtoastwhen the LCD remote displays track/etccings/etc info will be good
00:11:30Coldtoasterr... settings
00:11:48Coldtoastjust the file tree/track info will be enough for me
00:12:14ripnetukwould it be possible for now to just have a define to use the LCD display instead of the main display? should ensure we have no dependencies on certain screen size
00:12:24preglowColdtoast: well, not exactly low, it's pretty integral to the interface, but i'll wait until playback has gotten further
00:13:13preglowBagder: your patch destroyed ordinary plugins as well, it seems
00:13:33Bagderyes, I figure
00:13:38linuxstbYes, I've just noticed that as well.
00:13:39BagderI found some problems
00:13:54Bagderreading map files now
00:14:03Coldtoastyay! Bagder killed rockbox!
00:14:27preglowman, my reset button must be getting weary
00:14:55Coldtoastmy USB socket must be getting tired too
00:15:28preglowcan plugins loads fonts? :>
00:15:44preglowthe proportional font is wonderful for the text viewer, but i don't really like it for the rest of rockbox
00:16:02Bagderthere's a sysfont too
00:16:17 Quit hubbel ()
00:17:26linuxstbBagder: Not a big problem at the moment, but codec*.rock are getting installed in both the codecs/ and rocks/ directories.
00:17:27Coldtoastany chance at all crossfade will work when switching between playlists?
00:17:47Bagderlinuxstb: we'll fix the buildzip script to do right
00:17:48preglowColdtoast: it will always work between two tracks
00:18:30Coldtoastwell, I got to the end of a playlist, went to file tree, changed dirs, selected a track and it didn't fade
00:19:22Coldtoastgot a playlist going and just inserted a track from a different dir. see if it works
00:21:23preglowhow does seeking work on archoses?
00:21:37preglowdo you get realtime feedback?
00:21:58Bagderif you mean sound
00:22:21preglowi think i'd really like that for iriver rockbox, but i realize it wont be easy
00:22:36HCllike what?
00:22:44HClsound while seeking
00:22:45Coldtoasthmmmm. in my .wps, %pc will only update if I seek
00:22:51HClyea, that would be nice
00:23:48preglowBagder: same old error
00:23:55Bagdersame address too?
00:24:13 Quit muesli- (Read error: 148 (No route to host))
00:24:37preglowplugins work now, though
00:24:48Bagdera little gain at least
00:25:06Coldtoastsomething funky with .wps
00:25:43Coldtoast%pc only updates when I seek and %pt is COMPLETELY wrong; says this track is 35mins
00:26:12Bagderplayback.c fiddles with the address too
00:26:30linuxstbColdtoast: Which codec?
00:26:50*Bagder reads source code instead
00:26:52Coldtoastnow I can't seek properly
00:27:08Coldtoastthe progress bar moves a TEENY bit
00:27:23 Quit Aison ("( :: NoNameScript 3.72 :: )")
00:28:02Coldtoastwhen I release the stick, it keeps playing from the same oint I started seeking from
00:28:08BagderI need to rethink this approach...
00:29:02preglowColdtoast: sounds like a vbr file with no seektable
00:29:09preglowBagder: do you know how rockbox handles those? :>
00:29:45Bagderit is slightly different than I thought, I'll get some sleep and have a go at it tomorrow
00:29:53 Join hardeep [0] (
00:29:56Coldtoastnope. my .wps reports CBR
00:30:24preglowBagder: nitey
00:31:59Coldtoastok. it IS VBR but rockbox reports it as CBR
00:33:01preglowyes, i suspected that
00:33:11preglowit probably has no headers to indicate it is vbr
00:33:24preglowa vbr with no vbr headers is a pain to seek in
00:33:29preglowi don't know how rockbox handles them, actually
00:33:38Coldtoastwonder how that happened. I ripped the album with EAC/LAME
00:33:41linuxstbDoes the vbrfix plugin work on the iRiver? Will that do the job?
00:34:37Coldtoastpreglow: seemingly not well at all. Can't seek, It also broke crossfading
00:35:19preglowvbrfix inserts a xing header, yes
00:35:27preglowcan't imagine why it shouldn't work for iriver
00:35:53BagderVBR without proper VBR headers cause havok
00:36:06Bagdernot much to do about that
00:36:20Bagderunless we wanna scan the file
00:36:35Coldtoastit's causing all sort of weirdness actually
00:36:40preglowvbrfix for this is the correct approach
00:36:40HClhow about present an option to run vbrfix on it?
00:36:56HClif vbrfix is available
00:36:58Coldtoastone track ended, it crossfaded to the next, played about 5 secs then went to the next track
00:37:05Bagdervbrfix is always available
00:37:07preglowcodecmpa can't run another plugin directly
00:37:20Bagderbut asking things when playing music is not what I want
00:37:47Coldtoastis there a vbrfixer for Windows?
00:37:54ColdtoastI'll just process the files outside the player
00:37:55preglowvery probably
00:38:00BagderI believe perhaps the mp3 playback assumes too much knowledge
00:38:04Coldtoastthis will be the only album I have with the prob
00:38:05linuxstbpreglow: The metadata checking is done in playback.c (metadata.c), so in theory that could start vbrfix I guess.
00:38:22preglowlinuxstb: it would need to scan the mp3, sounds like a bit of a waste for each mp3
00:38:59linuxstbMaybe we could incorporate it into playback - a sort of "auto vbrfix".
00:39:04preglowthere's no way to find out whether an mp3 is vbr or not, apart from scanning the headers until you find a different bitrate in one
00:39:19linuxstbBut codecmpa does that whilst playing it.
00:39:19ColdtoastI must have eithe rnot enabled VBR headers in LAME or the build I used was slightly fubar
00:39:35Coldtoasthere's teh problem
00:39:36preglowColdtoast: lame should produce files with xing headers
00:39:51linuxstbThis is probably a silly question, but does the VBR header have to be at the start of the file?
00:39:54Coldtoast"i. A current limitation of LAME is that it does not write VBR headers
00:39:57 Quit Strath (Read error: 110 (Connection timed out))
00:40:06ColdtoastDec 2003
00:40:17ColdtoastI prolly did the album in 2003
00:40:27 Quit preglow ("CGI:IRC")
00:40:37 Join preglow [0] (
00:40:43 Join xen` [0] (
00:40:53 Quit xen` (Client Quit)
00:41:24preglowlinuxstb: is vorbis seeking hard?
00:42:35linuxstbpreglow: I don't know. The API has sample-accurate seeking functions, but I don't know how much work will go on behind the scenes.
00:42:39 Quit ripnetuk (Read error: 110 (Connection timed out))
00:43:13linuxstbOgg files should have a seektable in them, so it shouldn't be that bad.
00:43:40preglowmost sane formats have seektables
00:43:46preglowmp3 and musepack being the notable exceptions
00:43:59pregloweven mp3 has seektables mostly
00:44:00 Join TCK [0] (
00:44:23preglowmusepack not having a seektable has to be one of the greatest flaws i have seen in a while
00:44:28Rorianyone working on 22khz playback? :)
00:44:35preglownot actively, no
00:44:53Roribummer. at least I can use the iRiver FW for audiobooks
00:45:11Roribut no bookmarks in iRiver :P
00:45:21preglowand seeking is sooo fun
00:45:32Roriseeking is broken at presetn still is it not?
00:45:57preglowman, volume changes click like mad now
00:45:57Roriwell it will seek just not accurately :)
00:46:27RoriI have my volume set at 90 for the car
00:46:35Rori100 and it distorts ;)
00:46:52Coldtoastcool. fixed
00:48:15preglowmp3 is pretty much gapless
00:48:21Rorion lame
00:48:51Rorinon-lame glitches like a bitch. trying to re-download some rips in lossless
00:48:52preglowthat's the only kind of gapless i care about
00:49:00linuxstbpreglow: I don't think Ogg does in fact have a seektable - I misread the "seekable" variable. Looking at the source, it seems to do a bisecting search in the file...
00:49:27Rorichop chop search
00:49:32preglowlinuxstb: weee, oh well, as long as the frames are marked with some kind of sequence number, it shouldn't be too bad
00:49:50linuxstbBut it means random access to the file - something we don't want to give it.
00:50:17preglowwe don't have much of a choice
00:50:18linuxstbUnless for seeking purposes, we give the codes access to the raw file descriptor.
00:50:52preglowno, i think we should restrict them to the buffer
00:51:13preglowif we detect a codec is searching forward rapidly, we should start buffer the file further
00:51:22linuxstbBut the problem is that playback.c has to try and predict what data the codec wants.
00:51:38preglowindeed, in our case it will also have clear hints
00:51:45preglowwe can't do true random access seking
00:52:10preglowif the user keeps the button pressed for long and we're approaching end of buffer -> buffer more
00:52:45preglowhaving the seek block because we're busy buffering the part we're supposed to seek to isn't a bad thing
00:53:36Coldtoast%pc still doesn't work tho. not with any MP3s at all
00:57:19Coldtoastaaah ignore me
00:57:31Coldtoastsheesh! found why it wasn't working
00:57:49Coldtoasthad %s on the same line
00:57:52 Quit Sucka ("a bird in the bush is worth two in your house")
00:58:01HClwhy does that matter?
00:58:04preglowLinus Torvals is behind rockbox...are you sure??? The same guy behind linux kernel. It can't be...can it???
00:58:09preglowahh, misticriver...
00:58:12HClyea, the idiots.
00:58:21HClpebs is the only guy in that threat with a brain
00:58:28Coldtoastthe manual CLEARLY states:
00:58:29ColdtoastScrolling lines can not contain dynamic content such as timers,
00:58:29Coldtoastpeak meters or progress bars
00:58:43HCli haven't read the manual
00:58:53ColdtoastI'd read it just not READ it
00:59:21preglowMy understanding is that the iRiver CPU is a floating point only
01:00:40Coldtoastyou guys are tireless
01:01:14Coldtoastcrossfading seems to work perfectly now, backlight works perfectly
01:02:13HCland still.
01:02:20HCli haven't tried a recent build
01:02:30HClsince the xxx2wav builds
01:02:32HCl :P
01:02:46Coldtoastjust out of curiosity, why did you decide to limit the volume as bass increases rather than do what iriver did and drop he bass as volume increases to avoid clipping?
01:02:52HCli still haven't tried playback on iriver o.o..
01:03:01preglowhcl: why not?
01:03:05HCli dunno.
01:03:10HClbeen busy, i guess.
01:03:14ColdtoastI actually find your choice alters the sound a lot more than what iriver did
01:03:15HCli wrote runtime db support code today
01:03:21HCli need to get home and test it
01:03:26HCldon't have my iriver here
01:03:27preglowcoldtoast: well, we figure if you ask for more bass, you should have it
01:04:18Coldtoastmaybe have a setting where you can choose "Bass limiting" or "Volume limiting"?
01:04:19preglowi dont much care, i never use the eq
01:04:53preglowgot a sudden urge to code the eq
01:05:01ColdtoastYEAH! heh
01:05:07Coldtoastmake an old man very happy
01:05:17preglowbut i'd need to boot windows to find the eq coefs :/
01:05:40Coldtoasterrr... now don;t go doing anything rash now
01:05:50preglowbut no
01:05:54preglowi can't start that now
01:06:09Coldtoastyeah. seems like a lot for one day already
01:06:26preglowsomething'll go wrong, then i'll sit up until 6am to try to solve it, then i'll feel incredibly stupid, sleep two hours and be tired
01:06:28Coldtoastwhen the EQ is done should satisfy my needs I suppose
01:06:37preglowdepends on what your needs are ;)
01:06:46preglowi'm thinking of a 5 band affair
01:06:57Coldtoastwell, to be honest, I think I prefer bass limiting, if need be
01:07:15Coldtoastoh. for EQ? hmmm
01:07:22preglowyeah, software eq
01:07:37Zoom2wait. Linus Torvals is not the same Linus...
01:07:42Zoom2or was that sarcasam
01:07:43preglowzoom2: no shit
01:07:52Coldtoastmaybe just something like Winamp's EQ but with just 5 bands
01:07:59preglowzoom2: i was quoting someone on misticriver
01:08:09Coldtoastso 5 bands of EQ and a BOOST?
01:08:12preglowColdtoast: yes, that's what i'm going for
01:08:17Zoom2lol ok
01:08:56Coldtoastmaybe 7 bands? there's no difference between coding a 5 and 7 band one
01:09:07preglowi think five bands will be max
01:09:22Coldtoasthow about just 3?
01:09:26preglowi could of course just let you add however many bands you'd like
01:09:28ColdtoastBass, Mid, High
01:09:36HCli prefer that last option.
01:09:45HClit can save battery
01:09:48Coldtoast3 should REALLY be enough
01:09:50HClfor people who need less bands
01:09:52preglow3 bands or selectable?
01:10:09preglowyes, it'll just be gui deal
01:10:17preglowinternally it'll be much the same
01:10:24HCli'm gonna sleep
01:10:25Coldtoastselectable between how many bands are you thinking?
01:10:39preglowi don't know how adjustable the bands'll be yet either
01:10:50ColdtoastI mean how many choices would you have?
01:11:11preglowtheoretically, you can select filter type, center frequency, Q/bandwidth, and amplification/cut
01:11:19preglowusually, all but the last is hard coded
01:11:45ColdtoastWinamp uses -12db to +12db in 0.2db steps
01:11:51Coldtoastthat'd be fine, yeah?
01:12:12preglowyeah, sure, but that's just the last bit
01:12:18preglowwill people want to adjust the other parameters?
01:12:35Coldtoastadjust the frequence of each band?
01:13:16Coldtoastprolly need a bit of a surve on that one
01:13:26ColdtoastI wouldn't really myself
01:13:37preglowi wouldn't use the eq at all, so i don't really care ,)
01:13:42preglowi need to go to bed
01:13:45preglowsee you around
01:13:53 Quit preglow ("CGI:IRC")
01:13:53Coldtoastit's 9am here :)
01:14:00Coldtoastand I need to go to bed also
01:14:43 Quit Coldtoast ("Peace and Protection 4.22")
01:20:27 Quit crashd (Network is unreachable)
01:22:21 Join crashd [0] (
01:23:10[IDC]Dragonamiconn: do you know what linker sections *(.rodata.str1.1) and *(.rodata.str1.4) are for?
01:26:36BenacoolPretty calm there ... XD
01:27:36BenacoolNobody there ? >_> ... <_< ... -_-
01:41:47***Saving seen data "./dancer.seen"
01:46:09bill2or3I just got a new H340, any reason I shouldnt put the Korean firmware on?
01:47:06 Join muesli- [0] (
01:51:59RoriI prefer no EQ. If the sound was not right at the recording studio then the track sucks anyhow :)
01:53:03 Quit tvelocity ("Leaving")
01:53:41*Rori tries to find his favorite albums in lossless format to convert to Ogg
01:54:12Rorior rather any albums that were not ripped with gapless Lame that is
01:54:42 Quit Stryke` (Read error: 60 (Operation timed out))
01:54:42 Quit kenshin (Read error: 60 (Operation timed out))
01:54:42 Quit [IDC]Dragon (Read error: 60 (Operation timed out))
01:54:51 Quit Moos ("Bye ALL")
02:01:04 Quit muesli- ("ich will Khe!!!")
02:03:22 Quit DMJC-L (Connection timed out)
02:08:05 Quit DMJC (Connection timed out)
02:10:25BenacoolHi Bill2or3
02:10:37BenacoolGoodnight t0mas ^^
02:11:10BenacoolYou are always there bill ?
02:11:47Benacooloh Hello ^^
02:12:18Benacoolyou was talking of upgrading your H340 to Korean firmware ?
02:18:43bill2or3yeah, thinking about it
02:45:27Rorilol Tom Cruise got 0wn3d :)
02:46:27Zoom2not really, he told the guy off
02:47:27RoriI want to see that vid clip
03:13:42NibblerThe requested URL /b/src/1119310086608.jpg was not found on this server.
03:14:03 Join kenshin [0] (
03:22:03Rorianything posted on that section soon gets flooded off the site. can be pretty nasty stuff on there so avoid it if you don't wish to be offended.
03:41:48***Saving seen data "./dancer.seen"
03:42:25 Part N-Mi ("Ok, ok, je sors")
03:50:54 Join amiconn_ [0] (
03:51:29 Quit Andrew__ ("Leaving")
03:51:32 Join Andrew__ [0] (
03:58:27 Join Blair_42 [0] (
03:58:34 Quit kenshin (Read error: 110 (Connection timed out))
03:59:10Blair_42evening all. I have a bookmarking question
03:59:27 Quit amiconn (Read error: 60 (Operation timed out))
03:59:27 Nick amiconn_ is now known as amiconn (
04:00:14Blair_42it was working earlier. I can create bookmarks but when I attempt to go to one I just wind up back in the file browser
04:05:43 Join QT_ [0] (as@area51.users.madwifi)
04:10:01 Quit hardeep ("BitchX: it makes bathtime lots of fun!")
04:10:44 Join ashridah [0] (
04:17:25 Quit QT (Read error: 113 (No route to host))
04:29:14 Quit cYmen (Read error: 113 (No route to host))
04:35:48 Join Strath [0] (
04:40:00 Quit Zoom2 ("CGI:IRC (EOF)")
04:56:14Blair_42my bookmarks are working on my JBFM but not on my ondio
05:04:54 Join SPF19 [0] (~SPF19@
05:07:07 Join kenshin [0] (
05:08:19 Part Blair_42
05:13:18 Quit SPF19 ("sorry :(")
05:13:27 Join Zoom2 [0] (
05:20:01 Quit Rori ()
05:23:54 Join dropandho [0] (
05:23:58dropandhohey all!
05:24:09dropandhofigured i would check in
05:24:21dropandhoseems like there have been soooo many changes to the code lately
05:24:35dropandhojust wondering what has triggered all of this action
05:26:39Zoom2breakthroughs, teamwork, hardwork, alot of nice people who help out, etc.
05:27:05Benacooland soon a H3x0 firmware !
05:27:14dropandhohehe...woop woop!
05:27:22dropandhoi was just wondering tho...i am sorta confused
05:27:31dropandhomp3 actually works on it?
05:27:42Benacoolon the H1x0
05:27:49Benacoolnot on H3x0
05:28:11dropandhoso it is pretty close to public?
05:28:26Benacoolbut they are a always a lot of bugs ;-)
05:28:54dropandhoi noticed the wiki hasnt seen any changes....i was just curious what the deal was
05:28:56dropandhogreat news
05:29:00BenacoolI don't have a H1x0 so I can't really help you and I,m not from the team of rockbox so it'd all I know lol
05:29:06dropandhocongrats guys
05:29:19dropandhono worries
05:29:35BenacoolI have an H320 so I wait for the H3x0 port
05:29:47dropandhogot it
05:30:07dropandhowonder what is left to do
05:30:22Benacoolon wich ?
05:32:08Benacoolbug fixes and converting plug-ins taht are not already converted ... they are also work to do on codecs
05:32:28dropandhok k
05:32:31dropandhothis is exciting
05:32:35dropandhodidnt think it would be this close
05:32:42dropandhoand wps stuff?
05:32:44Benacoolnot sure for convertion but they are to be fix to run on the H1x0
05:32:48Benacoolwps ?
05:32:56kenshinand there's a few of us too new to do codes stuff that are trying our hands at config/wps/easier stuff.
05:32:57dropandhoand it is true that bat life is doubled?
05:33:05Benacooldon't know ^^'
05:33:09ashridahdropandho: uh, no, who told you that?
05:33:19Benacoolwhat ?
05:33:24dropandho<n0bby> i wouldnt be surprised if the battery life is anywhere near double at this rate
05:33:48kenshinbattery life is roughly equal
05:34:07dropandhok k
05:34:09ashridahwith what codec? because i'm pretty sure that people have benchmarked them to come out roughly the same at this point
05:34:16dropandhoworks for me!
05:34:31dropandhoi wish i could help
05:34:53Benacoolme too ^^
05:35:39kenshinanybody know where the logf() output goes? since i can't get the simulator working 100% i want to run it in debug on my h120.
05:35:42dropandhodo you guys know if a wiki is being updated to keep everyone in the loop where u are at?
05:37:02kenshindropandho: just keep track by watching the daily build changes
05:37:34dropandhogot it!
05:37:45dropandhojust hard to translate into english sometimes...hehe
05:37:50dropandhoi am a lameo!
05:37:55dropandhowell thanks guys
05:38:02Benacooli'm franch ;-)
05:38:31Benacoolin fact I talk french but i'm not form France ...
05:38:37Benacooljust to say it lol
05:39:35dropandhotake care and good luck!
05:39:48 Quit dropandho ()
05:41:49***Saving seen data "./dancer.seen"
06:06:50 Quit joshn_454 ("Leaving")
06:19:11 Quit Zoom2 ("CGI:IRC (EOF)")
06:22:26 Quit RotAtoR ()
06:23:00 Quit kenshin (Read error: 110 (Connection timed out))
07:00:39 Quit thegeek (Read error: 131 (Connection reset by peer))
07:22:36 Join B4gder [0] (
07:23:34 Join Pitman [0] (
07:41:53***Saving seen data "./dancer.seen"
07:46:54 Part Pitman
07:48:52 Join kenshin [0] (
08:12:46 Join Coldtoast [0] (
08:29:28 Quit Andrew__ ("Leaving")
08:29:48 Quit kenshin (Read error: 110 (Connection timed out))
08:45:18 Join Andrew179 [0] (
08:47:54Andrew179Is there anyone online who can reset my wiki password? I emailed Bjrn Stenberg, but haven't received a response yet.
09:03:40B4gderwhat's your wiki name?
09:05:49amiconnGood morning
09:06:20Bgermorning all
09:06:56Andrew179B4gder: my wiki name is AndrewHutchings
09:07:01Andrew179good morning
09:10:46B4gderI'll try to figure out how to reset it
09:11:14Andrew179thanks a lot
09:12:40 Join Chamois [0] (
09:20:44 Join bobTHC [0] (
09:22:07B4gderfill in and show me the crypt chunk, possibly in /msg
09:24:52 Join DaKi][er [0] (
09:25:12bobTHChi folks !
09:28:30 Join [1]Chamois [0] (
09:28:30 Quit Chamois (Read error: 131 (Connection reset by peer))
09:28:33 Nick [1]Chamois is now known as Chamois (
09:35:08 Join Coldt0ast [0] (
09:35:09 Quit Coldtoast (Read error: 104 (Connection reset by peer))
09:35:13 Nick Coldt0ast is now known as Coldtoast (
09:39:13 Nick QT_ is now known as QT (as@area51.users.madwifi)
09:41:14 Quit bobTHC ("Smoke Weed Every Day !")
09:41:54***Saving seen data "./dancer.seen"
09:47:06 Join Harpy [0] (
09:48:05 Quit DaKi][er (Read error: 145 (Connection timed out))
09:48:55 Quit ashridah ("Leaving")
09:52:05Coldtoastwhat's the native rockbox font?
09:54:11 Join DaKi][er [0] (
09:54:58 Join courtc_ [0] (
09:55:31 Quit courtc (Read error: 110 (Connection timed out))
10:00:16 Join kenshin [0] (
10:06:55 Quit linuxstb ("Leaving")
10:23:53 Quit B4gder ("Lämnar")
10:26:32 Join cYmen [0] (
10:30:45Coldtoastjesus CHRIST!
10:30:51Coldtoastgapless AND crossfading works perfectly
10:31:03 Join ashridah [0] (
10:32:25 Quit kenshin (Read error: 145 (Connection timed out))
10:32:41BgerAndrew179 ?
10:32:45Bgerjust a notice
10:33:02Bgerthere is a "hide attachment in normal topic view" checkbox
10:33:18Bgerand you won't get the table on the bottom of the page
10:33:26Andrew179oh, I was wondering about that....thanks
10:34:09Bgeru can hide them now too
10:34:21ColdtoastI only just now checked out gapless playback
10:34:24Bgeruse the link "action"
10:34:26ColdtoastI'm almost shocked!
10:34:38BgerColdtoast: why ?
10:34:47amiconnGapless isn't working perfectly yet, at least for mp3
10:34:59Coldtoastwell, I didn't think it'd really be anything to be honest. ppl always going on about it
10:35:18Coldtoastreally? seems to be here
10:35:30amiconnIt doesn't introduce gaps, but instead it sometimes chops off several frames
10:35:45BgerColdtoast: there are some missing frames at the end of some mp3 files
10:36:10amiconn...and that doesn't sound too good with dj mixes
10:36:15Coldtoastthe transition between tracks I have is pretty much all crowd noise
10:36:57Coldtoastmyself, I'll use crossfading. But gapless IS nice (even the way it is at the moment, I suppose)
10:37:52ColdtoastI just need EQ and LCD remote display and I have everything I want. heh
10:38:15Bgeri don't get it
10:38:34Bgerhow can i put some wiki page in Main/WebHome
10:38:34ColdtoastI think increasing bass is limiting the volume too much. but I imagine that's temp til the EQ is done
10:40:06 Join B4gder [0] (
10:40:53amiconnColdtoast: increasing bass will always limit the volume, otherwise you'll get distortions
10:41:13amiconnThis will happen always, even when eq is done in software
10:44:35Harpywould wider characters make fast scrolling text more readable?
10:44:49amiconnThe iriver firmware is cheating: when you have the bass boosted and increase volume too much, it will ditch the bass boost instead pumping up the volume changes the sound, unlike rockbox (which will simply not increase the volume when it's impossible due to too much bass or treble boost)
10:47:11ashridahColdtoast: yeah, there's a physical limit on how far you can drive the output before the output stage starts to clip.
10:53:53amiconn...apart from that I wonder whether people want to blow their ears away
10:54:22amiconnWhen I set rockbox to 100% volume, I can use my earphones as loudspeakers
11:01:26Bgeramiconn it depends on your phones ...
11:01:44Bgerand on the level of the current song also ...
11:03:08 Nick Lynx_awy is now known as Lynx_ (
11:03:30amiconnPerhaps... but I don't have any earphones that behave differently
11:04:00amiconnOnly if I connect my monitoring headphones (which have 600 Ohms) I need to increase the volume significantly
11:08:18Bgeryeah, if they aren't enough sensitive ...
11:08:50Bgerbtw, the neuros Open Source firmware is not open source atm, is it ?
11:09:58BgerAndrew179: what does "kaser" mean ?
11:10:23B4gderBger: nope, it isn't
11:10:32Andrew179Bger: I was trying to make out what is printed on the chip
11:11:00Andrew179But I can't quite do it
11:12:07BgerB4gder: i asked because i mentioned it in the NonArchos page
11:12:14B4gderI noticed
11:13:00B4gderhas been gone since jan 18
11:13:48B4gderand now they're discontinuing the player
11:14:53Bgergentoo's dep system
11:15:07Bgeremerge system - error "exec perl" ... not found
11:15:22Bgeremerge perl - > emerge autoconf -> perl is not found
11:17:41Bgerthat's becouse of my huge USE ...
11:17:44Bgerbacause ...
11:32:51BgerAndrew179: i see something like K*SER 1638D
11:32:57Bger* == unknown
11:33:28Bgermaybe 16S8D
11:34:36Andrew179I'm trying to make it out in Photoshop, and it looks more like K*SE8 1688D
11:36:05Andrew179or maybe 1638D
11:36:18Bgeri don't think it's 3
11:36:32Bgeri'm trying with GIMP ;)
11:37:20Bgercan't you see it on your own player ?
11:38:32Andrew179it's my brother that has one, 300 miles away :)
11:38:47Andrew179I could be acquiring one soon
11:39:20 Join Aison [0] (
11:41:38Bgerwhere is situated this chip
11:41:58***Saving seen data "./dancer.seen"
11:42:11Andrew179I'm using pictures from
11:42:30Andrew179this one specifically:
11:45:59Bgerthe chip is K4S281633D
11:46:07ColdtoastI have a request
11:46:32Bgertop : SEC 349 RN75
11:46:43Coldtoastcan you make all times have 2 digit numbers?
11:46:57Coldtoastso instead of a track being 2:54, have 02:54
11:47:10Bgerbottom Z0V860B2
11:47:11Andrew179Bger: great, I'll put that in the wiki
11:47:31BgerAndrew179 did you grab the file i mentioned yesterday ?
11:47:51Andrew179no, which file was that?
11:49:12Andrew179Bger: I'm getting a 404
11:49:22 Join webguest84 [0] (
11:49:54Bgerhm, never mind
11:49:56Bgeri have it
11:50:02 Quit webguest84 (Client Quit)
11:50:04Bgeri'll put it in the wiki on my own
11:52:42Andrew179Bger: just for reference, the file moved to
11:53:39Bgerhmm, possible
11:53:41Bgernever mind
11:53:46Bgeri have a picture of it
11:53:54Bgervery clear
11:54:14Bgerbut you've locked the topic ...
11:54:20Bgerare you editing it atm ?
11:54:51Andrew179nope, is there some way I can unlock it?
11:56:18Bgeryou can edit it and check "release edit lock" in the bottom of the preview page
11:56:27Bgeror we can wait
11:56:42Bgeras much as 30 min since last edit
11:57:22Andrew179ok, should be open now
11:58:01 Quit DaKi][er (Read error: 110 (Connection timed out))
12:00:21 Join kenshin [0] (
12:01:05Bgerhaha iAudio M3 has second "CPU" (if i can call it cpu) - PIC12F657
12:02:09Bgeri wonder what are they using it for
12:03:33 Quit Andrew179 (Read error: 54 (Connection reset by peer))
12:03:37Bgerthis is RAM ?
12:03:53Bger Description = K4S281633D 2M X 16Bit X 4 Banks Sdram in 54CSP ;; Organization = 8
12:04:15HClso yes.
12:04:34 Join Andrew179 [0] (
12:04:44Andrew179wireless conked out on me
12:04:56HClBger: 16mb ram
12:05:00HClif i got that right
12:05:35Bgeror 2?
12:05:56HCl2mbits * 16 bits * 4 banks / 8 bits to make it bytes... i think..
12:06:01Bger4M x 4
12:06:21Bgerso 16 :)
12:06:30HCldon't confuse mbit and mbyte though
12:06:35HClyour 4m is mbyte
12:06:38HClwhere the 2m up there is mbit
12:07:06Bgeri just can't make calculations .... :)
12:07:10HClsame :p
12:07:12HCli suck at math
12:08:35Bgerwow, Andrew179 is quick :)
12:09:12Andrew179just eating breakfast and editing the wiki :)
12:09:49Bgerbreakfast ? you're not from europe for sure :P
12:10:18Bgerhm... GMT 11:10
12:10:30Bgeror 10:10
12:10:37Andrew179I'm from the US, Georgia in particular
12:11:54Bgernow give me your home address, tell me when you're not at home and where do you leave the keys ... ;)
12:12:40Andrew179hah, that's all you're getting from me
12:13:32Bgerr u sure :P
12:14:11HCl ;p
12:14:23Bgeranyway, i suggest u to see the other chips
12:15:36Andrew179what do you mean?
12:17:29Bgerwith what ?
12:18:24Andrew179about the other chips
12:18:52Bgerok, i see one next to the memory
12:19:04Bgerworking on it atm
12:20:36Andrew179the TI chip, LD245A?
12:20:37 Join hicks [0] (~hicks@e0ca722efa4f38d1.session.tor)
12:21:17Andrew179that one's in the wiki already
12:21:18Bgerooo, no, you've added it
12:21:29Bgeri was looking at the top number
12:21:37Bgerwhich is not the same
12:21:49Andrew179I think I'll update the pictures with the higher-res ones you linked to
12:22:53Bgerwhat about U17
12:23:01Andrew179found the codec, it's on the daughterboard
12:23:11BgerU18, U19
12:24:15Andrew179haven't done those yet
12:24:24Bgerthese are something related to power supply, i think
12:27:33Bgerhm, this unit doesn't have main display?
12:28:34Andrew179nope, only the remote LCD
12:30:10amiconnI wouldn't want a unit without a display...
12:30:23Bgeramiconn: it has:)
12:30:30Bgerit's just on the remote
12:30:30Coldtoaston the remote
12:30:41amiconnYep, and I'm not much of a remote user
12:30:57amiconnI don't like 'em, just another box hanging around...
12:31:11Coldtoastever been a report of the remote causing noise on the M3?
12:31:16Coldtoastliek the iriver
12:31:45Bgeramiconn: what do you think about the chance of the "metal" chip being DSP ?
12:31:49Coldtoastmine doesn't do it but some peopl have had it
12:31:53Andrew179Coldtoast: I've never noticed that
12:33:35amiconnBger: I have no idea, but from the metal casing I'd think it's a chip with some more power dissipation
12:33:45amiconn..i.e. not a codec or such
12:33:59Bgerbut dsp ?
12:34:07amiconnIs there a scan of the whole board somewhere?
12:34:15Bgernot in the wiki
12:34:25Bger <= here
12:34:30Bgerbut it's 17MB
12:34:50Bgererror ... ~29MB
12:35:13amiconnWhy should there be a dsp?
12:35:26amiconnI'd think the M3 decodes in software just like the iriver
12:35:28Coldtoastamiconn: will you move some of the sound settings to the EQ? maybe even sterso width?
12:35:34amiconnThe cpu is capable of doing that
12:35:50Bgeramiconn: yes, it do, but there were some rumors about dsp chip
12:36:10Bgerit does...
12:36:19HCljust like how there were rumors that the uda had mp3 decoding ability >.>
12:36:23HClon misticriver, ofcourse >.>
12:37:12Coldtoastis there a touch of anti-misticriver in here?
12:37:29B4gdernot anti
12:37:32Coldtoastmisticriver looks HORRIBLE on Firefox
12:37:41B4gdernot for me
12:37:56B4gderthen I've never tried it with anything else
12:38:17amiconnColdtoast: I'm the wrong person to ask about eq; I'm not much into sophisticated audio processing
12:38:24B4gderwe're not anti misticriver, but we're aware of the amount of weirdos there
12:38:34HCl :P
12:38:36amiconnStereo width is currently non-functional, as is the channel selection
12:38:44amiconn(In fact they're closely related)
12:38:50Coldtoastyeah. I saw that
12:39:05Coldtoasttell you what I hate about misticriver
12:39:09HClColdtoast: ask preglow about eq
12:39:17Coldtoastthere really is WAY too much anti-ipod there
12:39:36B4gderI wouldn't know
12:39:43HCli'm not very anti ipod, i'm just not very fond of the drm it tends to promote a bit
12:39:49Coldtoastyou can't mention the ipod without ppl jumping up and down about how much it sux
12:39:57Bgeramiconn: <= see "Enjoy the powerful sound effects"
12:40:00HCland how its a closed platform
12:40:04HClthough ipodlinux do a great job
12:40:12HClthough they stil have a lonnnnng way to go...
12:40:27HClcow on america?
12:40:44Bgerhaha :)
12:40:45*HCl goes to eat lunch :)
12:41:42ColdtoastI hate threads like this, for example:
12:42:11amiconnHmm. From a first glance I can't say I like the cowon player designs...
12:42:18Coldtoastsomebody mentions ipodlinux have a firmware with vid support but no sound. So you then get 'tards saying crap like "Show them your h320"
12:42:31Bgerin fact i think that these effects can be done in software ... maybe this is just a headphone amplifier or ...
12:42:46Coldtoastyeah. nothing like embarrassing an ipod owner with video at 10fps, eh?
12:44:34Coldtoastoh! btw. I wanted to ask
12:44:47Coldtoastthe volume in RB. is that linear?
12:45:12Coldtoastalmost seems like it "decelerates" as it gets closer to max
12:45:33ashridahColdtoast: that'd be because it does, since it avoids clipping by just not increasing the volume anymore
12:47:01ColdtoastI dunno. I ALMOST tend to think the set clipping well on the conservative side. But anyway
12:47:20Bger <= haha does this girl come together with a bought M3
12:47:43HClColdtoast: they're just dumb...
12:47:52HCli don't understand who would want to watch video on an mp3 player anyways
12:48:00HCli have a 320x240 pda screen that runs at full fps
12:48:07HCli've watched the matrix on it in the train and stuff
12:48:12Coldtoastyeah. nothing like a 2.5" screen or whatever it is
12:48:17HClan mp3 screen would be way too small
12:48:26BgerHCl: it's nice to have, but just an add-on
12:48:26Coldtoastyeah. I have a 640x480 PDA for vids myself
12:48:45HClBger: oh come on. like you can watch anything comfortably on such a tiny screen with only 10fps
12:49:21Coldtoastthere's a MASSIVE difference between a 320x240 PDA and VGA one. I didn't realise just HOW big the difference is til I upgraded
12:49:38BgerHCl: in fact, i used this feature many times... not for watching movie, but showing some funny clip to a friend of mine
12:49:42HCli wouldn't really know, i don't mind my pda o.o;
12:50:04HClBger: just as useless a feature as those new mobile phones.
12:50:08HClmovies on your phone, please.
12:50:14HCllike thats useful.
12:50:16BgerHCl: it's better
12:50:22HClits pointless.
12:50:24HClthats what it is
12:50:31HCl :/
12:50:45Bgeras i said, it's a nice to have for free
12:50:57HClbut its not particularly useful.
12:51:05HClits just another extra
12:51:31Bgeri haven't bought my h340 to watch movies on it
12:51:36Bgerfor sure
12:51:56ColdtoastI think one of my friends is going to sell his h340
12:51:56amiconnColdtoast: Rockbox volume mapping is dB-linear, but it stops increasing at a certain value to avoid clipping if bass or treble are boosted
12:51:59Bgerbut it's nice to have this extra sometimes
12:52:00HClthe only point i could see in it
12:52:02HClis music clips
12:52:10Coldtoastamiconn: ok
12:52:15amiconnIf you have bass and treble set to flat, it is linear all the way up
12:52:22HCldoes it rescale videos and support xvid/divx?
12:52:33HCland even then. music clips would be a waste of battery o.o.
12:52:36Bgerh340 ?
12:52:57Bgerno, it can't rescale ... do you think it can with this coldfire @ 120MHz ?
12:53:09HCleven my pda doesn't manage to scale.
12:53:10Bger220x176x10fps mpeg4
12:53:13HClthats far too intensive
12:53:33HCli should save up for a slc3000
12:53:51HCli have the money for it O_o but. thats saved up for an tft tv
12:53:52Coldtoastone of the biggest irritations my friend has is when he's at work, where he plugs his h340 into speakers, when the battery runs out he has to wait to get home to his AC adapter to charge
12:54:17HClwhy doesn't he just take his ac adapter with him?
12:54:24B4gderor buy a second one
12:54:25HCldoesn't seem logical to me.
12:54:30Coldtoastcos he has USB set to Data, once the unit powers off due to low battery, impossible to charge via USB
12:54:45Bgerthat's not logical
12:54:57Bgerwhy don't he set it to charge from USB ?
12:55:10Coldtoastif he's been transferrign stuff
12:55:22Bgerthen he can't
12:55:29HClbut rockbox can make it switch at low battery, obviously, but he should watch battery level.
12:55:30Coldtoastit's sett o data and he often forgets to change to Charge
12:55:44HClout of sheer curiousity.
12:55:50HClhow long does he keep that thing on? O.o.
12:55:57HClor does the h3x0 have really poor battery life?
12:56:03Coldtoastall day. Thenthe next day
12:56:08BgerHCl: not so poor
12:56:17Bgerabout the life of h1x0
12:56:18HClthats not smart, he should take it home and recharge for the next day.
12:56:39Bgerand btw, it's better to charge LiIon more often
12:56:48Bgerand before the battery has drained completely
12:57:04Bgerbetter for batteri life i mean
12:57:07Coldtoasthe's an electrical engineer. takes it to work while he's buildign stuff
12:57:08*amiconn wonders why the h3x0 can't charge from usb when in data mode
12:57:12amiconnThat's silly...
12:57:20Bgeramiconn: no idea
12:57:21HClamiconn: yea, that to.
12:57:22Coldtoastit's the reality of the situation tho
12:57:32Coldtoastand anyway. the fact he doesn't charge it aside
12:57:36HClamiconn: maybe it drains more power while copying than it can charge?
12:57:40Bgerbut maybe the unit is using more power than the usb port can give
12:57:45HClyes :p
12:58:01Coldtoastit's odd it doesn't auto-switch at low voltage
12:58:02Bgerok, i'm going to lunch
12:58:05amiconnIt shouldn't, unless the hd is spinning continuously perhas
12:58:31amiconn(same problem we got on archos)
12:58:34Coldtoastso. I think that should be incorporated with Rockbox when you start porting to the h3x0
12:58:45Coldtoastswitch the USB mode
12:59:09amiconnHmm, I noticed that unlike with my archoses, my iriver hd is also spinning continuously as long as it is connected to usb
12:59:29ColdtoastI'll try mine
12:59:37 Join thegeek [0] (
12:59:50amiconnI guess that is actually a problem of the hd, maybe we should do the changes Linus suggested long ago...
13:00:08amiconn...setting power management mode to spindown before entering usb
13:00:40amiconnWindows doesn't seem to do this for all HDs
13:01:15Coldtoasthey wow. it does too
13:01:30 Join Moos [0] (
13:01:35MoosHi all
13:06:03 Quit hicks (Read error: 145 (Connection timed out))
13:08:23amiconnIt seems the iriver hardware does *not* prevent discharging the battery too far :(
13:08:34amiconnI'm at 2.73V now...
13:09:15amiconnSwitched off at 2.53V
13:10:28amiconnThe hd already had problems spinning up with the battery voltage being at ~3.1V
13:10:56Coldtoastshouldn't be a prob tho should it?
13:11:42ColdtoastRB shutting the machine off
13:11:53 Join preglow [0] (
13:12:00preglowamiconn: are you certain the reading is accurate?
13:12:08amiconnI'm pretty certain
13:12:10preglowthe voltage conversion part of it, at least
13:12:21preglowthen we need to cap it
13:12:37preglowdo we have a thread or something for battery monitoring?
13:12:52 Quit edx (Read error: 110 (Connection timed out))
13:13:51amiconnyup (in powermgmt.c)
13:14:25preglowthen adding a warning and subsequent shutdown shouldn't be too hard
13:14:36 Quit B4gder (Read error: 60 (Operation timed out))
13:15:00amiconnThe warning is the harder part. Only the UI thread should write to the display
13:15:06preglowand btw, somone raised a good point, not saving the settings if the disk doesn't sping down before we shut down isn't a problem for iriver
13:15:09preglowso we want to do that
13:15:16amiconn(Though we could introduce a sysevent SYS_BATTERY_LOW)
13:15:22preglowcould and should, imho
13:15:41preglowit's a pretty vital event
13:16:06amiconnHmm. If we do that, we should also do it for fm/v2 recorders
13:16:55amiconnI wonder whether the voltage table for these is correct though; the values are looking a bit too low
13:18:02Coldtoasthey. was there a build that kept playing audio while you seek?
13:18:44Coldtoast6-16-05 build, somebody said
13:19:16amiconnThere was no such build, on any platform
13:19:29Coldtoastsomebody said it in the forum
13:21:25 Join hicks [0] (~hicks@6e937a7f3968d77c.session.tor)
13:22:54 Join B4gder [0] (
13:29:53Bgeramiconn: can you check the battery voltage with a meter
13:35:33 Join Bgr [0] (~Bager@
13:35:34 Quit Bger (Read error: 131 (Connection reset by peer))
13:35:51 Nick Bgr is now known as Bger (~Bager@
13:42:01***Saving seen data "./dancer.seen"
13:52:03 Join GodEater [0] (
13:53:05 Quit GodEater (Client Quit)
13:54:49 Join bipak_ [0] (
13:55:05 Join RotAtoR [0] (
14:02:31 Nick courtc_ is now known as courtc (
14:10:25 Quit bipak (Read error: 110 (Connection timed out))
14:10:51amiconnW0w. I finally managed to implement a memset similar to the SH1 implementation (using longs, not yet line bursts) without needing to push any register.
14:11:53Bgerexcellent :)
14:12:42amiconnI'ts a bit faster than then C implemetation if the destinatin is already long aligned, and up to 4x as fast if not
14:13:19amiconn(The C implementation uses long stores only if the destination is already long aligned)
14:17:09preglowyou used only the scratch registers?
14:17:46preglowin such a manner that clobbering a few wouldn't have made it faster, no?
14:19:57amiconnAlso, I don't read the parameters more than once
14:20:07*HCl flops on the couch
14:20:18Bgerbtw, what's the chance of buffer not being long aligned ?
14:20:23preglowi wish they passed parameters in registers
14:20:26amiconnIt uses the same trick as the archos implementation (running down from end to start), but for a different reason
14:20:38Bgerdoesn't gcc take care about that ?
14:20:40preglowBger: if you depend on it, you should make certain it is long aligned
14:21:03B4gderthere's still that gcc patch floating around no one ever tested
14:21:04amiconnIn order to conform to ANSI, we need to return the start address. Much easier when running down from end to start
14:22:04preglowB4gder: yeah, but i really think it should be in mainline gcc for us to use it
14:22:24preglowrequiring everyone to patch their gcc is a bit troublesome
14:22:24B4gderbut it would be interesting to measure the difference
14:24:32 Quit B4gder ("Lämnar")
14:24:44HClif it matters a lot, we could hand out a pre-compiled gcc crosstool.
14:26:26preglowand get the patch put into gcc as soon as possible
14:26:52preglowi think it's pretty silly an arch with 16 registers doesn't allow parameter passing in registers as an option
14:28:32preglowamiconn: so, going to commit it soon?
14:28:50amiconnAdding comments now...
14:29:02amiconnI also added another slight opt for archos
14:30:38preglowmempcy will be a bit harder, right? :V
14:31:32amiconnI won't tackle memcpy now. I'll try adding burst mode support to memset(), then switch to gfx api coding
14:41:43 Join GodEater [0] (
14:43:46amiconnHmm. My memset() is a bit slower for medium sized blocks (8..100) which are already long aligned
14:44:39 Join bobTHC [0] (
14:46:09preglowi'm kind of surprised there was so little to gain, they must have done a good job of adapting the coldfire to c compilers
14:50:01 Join elinenbe [0] (~elinenbe_@
14:54:39 Join edx [0] (
15:07:44*amiconn is silly...
15:07:57amiconnNow asm memset() is always faster...
15:16:53Bgerwhat was the "slowing" thing ?
15:17:08amiconnMy loop end condition was off by 1 longword
15:17:58Bgerthese little ones ... :)
15:18:26amiconnYep. The main loop set one longword less, and the byte loop had to do 4 extra bytes because of that
15:19:48preglowoh, yes
15:19:50preglowbyte moves are slow
15:20:57amiconnI think I should commit now, before starting to fiddle with line alignment
15:24:39preglowdo so
15:27:54Bgerwhy "the recent cvs activity" table doesn't show it
15:28:09amiconnIt will, after it has been compiled
15:28:17amiconnBe patient... ;)
15:28:39Bgerafter compile ?
15:28:49Bgerhm... didn't know about this
15:31:27preglownow you do!
15:33:09preglowyou've forgotten whitespace behind your commas ;)
15:37:03amiconnMrf, I've lost the objdump disassembler command line :(
15:37:24preglowi've got it
15:37:36preglowif you mean the flat file disassembly thing
15:38:07preglowm68k-elf-objdump −−target=binary −−architecture=5200 -DS filename
15:38:08preglowworks for me
15:38:26preglowat least to the extent i can expect it to work, it gets thrown off track very easily
15:39:01preglowarchos models use a.out?
15:39:11amiconnNope, they use .elf as well
15:39:28amiconnBut there is a dedicated sh disassembler in tools/
15:39:30preglowhow come the function is called _memset?
15:39:54preglowelf doesn't allow underscores
15:39:56amiconnAll symbols get a _ prepended
15:40:15preglowthat shouldn't be the case for elf
15:40:17amiconnIt's a question of abi, I guess
15:40:28preglowmno, i think elf generally doesn't allow underscores
15:40:34preglownot 100% certain, of course
15:40:44amiconnIt has to, because it's working for archos
15:40:57amiconnsh-elf-gcc, sh-elf-objdump ....
15:42:04***Saving seen data "./dancer.seen"
15:42:13preglowthis is the first case of underscores in elf i've ever seen, at least
15:43:09preglowyeah, the elf spec does state c functions should not have leading underscores
15:43:40preglowbut not much of a problem for embedded environments, i guess
15:48:54 Quit bill2or3 ("Leaving")
16:09:19 Join asdsd____ [0] (
16:09:36 Part asdsd____
16:11:52 Join Sucka [0] (
16:11:54preglow/home/thomj/build/rockbox-devel/build/libFLAC.a(coldfire.o): In function `FLAC__lpc_restore_signal_16bit_order8_mcf5249':
16:11:57preglow/home/thomj/build/rockbox-devel/apps/codecs/libFLAC/coldfire.S:49: relocation truncated to fit: R_68K_PC8 against `*UND*'
16:12:00preglowthis is good, i assume? :V
16:17:42 Quit kenshin (Read error: 145 (Connection timed out))
16:21:59 Join webguest63 [0] (
16:28:55amiconnHmm. Maybe I can optimise more using lea, but that requires to push an address register
16:30:14 Join Febs_away [0] (
16:34:00 Part webguest63
16:35:50 Quit GodEater ("CGI:IRC")
16:44:15amiconnWow, burst mode is faaast....
16:44:32t0masamiconn? is there a reason to use the pitch_screen(void); and quick_screen(const int, const int); in screens.h only for the recorder?
16:44:44t0masor is that an example of wrong #if usage?
16:45:15 Join Kohlrabi [0] (
16:45:20amiconnThe quick screens are recorder only, because only the recorder has those F buttons
16:45:34t0masamiconn: but if I make keybindings for them on iriver...
16:45:41t0mascan I add them then?
16:45:41amiconnYou can't
16:45:48t0maswhy not?
16:46:14amiconnThe quick screens are meant to be used by pressing the quick screen button and the selection button within the screen at the same time
16:47:11amiconnAll other units except the recorder either have too few buttons (player, ondio), or don't allow pressing multiple buttons at once (iriver)
16:47:22amiconnSame goes for pitch screen, iriver does not yet allow pitching
16:47:40t0masah, maybe it's time to edit the menu then? :)
16:47:53t0massome people on the mailinglist wanted the F2 menu
16:47:59amiconnI know
16:48:18t0masand... on iriver... I can use joystick + on/play right?
16:48:21 Quit ashridah (Read error: 110 (Connection timed out))
16:49:05 Join spiralout [0] (
16:49:12amiconnYes, and play(on) is the only button suitable as a shift key. Stop(off) can also be used independently, but holding it will power down the unit...
16:49:29preglowamiconn: what was the difference in performance with burst mode?
16:49:48amiconnpreglow: I get a speedup of 2.5x for large blocks
16:50:01t0masamiconn: so holding select is possible... and then use on-play and stop for up and down?
16:50:13preglowamiconn: and what's keeping us from enabling it permanently?
16:50:30amiconnIt is enabled... I mean, using it in memset()
16:50:39 Join DMJC [0] (
16:50:59preglowahh, i kept reading 'page mode'
16:51:13preglowdinner next
16:51:19amiconnt0mas: Holding select is already taken, it brings up the context menu
16:51:56amiconnThe other way 'round (holding play and then use joystick directions) is also partially taken.
16:52:08amiconnIn the browser, play+up/down move page-wise
16:52:45t0masamiconn: but in wps I can do the holding select thing...
16:53:08amiconnYou could... as of today
16:53:29t0maswhy not add the F2 menu to wps then?
16:53:34t0masso people can use it while playing music?
16:53:35amiconnI'd prefer having a context menu for wps as well, and iirc Linus already had some code for it
16:53:45t0masah that would be cool...
16:57:41 Join Cassandra [0] (
16:58:23CassandraOh damn, I could do with a Linus. Is there anyone else who understands the iRiver bootloader around?
17:04:43CassandraOK. Anyone know how the Reset hole on the iRiver is supposed to work? I stick a pin in mine and all I seem to hit is empty air.
17:05:09 Quit Seed (Nick collision from services.)
17:05:18 Join Seed [0] (
17:08:42crwltry something slightly thicker, like a paper clip
17:09:15CassandraThat's what I'm using. Does it have to be metal?
17:09:38CassandraWhat am I supposed to feel?
17:09:49amiconnIt's just a button. I'm using a paperclip, and it's working perfectly
17:10:10zetry pushing further, or sticking it at different angles till it hits something
17:10:11amiconnYou should feel a slight click, as with the side buttons
17:10:57CassandraRight. Yeah. Got it. It's not helping. Looks like my iRiver is screwed. :(
17:12:15amiconnWhat does happen?
17:12:39CassandraWhen you push on the first time, brief flash then turns off.
17:13:22zebatteries got a charge?
17:14:03CassandraPress it a couple more times, or for a long time, you get the iRiver logo for a couple of seconds followed by reboot. Then LCD flash, pause, logo, reboot, LCD flash, pause, logo, reboot
17:14:10 Quit DMJC (Read error: 110 (Connection timed out))
17:14:10CassandraWhich goes on for a while.
17:14:20CassandraShould be. I charged it last night.
17:14:50amiconnDid you already put rockbox on the unit?
17:16:32CassandraJust tried it on mains. Same result.
17:16:33preglowi've never even heard of something like this
17:16:40CassandraYep, ami.
17:17:10CassandraIn fact, if you RECORD + on, it prints up the message about starting iRiver firmware then does the same thing.
17:17:36CassandraThe boot rom won't go into USB mode either, so I can't see if I have a corrupt rockbox.iriver.
17:18:02amiconnIf you connect USB first, then power on, what happens?
17:18:59preglowso the bootloader wont go to usb mode?
17:19:08CassandraSometimes I get 2 lcd flashed with USB plugged in.
17:19:58preglowthen things aren't looking very bright
17:20:00 Join balor [0] (~balor@
17:20:16CassandraNo. No they aren't.
17:20:51CassandraMy next stage would be to disassemble and give it a really hard reset.
17:21:07CassandraExcept I don't seem to have the right screwdriver for that.
17:21:34balorI'm looking for a good ogg/flac/mp3 player. The ipodlinux guys said to check you guys out. I'm a bit dim when it comes to the numbering of irvers though. Will the H10 play ogg and flac?
17:21:54CassandraNope. Just the H120 and H140.
17:22:13Cassandra(Maybe the H320 and H340 in a few months)
17:22:40zebalor: i like my rio karma except for the fact that it doesn't have opensource firmware... its also discontinued now i guess
17:22:51balorCassandra, thanks. would you recommend either of those for someone looking for an ogg/flac player?
17:23:10CassandraThey're really about your only option.
17:23:35zewell neundo or whatever it is plays ogg/flac too i think, and so does the karma
17:23:51CassandraThe iRiver stuff is quite new, and not yet stable.
17:24:01zeif the karma had rockbox or something like it, it'd totally own, but it doesn't seem likely to happen :/
17:25:24zethe karma firmware isn't too bad, but its quite limited compared to the hw capabilities of the device, and its a bit buggy, plus they apparently aren't releasing any new versions
17:25:27balorCassandra, so I should buy an archos?
17:25:37zebalor: not if you want ogg/flac, afaik
17:25:45CassandraArchos don't do ogg/flac. iriver H120 is literally my _only_ option?
17:26:25zebalor: that or neundo or karma
17:26:42Andrew179the iAudio players support ogg and flac
17:26:45zei think neundo's what its called, i'm not sure... i think its opensource too iirc
17:27:10*ze checks
17:27:25zehmm no i guess maybe its not neuendo
17:27:34zeits like neu something i think
17:27:45CassandraYou're probably thinking of the Neuros.
17:27:49zeyeah i think so
17:28:07zeyeah that looks like it
17:28:28zehmm their latest version apparently comes in 20 30 40 and 80G models with usb2
17:29:10zegrr i hate sites with broken links
17:30:28balorze, looks like it's discontinued aswell
17:32:00balorze, so the square looking Rio Karma plays ogg & flac?
17:32:12zesquare looking?
17:32:15Andrew179balor: here's the link to iAudio, be sure to check out the M3 and X5 players:
17:32:25zeits a bit curvier than a square but i guess it is kindof squareish
17:33:02zethat and yeah it plays flac and ogg
17:33:26balorze, it's cheap £120 ~= $170
17:33:40crwlthe iAudio X5 seems to be really good
17:33:56crwlat least it's damn thin, plays all the necessary formats and supposedly has coldfire in it
17:33:56zehmm odd this site lists some cons for it like "recharges only in cradle" which isn't true, "not mac compatible" which also isn't true since it has a java client that works over ethernet
17:34:04balorcrwl, except the site requires flash to navigate it :(
17:34:12zeit is thick, i dunno about the warranty, and it doesn't have a belt clip
17:34:13crwlbalor, hehe, as if iriver's site didn't :D
17:34:37preglowbalor: the iaudio x5 plays ogg and flac, but i don't think it's released anywhere but korea yet
17:34:50crwlpreglow, bah
17:34:55balorpreglow, lol
17:35:00crwlpreglow, my friend (here in finland) has one, he ordered it from germany
17:35:06crwlit's definitely available in the EU at least
17:35:06balormethinks I'll buy a rio karma
17:35:14preglowaight, didn't know that
17:35:18*preglow huggles his h120
17:35:25zei think there's a karma v2 out now too
17:35:28zei dunno anything about it though
17:35:56preglowi wouldn't consider anything that isn't running a risk of carrying rockbox in the near future
17:36:03zei heard there was going to be one anyway
17:36:23zepreglow: port it to the karma!!!
17:36:55zethis thing is capable of so much if only its fw was opensource
17:37:05 Join kenshin [0] (
17:38:08zeit runs ecos as it is
17:38:17zeas in thats what the closed-source firmware is
17:39:21baloriAudio X5 is $299, ~$170 more than the Rio karma
17:39:23HClkarma.. only 20gb
17:39:32preglowkarma uses a dsp chip, yes=
17:39:42HClit supports flac
17:39:52preglowif so, it significantly hardens the porting work
17:39:56zethere's hacks to upgrade the karma's disk
17:40:16preglow20gig is more than enough for me at the moment
17:40:24ze and (and other pages on that site) have various details
17:40:30*HCl already has space trouble on his iriver
17:40:36CassandraLucky you. 40 is getting tight for me.
17:40:54HClrio karma is portalplayer
17:40:54CassandraOf course, since I now have a brick, it's kind of academic. :(
17:41:00HClwhat happened?
17:41:21crwlmy h120 has been getting quite full, but i could re-encode all those ~200 kbps oggs with -q4 and have plenty of space again
17:41:27CassandraWon't boot. Not sure why. Can't even get bootloader USB mode up.
17:41:27HClew! and you can't use the rio karma as a harddisk.
17:41:40HCli'd even forgotten that there are players that don't support that.
17:41:52zeheh yeah thats the one annoying thing to me about the karma
17:41:55HClCassandra: send to linus?
17:42:01zethe only way to use it in linux is with a java app via ethernet
17:42:05***Saving seen data "./dancer.seen"
17:42:06HClze: horrid.
17:42:10zefortunately its about the fastest and most stable java app i've ever used
17:42:14CassandraMay be my only option. Have to see what Linus says.
17:42:21zeit actually works quite well
17:42:45HClnot as well compared to a portable harddisk
17:42:52zei dunno
17:42:59zei had plenty of issues with my archos via usb
17:42:59HCllol, i can guarantee you that.
17:43:08zeusb flakyness in general
17:43:16preglowamiconn: if you exchange the first b in a conditational branch opcode with a j, you don't have to specify a size prefix, gas will use the smallest possible version
17:43:28zeat least if the ethernet flakes out the filesystem doesn't get funky
17:43:36zethough the karma doesn't actually have a filesystem
17:43:42CassandraI'm going to go leave my iRiver plugged into the mains for a bit. Just in case.
17:43:45HClit has one, its just propretiary
17:43:56zewell its more like a database if i understand correctly
17:44:00HClThe onboard filesystem is a proprietary journalling FS, and Rio's been tight-lipped about its specs, but they are planning a migration to FAT32 for the sake of USB Mass-Storage Capability (MSC). Usually, this would spell disaster, as it does with the iRiver, but Rio's engineers have sworn that they're going to get it absolutely right. Given their track record so far, I trust them, but I'd be very curious how they manage it without adding fragmentation
17:44:01zeit doesn't have files, just entries
17:44:03amiconnpreglow: Nice to know, but I strongly doubt any of my routines will get so large that .b doesn't suffice
17:44:12HClexactly what disaster with iriver are they talking about?
17:44:14HClit works fine?
17:44:29zei dunno if that statement is accurate
17:44:30zecould be though
17:44:38t0masHCl: boot time
17:44:38preglowamiconn: sure, just good to know
17:44:48t0massome people really hate that 15 secs waiting
17:44:48preglowamiconn: you've also got a mov.l hanging around, btw :>
17:44:49HClt0mas: thats just a poor choice in the firmware of iriver itself.
17:44:49zei don't know if they're even releasing new firmware for it at all
17:44:52CassandraOK. Welcome to planet bizarro.
17:44:54t0masHCl: it is
17:45:01HClyou can't blame that on fat32 O.o.
17:45:01CassandraI just turned it on again by accident.
17:45:05zeespecially since its discontinued
17:45:05CassandraIt booted.
17:45:07HClCassandra: :p
17:45:18HClgotta love computers :p
17:45:22t0masHCl: I don't care about that 15 seconds anyway... just leave it on
17:45:30zebut hey if they do release new firmware and make it usb mass storage compliant
17:45:39zethen thats just another thing in its favor i guess
17:45:41HClIf it stinks, it's chemistry. If it moves, it's biology. If it doesn't work, its computerscience!
17:45:57HClt0mas: i've found it annoying from time to time, but i have known that its not cause of fat32
17:46:01amiconnpreglow: Fun. The assembler didn't complain....
17:46:17zealthough i think i prefer whatever its got currently to a fat32 fs
17:46:17amiconnmov.l is better anyway. It's shorter....
17:46:32zeit seems to work well and never seems to get any errors
17:46:41CassandraWoohoo. That is so surreal.
17:46:52preglowamiconn: i believe it's legal, but i also believe move.l is the official opcode name
17:47:07preglowconsistency with the rest of rockbox source doesn't hurt
17:47:36zei think filesystem's are poor way of organizing data anymore too... being able to reference data via multiple paths of metadata makes more sense these days, especially for media
17:47:40amiconnThat's right. Changed it locally, in my burst mode enabled version :)
17:48:08zeHCl: whats portalplayer?
17:48:23HClits a type of cpu used in various portable players
17:48:30HClrio karma, ipod..
17:48:33HCland there are more
17:48:35preglowit's more a system on a chip
17:48:37HClits a very closed platform
17:48:38preglowwith a couple of cpus on
17:48:39zethis thing says its got 2 ARM7's that can go up to 90Mhz each
17:48:48HClits extremely hard to build anything opensource on it
17:48:56HClthe ipodlinux project has existed for over 2 years
17:49:10HCland only a month ago or so they finally managed to figure out how to set the clock speed of 1 of the two cpus
17:49:10zelast i heard they could do software mp3 decoding in realtime
17:49:31HClthey only have access to 1 of the two arms, and only recently they figured out how to set the clock speed.
17:49:37t0masnot at 320 kbits
17:49:39HCland thats after more than 2 years time
17:49:42zei see
17:49:44CassandraGood for them. Must port Rockbox to Linux at some point.
17:49:53HClto give you an idea how hard it is to develop for portalplayer
17:49:53t0masCassandra: why?
17:49:59zewell maybe some of their progress can apply to the karma
17:50:30zewow the karma's battery is only 1600mAh?
17:50:36zethats amazing, it gets great battery life
17:50:42CassandraIt'd allow you to do cool things like build your own jukebox for insertion into a stereo system.
17:51:05CassandraPlus it'd simplify Rockbox debugging no end.
17:51:07CassandraOr a lot of it, anyway.
17:51:11zeCassandra: um rockbox and linux are totally different animals
17:51:18zeCassandra: its like saying should port linux to windows almost
17:51:22zeCassandra: or something like that
17:51:32zeCassandra: rockbox effectively is an operating system
17:51:41preglowze: they can decode aac and vorbis as well now
17:51:43zeCassandra: there is a simulator though
17:52:05CassandraNot really no. Rockbox is agnostic about what hardware platform it runs on.
17:52:39zeCassandra: um, what?
17:52:42CassandraA cleaner approach might be to port it as a standalone OS for i386, but that's kind of a lot of effort.
17:53:11zeCassandra: you're not making any sense to me heh
17:53:16preglowamiconn: hahah, i love the seven line sequence that only involves two registers
17:53:20CassandraSorry about that.
17:53:41amiconnpreglow: Any better way to do that? I couldn't think of one...
17:55:07preglowi know how to expand it to eight instructions and save a word of space :P
17:56:30amiconnI found that the instruction length influences execution time less than expected
17:56:52amiconn(perhaps that's due to the icache, and memset being in iram)
17:57:17amiconnBtw, how would you manage to save space?
17:57:29zeso it seems like wma and its DRM is the main thing preventing rio from considering opensourcing the karma firmware and also making it illegal to reverse engineer
17:57:32preglowmemset is in iram?
17:57:35zepeople who use wma on the karma should die
17:57:58preglowthen instruction size doesn't matter much at all, as long as it doesn't take longer to execute at a bigger size
17:58:03HClpeople who made drm should die.
17:58:05amiconnThe longest instruction is and.l #0xFF,%d0
17:58:21preglowyes, there's another and as well
17:58:22zeHCl: yeah
17:58:23preglowwith a literal
17:58:50amiconnNot in the sequence you mentioned...
17:59:27preglowforget it, btw
17:59:30preglowi can't save a word
18:00:17amiconnThe SH1 does the same in 5 instructions, and it doesn't need a literal
18:01:06 Join einhirn [0] (
18:01:39preglowbtw, can't we depend on the data being constrained to a byte?
18:02:00amiconnThe argument is an int, and an int is signed, so no
18:02:01preglowit may be larger, yes, but it shouldn't be ;)
18:02:47preglowoh well
18:03:16zehmm according to this page, the karma fs isn't journaled, but the implimenters of it made others at rio think it was anyway
18:04:07thegeekwhy would you need a journaled fs on a portable device?
18:04:15thegeekit's not like you would be writing all that much to it
18:04:19zeto protected against unexpected power failures of ocurse
18:04:31thegeekyes, but when you are not writing
18:04:35thegeekwhat is the point of journaling?
18:04:53zeyou still gotta transfer files to it
18:04:53thegeekseems like it would just suck cpu power
18:04:58thegeekbut damn
18:05:04thegeekif my pc reboots
18:05:09thegeekI'll just copy over again
18:05:16zewell if its just reading it doesn't need to write to the journal
18:05:17thegeekIts not like I would loose anything vital
18:05:22 Quit bobTHC ("Smoke Weed Every Day !")
18:05:24zenot the pc, the karma
18:05:25thegeekbut still
18:05:31thegeekwierd for a portable
18:05:37zei don't think you can fsck it either
18:05:59zei dunno, maybe the fw can do an fsck
18:06:22zehmm this page says the karma firmware is built with gcc heh
18:07:05zeah yeah there's also writing playlists, bookmarks, resume details, settings, etc to the hd
18:07:22zethe firmware itself loads out of flash memory though
18:07:26thegeekI was ironic
18:30:19 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
18:41:11 Nick Febs_away is now known as Febs (
18:42:04FebsHi folks. Quick question about playlists. When using the "create playlists" command, in what order are directories added to the playlist? I've searched through the manual and IRC logs, but haven't been able to find an answer.
18:46:24 Quit kenshin (Read error: 110 (Connection timed out))
18:47:42FebsExample: when using "create playlist" at the root level, my root-level folders are added in this order: RECORD, Other, RandB, Classical, Rock, Jazz, Soundtracks, etc.
18:48:20 Nick Lynx_ is now known as Lynx_awy (
18:49:19amiconnFebs: I guess this adds subdirs unsorted, i.e. the order in which they appear in the parent directory
18:55:50FebsI just confirmed that the order does not match any of the "Sort Directories" options (alphabetical, by date, by newest date).
19:00:36FebsLooking at the directory unsorted in DOS ... you're right, amiconn. I knew I was missing something obvious.
19:04:52KohlrabiJust a small question: Is the low voltage battery issue solved yet?
19:09:47 Quit hicks (Connection reset by peer)
19:10:40 Join hicks [0] (~hicks@
19:13:39preglowwhy don't even know if there IS an issue, how does suddenly everyone else know?
19:14:20preglowthere might not be an issue
19:14:25preglowbut yes, we're going to fix it anyway
19:14:33preglowjust don't drain your battery until your unit dies
19:14:43KohlrabiThat i wanted to know :)
19:15:16zethats funny... issue's not even concerned and people come around asking if its fixed yet
19:15:18KohlrabiI didn't even install rockbox yet, will wait for that fix maybe
19:15:22zenot even confirmed rather
19:15:33*ze is tired
19:15:37SlasheriInteresting, i got *panic* Stkov audio
19:15:57KohlrabiWell if you're gonna fix it, there is actually some kind of issue ;)
19:17:10preglowKohlrabi: it's not dangerous, it MIGHT happen if you drain your player completely, completely empty
19:17:23preglowjust don't do that
19:17:29preglowrockbox has a nice battery indicator
19:17:34Kohlrabiwell, OK
19:18:16 Join Teslacuted [0] (
19:18:48crwlis it safe to use rockbox till there's 0% battery left according to the indicator?
19:19:07preglowaccording to a readin amiconn got, it might drain the battery below 3 volts some times
19:19:13preglowlike if the drive spins up when its about to die
19:19:36TeslacutedI'm using Rockbox on my Iriver and I noticed any 48khz mp3s are playing at 44.1khz speed, is this a known bug?
19:19:43preglowTeslacuted: yes
19:19:56preglowiriver hardware doesn't support 48khz, and we haven't coded resampling yet
19:20:29TeslacutedDid the original firmware resample OTF?
19:20:44Coldtoastcan you make things like play position have 2 digits instead of 1? so 2:54 is 02:54?
19:21:08Coldtoastall the times like track length, etc
19:21:21preglowTeslacuted: yes
19:21:22amiconnpreglow: Burst mode memset is looking good, I found a way to make small block memset a bit faster as well
19:21:35preglowTeslacuted: the original firmware resamples everything to 44.1khz
19:22:06amiconnColdtoast: Why? It'll automagically become 2 digits when it's above 9:59
19:22:30preglowpeople are obsessed with everything staying statically positioned at all times
19:23:02Coldtoastamiconn: when formatting .wps
19:23:06preglowooh, they've found the audio codec for m3
19:23:18amiconnI prefer the data nicely grouped together, and 5:30 is more readable than 05:30, imho
19:23:28CassandraWho has. And what's M3?
19:23:40Coldtoastwell, if I position something it's irritating that later on I might get a track that pushes things off the edge
19:23:56Coldtoast%s doesn't work with timers, so that's no solution
19:24:11CassandraOr you could just make sure there's a space at the end of your line.
19:24:28amiconn2 digits won't help you if you are positioning with extraneous spaces
19:24:34preglowCassandra: m3 is another coldfire based dap
19:24:44preglowCassandra: might easily end up having rockbox on it
19:24:54amiconnIf the track is longer than 59:59, it'll display hours as well
19:25:03Coldtoastoh well. I'll just take that as a no then
19:25:42amiconn...and there is another reason why it is implemented like it is. Most archoses don't have exactly much screen space, so at least any 2 digit formatting should be optional
19:26:05Coldtoastoptional is fine with me. heh
19:26:14amiconnI think the clean solution is to have text justification, but that's not yet implemented
19:26:19CassandraWell, if you wanted to supply alternative labels for track time formatted as you wanted, a patch would almost certainly be accepted.
19:26:23Coldtoastyep.that WOULD solve it
19:26:32 Quit Teslacuted ()
19:27:20amiconnColdtoast: I agree that we should keep that in mind, but as you've probably been told already, iriver rockbox is far from finished
19:27:27amiconnI'd consider it not even beta
19:27:51Cassandra*nods* Definitely still alpha level code.
19:27:59Coldtoasteven if you could specify somethinand therefore this is the perfect tiem to suggest such things, right?
19:28:03CassandraBut pretty good alpha. :)
19:28:18CassandraNah. The WPS code is mature.
19:28:30amiconnI think we should concentrate on getting all existing features to work on iriver, then start extending
19:29:07ColdtoastCassandra: almost seems like you're saying "I'll submit it to the wishlist
19:29:23Coldtoasterr, ignore that Cassandra. that wasn't fo ryou. it was a general comment
19:29:37preglowamiconn: are you gonna replace the current memset by the burst memset, or
19:29:38Coldtoastgot halfway thru opne sentence then changed my mind. heh
19:29:41CassandraI just don't think it's going to be a priority for anyone except you, I'm afraid.
19:29:49amiconnpreglow: Yes that's the plan
19:30:04amiconnI have to comment it first, the usual thing...
19:30:08preglowthen what was the point of not touching the stack? surely you'll have do that now
19:30:18ColdtoastI dodn't say it was a priority. it was a suggestion, plain and simple
19:30:26amiconnI'm touching the stack only when actually bursting
19:30:55preglowwriting four registers at a time, yes+
19:31:08preglow?, that is, my shift key is broken, it seems
19:31:08amiconnNot touching the stack saves time. The stack might be in dram depending on the thread
19:31:33preglowstack should pretty much always be in iram
19:31:37preglowif we can help it
19:31:43amiconnNot all stacks are
19:32:01*amiconn checks
19:32:30 Join tvelocity [0] (
19:34:11 Quit Bger (Read error: 148 (No route to host))
19:37:15 Join Bger [0] (~Bager@
19:37:20amiconnpreglow: The *only* stacks that are in IRAM are main and codec. All others are in DRAM
19:38:10preglowthat's not good
19:38:23preglowhow much stack in average have threads been coping with on archos?
19:38:45preglowplugin stack is main thread stack, yes?
19:39:09amiconnThe default stack size is 4 KByte for all platforms.
19:39:18amiconnMain stack is 8 KByte on archos
19:39:26amiconn(32 KByte on iriver)
19:39:42amiconnPlugins are running in the main thread
19:40:08amiconnWe can't put any stack in IRAM on archos
19:40:14preglow32kbyte is a bit overkill for main stack
19:40:42amiconnMaybe. Iirc this was changed for rockboy and the xxx2wav plugins
19:41:10preglowi think we should get rid of the xxx2wav plugs soon
19:41:28preglowand implement a wav rendrer somewhere else, using the ordinary codecs
19:41:45preglowahh, no, that doesn't work right now, since the codecs actually talk directly to the audio driver
19:41:48preglowwhich is badness
19:41:51amiconnYepp. I think it would be a good mark to do this when codec api gets separated
19:42:04preglowthat sorely needs to change
19:42:09***Saving seen data "./dancer.seen"
19:42:28amiconnDo the codecs call a function for this?
19:43:09amiconnIf so, we could hand them a function pointer instead, allowing us to have different target functions
19:43:11preglowwe've still not agreed on how much the codec should do and how much the codec loader should do, though
19:43:19preglowi'd like most of the processing to be placed in the codec loader
19:43:31preglowthey do access it through a function pointer
19:44:02amiconnAnother function could mimic audiobuffer_insert() and render to a wav file
19:44:30 Join Stryke` [0] (
19:47:09preglowi think the codecs should hand the buffer to the codec api no matter what
19:47:15preglowso that it can do what it wants
19:49:03 Nick Febs is now known as Febs_away (
19:49:05preglowplus apply eqs, etc
19:54:49CassandraOh, that reminds me. Something I've been wondering for a while. Does the iRiver's SRS WOW support get implemented in hardware or software.
19:55:09CassandraNot that I'm actually sure what all those bloody acronyms mean.
19:56:32preglowSeed: i assume people willing to rewrite parts of libmusepack haven't exactly been popping out the woodwork lately?
19:56:38preglowCassandra: software
19:56:56preglowCassandra: and i bet it's patented
19:57:07CassandraThought that might be the case.
19:57:25CassandraAlmost certainly. We can however implement MDBlike stuff, I expect.
19:57:48CassandraPsychoauccoustics has a hell of a lot of prior art.
19:58:28 Join rubberglove [0] (
19:58:58preglowsure, we can do our own processing, fun part is coming up with something that works and doesn't require advanced processing
20:00:28Seedpreglow: there's barely any woodwork to pop out of :)
20:01:45preglowi decided to try some prescaling the other day, but promptly gave up, since you seem to do different scales everywhere
20:01:59preglowin some cases the scale factors are in arrays as well, which doesn't exactly simplify my work
20:03:02amiconnpreglow: Resampling code would be even more appreciated imho
20:03:02 Quit rubberglove (Client Quit)
20:03:07preglowi also had a look at the code required to simulate a 64 bit multiply, it's a miracle libmusepack works at all on h1x0
20:03:51preglowresampling requires more than half an hour :/
20:04:53amiconnOkay, about to commit burst-mode memset(). Speedup is 2.5x for large blocks
20:06:35elinenbeamiconn: is that a 2.5x speedup over the old 4x speedup you already committed today?
20:06:48amiconnelinenbe: yup
20:06:49 Join Christi-S [0] (
20:07:02amiconn...and it happens for all alignments
20:08:15Christi-SBah. Housemate unplugged ADSL router to hoover.
20:08:26Christi-SBad housemate. No biscuit.
20:08:51 Quit Cassandra (Nick collision from services.)
20:08:58 Nick Christi-S is now known as Cassandra (
20:09:38preglowand to hoover!
20:09:47elinenbeamiconn: nice work!
20:18:52 Join kenshin [0] (
20:22:13preglowi'm giving a quick pass at resampling now
20:23:14preglowlooks like i need to introduce a delay of one sample, i hope that doesn't matter too much...
20:24:20amiconnOooh, one sample, rockbox delays like mad! ;-)
20:26:31preglowthe latency might be unbearable, yes
20:28:28amiconnHmm, that one .flac ios still skipping....
20:29:29dwihnoHow many mA does the original archos charger provide?
20:29:35dwihno(rec 20)
20:30:33amiconnOriginal archos charger (rec v1 and player) is 9V 600 mA
20:31:00 Join webguest20 [0] (
20:31:39dwihno180 mA is not enough then :)
20:31:46dwihnothanks for checking amiconn
20:32:53amiconnpreglow: Didn't you write some asm routines for FLAC today?
20:33:04 Join Tangleding [0] (
20:33:31Coldtoasttha's a fun nick to say
20:33:40TangledingI've just encountered the HD loop access matter
20:33:50Tangledingsince my battery was to low
20:34:26TangledingHi coldtoast, talking to me or other?
20:34:38Coldtoastto you. like the nick
20:34:50 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
20:34:54 Join Stryke` [0] (
20:34:56preglowamiconn: i did yesterday, but i need to write a 32 bit version of them as well, before i can see if i sped things up
20:35:03 Quit webguest20 (Client Quit)
20:35:08preglowand that wont happen yet
20:35:23HCli'm a train i'm a train i'm a chookatrain...
20:35:40preglowHCl: lucky you
20:35:44Tangledingoh okay
20:35:51amiconnpreglow: Hmm, I thought you could help my poort ABBA track :(
20:36:01preglowamiconn: and yeah, i stumbled upon a problem trying to use PC relative addressing
20:36:04Tangledingstupid nick in fact:
20:36:05preglowamiconn: what's wrong with it?
20:36:10Tangledingit means: Tang The mad
20:36:31Coldtoastreally? it kinda sounds dirty
20:36:34amiconnpreglow: It's skipping like mad due to the flac decoder being less than realtime
20:36:39Tangleding'cause my first name is Tanguy, and I'm known to be a little bit crazy ;) )
20:37:06Coldtoastaah. I know that nick from the rockbox wps gallery. heh
20:37:31Tangledinggonna actualise the scrennshot indeed
20:40:05preglowamiconn: my opt wont help that, i think, libflac needs iram enhancements
20:40:37amiconnWhat does the opt do then, if it doesn't increase performance?
20:40:51amiconnIt only happens with this one track btw
20:41:13amiconn(re-ripped an ABBA album as FLAC just to test FLAC on iriver)
20:42:09amiconnI'm not that impressed about the compression though; track size is clearly above 50% compared to .wav for all tracks
20:42:18amiconnI should test wavpack too...
20:44:27preglowwavpack is much, much more efficient
20:45:19preglowit'll help, but just not much, i don't know how much yet, since i need to make 32 bit versions of the lpc routine (currently i only have 16 bit version)
20:45:21amiconnI know that wavpack performs better in rockbox, I'm interested in the compression compared to FLAC
20:45:34preglowthen you'll be pleased to hear the compression too usually is better
20:45:38CassandraIs wavpack lossless?
20:46:51crwl is this too old nowadays?
20:47:18 Join webguest00 [0] (
20:47:31 Part balor ("Leaving")
20:48:13webguest00what do i need to view fnt files?
20:48:29Andrew179Cassandra: wavpack has lossless and hybrid modes
20:48:54preglowCassandra: you can choose, by default it writes a single lossless file, like flac, but it can also write two files: one which is lossy, and another which acts as a correctional file, making the file lossless again
20:49:13webguest00i have korean fnt files which were converted from bdf files i've found and i want to add korean using them
20:49:15amiconn" WavPack is a completely open audio compression format providing lossless, high-quality lossy, and a unique hybrid compression mode."
20:52:36CassandraWhat's the point of the hybrid mode?
20:56:56amiconnAnother skipping track...
20:59:59webguest00hm.. can anyone answer my question?
21:01:11 Join TCK- [0] (
21:01:11 Quit TCK (Read error: 54 (Connection reset by peer))
21:04:18 Join LinusN [0] (
21:07:08webguest00ok uhm
21:07:14webguest00 what do i need to view fnt files?
21:07:30webguest00 i have korean fnt files which were converted from bdf files i've found and i want to add korean using them
21:09:12TangledingI've updated my WPS with good multiple display screenshot
21:09:23LinusNas far as i can tell, korean requires unicode support, something we don't yet have
21:10:56webguest00hmm so the fonts i have are useless for now then?
21:11:02 Quit spiralout ("ChatZilla [Firefox 1.0.4/20050511]")
21:11:48CassandraHeya Linus. Had a weird problem with the bootloader earlier, but it went away again.
21:11:56LinusNwebguest00: we would still need the fonts eventually
21:12:02LinusNCassandra: huh?
21:12:21CassandraThe bootloader failed for me for about half a day.
21:12:37CassandraTo the extent I thought my iRiver was buggered.
21:12:50CassandraThen for no apparent reason it started working again.
21:13:05webguest00hmm ok so when unicode is ready, ill do the translating and etc i guess
21:13:46LinusNCassandra: "failed"?
21:14:02amiconnpreglow: Next skippy track... I'm currently encoding as .vw btw :)
21:14:25 Quit CoCoLUS (Read error: 110 (Connection timed out))
21:14:47CassandraWhen the unit turned on, the backlight came on briefly. There was nothing on screen. It went off again. A couple of seconds later it came on again with the iRiver logo screen, which died after some disk access.
21:15:02CassandraAfter that it kept doing the same thing over and over again.
21:15:26CassandraHardware resetting stopped it, but pressing the on button again started off the whole cycle again.
21:15:55CassandraStarting with REC+ON gave a "loading original firmware" message, then did the same as above.
21:16:06CassandraStarting with USB in made no difference.
21:16:14 Join CoCoLUS [0] (
21:16:41LinusNCassandra: battery status?
21:16:54CassandraFine. Charged OK the night before, working fine now.
21:17:06LinusNweird indeed
21:17:06webguest00also, it seems that rockbox's volume is slightly lower than original firmware. is this due to unfinished codec arc or unavoidable problem?
21:17:41LinusNand different compensation for high bass/treble settings
21:17:43Cassandra*nods* It started when Rockbox bombed out after "A Momentary Lapse of Reason"
21:17:53amiconnwebguest00: Volume is the same as with original firmware, unless you've boosted treble or bass
21:18:31Cassandrawhich caused a spontaneous reboot to the state described above.
21:18:43CassandraRockbox hates Pink Floyd - who'd have thought it.
21:18:46amiconnRockbox caps the volume in order to keep your desired bass / treble boost, while iriver fw ditches bass and treble if you pump up the volume
21:19:08Tangledingargh 1st Rbx critical failure?
21:19:23CassandraSort of. It went away again.
21:19:37Tangledingwonder if i wont come back to iriver fw...
21:19:52amiconnLinusN: Did you read about my battery check? Turns out we can draw lower than 3 volts, although the HD isn't able to spin at such a low voltage
21:20:01LinusNbtw, i think the default file view option should be "Supported"
21:20:10LinusNamiconn: figures
21:20:34CassandraRockbox is alpha code on the iRiver. It's going to be risky to use. If you don't like taking risks, wait till it's stable.
21:20:57amiconnLinusN: It went as low as 2.63 V before powering off
21:21:02Cassandralinus: Agree wrt default file option.
21:21:03amiconn2.53 V even
21:21:13LinusNamiconn: wee
21:22:02Tangledingokay thanks cassandra
21:22:05Tangledingsorry for you
21:22:07Coldtoastwith volume at max, bass and treble boost at max the iriver firmware sounds better the rockbox does. Just my 2cents
21:22:16Tangledingi reported the IRC quote on my french board
21:22:29Tangleding(i've to go)
21:22:33 Quit Tangleding ("Chatzilla 0.9.68a [Firefox 1.0.4/20050511]")
21:22:40CassandraArgh. Don't quote me! I know nothing!
21:23:07amiconnColdtoast: If you set it volume to max in iriver fw, you don't have any bass or treble boost, regardless what value you set it to
21:23:12ColdtoastI know
21:23:31Coldtoastoh. sorry
21:23:42amiconnSo set bass and treble to flat in rockbox, volume to maximum, and it should sound the same
21:23:55Coldtoasthmmm. there's more bass than rickbox woth 0 bass/treble boost tho
21:24:02Coldtoastjust comparing now
21:24:17CassandraYou got SRS turned on.
21:24:23ColdtoastI'll check again tho. listenign to the iriver fw at the moment
21:24:26CassandraThat will give more bass.
21:24:29Coldtoastnope. I don't have that on
21:24:32ColdtoastUser EQ
21:24:55webguest00hmm this is weird. i have a directory named "korean" which has both korean and english subdirs in it. but i only see broken korean dirs and don't see any english dirs. any reasons for that?
21:25:22Coldtoastyeah. I can definitely hear a difference. the iriver fw sounds "thicker"
21:26:19Coldtoasterr, sorry. hold on. I was 1 notch down in the iriver fw
21:26:24webguest00another funny thing is, i have another dir named "christian" which is mostly composed of english dirs and like one or two korean dirs. but in that dir, i see all my dirs and files, whereas in "korean" dir, i do not see any dirs labled in english
21:26:29Coldtoastyeah. it does sound the same
21:27:38webguest00hmm for me, when i put the volume at 25%, it's faintly audible
21:27:52webguest00whereas in iriver fw, vol lvl of 10 is enough for me
21:28:10Coldtoastvolume 10?!?! man!
21:28:12ColdtoastI use 32
21:28:16CassandraThe scales aren't linear.
21:28:25amiconnI use 50..60% in rockbox
21:28:34CassandraAnd Rockbox and the iRiver have different ones I believe.
21:28:38webguest00so it's not the same then
21:28:49webguest00but near the end, it is
21:28:50amiconnRockbox scale is dB-linear and uses the full range of the audio chip
21:29:18webguest00so 25% in rockbox != 10 in original
21:31:02Coldtoastthink maybe you could make bass-limits-volume and volume-limits-bass switchable?
21:31:52Coldtoastiriver mode or rockbox mode. heh.
21:35:34ColdtoastI can't imagine rockbox sounding good on anything with the bass >~16 at the moment. It completely muddies the sound
21:35:49thegeekI much prefer the volume control in rockbox
21:35:54webguest00oh and can anyone answer my weird directory question?
21:36:04webguest00like,.. some dirs not showing?
21:37:18amiconnW0w, wavpack surely is efficient :)
21:37:59tvelocityhow much efficient?
21:38:18CassandraNone of us use Korean Windows. Sorry.
21:38:21 Quit Cassandra ("If I were actually witty, this quitline would be funny.")
21:38:50amiconntvelocity: Didn't measure. I encoded the same album with wavpack that I did with flac.
21:39:06amiconnCompression is 2.5% higher, and the compressor is blazingly fast
21:39:15amiconnDecoding doesn't boost the cpu for long
21:39:25tvelocitysounds good
21:39:51amiconnWith flac, some tracks don't even play continuously
21:42:13***Saving seen data "./dancer.seen"
21:43:52 Part Moos
21:45:04webguest00ok, so i've moved the dirs that werent showing to somewhere else-> booted up rockbox-> usb -> moved back the dirs-> rockbox- and they show now. is that a known bug?
21:51:26Coldtoastwith crossfading, is it possible to get it working with unbuffered tracks? sounds like a silly question I think. But once you get to the end of the buffered tracks, crossfading doesn't work
21:51:37LinusNwebguest00: now it is :-)
21:53:00Coldtoastmaybe have it so if you move along tracks closer than, say, 1 sec it doesn't crossfade?
21:53:23Coldtoasttha way you'd never notice it doesn't crossfade when it gets to the end of buffered tracks
21:53:57Coldtoastdoesn't seem much point crossfading when you're skipping along a playlist quickly
22:02:07LinusNColdtoast: the crossfade is done several seconds ahead of time
22:02:17LinusNin the pcm output buffer
22:02:26 Quit RotAtoR ()
22:02:48ColdtoastLinusN: once it gets to the end of the buffered tracks and loads one from the HDD, no crossfading
22:03:57Coldtoastit's kind of odd when you skip along tracks and the first few fade then it runs out of buffered tracks and abruptly starts playing what it's loading
22:04:27SlasheriHmm, i found that probably the crash i had today was caused somehow while i was walking and hdd had problems to read the data. When i got found, i noticed that hdd was still spinning and the panic message (stack overflow) was on screen
22:04:52Slasheri*got home
22:05:11 Join l8bi [0] (
22:05:16 Part l8bi
22:06:07 Join eightbit [0] (
22:10:08SlasheriColdtoast: it might be possible to enable crossfading for unbuffered tracks also. But that might require buffering small amount of the next track before doing the actual track switching
22:11:18ColdtoastI think not crossfading if you're skipping quickly would be better
22:12:09 Part LinusN
22:12:20Slasheriyep, crossfading is anyway impossible before the pcm buffer is fully filled
22:13:13ColdtoastI mean, it's not just my imagination is it? once it runs out of buffered tracks and you're skipping to tracks it has to load from the HDD, crossfading doesn't work?
22:13:44Slasherithat's true.
22:14:07 Quit t0mas ("c ya tomorrow")
22:14:56Slasherieven if tracks are buffered, it doesn't guarantee that crossfading should work. Almost completely filled raw pcm audio buffer is required too
22:15:31Slasheriand that buffer starts to fill after buffering has finished
22:15:55Slasheriwith some high bitrate files (ogg q10 for example), it might take a long time to fill the pcm buffer
22:17:20Coldtoastit's definitely nice having tracks crossfade as you're listening but seems a bit rough when you're skipping tracks and only the first few fade
22:19:42 Quit Benacool ()
22:20:05KohlrabiJust a cautious question: Is replaygain implemented yet?
22:20:29KohlrabiOn the HP I read that it is *possible* - is it made?
22:21:02 Quit eightbit ("CGI:IRC")
22:21:04 Join eightbit [0] (
22:21:21Slasheriit's not yet made but i think that will be implemented in future
22:22:02Coldtoasthmmm. tho.... thinkign about it.... if you went by something like a "less than a second" rule, the first time you skip, the first track will fade to the next regardless. You'd need to assume by default that no crossfading is used and only after playing, say15secs of music that the next track switch is a fade
22:22:15KohlrabiOK, thanks, looking forward to it
22:24:10Coldtoastmeaning you'd have a "fade flag" that gets set to false when a track loads then true after some seconds
22:24:34Coldtoasterr.. well, I know what I mean. just can't explain it. heh
22:25:55Slasherii think i know what you mean :) the track skipping will be improved to handle better those situations
22:29:24Coldtoastwhat I mean is if you're listening to a track and start skipping along using a "<1sec" rule, the first time you skip, from the initial song to the next one, it'll still fade cos that <1sec isn't there. If for the first 10secs of a track the crossfade flag was false then it changes to trues, it'd pretty much work where fast-skipping tracks doesn't fade but listening to more than 10secs will mean it fades on track switch
22:31:13Coldtoastanyway. laundry time I think
22:33:31 Join RotAtoR [0] (
22:35:21HCldoes replaygain need to store per-song settings or is it completely dynamic?
22:35:59Kohlrabireplaygain is metadata in the songfiles
22:36:11KohlrabiSo you have to analyse a song first
22:36:14KohlrabiThen it's written
22:37:27Stryke`volume is gonna be low on a portable if everything is replaygained to 89 dB
22:38:20Kohlrabimaybe they'll implement a preamp, too :)
22:38:27Kohlrabiso that you can boost it a bit
22:38:50Stryke`it depends on what people are using replaygain for
22:39:05Stryke`clipping prevention or a sort-of normalizer for volume
22:39:11 Quit cYmen (Read error: 104 (Connection reset by peer))
22:41:31 Join cYmen [0] (
22:44:39Stryke`it seems a little odd having a preamp when you can just up the volume a few ticks
22:45:08Kohlrabiwell, yeah
22:45:28Stryke`but i suppose the higher volumes would be unreachable with replaygain
22:49:01Bagderalmost there now
22:49:18BagderSlasheri: here?
22:49:33SlasheriBagder: yes, hi :)
22:49:47Slasherihmm, checking that patch
22:49:55Bagderit can play songs
22:49:57Bagderor plugins
22:50:00Bagderbut not at once
22:50:11Bagderbut I don't really understand why (yet)
22:50:25Bagderthey are now in separate areas
22:53:14Slasherihmm, i will try that soon :)
22:53:45BagderI basicly just fixed the linking and patched the codec_load_ram
22:56:24preglowsuddenly the neighbour lady comes by with a couple of litres of altbier
22:56:28preglowthis did not make me more productive
22:58:48 Quit webguest00 ("CGI:IRC (EOF)")
22:59:59 Join linuxstb [0] (
23:04:31 Part bipak_
23:07:42preglowamiconn: seems you helped spur along vorbis playback as well, heh
23:08:11preglowamiconn: and are you saying you've got a skippy wavpack track?
23:10:05 Join Christi-S [0] (
23:12:11eightbitHi, I would like to learn how to output sound. I have made myself a little step sequencer and it is kind of pointless without sound
23:12:36eightbitI have only worked with sound before in VSTs where everything is floats
23:13:48eightbitI am guessing that I must eventually use the function: mp3_play_data?
23:16:25linuxstbeightbit: Which player are you using? iriver or Archos?
23:16:58 Join einhirn [0] (
23:18:40preglowi THINK plugins can generate sound, somehow
23:18:43preglowrockboy does this
23:18:51eightbitSo does the metronome
23:19:02preglowwell, yeah, but that's a archos plugin for now
23:19:05preglowit generates mp3 data, i guess
23:19:11eightbitah, it does
23:19:48preglowon iriver, you can generate pcm data directly
23:19:58preglowlike i said, rockboy does this, but i don't know exactly how
23:20:13preglowcodec plugins are expected to be fed data as of now, so that's probably not a good way to do it
23:20:21eightbitI have only worked with audio data in floats so I need to learn how to represent it in fixed point or ints
23:20:34eightbitI shall scan through rockboy...
23:20:45preglowit's much the same if you learn how to use the emac unit and can get by with numbers in the range -1 to +1, heh
23:21:00eightbitbtw, the rockbox as it is is amazing!
23:21:03HClrockboy is dead
23:21:11HClever since the playback was introduced
23:21:11preglowthat's true
23:21:16preglowyou can't do it that way anymore
23:21:32HClthe sound driver needs to be rewritten for rockboy
23:21:34 Join spiralout [0] (
23:21:44HCli should update the wiki about that.
23:21:45 Quit TCK- (Read error: 131 (Connection reset by peer))
23:21:54spiralouthi:) q: what means " Made Supported the default File View option"
23:22:00eightbithmmm, how is it done in normal playback?
23:22:52Bagderspiralout: it means supported is now the default file view option ;-)
23:22:58eightbitI messed around a little like a month ago using the old test wave source to play a drum sample but that option will not work on me
23:23:22spiraloutand what is the default file view option
23:23:25pregloweightbit: so, what're you making? a tracker thing?
23:23:32Bagderspiralout: the view option used from the start
23:23:40Bagderwhen not modified
23:23:52Bagderwhat files to show in the dir browser
23:24:11spiraloutah ok
23:24:17eightbitpreglow: A little step sequencer... every used reason?
23:24:29pregloweightbit: sure
23:24:31eightbitif so think redrum
23:24:38preglowsounds good
23:24:51preglowbut yes, plugins should be able to make their own sound
23:24:53preglowsome time...
23:24:55preglowin the future...
23:24:55eightbitthe interfave works nicely for me now
23:25:13eightbitI need to update it for the remote though...
23:25:24spiraloutis it possible to run the backlight in about half brightness?
23:25:53eightbitWell, is the old wave test code still in place? I havent looked recently
23:25:58spiraloutnot now i know but is it possible
23:26:14preglowspiralout: it is possible
23:26:33spiraloutwould be great for car or something...
23:27:24cooleocrap did i miss Linus?
23:27:30preglowcooleo: indeed
23:27:49cooleojust my luck!! jeje
23:27:54pregloweightbit: but yes, i'm very interested in making rockbox a dsp platform myself ;)
23:28:08spiraloutat night in car it is a little bit to bright...for me option would be great
23:28:14spiraloutwhat do you think?
23:28:41eightbitI was thinking that too
23:28:42preglowspiralout: it is possible, someone mentioned measuring power consumption before it is put in rockbox, though, but perhaps you'd use it even if it consumed more power?
23:28:57cooleoare you guys talking bout H100?
23:29:02preglowcooleo: aye
23:29:08eightbitusing the digital in and out it would make a nice fx box
23:29:24pregloweightbit: oh, i'm planning on using it as an fx box with analogue line in and out as well
23:29:26SlasheriBagder: i will look that more tomorrow.. it crashed if i tried to play something :/ are you sure that codecs are linked in right way?
23:29:47BagderI'm listening to music right now
23:29:57eightbitout of interest how god are the converters relative to the pro stuff?
23:30:09BagderSlasheri: and I checked the map files and they looked fine
23:30:10pregloweightbit: probably not too good, and we're locked to 44.1khz
23:30:11spiraloutyes i need it only for car...perhaps i can use a own carconfig for the night
23:30:22pregloweightbit: anywho, i reckon it'll make a decent vocoder, and that's all i want from it ;)
23:30:35eightbitthat would be cool!
23:30:44preglowoh yes
23:31:02BagderSlasheri: I'll check it closer and see if I figure it out
23:31:10SlasheriBagder: ok, good :)
23:31:42spiraloutok just a thougt time to sleep for me now bye
23:31:45 Quit spiralout ("ChatZilla [Firefox 1.0.4/20050511]")
23:32:04eightbitDo you know if this would be possible? : using the digital input as a kind of midi sync that I could potentially hook my stepsequencer to
23:32:19 Quit Coldtoast ("Peace and Protection 4.22")
23:32:25pregloweightbit: well, yeah, as long as you can generate a decent sync pulse
23:32:37eightbitI was thinking if you made a loop and ouput it to the h100 in something like ableton live
23:32:41pregloweightbit: i believe the digital and analogue inputs are seperate
23:32:55eightbitThat would be very useful
23:33:45preglowhaha, doesn't matter much if they're separate as long as they share the same socket, though
23:33:46eightbitMy pathetic dream would be to someday own an H300 with rb and then plug in a usb midi port
23:33:52*preglow slaps himself
23:34:00preglowwe can't control usb
23:34:10eightbitH300 does though?
23:34:14preglowi don't know
23:34:31preglowthat would rock very much
23:34:36eightbitI think so because it can get photos from a camera
23:34:38preglowoh, indeed it would
23:34:59preglowif that works, i will have to get a h3x0 as well
23:35:05eightbitImagine that, brining on a litte iriver synth!
23:35:08preglowportable softsynth i can put in my pocket
23:35:39eightbitI am sure you could make a nice mono base synth with it
23:35:49preglowoh yes
23:35:53cooleotheoretically it could
23:35:57cooleoi've got one
23:36:02cooleothey rock
23:36:02preglownot just theoretically
23:36:14cooleoyeh but are you guys good enough?
23:36:15preglowi'm certain the coldfire can put out several voices
23:36:21eightbitthat would be crazy though
23:36:29preglowcooleo: i'm fairly good at synth programming
23:36:40cooleoi love you guys
23:36:45cooleoyou dont even own this thing
23:37:04eightbitI have done a little bit with vst's, but a lot more with reaktor and the like
23:37:13preglowi've done both
23:37:39cooleoi've dabbled in vsts in cubase
23:37:48preglowi've programmed a couple
23:38:12eightbitanyway you mentioned 'emac units' above somewhere I think...
23:38:23pregloweightbit: well, yeah, it's the fastest way to do maths on h1x0 and h3x0
23:38:32pregloweightbit: and it's very well suited to filters and the like
23:38:41eightbitdo you know a good resource on that?
23:38:54preglownot really, not apart from the manual
23:39:00preglowand then there's me
23:39:20preglowif you know fixed point math, it's a breeze
23:39:26preglowif not, you need to learn some of that first
23:39:33cooleoim off
23:39:38cooleoexam tomorrow
23:39:41eightbitnah, I need to learn that as well
23:39:41preglowait, have nice
23:39:45cooleocya later
23:39:58 Quit cooleo ("Azureus")
23:40:43eightbitI have only used fixed point a little in shaders
23:40:53eightbitand I have not dug deep into them
23:40:58pregloweightbit: it's quite simple, you do math as usual, but just need to keep track of your dynamic range at all times
23:42:17***Saving seen data "./dancer.seen"
23:42:37preglowi have a small plugin that generates a filtered sawtooth
23:42:44preglowbut i haven't really commented it well...
23:43:48eightbitbut it does it output the sound aswel?
23:44:02preglownope, writes to a file, it's from before rockbox on iriver did sound
23:44:11preglowwhen i tried to figure out how emac worked
23:44:36eightbitOh nice! I would like to see it if I could. Might be able to figure it out from you coding.
23:45:57preglowsure, gimme a sec
23:46:51 Quit Chamois (" HydraIRC -> <- Go on, try it!")
23:47:13eightbitanyway, in a float you go the exponant that tells you were the dot is. I guess in fixed that exponent is set leaving far more room for the mantissa?
23:48:15eightbitand calcuations thus are far simpler. I guess the processor can treat them just like ints
23:49:21preglowthe exponent is constant
23:49:37preglowfor the emac unit, the dot is always at the front
23:49:48preglowso your numbers can be in the range -1 to 0.99999999whatever
23:50:05preglowif you need numbers outside this range, you need to do some scaling before you multiply, then some scaling after the multiply
23:50:58eightbitI am a little off in my low level stuff... what bit sequence would give me -1?
23:51:29preglow0x7fffffff is 0.999999whatever
23:52:17preglowthere's dsptest.c, sinetab.h and emac.h
23:52:28preglowi don't even know if it makes good output anymore, haven't tested it in yonks
23:52:34eightbitThanks a lot!
23:52:46eightbitOutputs to a wave?
23:52:55preglowto a raw file
23:52:59preglow16 bit mono big endian
23:53:17preglowso you need to convert it somehow, i used adobe audition/cool edit
23:54:58 Quit cYmen (Read error: 113 (No route to host))
23:55:02preglowthink so
23:55:17preglowyeah, it's a define at the top
23:55:20eightbitAnother question...
23:55:29eightbitDo you know much about resampling?
23:56:06preglowi know something, at least
23:56:13eightbitit would be nice to eventually add a pitchshifter into my plugin
23:56:28preglowa pitch shifter doesn't really resample, it does more fancy stuff
23:56:49preglowunless you don't mind the length of your sound being altered as well
23:56:52eightbitno, just the basic one where it is linked to tempo
23:57:02preglowthat's simple, yes
23:57:24eightbitjust linear interpolation?
23:57:35preglowdepends on the resampling ratio
23:57:50preglowthat is, how much you want to pitch shift
23:58:02eightbitwell I doubt it would be much more than an octave either way
23:58:10eightbitIe doubling/halfing speed
23:58:18preglowwell, linear interpolation will sounds okish
23:58:21preglowno harshness
23:58:26preglowbut of course, it wont be high quality
23:58:50eightbitWell, we are talking about a step sequencer on an mp3 player!

Previous day | Next day