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 2005-03-17

00:05:05 Quit amiconn (Read error: 60 (Operation timed out))
00:05:05 Nick amiconn_ is now known as amiconn (
00:05:34hardeepis anyone actively working on the sound apis?
00:11:30hardeeppreglow: you're working on the codecs right?
00:12:53 Join Aison [0] (
00:22:54preglowhardeep: well, yes, when i have the time
00:25:15 Quit crash__ (Remote closed the connection)
00:25:39 Join crash_ [0] (
00:42:12*HCl smacks windows
00:42:43***Saving seen data "./dancer.seen"
00:45:37 Part MoosCamaro
00:46:18 Quit crash_ (Remote closed the connection)
00:47:13hardeephmmmm, does libmad require the stream buffer to contain complete frames?
00:47:23hardeepspecifically, can it handle an incomplete last frame?
00:48:22preglowit can handle incomplete frames of all kinds
00:48:23preglowwhy the last frame?
00:48:50preglowthat's is not done that way at all
00:49:03preglowyou always hand libmad complete frames
00:49:15preglowin the case of gapless playback, you cut away a part of the decoded output
00:49:19preglownot the input
00:49:29rasherthere you go
00:50:35hardeepi was thinking of a pipelined system where you pass the decoder fixed sized chunks at a time
00:50:50preglowyou don't hand the decoder anything, the decoder asks what it needs of you
00:51:03preglowat least i'm pretty sure that's the way we'll end up doing it
00:51:31preglowand no, fixed size chunks wont do, imagine what would happen in the case of a corrupt stream
00:51:41preglowthe decoder will need to eat lots of data quickly, to resync
00:52:08preglowand if the data you just handed it as a fixed chunk is not enough, then you get a nasty gap in the output
00:52:52preglowwe will most certainly need a system where the decoder can tell us how much data it needs
00:53:09hardeepdoes libmad do that?
00:53:16preglowall codecs does that, more or less
00:53:17hardeepfrom what i can tell, it just asks for more data
00:53:24hardeepnot an amount
00:54:03preglowyes, but it also tells you how much data it just produced
00:54:16preglowif that data is enough to fill the buffer, it has obviously gotten enough data
00:54:28preglowif not, you need to hand it more
00:54:31preglowit varies from codec to codec
00:54:33rasherplus we need (want) to support non-streaming formats
00:54:49preglowand imagine the case of vorbis, that needs to eat a lot of data at the start of each stream
00:54:51rasherwhich would just want the entire file
00:55:29preglowyup, we'll need to treat those cases differently somehow
00:55:31preglowif only subtly
00:55:31hardeepi'm guessing the non-streaming ones will access the file directly though, instead of needing us to load it into memory
00:55:43preglownah, we'll load it into memory
00:55:48preglowto minimize disk accesses
00:55:52hardeepwhat if it doesn't fit?
00:55:54rasherit will be anyway
00:56:00preglowthen there is a problem
00:56:02rasherthen we cry
00:56:10preglowif it doesn't fit at all in the buffer, there's nothing we can do
00:56:27preglowif it doesn't fit in the remaining buffer, we just wait with the loading until all the other data is consumed
00:56:41hardeepthat's why i was thinking of doing it the other way, where we don't load the entire file right away
00:56:56preglowthat would be wasteful, we have to load all the data we can all the time
00:57:17preglowthe only exception would be non-streaming codecs that can't fit their data in the remaining part of the buffer
00:57:58hardeepthe iriver's have 32mb of memory right?
00:58:19hardeepwell, i have mp3s that are larger than 32mb
00:58:26preglowthat's not a problem, they're streaming
00:58:30preglowwe can load just what fits
00:58:47preglowthe non-streaming codecs, like .mod, .xm, etc, is the problem
00:58:50rasherI seriously doubt there are many midi/mod/sid files out there larger than 32mbs
00:58:57preglowi have never seen a mod above 32 meg
00:59:04preglowand we just can't play those
00:59:24hardeepyeah, perhaps
00:59:30preglowstreaming codecs eat their input a chunk at a time, non-streaming codecs need the entire file at once
00:59:36 Join geoff_o [0] (
00:59:52preglowa Ãproblem will most certainly be non-streaming codecs that need to uncompress their data
00:59:56preglowthat will be a tricky problem
01:00:23preglowa very tricky one
01:00:42 Join crash_ [0] (
01:01:30preglowwe will perhaps have to just let those access the file directly
01:02:09preglowi can think of none that requires it off-hand, but there might be some
01:02:11 Quit matsl ("Leaving")
01:05:11preglowwavpack will be another tricky case, since it sometimes requires two files, but luckily, the wavpack developer himself seems interested, we'll let him handle that :P
01:06:19thegeekI don't get that, it should atleast be possible to put those two files into some kind of container
01:06:29thegeekhaving two files for every "song" is just wierd
01:06:55 Quit midk ("Leaving")
01:07:17preglowit's not that far-fetched, it can play the one file by itself, it will just be lossy
01:07:24preglowthe other file is there to complement it and make it lossless
01:08:07preglowdeed, not something you see everyday
01:08:15preglowhaven't tried it myself, but read about it briefly
01:08:27rasheroh god
01:08:44*rasher only now sees the parent thing in twiki
01:09:43rashernote the two latest edits
01:09:48 Join WeaZzz [0] (
01:10:00rasherand the "location" at the top of each page
01:11:04preglowi don't get it, everything seems normal
01:11:37rasherit registers other pages as "parent" of this
01:11:53rasherthere should be a tree listing somewhere...
01:12:52preglowi'm daft, i still don't see anything that worries me
01:13:01rasherwell it's not worrying
01:13:30rasherRockbox > Main > PluginIndex > PluginLogo
01:13:33preglowi don't see anything out of the ordinary either
01:13:40rasherrather than Rockbox > Main > PluginLogo
01:13:46rasherthat's how it was
01:14:38rasherbecause I hadn't set a parent
01:15:14preglowi misunderstodd everything you said in the last five lines
01:15:23rashernow I'm wondering how the pages are ordered
01:15:24preglowall's fine now
01:16:01preglowseems i should go to bed to recharge my brain
01:16:12*HCl reads up
01:19:03HCli'd almost say we shouldn't use an input codec buffer, and just let the codec handle that.
01:19:07HClbut ok.
01:19:32HClfor non streaming codecs, its almost pointless to have a codec input buffer
01:19:37 Quit Seed (Read error: 60 (Operation timed out))
01:20:13HClsince they need the whole file anyways, it barely serves a purpose?
01:20:33 Quit mecraw ("Trillian (")
01:20:34preglowwell, if we at one point had the disk spinning and was loading files anyway
01:20:36HClofcourse its nice to be able to feed music in various formats through our api, regardless of where its from
01:21:07HClah well.
01:21:16preglowbut we might have to comÃpromise
01:21:23preglowi'm thinking of various mod formats
01:21:36preglowthey need to unpack their pattern data, ok, they can't use it like its stored in the file
01:21:40HCli'm just thinking it would sometimes make more sense to give a codec a file it needs to play and its output buffer
01:21:46preglowbut the sample data as well would have to be decompressed
01:21:56 Join bagawk [0] (Lee@bagawk.user)
01:22:24bagawkHello amiconn & Bagder
01:22:27preglowi think we should always load the file to the mp3 buffer if it's feasible
01:22:44preglowbut i'm starting to see it's not really feasible for most mod formats
01:23:59 Join bippy [0] (
01:24:57HCli'm just thinking that for streaming formats, you can use an api like that, but for non streaming ones, there's really no point and you're better off just giving the codec the file its supposed to play and the buffer to which its supposed to write its output
01:26:12preglowwhen you're seeing you're about to load a non-streaming file, you could just tell the codec "hey, you'll be playing this file in the near future, load what you need to play it soon, you'll never see this file again"
01:26:43preglowyes, i think this i show we should do it
01:26:46preglowmost definitely
01:27:03hardeepwhy the "you'll never see this file again" ?
01:27:13preglowhardeep: so we don't have to spin the disk up needlessly
01:27:14hardeepthe prefetch idea sounds good though
01:27:30preglowhardeep: spinning the disk up is a design constraint, you have to do it as seldom as possible
01:27:47hardeepi just don't understand what you mean by that statement
01:27:56hardeepdo you mean that they only get once chance to load everything they neeD?
01:28:10 Quit crash_ (Read error: 60 (Operation timed out))
01:28:15hardeepis that really reasonable?
01:28:25preglowwhy not? can you think of an instance where it's not enough?
01:28:55hardeepnot enough buffer for whatever they need to do?
01:28:58bagawkHello hardeep I have not seen you in a very long time.
01:29:08 Join webmind_ [0] (
01:29:15preglowyes, then they need to return an error message, and the file will be retried when there is more buffer to spare
01:29:30preglowif it returns an error message when the entire buffer is free, then that file can not be played, plain and simple
01:29:53preglowbut that's the way it _has_ to be with non-streaming codecs, no matter how you do it
01:30:26hardeepnot if we remove the disk access limitation
01:30:31hardeep(for the non-streaming codecs)
01:30:39preglowi really think thats's non-negotiable
01:30:40hardeeplet them seek in the file if needed
01:30:53preglowno way, if they can seek however much they want, they still need the entire file in memory at once
01:31:06hardeepno, they can seek on disk
01:31:13hardeepit's expensive
01:31:21preglowso you propose a mod player reads it samples off of disk when it needs them?
01:31:23preglowthat will never work
01:31:32hardeepwhy not?
01:31:36preglowbecause the disk is too slow
01:32:02preglowa pÃc disk is slow, the disk in the 1x0 is slower by far
01:32:24preglowplease tell me if my sentences are clobbered by strange characters, i've got utf8 problems, it seems ;)
01:32:38hardeepi think that might be worth a test
01:32:45thegeeka pÃc disk i
01:32:47preglowi can tell you right now, it won't work
01:32:48hardeepi'm not all that familiar with mod
01:32:50preglowbu tfeel free to test it
01:33:13hardeepeither way, that's definitely farther into the future =)
01:33:23preglowindeed, but it's worth contemplating
01:33:26hardeepbut your idea of a prefect and pass a file sounds good
01:33:36preglowyes, i happen to agree :)
01:33:48preglowit solves the compressed file deal well also
01:34:30*preglow starts to put down codec api notes
01:34:35hardeepback to the streaming case for a sec
01:34:43HClyou don't want to read anythingfrom disk while playing music.
01:34:56preglowHCl: you can't, so it's a non-issue :)
01:34:59HClwith non streaming formats you just want to load the entire file into ram..
01:35:03HCl :P
01:35:04preglowthey playback would crackle
01:35:20preglownot only want, you have to
01:35:44preglowi'm quite firm in my belief that you can in no way stream data off the disk as you need it for non-streaming codecs
01:36:23hardeepdo we really want to deal with a dynamic buffer or would it be better to have a fixed size input buffer that we replenish when it goes below a certain size (similar to how it's currently done for the Archos players)
01:37:38preglowwell, we'll have a fixed size buffer as well
01:37:41preglowsized 32 meg
01:37:59hardeepbut the 32 meg is for both input and output no?
01:38:13preglowa tiny portion will be reserved for output
01:38:36hardeephmmm, the decoders are that fast?
01:38:37preglowhalf a second of audio will steal about 22000*4 bytes
01:38:58preglowthat's not much compared to 32 megs
01:39:03hardeepwhat about any post-processing we may want to do?
01:39:09hardeepmaybe normalization
01:39:15preglowwhat about it?
01:39:23 Quit webmind ("leaving")
01:39:29preglowin all cases but crossfade, that buffer will still be the same size
01:40:21hardeepi'm just trying to think of how large an output buffer we would need to make sure we can sustain playback
01:40:30hardeepwith all the processing that we're doing
01:40:48preglowthe codec alone should leave a wide margin
01:40:58preglowbut i don't really know
01:41:11preglowi'd like smaller buffers myself, to be able to do realtime stuff
01:41:20hardeepbut if the output buffer is small, the codec will be idle a lot of the time
01:41:37preglowthe codec will be active the same amount of time no matter what the output buffer size
01:42:00preglowbut longer buffer sizes are preferrable, so we can use our cpu_boost trick
01:42:10rasheryeah, I was just about to say
01:42:23rasheryou'd want longish chunks of cpu activity
01:42:30preglowcpu_boost eats 10 miliseconds, so we have to use at seldom as possible
01:43:26hardeepah, didn't know about cpu_boost, interesting
01:43:36rasherit takes a bit to kick in
01:43:43preglowand out, heh
01:44:08preglowmy guess is it'lÃl be fast to switch the freq up, but slow down
01:44:51preglowand as we've seen with rashers latest test, cpu frequency matters a lot
01:45:04preglowheaps more than i'd guessed
01:45:32hardeepresults in the irc logs?
01:45:38rasherresults in the wiki
01:46:03preglowsearch for 'graÃph'
01:46:18preglowyou'll find total running time at 120mhz vs. 48mhz
01:50:13rasherquite surprising
01:52:07thegeeknice work preglow:)
01:52:16preglowon what? :P
01:53:01bippycongrats preglow
01:53:03thegeekwrong guy;)
01:53:13thegeekI meant rasher ofcourse;)
01:53:47 Quit bagawk ("Leaving")
01:53:48preglowi don't mind the occasional pat on the back, just not used to it when i haven't done anything, heh
01:53:54rasherMy work wasn't a lot
01:54:39rasherthe real work was getting the battery status available, I didn't do that
01:54:41bippyhello ?
01:54:52rasherI merely printed it to a file and produced pretty graphs
01:54:55rasherhi, bippy
01:55:12*preglow brushes teeth
01:55:12bippyHi, just wondering if my connection had gone sorry
01:56:33bippyOh rasher, you said you werent at risk from volcanoes yesterday...
01:56:59pregloware there even _mountains_ in denmark? :P
01:58:01bippyThere doesnt have to be a so called mountain for a volcano
01:58:33rasherRisk level = 1
01:58:50rasher"No immediate risk"
01:59:02rasherI expect that to be "in the coming million years"
01:59:09bippyI know, thats the part that upset me
01:59:44HClyou can't have a volcano unless you're at the edge of one of the moving whatdoyoucallthemplates.
01:59:55preglowkiss your plans for eternal life goodbye, rasher
02:00:02HClwho wants to live forever anyways :)
02:00:03preglowyou just might be killed by a volcano some time "soon"
02:00:06HCl :p
02:00:40bippyTetonic plates ?
02:02:07bippyOnly caldera you should be worried about is yellowstone, will blow in the next 100,000 its debated wether its 40,000 overdue tho
02:02:35rasherI've never seen that happen
02:02:54rasherthe imagesearch suggestion
02:02:58preglowthat's new
02:03:00preglowi just saw it today
02:03:16preglowi use google a lot, and i've just seen it today, so think it is, at least
02:03:27bippyImage search?
02:04:05bippyImage search has been there for about 5 years :|
02:04:07thegeekkinda evil that preglow
02:04:15rasheryeah but it suggests
02:04:18thegeekthat it automatically shows relevant image
02:04:22rasherSee 676 image results for tectonic plates »
02:04:22thegeekI mean
02:04:31thegeekit would have to really suck bw for those two poor sites
02:04:32rasherit says on a regular search
02:04:46thegeekand some .gov
02:04:55bippyi got 758 images...
02:05:02bippyDont work so well eh
02:05:11rashergoogle is sortof inconsistant
02:05:29preglowoh well
02:05:39preglowi'll just let that imminent volcano be a surprise
02:05:51preglowsounds like great way to die anyway
02:05:52bippyWhy on earth did you notice somthing like that anyway :|
02:07:05bippyNot so imminent, although there has been a rise of 74cm in the last century in the area. Depends how bad solar flares are.
02:07:06rasherI'm used to how goole looks
02:07:25rasheror.. what are you on about now
02:07:38bippyAnyway, im off keep up the good work
02:07:54bippynvm lol too complicated
02:07:56 Quit bippy ("CGI:IRC")
02:09:40preglowi think i'll ignore it anyway
02:09:53rashercouldn't make sense of that..
02:10:04preglowi'd rather be kidnapped by green men, but hey, a volcano will do
02:10:25rasherI'm thinking spontaneous combustion
02:10:53preglowhaving high hopes, are we!
02:11:03rasherSays the volcano guy
02:11:06 Quit markun ()
02:11:25preglowno, if i'm going to burn, it has to be quick, hence the volcano
02:11:40rasherI'm sure there have been more spontaneous combustions here than volcano eruptions in the last 1000 years
02:11:41preglowspontaneous explosion sounds good, though
02:12:23preglowspontaneous teleportation
02:12:31preglowi'll get some shuteye
02:12:41preglowlater, all
02:12:57 Quit preglow ("leaving")
02:13:04rashergetting telefragged
02:16:33 Quit WeaZzz ()
02:20:32 Join StrathAFK [0] (
02:22:00 Quit Aison ("( :: NoNameScript 3.72 :: )")
02:26:16 Quit Strath (Nick collision from services.)
02:26:48 Nick StrathAFK is now known as Strath (
02:27:12 Quit Strath ("Client closed")
02:33:03 Join Seed [0] (
02:37:20 Quit tvelocity[away] ()
02:42:45***Saving seen data "./dancer.seen"
02:57:21 Quit Seed (Nick collision from services.)
02:57:29 Join Seed [0] (
03:07:59*rasher wades through misticriver to find faqs
03:16:28 Part geoff_o ("Kopete 0.9.1 :")
03:16:32 Quit Seed (Read error: 60 (Operation timed out))
03:17:58 Join crash_ [0] (
03:19:06 Join Seed [0] (
03:26:06 Quit Seed (Nick collision from services.)
03:26:13 Join Seed [0] (
03:44:00 Quit hardeep ("[BX] Uh oh... I'm getting a blip on the gaydar. Richard Simons is coming!")
03:44:01 Quit DMJC (Read error: 54 (Connection reset by peer))
03:44:33 Join DMJC [0] (
03:46:01 Quit lostlogic (
03:46:41NJoinlostlogic [0] (
04:05:22 Join QT_ [0] (as@area51.users.madwifi)
04:13:54 Quit QT (Read error: 60 (Operation timed out))
04:42:49***Saving seen data "./dancer.seen"
04:52:19 Quit ze (Read error: 104 (Connection reset by peer))
04:57:11 Join ze [0] (
06:00:02 Join midk [0] (
06:00:29 Join Strath [0] (
06:01:42 Quit Strath (Client Quit)
06:08:35 Quit pill (Read error: 60 (Operation timed out))
06:10:03midkwhy aren't snake levels included in the plugin?
06:32:02 Join Strath [0] (
06:42:50***Saving seen data "./dancer.seen"
06:57:19 Join pill [0] (
07:07:21 Join LinusN [0] (
07:10:32 Join mecraw [0] (
07:29:35 Quit mecraw ("Trillian (")
07:40:51LinusNdamn! it looks like we will have to resample 48000hz to 44100hz, since the hardware can't handle 48000 :-(
07:41:34LinusNexperiments show that the iriver firmware resamples *everything* to 44100hz
07:43:55LinusNunfortunately, we might not have enough cpu to to a really high quality resampling
07:44:43LinusNstill, we would have to resample anyway to listen to lower sampling frequencies on S/PDIF
07:46:25dwihnonot supporting 48000hz is bad :(
07:50:38LinusNindeed, but it's not the end of the world
07:51:27dwihnoMy dearest treasure torn away
08:13:32amiconnLinusN: Didn't I say that every mp3 player seems to have its design flaws? You answered that you didn't find any so far then, unlike with the archos. Now there's that 44.1 kHz restriction...
08:13:52LinusNand the slow cpu
08:14:05dwihnoSlow CPU?
08:14:09dwihnois it that slow?
08:15:55LinusNand the lame battery charging
08:16:49LinusNthey use a battery charging chip that is specifically designed to charge from usb
08:17:07LinusNbut they haven't connected the usb voltage to it
08:17:50LinusNand they use an external spi a/d converter for the buttons when there is an internal a/d
08:20:05LinusNthey bitbang the eeprom i2c when there are two i2c controllers that work just fine
08:20:30LinusNi can't understand why they didn't use the same bus
08:21:01LinusNto name a few design issues :-)
08:21:12LinusNdwihno: yes, the cpu is slow
08:21:21LinusNbecause is isn't superscalar
08:21:26 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
08:21:44LinusNso the low memory bandwidth seriously hampers the performance
08:21:57LinusNnot to mention the lack of a data cache
08:30:33amiconnThat's another issue that puzzles me. The 680x0 cpus do have a data cache since the 68020, and coldfire is 68k based...
08:30:55LinusN68k, as in 68000 methinks
08:33:12amiconnThen it wouldn't have an instruction cache either...
08:33:21LinusNtrue, true...
08:35:08amiconnA probably related question: You told that you'd better not enable page mode, beause a page "burst" is 5-4-4-4 (or such), and a burst mode burst is 5-1-1-1 (or such). Isn't it possible to enable both modes?
08:35:39amiconn...because if it is, enabling page mode would still give an advantage, when the data to be read is non-contiguous
08:42:51 Quit gromit` (Remote closed the connection)
08:42:54***Saving seen data "./dancer.seen"
08:45:57LinusNno, the mode setting is a bit, selecting either page mode or burst mode
09:01:55 Join einhirn [0] (
09:02:25 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
09:04:48 Nick Lynx_awy is now known as Lynx_ (HydraIRC@
09:13:16 Join Schnueff [0] (
09:15:16Schnueffgood morning
09:33:53 Join cYmen [0] (
09:43:28 Nick QT_ is now known as QT (as@area51.users.madwifi)
09:49:00 Join kurzhaarrocker [0] (
09:50:39kurzhaarrockerIf you are forced to resample - does that mean unnecessary power consumption -> lower runtime?
09:56:50LinusNin the long run, yes
10:00:42kurzhaarrockerMaybe there's a break-even point. Like "22kHz + resampling" is as expensive as "44.1 kHz without resampling" because of reduced data flow...
10:02:51LinusNthere are ways to conserve power, like only resample on "odd" frequencies
10:03:41LinusNfor instance, resampling is only necessary for the even frequencies, like 8000, 12000, 24000, 32000 and 48000
10:04:20LinusNand also if you want to output to s/pdif on other frequencies than 44100
10:04:57LinusNso we can play 22050 in the earphones without resampling
10:05:09kurzhaarrockeryes, I understand that
10:06:36LinusNworst case is probably downsampling 48000->44100
10:09:57 Quit Schee4 (Read error: 60 (Operation timed out))
10:11:16 Join Patr3ck [0] (
10:22:50*Bagder builds yet another cross compiler
10:23:06 Nick kergoth is now known as kergoth`zzz (
10:24:33 Join preglow [0] (
10:24:44preglowso it actually doesn't support 48khz
10:24:53preglownow, isn't that a kick in the groin
10:26:09preglowdoing high quality resampling will consume quite abit of cpu power
10:27:08preglowbut yes, the resampling method can be made selectable
10:27:44LinusNi think linear interpolation will do fine
10:27:49kurzhaarrockerWell - When there's _good_ 16 Bits 44.1kHz I wouldn't need 48 kHz anyway. What makes me shy away from the irvier thingies ist that they don't work on mignon AA batteries.
10:28:20BagderI don't see any AA-battery players these days
10:28:37preglowlinear interpolation will probably do for some instances
10:28:53preglowwell, it probably pretty much has to
10:28:58kurzhaarrockerMe neither - unfortunately I seem to be stuck with the archos
10:29:23 Join gromit` [0] (~gromit`
10:30:56kurzhaarrockerA few years ago I'd have said "Noone needs 48 kHz". But nowadays with DVDs audio at 48 kHz things are different.
10:32:31amiconnDAB also uses 48 kHz iirc
10:33:24preglownot supporting 48khz is stupid
10:33:43preglowmay iriver burn for it
10:34:11kurzhaarrockerIt could be worse - imagine 48 kHz was the only frequency.
10:34:20preglowhey, i like 48khz ;)
10:34:24 Nick Strath is now known as StrathAFK (
10:42:57***Saving seen data "./dancer.seen"
10:54:31LinusNit's not really iriver's fault
10:55:25LinusNthe mcf5249 has a very limited set of dividers for the audio clock
10:55:57LinusNso you really can't have both frequency sets available
10:56:34LinusNeither you have the 11.025, 22.050, 44.100 etc, or you have 8.000, 12.000 etc
10:56:39preglowthey chose the frigging thing
10:57:00LinusNi think freescale were extremely lame when they designed the 5249
10:57:28LinusNwho could imagine that they made it so limited
10:57:36preglowit's most certainly not as leet as i thought
10:57:49LinusNi feel very disappointed
10:58:06LinusNit had so many possibilities "on paper"
10:58:13 Join SeeSchloss [0] (
10:58:24LinusNbut not when you examine it under the hood
10:59:24LinusNbtw, philips is not entirely against releasing their code, but it will probably end in some legal matters
10:59:42LinusNi wouldn't hold my breath
10:59:44preglownot entirely against releasing the codec code?
10:59:49preglowi did not see that coming
11:00:14LinusN"It is possible that we could share our embedded software with our customers, but it is not a usual practice"
11:00:31LinusN"our customers" being the key word here...
11:00:44LinusN"The sales team would have to know the details of what you plan on doing, etc. I'm sure there would be a legal agreement before any such source code could be divulged."
11:00:48kurzhaarrockerWell compare the lack of sample frequencies with the design flaw that forced bitswapping in the archos..
11:01:07preglowi am also sure that will be the case
11:01:46LinusNi've been straight with them, explaining that we are doing open source development
11:02:12LinusNbut i think he still misunderstood, thinking that we cooperated with iriver
11:02:39Rick"If he didn't ask..."
11:03:02LinusN"I would suggest that you get in touch with iRiver and find out who their Philips Semiconductors Sales person is that they worked with to get the SAA7750 designed in. You should then work with that person to see if it is possible to obtain the source code and the possibility of sharing the internal programming features."
11:04:00preglowyes, us open source firmware developers often enjoy good cooperation
11:04:33LinusNi'm not sure i want to contact iriver about it
11:05:09LinusNat least not until we have a track record on the H-series players
11:05:12preglowi think it'd be pretty futile anyway
11:05:33preglowthat's true
11:05:53*LinusN is listening to Anna Ternheim on his iriver with rockbox
11:06:44LinusNjust a modified version of the wav playback hack
11:06:48*preglow applaudes
11:07:01LinusNi'm writing the pcm playback api
11:07:40preglowthe codec api will be tricky, btw
11:07:41kurzhaarrockerincluding peak meter? :)
11:07:45preglowin case you didn't know that :)
11:08:32LinusNkurzhaarrocker: the peak values will have to be "manually" calculated
11:09:02kurzhaarrockerCould that be done while resampling?
11:09:05*kurzhaarrocker hides
11:09:22preglowwe won't always resample
11:09:23preglowbut yes
11:09:36preglowshouldn't be too hard anyway
11:09:40kurzhaarrockersorry, I was just kidding
11:10:01LinusNkurzhaarrocker: we can get coarse peak info from the mp3 decoding as well, can't we?
11:10:11LinusNi mean preglow
11:10:52LinusNbetter get in on a pcm level anyway, so we can do it independent of the codec
11:10:58preglowhmm, i don't know, the info would have to come from some place after the subband recomposition
11:11:33LinusNi've been reading up on psychoacoustic bass generation
11:11:48LinusNmissing fundamentals and stuff
11:11:51kurzhaarrockerIf you have to output pcm I don't see any reason why anyone would wants to find peaks anywhere else.
11:12:18LinusNkurzhaarrocker: because it takes cpu
11:12:51LinusNinteresting, the uda1380 has a silence detector
11:12:58LinusNand automatic gain
11:13:29kurzhaarrockerIs there docu that describes how the silence detection works?
11:14:29LinusNit looks pretty dump
11:14:41LinusNnot usable for recording triggering
11:14:49kurzhaarrockersimple threshold?
11:15:53LinusNit triggers when it sees X samples with digital 0
11:16:05LinusNpretty useless
11:16:32LinusNX being 3200, 4800, 9600 or 19200
11:18:12preglowthat's useless
11:18:36kurzhaarrockeryes, thats complete crap
11:19:02preglowi wonder how much cpu we'll have left for dsp
11:21:00Hadakaoh btw, I have a problem that some of my MP3:s do not play over the digital output with Archos Recorder 20 (and rather recent Rockbox) - I think it is simply because the sampling rate is something wacky, like 22kHz or 11kHz, but can somebody confirm this?
11:22:54kurzhaarrockerIt's been a while since I last used the digital output of the archos.
11:23:38LinusNHadaka: s/pdif can only play 44100 and 48000 on the archos
11:24:11LinusNand we can't resample, even if we wanted :-(
11:24:21HadakaLinusN: thank you, that's what I wanted to know
11:26:59preglowi'm wondering how we'll handle stuff like mp3's in wav files and cue sheets
11:28:00 Join [av]bani [0] (
11:28:35LinusNpreglow: we'll deal with that later
11:28:40LinusNlunch time
11:28:49preglowyes, sure
11:29:24preglowcodecs will have to load their own files
11:29:59LinusNimho, the codecs should be fed
11:30:16preglowwe discussed this last night, that goes poorly with non-streaming formats
11:30:18LinusNthe codec wouldn't know where the data comes from
11:30:48preglowlike .xms, which will have to unpack both pattern data and sample data from the file
11:31:03LinusNthe codec wants a read() function
11:31:04[av]bani <- doesnt that just inspire confidence ;)
11:31:15preglowonce that's done, it doesn't need it anymore, and requires the same amount of space again for unpackad and processed data
11:31:17 Join midk_ [0] (
11:32:19preglowmacromedia hasn't ported flash to amd64 yet, so web sites are flash free whether i want it or not these days
11:32:20LinusNmy point is that the codec doesn't need to know about cue sheets, or wav-file containers
11:32:24preglowahh, nio
11:32:25preglowno <-
11:32:38preglowsome meta-codec will have to handle that
11:33:06preglowand yes, they should most definitely just have a read hook they get from us
11:33:20preglowbut they should also be able to request opening of other files
11:33:33preglowlike soundfonts and patches for midi, the two file system of wavpack
11:36:27amiconnLinusN: Slight correction: s/pdif can play 48000, 44100 and 32000 on the archos
11:42:31[av]banitimidity is pretty cpu intensive
11:43:49[av]baniif the cf has bandwidth problems to main memory, timidity might be out of the question
11:44:49 Join midk__ [0] (
11:45:10 Quit midk (Read error: 113 (No route to host))
11:45:58 Quit midk_ (Read error: 104 (Connection reset by peer))
11:48:27preglowwho mentioned timidity?
11:48:43preglowif you're referring to the midi engine, that's being written from scratch
11:50:29[av]baniyes, again it may be out of the question if the cf has memory bw problems
11:50:41[av]banihow much bw does it have?
11:50:52 Join midk [0] (
11:50:52 Quit midk__ (Read error: 113 (No route to host))
12:02:57 Join MoosCamaro [0] (
12:03:06MoosCamaroHey all
12:09:03 Quit StrathAFK (Read error: 110 (Connection timed out))
12:14:18preglow[av]bani: it has 32 megabuytes
12:14:51preglowit'll be fun to see how the sample based codecs perform
12:15:12[av]baninot total memory
12:15:27preglowi misread...
12:15:35preglowlow bandwidth
12:15:38preglowhaven't got a number
12:16:30preglowdepends on how you use the ram as well
12:23:52 Quit DMJC ("Leaving")
12:26:32 Quit Heidelweg (Read error: 110 (Connection timed out))
12:27:51 Part [av]bani
12:39:42 Join ghode [0] (
12:40:26 Quit ghode (Client Quit)
12:42:59***Saving seen data "./dancer.seen"
12:56:35rasheroooh, sidplay2 is all integer math?!
13:03:22preglowthere's not reason for the emulation engine to use floating point
13:03:25preglowand resid is all integer
13:03:38rasherwell you never know
13:04:27rasherresid is c++
13:05:02rasherand also, "don't expect a complex SID emulator like this to run on a machine with less than 233-300 MHz"
13:05:29preglowthe emulation engine doesn't have to be cycle accurate
13:05:36preglowresid is c++?
13:05:41preglowit's been a while since i've had a look at it
13:06:30preglowaghh, bummer
13:06:41preglowcause resid would fit perfectly otherwise
13:07:40rasheryou'd probably need to cut a lot of corners
13:08:18preglowwell, not in resid
13:08:22preglowat least i don't think so
13:08:31preglowthe major hurdle would be the cycle accurate emulation of the 6510
13:08:42preglowand it doesn't have to be that accurate
13:09:32*rasher inspects
13:11:51rasherit is indeed c++
13:12:18rasherbut from a quick glance, it look spretty C-ish
13:14:05rasherconfigure:#include <float.h>
13:14:12rasherwhat does that mean?
13:14:19rasher(I just grepped for float :)
13:15:58rashersidplay is also cpp
13:16:22rasheruh.. nevermind
13:17:06rasherthis was the actual sidplay
13:17:27 Join Patr3ck_ [0] (
13:19:41preglowi can't think of a reason why it should use floating point
13:20:33LinusNfloating point makes life easy sometimes
13:20:45LinusNi can imagine that they use fp for the filters
13:21:02preglowthey don't
13:21:05preglowit's fixed point
13:21:14preglowi know, 'cause i did a floating point version of it ;)
13:25:11rasherexcellent then.
13:25:23rasherand the c++ is only minor it seems
13:25:25preglowit's a very simple filter
13:25:47preglowand surprisingly difficult to emulate properly
13:26:50preglowwhich is why i love it so much
13:26:58*preglow strokes his sidstation
13:27:54rashergentoo still installing..
13:28:18preglowthat's what gentoo does
13:28:36rasherespecially when running on a 100MHz laptop
13:29:08rasher(actually 400mhz, but speedstepping is messing up)
13:30:49preglowi prefer binary stuff on my desktop
13:30:54preglowserver's running gentoo
13:32:00rasherI'm desperately hoping that some optimizations will make a difference
13:32:09rasherbecause this laptop needs all the help it can get
13:32:21preglowwoah, i just emerge sync'ed, and this popped up: app-emulation/coldfire/files/
13:32:32preglowme try
13:32:40 Quit Patr3ck (Read error: 110 (Connection timed out))
13:33:03rasherStill waiting for a cellphone with a 6581 chip.. now THAt would be excellent ringtone-melodies :)
13:33:08rashernone of this midi crap
13:36:55preglowgod knows how many times i've wished for the very same thing
13:37:14preglowsweet pulse wave arpeggios calling my attention
13:37:18preglowthat would be something
13:38:11rashersettled for midi renditions of c64 tunes :-\
13:38:20 Quit Rick (Read error: 60 (Operation timed out))
13:50:25 Quit Schnueff ("leaving")
13:50:28rasherthe battery graphs are nearly identical
13:50:57 Join Aison [0] (
13:50:59rashereven down to the longer period around 222
13:52:56LinusNrasher: it would be interesting to see how long it lives in 11MHz
13:53:14LinusNand how long it lives in 120MHz without the spinning disk
13:53:15rasherI was afraid you were going to say that :)
13:53:25rasherbut yes
13:53:29rasherI'll get right to it
13:53:34LinusNpcm sound api works fine here
13:54:23rasherso to get 11mhz I just make sure never to call cpu_boost() right?
13:54:35preglowrasher: yes
13:54:45LinusNyes, or call set_cpu_frequency(0)
13:55:12preglowany thoughts on whether metadata handling should be in the codecs?
13:55:46LinusNi don't think it should
13:55:49preglownor do i
13:56:04dwihnometadata API?
13:56:12preglowhaving to load a codec just to look up title/artist is not very clever
13:56:15LinusNon the other hand, the codec knows best how to extract it
13:56:36preglowindeed, and it will sometimes be necessary to have the codec do it as well
13:56:38preglowseek tables, etc
13:57:01preglowand replaygain data
13:57:26*rasher waits for the iriver to fully charge again
14:00:38rasherwell that should take over 5 hours possibly :-\
14:01:23 Join ABerg [0] (
14:02:58ABergHi all !
14:03:27preglowwhy, hello
14:03:41 Join Heidelbaer [0] (
14:03:56ABergI have reading the IRC logs since serious work has begun on the iRiver port
14:04:09ABergGreat work !
14:04:57preglowmost of us agree on you with that
14:05:05 Join R3nTiL [0] (~zorroz@
14:05:37rashermisticriver really seems full of people unable to navigate a wiki
14:05:59ABergNow, I have had several ideas on the codec audio API and tried to organise them in a design document
14:06:56preglowABerg: goodie
14:07:13ABergCould some old-timers have a look at it and tell me if its worth putting in the Wiki
14:07:27ABergor if I am completely off..
14:09:25ABergSorry for the bad english, I am swiss
14:09:26rasherI'm in no position to comment on it, but lots of thoughts won't hurt
14:09:33BagderABerg: the loader loads compressed data, not uncompressed
14:10:00ABergYeah right of course
14:11:20BagderI like it a lot
14:11:34Bagderhaven't read all of it yet, but it surely is a very good start
14:12:23Bagder(there's always a but)
14:12:36preglowone gripe, the codecs will probably have to read from disk, not a buffer
14:12:40Bagderthe loader can't be activated by the codec
14:12:54ABergWhy ?
14:13:18Bagderpreglow: ugha, we don't want that
14:13:33Bagderwe want to be able to load N songs at once
14:13:47preglowBagder: i mean the codecs will have to do the loading to the mp3 buffer
14:13:54preglowBagder: they know best what they need loaded
14:14:15Bagderbut that'll be... ugly
14:14:18preglowthis is a must in the case om non-streaming codecs like mods
14:14:22Bagderwhen 5 songs using different codecs are loaded
14:14:26preglowthey can't use the data as it is
14:14:26 Join IRCMonkey [0] (
14:14:42preglowthey have to unpack pattern data, unpack samples
14:14:59preglowbetter to have that done at load time
14:15:05IRCMonkeyMorining. Any chance to get rockbox to work with creative shit?
14:15:15preglowIRCMonkey: perhaps, get specifications
14:15:22BagderIRCMonkey: if you are skilled and have lots of time
14:16:07preglowBagder: yes, i'm pondering if we might not need to split the codecs up somewhat, into a loader part and an actual codec part
14:16:23preglowBagder: the loader part would be loaded in memory all the time, and handles file and metadata loading
14:16:35preglowboth those things will be needed often
14:16:46ABergBagder, preglow: the idea is that the codecs could give hints to the loader
14:17:11Bagderwe want to fill the whole buffer with songs, at load-time
14:17:13ABergso that it can adopt the best reading strategy
14:17:38preglowBagder: a simple codec caching system where the user can select which codecs he wants to use the most might be neat at some point
14:18:37BagderABerg: my point being that the loader only needs to know when the codec has N bytes left, then it starts to fill the buffer again
14:19:20ABergThere might be several file in loader cache, some that have not even started decoding
14:19:23IRCMonkeyCreative firmwares are real crap for real. Best of luck with your development. Bye bye
14:20:29preglowcreatives products are mostly crap
14:21:56IRCMonkeyyep but cheap
14:22:48rasherthere really are an awful lot of patches
14:23:01 Join DMJC [0] (
14:23:18ABergIt is really when the loader detects that its cache is nearly empty, regardless of which track/codec is currently playing, that it must refill
14:23:18IRCMonkeyg2g byez
14:23:19 Quit IRCMonkey ("Chatzilla 0.9.67 [Firefox 1.0.1/20050223]")
14:23:39BagderABerg: in my head it is
14:24:45 Part MoosCamaro
14:25:10rasherweird.. still charging
14:26:34rashermaybe it takes a while to realize the battery's full
14:26:46 Part kurzhaarrocker
14:29:11rasherbattery test running..
14:30:41rashereta, ~20:00
14:31:34 Quit preglow (Read error: 104 (Connection reset by peer))
14:32:51 Join preglow [0] (
14:32:56*preglow kicks his server
14:35:33rasherI thought of a good one last night
14:35:45rashergetting telefragged would be cool way to go
14:36:11preglowmost definitely
14:36:24*preglow suddenly misses quake
14:37:31LinusNHCl: you there?
14:38:13HClwebmind_: not window 19, screen window 1
14:38:32LinusNjust curious how the rockboy sound emulation works
14:38:48webmind_HCl, screen ?
14:38:50HClummm. it has this pcm_submit function.. there's a little documentation on it.. let me fetch it...
14:38:53HClwebmind_: yea.
14:39:01webmind_HCl, you have switch screens to switch from channel?
14:39:08HClwebmind_: servers, yes.
14:39:13rasherwow, this old ipaq forgot everything after lying without power for a few weeks
14:39:20webmind_HCl, hm offtopic hold on
14:39:51webmind_HCl, see sceen window 2?
14:40:11HClLinusN: hmm, there's a document on it thats not in cvs, dunno why, let me check it in.
14:41:19HClwebmind_: in a bit, hold on.
14:41:40HClThe main sound module interfaces with the system-specific code through
14:41:40HClone structure, pcm, and a few functions: pcm_init, pcm_close, and
14:41:40HClpcm_submit. While the first two should be obvious, pcm_submit needs
14:41:40DBUGEnqueued KICK HCl
14:41:40HClsome explaining. Whenever realtime sound output is operational,
14:41:46HClhm, too large a paste
14:41:54HClcheck out cvs
14:42:00HClopen up the HACKING documentation
14:42:03HCland go to the sound subsection..
14:42:27HClrockboy already emulates sound
14:42:30HClits just not submitting it
14:42:32CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
14:42:32*LinusN is reading the HACKING file
14:42:34HClto the dsp
14:42:41preglowHCl: so it'll be just as fast as it is now?
14:42:45preglowHCl: kick ass
14:42:51HClaside from the delay needed to play the sound
14:42:54preglowHCl: kick ass in the extreme, as a matter of fact
14:43:00HClits probably too slow to play sound properly, as of yet
14:43:01***Saving seen data "./dancer.seen"
14:43:07HClan assembly interpreter core should fix that.
14:43:19preglowwhat, you given up on dynarec? :P
14:43:33HClwell. not yet.
14:43:45HClbut seeing the speed as it is at the moment with a c interpreter core
14:43:54HCldynarec isn't very needed yet.
14:43:58preglowbut yes
14:44:02HCleventually it might be nice to get it to run at a lower mhz
14:44:05preglowfaster core, and put some stuff in iram
14:44:12HClmaking an interpreter core in assembly
14:44:17HClthen expanding it with dynarec eventually
14:44:21HClwould probably be more productive
14:47:09BagderHCl: add the file in apps/FILES too
14:47:34Bagderit controls what goes into the tarballs
15:00:25 Join t0mas [0] (
15:06:59preglowLinusN: what size buffers do you use?
15:07:34t0maspreglow? you think Linux know's more about the timer trick we talked about?
15:07:48LinusNpreglow: size buffers?
15:07:53t0masOh.. sorry... linuS not X :)
15:08:14LinusNhehe, i even write linux myself sometimes :-)
15:08:39LinusNtimer trick?
15:08:59 Quit R3nTiL (Read error: 60 (Operation timed out))
15:09:03t0masWe were thinking about making the coldfire cpu run at the lowest possible speed
15:09:12t0masbecause rasher found out it saves a lot of power
15:09:20t0masand that way making a wakeup clock
15:09:38LinusNi and björn talked about that over lunch
15:09:42preglowLinusN: why yes, for passing audio to the driver
15:09:43t0masbut we could nowhere foud the minimum speed
15:09:54LinusNminimum speed is 11mhz
15:10:02t0massure? can't get lower?
15:10:10preglowwouldn't be minimum if it could
15:12:03LinusNt0mas: that's the frequency of the external oscillator
15:12:28t0masok, and is there a way for rasher to get it at that speed from within a plugin?
15:13:27t0mashow? I only found the boost mode option...
15:14:10 Join Mong0 [0] (
15:14:19LinusNpreglow: the pcm playback api has no opinion about the pcm buffer size
15:14:31LinusNyou give it a pointer and a length
15:14:48LinusNand a callback
15:19:41LinusNregarding the cpu speed, i and björn have an idea about letting the cpu go to the idle frequency when it has absolutely nothing to do
15:20:09t0masok, sounds intresting too
15:20:11LinusNand go to the normal frequency when the user presses a key
15:20:24LinusNor anything else that would need more cpu
15:20:41LinusNso we would have 3 states: IDLE, NORMAL and BOOST
15:21:04HClwouldn't idle be enough while users are browsing for music?
15:21:14LinusNprobably not
15:21:20t0maspreglow and I were talking about a timer mode, so disable lcd and hd... get cpu to minimum... and then just wait for a specific time
15:21:34LinusNdisable lcd?
15:21:34preglowyes, that'd be a great idea, and btw, you probably know more about plls than me. wouldn't you guess that the 10ms worst case lock time is for when it is set to 11mhz?
15:21:59LinusNgoing to 11mhz is fast
15:22:01t0masyeah, isn't it possible to clear the screen and make it stop updating?
15:22:10LinusNt0mas: when and why?
15:22:20LinusNpreglow: the 10ms is for the pll to lock
15:22:25t0masfor the timer mode... to wait for say 7 am to wake me up ;)
15:22:28LinusNthe pll is disabled in 11mhz
15:22:37preglowLinusN: yeah, i know, i just thought it would lock faster at higher clock rates
15:22:37LinusNt0mas: aaaaaaaaah
15:22:54LinusN11mhz doesn't need locking at all
15:23:21preglowno, pll is probably disabled then, but is what i say generally true?
15:23:38LinusNi doubt it
15:23:52LinusNit all depends on the pll itself
15:23:54t0masanother one...
15:23:59t0mas21.20.57 # <rasher> B8
15:23:59t0mas21.21.07 # <rasher> C2 now :-\
15:23:59DBUGEnqueued KICK t0mas
15:23:59t0mas21.21.09 # <preglow> b8?
15:23:59t0mas21.21.17 # <preglow> it shouldn't ever reach b8
15:23:59***Alert Mode level 1
15:23:59t0mas21.21.17 # <preglow> heh
15:24:01t0mas21.21.37 # <rasher> it did in the last test as well
15:24:03t0mas21.21.51 # <preglow> queer, linux said the range was from c0-ff
15:24:07LinusNt0mas: saw that
15:24:49preglowyou could have pasted me correcting 'linux' as well!! :P
15:25:03t0maswanted to paste as less as possible...
15:25:32preglowyea, i'm just kidding
15:27:01t0masbut is there a simple explanation for the B8 value?
15:27:22HClevil gnomes.
15:27:24CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
15:27:24*HCl nods.
15:27:38preglowterrible, fearless trolls from the cold north
15:29:05LinusNt0mas: the adc range is 00-ff
15:29:31LinusNthe b8 value is simply representing the voltage where the hardware fails to run
15:29:36 Quit Mong0 ("CGI:IRC")
15:30:07t0masoh ok, and your's failed at C0 ?
15:30:12LinusNnot mine
15:34:00***Alert Mode OFF
15:35:28rasherit went even lower
15:35:38rasher9A I think
15:36:33t0mas9A = 100 isn't it?
15:36:35rasherwould probably go lower yet without harddisk activity
15:38:56preglow9a is far above 100
15:39:32t0masoh lol
15:39:39t0mas9 * 16 + 10 != 100
15:40:14t0masdamn... since the calculator I got a little slower calculating myself :)
15:42:26preglowdamn gnome calculator doesn't do conversion!
15:42:30preglowi'll kill it
15:43:02Bagderuse google ;-)
15:43:31t0masyeah, this time i did it right ;)
15:43:36rasherpreglow: I know!
15:43:38rasherit's horrible
15:44:12rasherI wonder what the reasoning behind that is
15:44:30preglowBagder: hahahah
15:47:39 Part LinusN
16:17:43 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
16:18:49 Join Rick [0] (
16:31:49 Join mecraw [0] (
16:33:15 Join markun [0] (
16:33:22HClhey markun
16:33:35HClmy flatmate didn't have an optical in, only out :/
16:33:44markunMy flatmate has one :)
16:33:49markunI didn't know it
16:33:58HCli guess you'll be retrieving your cable again then, hm?
16:35:59 Quit ABerg (Read error: 110 (Connection timed out))
16:36:00markunI'm not sure if his set shows the sample rate
16:36:27HCli thought i saw something about linus saying 48000 wasn't done..
16:36:53markunLinus this morning said experiments showed only 44.1 was possible, do you know what experiment?
16:37:32preglow44.1 and 22.05 and 11.025
16:37:37preglowall the multiples
16:37:58preglowno need to do the optical thing any more
16:38:02preglowhe's found out what he needs to know
16:38:27markunHCl: I'll come by to fetch the cable anyway.
16:40:01preglowmarkun: any plans working on the grayscale patch? :)
16:43:02***Saving seen data "./dancer.seen"
16:51:46 Join tvelocity [0] (
16:52:12rasherbattery level still E4 ...
16:52:24rasherthis'll take ges
16:54:05preglowit'll probably have to stay on tonight
16:54:10preglowthis is with hard drive on?
16:54:54rasheryes and 11mhz
16:55:04*preglow goes woot
16:55:32tvelocitystill hacking on the battery?
16:55:40 Join Stryke` [0] (~Chairman8@resnet-241-86.resnet.UMBC.EDU)
17:01:57markunpreglow: don't know yet. I will try to make lcd_bitmap a bit faster.
17:02:11rashertvelocity: just collecting graphs
17:06:21preglowwhat's slowing it down?
17:07:49rashergetting the last functions done would be much nicer :>
17:08:04preglowi think linus wants it commited
17:08:14preglowslowness and all
17:08:43markunok, I will implement the missing functions first then.
17:11:34 Nick tvelocity is now known as tvelocity[away] (
17:28:51rasherwait wait wait...
17:29:13rasherpreglow: I don't think cpu is that important
17:29:15preglowwork work
17:29:32rasherit's just that the hdd activity is spread out
17:29:45rasheror not
17:33:19 Join ABerg [0] (
17:33:21 Join stevenm [0] (
17:33:55stevenmhey preglow, what was the code for that cubic spline sampling algorithm again?
17:34:53preglowi haven't got it off-hand, but i bet you can find a code snippet at
17:35:35preglowlike that, for example
17:35:47rasheremerge system finished!
17:36:04stevenmrasher, woohoo! Now onto X :_
17:36:06stevenmpreglow, thanks
17:36:24rashernow I really really need to build a kernel with toshiba support so I can use the last 3/4 of the processor speed
17:36:38stevenmrasher, a laptop with frequency scaling ?
17:36:48stevenmrasher, emerge cpuspeedy
17:37:02rasherthis is non-standard toshiba crap
17:37:08rasherneeds toshiba-utils afaik
17:37:16rasherand a kernel option
17:37:33stevenmeew. And Intel cpu performance states dont work with it?
17:38:36rasherI don't think so
17:38:48rasherbeen running at 100MHz until now :-\
17:38:54stevenmrasher, eeeew
17:39:13stevenmrasher, echo something > /sys/devices/system/cpu/.../cpufreq ?
17:39:26preglowstevenm: the same guy has written a good, big paper on interpolators if you're really interested ;)
17:40:04rasheris there really no middle-ground between gentoo-dev-sources and gentoo-sources?
17:40:17rasherI'm an idiot
17:40:19rasherignore me
17:40:28stevenmpreglow, yea, my roommate has this huge BOOK on it. but books don't exactly have a section titled "for optimizing it for an iRiver, use this code"
17:40:33tvelocity[away]gentoo-dev-sources are 2.6 sources
17:40:53preglowstevenm: why cubic spline, btw? have you tried linear?
17:40:55rasheryeah, I read the version wrong
17:42:10tvelocity[away]i really think they should stop calling 2.6 sources dev-sources. it's confusing IMO
17:43:44rasherit'll be fun to see if I can remember how to configure a kernel for this old beast
17:43:49stevenmpreglow, stripwax tried it but he said it sounded like crap
17:43:57preglowshouldn't sound like crap
17:44:02preglowbut of course, hermite will sound better
17:44:30stevenmpreglow, I just commented out my ghetto lowpass filter... it sounds much better than when you muffle the sound.. but high string instruments have this creepy sampling noise
17:44:45stevenmhermite? What would you recommend for iriver ?
17:45:27preglowwhat i pasted you is hermite
17:45:54preglowi would recommend linear interpolation for iriver
17:46:21stevenmhmm.. so how would you DO linear?
17:46:31stevenmAll I dont understand is, where do I get the 'fractional' value ?
17:47:02preglowand btw, you can't do it like that, of course
17:47:05preglowthat assumes floating point
17:47:27stevenmyes ... that's a bit of a problem
17:47:33preglowthe fractional value should be part of your sample playback iterator thing
17:47:39preglowi've made a linear interpolator using the emac
17:47:54stevenmso, what would fractional be in terms of the delta value ?
17:48:16preglowwell, that depends on how you've done it
17:48:21stevenmlike, there's current position, and the delta value.
17:48:34stevenmand each time, it adds delta to current position, and outputs that sample.
17:48:35preglowyou shift current position right some places before you index the sample, yes?
17:48:46rasherI think we need battery tests without hdd activity
17:48:49stevenm9 bits I htink
17:48:52preglowyes, the bits you shift away are the fractional bits
17:49:09preglowi do linear interpolation there
17:49:22preglowand you see how i extract the frac part
17:49:25stevenms=getSample(wf, (so->cp>>10))
17:49:29stevenmthats my code
17:49:33stevenmaah, thanks.. makes sense
17:50:03preglowi have a 25 bit frac part, since a sine table really only reaquires a 7 bit index
17:50:19stevenmso, then, I'd have to AND the position with 2^10 - 1 and that would be the 'fraction' ?
17:50:29stevenmoh snap
17:50:38stevenmI have a class in 30 min.. and no breakfast yet
17:50:51stevenmI'll be back in a few hours to finish this. Thank you for the info
17:50:55preglowgo have your breakfast
17:51:08preglowto do the actual math, you need to know some fixed point, though
17:51:11preglowi wish i knew of a good guide
17:51:20preglowi don't have time to write one right now
17:51:27preglowi'm going home this easter, and am packing
17:51:28rasherpreglow: I have a feeling that the spread-out reads cause the longer battery life
17:51:38preglowspread out reads?
17:51:54rasheryeah, because of the longer processing time
17:52:09preglowof course
17:52:13rasherso reads happen less often
17:54:00preglowwill be fun to get the results from the nodisk read test, though
17:54:06preglowyou've got your work cut out for you ;)
17:54:10preglowbut anywho, more packing
17:54:50 Part stevenm ("'food, class'")
17:54:52 Join elinenbe [0] (~elinenbe_@
17:56:49rasheryes, nodisk will be more interesting
17:56:52rasherand a longer test :(
17:58:38 Quit Patr3ck_ ()
17:59:37rasherJESUS CHRIST
17:59:41rashergentoo uses devfs?!
18:01:12preglowit is udev ready
18:01:20rasheryeah.. but..
18:01:36rasherI mean even debian uses udev
18:01:38preglowyou can smack udev in in a jiffy
18:01:41preglownot stable
18:03:54 Quit DMJC (Read error: 54 (Connection reset by peer))
18:04:30 Join DMJC [0] (
18:04:47 Join wacky_ [0] (
18:05:34 Quit mecraw ("Trillian (")
18:06:32rasheroh my
18:06:33rasherstill at E0
18:06:48preglowhow long's it been running?
18:06:54rasher3:40 hours
18:08:59rasher5 hours wasn't nearly enough
18:09:03rashermore like 8 hours
18:09:26wacky_still no testing of the 1.65 firmware ??
18:09:57preglownone that i know of
18:10:05wacky_I can put a ready-to-use .hex file to someone with a BDM to test.. in the case where compiling the thing would be the problem
18:10:22preglowthat's not a problem
18:10:42preglowus not seeing a reason to spend time testing it would be the problem
18:11:29wacky_it's a question of minutes, isn't it ? if the thing boots or not, isn't that all ?
18:11:46preglowthere's only one of us that has a bdm interface, and he is very busy
18:11:54wacky_:P yeah I know :) heheh
18:12:10preglowi'm not about to test it because of features i really don't need
18:16:54 Nick Lynx_ is now known as Lynx_awy (HydraIRC@
18:20:08elinenbeI only run rockbox on my iriver now... screw the iriver cruft
18:20:44wacky_elinenbe - do you use your iriver to play music ?!
18:21:05wacky_what's the use if you don't use the original firmware ?
18:21:17wacky_since rockbox doesn't play anything yet
18:21:51*wacky_ thinks, maybe elinenbe is a Sokoban freak?
18:25:42 Join _Lucretia_ [0] (
18:34:06elinenbewacky_: is correct!
18:43:07***Saving seen data "./dancer.seen"
18:49:27amiconnpreglow: elinenbe actually wrote rockbox' sokoban
18:49:59preglowdidn't know that
18:52:09 Quit pill (Read error: 110 (Connection timed out))
18:52:38 Join einhirn [0] (
18:57:45 Quit wacky_ ("see ya")
18:58:26 Join Strath [0] (
19:02:32amiconnpreglow: Btw, this was also before rockbox was able to play music - on the archos
19:02:50preglowyes, games do seem to be prioritized ;)
19:07:34HClyou can play gameboy games with it :P
19:09:16amiconnI just ordered a H-140 :)
19:10:08preglowlooking forward to seeing the ridiculously optimized lcd driver
19:10:21HClwith the grayscale lib?
19:10:24amiconn49-level greyscale...
19:14:13HClshould be nice.
19:14:18HCli kind of want to look at dumb
19:14:26HCli guess i'll go do that
19:14:28HClmy anime can wait
19:14:29HCl :)
19:21:23preglowgood call
19:24:03rasherI think dumb is... dumb
19:25:48HCli really want mod playback
19:25:58HClpreferably with lots and lots of useless stats
19:26:39rasherafaics dumb only supports like 5 formats
19:28:54 Join webguest62 [0] (
19:29:43 Join Mong0 [0] (
19:31:48webguest62hey ami are you still there somewhere? i read that you had ordered a h140. where from? i search all the bloody shops and on the net too but all are out except ebay which I'm not sure I trust has a h120
19:33:58rasherand h140
19:34:18rasherI... think, let me check
19:34:22rasheroh.. no
19:34:29webguest62last shipment entering uk i think went to advanced mp3 players or they got some of them
19:35:02rasherh120 is in stock
19:35:06webguest62i checked the common websites like amazon advanced (/) mp3 players all 3 and no h140
19:35:11webguest62i used to have an h140
19:35:49webguest62decided i wanted a colour screen of h340 and put it up for sale. the day i decided i needed to play music only i came home from work and some bugger had bought it!
19:36:13rashercouldn't you cancel the sale?
19:36:20webguest62so i thought lets get a h340 instead otherwise buying another h140 i had sold made no sense
19:36:25webguest62dont think so
19:36:28webguest62amazon marketplace
19:36:52rasherwell you could've asked him privately :)
19:36:59webguest62hmm didnt think of that :(
19:37:33webguest62wish i had kept it then i could go and try rockbox even in current non music playing state
19:37:43rashershops in denmark have them it seems
19:37:55rashernot sure if any of them ship internationally
19:38:00rasherwhere are you, btw?
19:38:06webguest62newcastle uk
19:38:30webguest62hmm whats the thing to let me change my name? or do i have to register a nick first?
19:38:36HCldumb is handy.
19:38:48HClit doesn't play mods by default, but writes a .pcm file instead
19:38:56rasherwebguest62: /nick
19:39:02webguest62been years since i used mirc
19:39:06HClseems like i just need to get it to compile on the iriver
19:39:14 Nick webguest62 is now known as dukezed (
19:39:31rasherHrm.. the only shop with a h140 lists delivery time as "over 7 days"
19:39:39dukezedin denmark?
19:39:40rasherwhich, knowing this shop, means out of stock
19:39:56rasherat least, pretty likely
19:40:42dukezedmaybe ill pop down to the shops in the next town during weekend. the clueless sale assistants probably been flogging ipods to ppl like my workmate who chose it because 'its so small!'
19:41:00dukezedso hopefully some irivers still in stock
19:41:11rasherquite possibly
19:41:49dukezednot to say the ipod is a bad player but choosing a player because it is small is stupid
19:42:09dukezedif thats the only reason
19:42:36rasherwell if that's the most important to you
19:43:54rasherthe ipod has 3 things going for it
19:44:00rasher1) other people have it
19:44:03rasher2) design
19:44:05rasher3) size
19:44:16rasherI think that's it :-\
19:44:22rasherat least from my point of view
19:44:29dukezedit does what it does well, though I do not like the itunes tie-in
19:44:32 Join prpplague [0] (
19:44:40dukezedor the marketing. why not show us what it does
19:45:21dukezedinstead of showing some colourful 'hip' ppls silhouettes jumping around with ipods on
19:47:01prpplagueamiconn: found an irc log of #rockbox where you guys were discussing using mmc/sd cards via spi mode
19:47:13prpplagueamiconn: you guys still doing some dev on that?
19:50:01dukezedhmm well im off to play eq2 until the time comes to make dinner
19:50:21amiconndukezed: It seems no shop has H1xx in stock here in germany too. I ordered it from ; they tell a delivery time of 2..14 days
19:51:22dukezedk thx i will check that site out
19:51:26amiconnprpplague: The archos Ondio uses MMCs via spi. The MMC driver in rockbox is mostly my work
19:52:22prpplagueamiconn: is the socket mmc only? or does it support sd cards as well in the socket?
19:53:29amiconnThe socket is suitable for MMCs and thin SD cards. Ordinary sd cards don't fit, as they are 2.1 mm thick compared to 1.4mm of MMC (and thin SD)
19:53:49prpplagueamiconn: ahh ok
19:54:25amiconnThe usb bridge supports both mmc and sd, and spi mode seems to be so similar that several users reported sd cards working.
19:54:26prpplagueamiconn: who's in charge of the mmc code?
19:54:49prpplagueamiconn: yea, the sd and mmc in spi are identical except for the init sequence
19:55:01dukezedouch its all in german ill probably end up ordering the wrong things due to not knowing much german.. hmm i wonder what price it is. must be euros? damn uk customs probably charge me vat for it unless they will try listing it as a gift?
19:55:10prpplagueamiconn: the sd uses a different card init command
19:55:53rasherwill you get charged for a trade inside the eu?
19:55:57amiconnprpplague: Maybe in spi mode the command is identical, otherwise it won't work (I didn't implement anything special for sd support)
19:56:38*HCl slaps the codecs makefile
19:56:43dukezedhmm to be honest not sure. things ive ordered from america there was possiblity of it
19:56:44HCli'm not getting it to execute make in my dumb dir
19:56:59prpplagueamiconn: hmm, not from what ive seen, the mmc uses CMD1 whereas the sd is suppose to be using ACMD41
19:57:01amiconndukezed: Yes, the price is in euros. Here's the direct product link:
19:57:08dukezedrasher: but ive never ordered from another eu country
19:57:31rasherI ordered my h120 from the uk and didn't get charged.. but..
19:57:38rasherI don't know if I might've
19:57:38dukezedk im there already ami
19:57:38amiconnprpplague: It seems that this is actually the same command, only different naming
19:57:47amiconnMMC CMD1 is 0x41
19:58:33*prpplague doubles checks his notes
19:59:08 Quit NibbIer (Read error: 113 (No route to host))
19:59:44dukezedhmm what does "Ihr Warenkorb ist noch leer." mean? "your basket is empty"?
19:59:49 Join Nibbler [0] (
19:59:50 Join NibbIer [0] (
19:59:52 Join xNibbler [0] (
19:59:53amiconndukezed: yup
19:59:57 Quit NibbIer (Read error: 104 (Connection reset by peer))
20:00:00 Quit Nibbler (Read error: 54 (Connection reset by peer))
20:00:49dukezedami:im kind of worried ill end up ordering the wrong things or wrong order details even.
20:01:02dukezedah so menge must be # i want
20:01:05dukezedi left it blank
20:01:09rasherdukezed: according to the Danish customs website you don't have to pay for things bought in another eu country (at least not if you were visiting it)
20:01:18amiconnyes, "Menge" is the number
20:02:04dukezedbut for things coming into the country which i didnt have on person because it was sent ?
20:02:23rasherI can't find that... but shouldn't that be the same?
20:03:23rasherwell you never know..
20:04:14dukezedi was under impression there was a difference between bringing something into the country thats been used, the customs cant tell anything about history of where you bought it etc, and taking something new back importing it into the country?
20:04:32rasheroh, I missed an entire "internet purchases" category
20:04:47rasherwell this was for "new" things
20:05:59dukezedhmm "einkauf fortsetzen" and "zur kasse"
20:06:09dukezedok lets click one
20:06:56amiconndukezed: "zur kasse" is what you probably want
20:07:18rasherright, this looks to me like there's no customs paid on purchases inside the EU, and VAT in seller's country
20:07:28 Join matsl [0] (
20:07:35rasherat least for Danes.
20:07:48rasherbut guessint it'll be the same for other eu citizens
20:08:10dukezedSie sinde bereits Kunde?:email pass needed Ihre erste Bestellung?:email only needed
20:08:37amiconn"Ihre erste Bestellung". Then you need to give some more data on the next page
20:08:48amiconn("your first order")
20:09:04prpplagueamiconn: ahh the mmc/sd code in rockbox is really good
20:09:14dukezedhmm must also find out what euro into pounds is
20:09:27dukezedim sure last time i checked it was 1.5 ish?
20:09:44HClits really kind of scary
20:09:47HClto how clean dumb compiles
20:09:49HCleven for iriver
20:10:12amiconnprpplague: :)
20:10:15HClokay, so i have a few undefined references to read, write, lseek and exit..
20:10:17rasherHCl: wow, it Just Worked?
20:10:30HClrasher: yea, it compiled completely clean, aside from a few undefined references
20:11:19dukezedhehe, i think ill choose 'herr' for my 'anrede' but what does 'firma' mean if i chose it?
20:11:31amiconndukezed: british pounds (£) ?
20:11:37rasherthat you're a company :-\
20:11:42rasherI'm guessing.
20:11:44elinenbeHCl: what are you talking about?
20:11:54rasherHCl: mikmod!!!1!!1!!!!ONEONE!!!
20:11:58dukezedy ami
20:12:07amiconndukezed: "Herr" means mister, "Firma" means company
20:12:07HClit compiled completely clean.
20:12:17HClthe undefined references were just for the example program
20:12:51dukezedk. vor = sur name nach = first name?
20:12:54elinenbeHCl: what compiled?
20:13:01rasherother way round
20:13:03elinenbeI was just reading the logs and was a little lost...
20:13:05amiconndukezed: the other way round
20:13:13dukezedthx for your patience
20:13:19rasherwee, I'm beating the german to it :)
20:13:28HCldumb, its a tracker format player
20:13:32HClits said to be really good
20:13:32amiconnNachname = surname, Vorname = given name
20:13:42HCli have a feeling it won't run at realtime though
20:13:56dukezedhmmm before i go any further do they actually say they will deliver to other countries?
20:14:02elinenbehow many people have the h140 over the h120?
20:14:13elinenbeit seems that the h140 was a pretty rare player
20:14:16t0mashm.. I have a H120
20:14:23rasherI think it's about 2:1 in favour of h140
20:14:26t0masand I only know people having a 120
20:14:38rasherat least it was when people were flashing first
20:14:51elinenbeI have a h120... is there anywhere to even buy an h140 in America?
20:15:10t0maselinenbe: you have a black one?
20:15:39amiconndukezed: They ask for "Land" (country) within the form your are at, so I guess the answer is "yes".
20:15:53dukezedStraße /Hausnr.* = street house number Postleitzahl (Zahl ohne Leerstellen)* = post code Stadt* =state/county Land* = country..? correct me if im wrong
20:16:14 Quit DMJC (Read error: 54 (Connection reset by peer))
20:16:15rasherstadt = city
20:16:15 Quit Mong0 ("CGI:IRC")
20:16:21elinenbet0mas: yes I do.
20:16:23amiconnrasher: yup
20:16:35 Join DMJC [0] (
20:17:21amiconndukezed: The light blue part you only need to fill in if you want the delivery to go to a different address
20:17:22dukezedso street housenumber/postcode?/city/country
20:17:38prpplagueamiconn: very nice code indeed
20:17:50amiconndukezed: yes
20:17:53dukezedi see gro'B' britannien on there with the funny 'b'. thats me then i guess
20:18:00rashersure iss
20:18:14dukezedwhat do you mean by blue bit?
20:18:17rasherfunny b is a double-s
20:18:27dukezedi see grey boxes below if thats what you mean
20:18:41dukezedi know that. gross /great?
20:18:59 Join hardeep [0] (1098@
20:19:13amiconnrasher: ypuip?
20:19:19rashertypo for yup
20:19:27rasherquite impressive, if you ask me
20:20:41dukezedwhats Zahl ohne Leerstellen after the Postleitzahl mean?
20:20:56dukezedbefore i press the next button :)
20:21:00amiconnThat means digits without spaces
20:21:20amiconn(German postcodes are 5 digits)
20:22:03dukezedk so i can hopefully assume they will ignore that rule since they are sending it to another country
20:22:29 Nick kergoth`zzz is now known as kergoth (
20:23:06amiconndukezed: 350 EUR == 243.23 GBP
20:23:23dukezed3 options
20:23:51 Quit tvelocity[away] (Remote closed the connection)
20:24:02dukezedhmm even less than 1.5 euros
20:25:57amiconnRechnung == invoice, Vorkasse = pre-payment (?), Nachnahme = COD
20:26:19amiconninvoice isn't possible for first-time orders
20:26:21dukezedCOD = ?
20:26:28amiconncash on delivery
20:26:44amiconnPretty common in germany
20:27:03dukezedso they want to trust me first before offering invoice and only if i pay beforehand
20:27:33dukezedok vorkasse it is
20:28:29 Join tvelocity [0] (
20:30:32dukezedhehe the address label unsurprisingly follows the german convention of postcode before city, i wonder if i can switch the information in both fields so it is more like the uk
20:30:54dukezedi pressed the button it has taken me to a screen with the following seems to be some final message
20:32:14dukezedVielen Dank für Ihren Einkauf bei TheFractionPrice. Abhängig von der Wahl der Zahlungsart haben wir Ihnen eine Email für die Zahlungsmodalitäten zugeschickt. Bei Rückfragen zu Ihrer Bestellung geben Sie bitte immer die Bestell-Nr.: *somenumberhere (dunno if its important)* an. Mit freundlichen Grüßen Ihr TheFractionPrice - Team
20:32:23HCltitania:/home/hcl/rockbox/rockbox-devel/iriver# ls -al libdumbd.a
20:32:23HCl-rw-r−−r−− 1 root root 102234 Mar 17 20:31 libdumbd.a
20:32:28 Join Sucka [0] (
20:32:33HClseems to have compiled cleanlny
20:32:34 Quit tvelocity (Client Quit)
20:33:51amiconndukezed: You should have received an eMail as well. Bestell-Nr. means order number. You need it for reference
20:34:11 Quit Liehbeth ()
20:34:14*HCl is talking to the author of dumb a bit...
20:34:35dukezedno option exists to take me forward to a method of payment?
20:34:41amiconnIf you chose paying beforehand, there should be given a bank account where to send the money
20:34:49dukezedfrom within email or am i misunderstanding 'prepayment'?
20:34:58dukezedyes i chose vorkasse
20:35:12amiconnyes, in the mail
20:35:20dukezedhmm bank account to 'send the money'?
20:35:51amiconnYes, for prepayment
20:35:58amiconnI chose cash on delivery so I don't know exactly how this email looks like
20:36:01dukezedif you are talking 'western union' kind of sending doesnt seem very secure
20:36:42dukezedunless i can pay with credit card i wouldnt be going through with this i think. but i need to check email out
20:37:52amiconnI don't know whether they offer credit card payment. In germany it is quite common to transfer the money from one bank account to another. Dunno how this is called exactly in english, but certainly has nothing to do with 'western union'
20:38:01dukezedhmm sells 2 2nd hand ihp140s for 220 and upwards... but condition might be crap
20:38:33dukezedor just use creditcard like with say amazon?
20:38:34preglowrasher: how's the battery doinÃ' ?
20:39:02dukezedhmm sorry i should of payed more attention to "In germany it is quite common to transfer the money from one bank account to another"
20:39:35dukezedoh brb
20:39:36 Quit dukezed ("CGI:IRC")
20:41:56 Join webguest02 [0] (
20:42:17 Nick webguest02 is now known as dukezed (
20:42:56dukezedyep you were right ami
20:43:02dukezedbank acc details
20:43:08***Saving seen data "./dancer.seen"
20:44:14dukezedi should know, i see some foreign cheques sometimes working in the inland revenue
20:44:18amiconnI searched a bit; didn't find a more precise word than '(money) transfer'. You will need (or should use) IBAN and BIC within the EU
20:44:19dukezedhmm how to proceed
20:44:46dukezediban and bic..? money transfer services like paypal but with bank accounts?
20:45:24rasherpreglow: oh, it shut off!
20:45:31amiconnHmm, I don't know about paypal either.
20:45:41 Join stevenm [0] (
20:45:48preglowrasher: time?
20:45:56rasheruhm.. not sure
20:45:59rasherlet me see
20:46:01dukezedhmm not sure what those iban and bic references are for
20:46:08stevenmpreglow, I just got the linear thing wokring.. did it fast, using double math, and it sounds all right.
20:46:13dukezedthere are two within the email 1 of each ami
20:46:27stevenmnow gonna see if I can get it working w/ integer only
20:47:15HClwhats the proper way of adding a dir in cvs
20:47:20HClcvs add isn't working recursively
20:47:27HCli want to add dumb to cvs
20:47:29amiconndukezed: International Bank Account Number (IBAN), Bank Identifier Code (BIC)
20:47:43preglowstevenm: if the fractional part is less than sixteen bits and the samples are sixteen bits or less, you can even do it without the emac unit
20:47:46rasherpreglow: not a lot of points though :(
20:48:21stevenmpreglow, emac unit? Samples are 16 bits, and the fractional part is 10 bits
20:48:28amiconnHCl: You first need to add the dir explicitly, then add the files within
20:48:30preglowrasher: that's to be expected
20:48:53preglowstevenm: the linear interpolator thing i coded uses emac
20:49:01preglowstevenm: but you won't need that, then
20:49:26stevenmpreglow, I see.. I just had a huge discussion w/ a friend of mine about it, in class.. haven't even read thru that code
20:49:55 Join pill [0] (
20:49:58stevenmah, emac, I see. not sure what that does though
20:49:59amiconndukezed: It seems although they offer shipment to other countries, they didn't really think about the consequences/ necessities. PM me if you want
20:50:04preglowstevenm: then there's no need to do so
20:50:25HCli'm gonna checkin a clean dumb 0.9.2 codec, ok?
20:50:27stevenmah all right
20:50:29preglowstevenm: it just multiplies and adds, nothing magical, but your numbers are sufficiently low in resolution, you don't need it
20:50:30HClit compiles really cleanly
20:50:39HClits untested and straight code of the 0.9.2 release
20:50:39preglowgo ahead
20:50:47preglowdoes it use floats?
20:50:53HCli don't know.
20:50:54HCllet me ask
20:51:11HCli don't think it does, or it would've had problems compiling, right?
20:52:29dukezedami: thx i think i will go explore the other options fully before doing this now. like i said earlier there is still the other shops in the town close by perhaps they have some. all i need is 1 :) maybe hidden behind some boxes on a shelf just waiting for me to come along
20:53:18preglowHCl: gcc can make fpu emulation code, it might be doing that
20:53:27amiconnHCl: gcc handles fp, even on archos/sh1, but then the fp lib gets linked to the binary, and it's sloow
20:53:37HClmmm, i think its not using the proper compiler flags either.
20:55:44preglowno matter anyway, if it's using fpu code, we'll replace it
21:02:35rasherpreglow: battery wikipage
21:02:41preglowrasher: excellent
21:03:15pregloweven longer battery time
21:03:37rashercan't that just be because the harddisk draws less power?
21:04:00rasherI think we'll need a test without hdd
21:04:49HClmeh, i think i'll delay committing this. wasn't compiling it with the proper coldfire flags
21:05:23preglowrasher: of course we will :P
21:05:36preglowrasher: linus ordered a 120mhz test with no hd this very morning
21:05:59rasherdo I have to do *anything*?
21:06:15rasherdoes it draw more power if I set it to calculate things
21:06:33preglowi think he meant the same decoding loop you do now
21:06:41preglowjust with no disk accesses apart from the first load
21:06:41rasherbut that does hdd..
21:06:52preglowand your occasional writing
21:06:59preglowbut i don't know
21:07:00rasheryes.. let me see..
21:07:04rasherwell sounds likely
21:07:45t0masrasher? is that low power test finished?
21:07:57t0mashow long did it run?
21:08:03rasherbut it's hard to say much
21:08:04stevenmpreglow, okay, linear interpolation works with integer math.. and sounds dang nice too
21:08:14rasherbecause we're using the hdd
21:08:19stevenmgonna try it at higher pitched notes now
21:08:27preglowstevenm: cool
21:08:30rashernot really sure how much the time differs because of hdd activity
21:08:41t0masno, but what was te running time?
21:08:51rasher5:30 I think
21:09:05amiconnrasher: I think you need to keep the cpu busy for the no-hd test. The rockbox kernel puts the cpu to sleep if it has nothing to do
21:09:10t0maswell... that's almost twice the time at 120 mhz isn't it?
21:09:28rashert0mas: yes.. but I'm thinking it might be because the harddisk is working less as well
21:09:36 Quit _Lucretia_ (Read error: 110 (Connection timed out))
21:09:39amiconnAnd my guess is the cpu doesn't draw significantly more power when sleeping at 120 MHz compared to sleeping at 11 MHz
21:09:42rashereach read has 10 minuts between them
21:09:46preglowrasher: only one way to find out!!
21:09:51rasherI know...
21:10:28rasherlet me see if I can pull out the writes and have it only open once
21:11:10t0maswhat's steven writing?
21:11:14t0massound improvement?
21:11:22rasherI think
21:11:32stevenmyea, MIDI
21:11:42preglowstevenm: better yet, make a reeeeeally low pitched sound, and see if it really connects the samples with lines
21:12:13stevenmtried it on the really really really high strings file... worked perfectly, and without the ghetto lowpass filter to muffle it
21:12:36preglowHCl: what did you do to compile dumb?
21:12:38*rasher stares
21:12:45preglowHCl: making it fit in the codec make system is pretty easy
21:12:47HClpreglow: i just removed it, heh.
21:13:01HClit uses floating point
21:13:05preglowfor what
21:13:07stevenmpreglow, hmmm... let me try that. gotta make a midi file tho
21:13:16HCl20:50 < HCl> does dumb use floating point?
21:13:16HCl21:03 < entheh> it does, but not in the inner loops
21:13:18HClno idea.
21:13:21HClentheh = developer.
21:13:24 Join tvelocity [0] (
21:13:26HClefnet, #dumb
21:13:37rasherwait... I'll be spinning the harddisk up each time I write the battery info
21:13:59preglowrasher: yes, of course, but choose a long mp3
21:14:14rasherhow long?
21:14:20preglowi dunno
21:14:28preglowlong enough to get a decent amount of samples
21:14:30rasheroh DAMMIT
21:14:32rasherthat won't matter
21:14:51rasherI'll just move the statuswriting outside the decoding loop of course :)
21:14:52preglowremember that libmad is about twice as fast without disk activity as it is
21:15:11rasherbut I was writing status each time the buffer filled
21:15:16rasherno need for that
21:15:32rasheroh, I'm not
21:15:56*preglow pats rasher. as long as you figure it out in the end
21:18:01rasheris any writing happening in xxx2wav?
21:18:07 Quit ABerg ()
21:18:08rasherapart from opening the file
21:19:56 Join _Lucretia_ [0] (
21:20:58 Join hi [0] (
21:21:54 Part hi
21:21:59rasherjust so you know, I'm NOT going to run a 11mhz test without hdd :)
21:24:50preglowof course you are, of course
21:25:10stevenmpreglow, I just tried the cubic/hemite one.. linear produces less noise.
21:25:32preglowstevenm: then you're doing something wrong
21:26:12stevenmpreglow, I copied the code down, exactly.. and all using double math. It produces more of a 'background hiss'
21:26:16rasherpreglow: it'll run for days :<
21:26:59*rasher emerges some more
21:27:03preglowstevenm: i'd appreciate some samples
21:27:09rasherI wonder how many days until I have a running system
21:27:14stevenmpreglow, of code or of sound ?
21:28:28preglowstevenm: sound
21:29:01stevenmpreglow, all right.. this will take some time
21:29:37dukezedwell i am registered my interest to buy off ebay a h140
21:30:05dukezedguy says he used it on his honeymoon but since then has been busy with his wife, whether i believe him or not i dont know :)
21:31:17rasherhe'd have to make it up to her for nerding it out on their honeymoon
21:31:57stevenmaww great, wine is broken. what say you guys we port Goldwave to linux? :-P
21:32:11dukezedso i told him i was a previous owner of a h140 (kind of suggesting "if you lie i will know about it!" without actually saying it to him) and asked him some questions
21:32:32*HCl needs an assembly cpu core for gnuboy...
21:32:48*HCl goes to make another attempt at dumb
21:39:03HCldoh, i'm dumb :p
21:40:15stevenmpreglow, stevenm/mi1_nointer.wav">
21:40:18stevenmpreglow, stevenm/mi1_lin.wav">
21:40:22stevenmpreglow, stevenm/mi1_cubic.wav">
21:40:36stevenmpreglow, watch out, each is like, 9 megs in size
21:43:59preglowno problem
21:44:41stevenmpreglow, there's also an mi1.mp3 in that directory.. it uses no interpolation, but has a ghetto lowpass filter, which gets rid of the noise (and half the sound with it)
21:45:29 Quit SeeSchloss (Connection timed out)
21:46:13 Join SeeSchloss [0] (
21:47:43HClokay, i got dumb to compile for real now, with the proper flags
21:48:09preglowstevenm: linear one MIGHT be better
21:48:13preglowstevenm: have you fixed guard samples?
21:48:16rasherHCl: wee
21:49:23rasherI'm waiting for the battery to charge.. :(
21:49:26preglowlinear one will need a guard sample as well, but just one
21:49:47stevenmpreglow, what is a guard sample ?
21:50:32HCli probably broke out hell, and lots of daily builds are gonna get broken that i need to fix and stuff *hides under the couch*
21:51:03 Join prpplague^2 [0] (
21:51:07HClbut preglow said he could fix the floating point stuff of dumb
21:51:11HCland it compiles for iriver for me
21:51:16 Join MoosCamaro [0] (
21:51:25MoosCamaroHello all
21:51:36 Join webguest23 [0] (
21:51:49 Quit dukezed ("CGI:IRC (EOF)")
21:52:02MoosCamaroYou have got very good work
21:52:10 Nick webguest23 is now known as dukezed (
21:52:20HClrockbox should be sweet once we get all the codecs to work..
21:53:04MoosCamaroI see the iriver port status and irc, and congratulations
21:53:19rasherjesus.. emerge −−help is really dumb
21:53:26rasherit's called manpages, write them!
21:54:03dukezedwith something like portal player is it possible to do all the things you do with rockbox right now, including the things you are making for iriver or is it only due to the irivers cpu that you can do more things?
21:54:14preglowstevenm: if you're at the end of the sample, the interpolation will read outside it
21:54:25preglowstevenm: since it does buf[i + 1], buf[i + 2]
21:54:57preglowstevenm: get it?
21:55:16preglowunless it never reaches the end thanks to loop points, of course
21:55:44preglowHCl: i just hope you don't expect it to be done fast, i'm going away for a while now
21:55:52HCllol. nah.
21:55:53preglowi'm going to 28k8 land for about a week and a half
21:56:02preglowhence, i won't be using the internet
21:56:15stevenm28k8 ?
21:56:30dukezeddodgy modem?
21:56:40preglowyes, you know, a modem, the old noisy buggers you had to use before you got your adsl or fiber connection :)
21:56:48*amiconn is already annoyed by using isdn speed (64k)
21:56:54dukezedsound like a spectrum
21:56:58dukezedloading games
21:56:59preglowi'm on 100mbit, imagine the contrast
21:57:04stevenmpreglow, hey now.. my "laptop" has a 300 baud modem. with only pulse dialing
21:57:19preglowgod, that's _got_ to be old
21:57:25stevenmTandy 102
21:57:31tvelocitypreglow, please never mention your net connection again when i'm online :P
21:57:41stevenm40 x 8 LCD, 4 AA batteries, 1Mhz CPU
21:57:54dukezedwhat speed is yours tvelocity
21:58:18HCl40x8 lines, i hope.
21:58:22HClas in text, not pixels
21:58:25tvelocity384k down 128k up
21:58:37dukezedhmm almost same as me
21:59:00tvelocityyeah but you don't pay a fortune for it:P
21:59:09*t0mas is at 320KB/s down and 70 KB/s up
21:59:09HClwell, i have 100mbit up/down :)
21:59:17preglowtvelocity: what, my 100mbit full duplex connection?
21:59:19t0masHCl: you are at studentnet?
21:59:27HCland markun has 100mbit up/down too
21:59:28HCl :)
21:59:38t0masyeah, me too in one year ;)
21:59:40tvelocitywill you all die please?
21:59:43preglowwith no firewalling and as many ips as i like?
21:59:45dukezedonly maybe 33 euro/month
21:59:46preglowok, i won't mention it
21:59:52HCltvelocity: sure! but not just yet :)
21:59:55preglowit's probably one of the bestconnections in the world
22:00:20t0maspreglow: university connection?
22:00:35preglowi pay about 15$ a month for it
22:00:36preglowt0mas: indeed
22:01:04t0masno datalimits?
22:01:15HCli have a 15gb/week upload limit
22:01:20t0mascuz that's the problem with some university connections here...
22:01:26HClbut on wireless, with 11mbit, i have no limits
22:01:31t0masHCl: that's anti warez ;)
22:01:39preglowt0mas: i have recently gotten a 10gig/day limit
22:01:44HCl10gig a day!
22:01:50preglowin + out
22:01:53preglowso 20 gigs total
22:01:54HClthats not much :/
22:02:08preglowi can easily break it, but i like my connection
22:02:10preglowso i wont
22:02:13t0masyeah, my father works at the university of Utrecht, and they recently set a 10 GB / month upload limit
22:02:25preglowthat, though, is not much
22:02:33t0masjust because of bittorrent... and other uploading tools
22:02:45*HCl is running bittorrent.....
22:02:48preglowyes, this is a warez nest
22:02:49t0masunlimited down... at 100 mbit... but the uploadlimit...
22:02:57HClohh. preglow, wanna link with our warez net? :P
22:03:02preglowwe used to have unlimited up and down here, but people abused it to an amazing degree
22:03:09t0masyeah, there too
22:03:13t0masFTP's running...
22:03:22t0masuploading warez...
22:03:26preglowsome people had the 100mbit maxed out 24/7
22:03:26preglowthat's bad
22:03:28HClwe have unlimited down, and practically an unlimited up, and we still abuse it :)
22:03:31HClpreglow: true.
22:03:41t0masHCl: tutwente?
22:03:50HClwell, 15gb up by standard line
22:03:59HClunlimited up/down with wireless, 11mbit
22:04:14HClthough there's recently some 54mbit wireless section
22:04:25*t0mas is thinking about study-ing in delft :)
22:04:36HCli heard delft sucks warez-wise
22:04:38t0masBut I'll have to check the internet connection first :P
22:04:46t0masyeah, heard the same
22:04:50HClthey have 100mbit, but no sharing network like we do
22:05:08t0masyeah, maybe leiden is better...
22:05:10HClso files only get shared on a per-flat basis
22:05:12HClwhich is just crap
22:05:19HCli'm not supposed to talk about this stuff.
22:05:23preglowstevenm: i really think you've got a guard sample problem, yes
22:05:26HClwe don't really have a sharing network here
22:05:35t0masvpn ;)
22:05:43preglowHCl: oh, we've got a hub going for the entire student "village" here
22:05:49preglowno limits internally
22:05:52t0masjust connect all your firewalls/routers together with vpn...
22:05:54HClpreglow: wanna link? same thing here :P
22:05:57t0maswe do it in highschool too now ;)
22:05:59 Quit prpplague (Read error: 110 (Connection timed out))
22:06:12preglowHCl: well, then the limit would kick in, now wouldn't it :P
22:06:18preglowi never use it anymore anyway
22:06:23t0masDC++ ?
22:06:40t0mashow many gig's normally shared?
22:07:39 Nick prpplague^2 is now known as prpplague (
22:07:40HClpreglow: do you mind if occasionally i ask whether there's a certain thing on your network if we don't have it? :p
22:07:45preglowHCl: sure
22:07:49HClthanks :)
22:08:00HClwe have 100mbit anyways, its not like it clogs our connection
22:08:03preglowdon't know, as i've said, i never use it anymore, the odd chance i want something, i usually join a great, big swedish hub that's got 100s of terabytes
22:08:11stevenmpreglow, all the samples out of range are automatically set to 0
22:08:16preglowstevenm: that's not correct
22:08:33preglowstevenm: when you read past the loop point, it should wrap around to the loop point start
22:08:40rasherI just inserted a random cd because gentoo was barking at me when rebooting
22:08:50rasherso when the laptop came back up, it booted the cd
22:08:54rasherturns out it was RedHat 5.2
22:09:40rasherImagine my chok after having installed gentoo for days, and when it comes back up after the first reboot, it tells me "Welcome to RedHat 5.2!"
22:10:02stevenmpreglow, yea, I suppose so.
22:10:02stevenmpreglow, that'll take some tinkering, as there's like, 4 different looping modes
22:10:18rasherwow, the kernel booted
22:11:08preglowstevenm: yup, but that is the way it's supposed to be
22:11:45preglowwe'll have to handle that some fancy way with hardware
22:11:58preglowlike actually modifying the samples after the loop stop point
22:12:03preglowit's never supposed to be reached anyway
22:15:03stevenmpreglow, but wait a second.. if there is a loop point, then what is wrong with it reading one sample past the loop ?
22:15:18*t0mas is going to bed... have to get up at 6 tomorrow :X
22:15:31 Quit t0mas ("bye and goodnight")
22:15:35preglowstevenm: because that's not the next sample to be played
22:15:51preglowstevenm: the next sample to be played is around the loop start point
22:15:54stevenmpreglow, problem 2: some files specify "loop fractions"
22:16:00*HCl goes to peek at the dreadful consequences of committing dumb
22:18:14rasherno change afaics
22:18:35HClit hasn't recompiled yet
22:19:50 Join mecraw [0] (fwuser@
22:19:50amiconnIt has, otherwise it wouldn't be on the front page
22:21:01*HCl forgot one file..
22:21:56preglowa sec, eating
22:26:35preglowstevenm: ehh? loop fractions?
22:26:35 Quit dukezed ("CGI:IRC (EOF)")
22:27:02preglowahh, so _THATS_ why cvs commits are slow in showing up there
22:27:25rasherthey only appear after a build?
22:27:39HCl helpers/stdfile.c \
22:27:47HClaccidental paste
22:27:50preglowyou're going to hell for that
22:27:58HCli know :p
22:28:02HClah well, can't be helped
22:28:18rashergar, network not coming up :(
22:29:48 Quit elinenbe (" HydraIRC rocks! -> <-")
22:32:18 Part MoosCamaro
22:35:01stevenmpreglow, there is a specification SOMEWHERE that a loop starts at "xxxxx+ xx/16" samples
22:35:09stevenmfiles hardly ever use it and there is no doc on it
22:35:26stevenmi dont think loop fractions are the issue. but I just added a guard sample.. must go for a bit
22:35:52HCl12/13 more warnings.
22:37:05 Join Dicko [0] (
22:37:06 Quit einhirn (Read error: 104 (Connection reset by peer))
22:38:11HClsome of which need obvious fixing, references to printf etc.
22:39:45 Join prpplague^2 [0] (
22:42:22preglowaye, yes
22:42:44HCli see a usage of log, thats a bit disturbing..
22:43:11***Saving seen data "./dancer.seen"
22:43:46HCli don't even know whether it'll run fast enough
22:45:19preglowworth a shot
22:45:31HClits supposed to give really good quality
22:45:31preglowi'll have a quick look
22:46:21 Quit prpplague (Read error: 60 (Operation timed out))
22:47:57preglowdoes the rockbox disk io do caching?
22:47:58preglowi guess not
22:49:09rasherACPI wakeup devices:
22:49:19rasher USB VIY0 VIY1 HORN LID
22:49:35rasherwhat, if I honk at my computer it'll wake up?!
22:50:03HCl xD
22:50:28thegeekgod damn rasher, can't you see it's hornY....
22:50:36thegeekhad to say it
22:51:10rasherah dammit
22:51:13thegeekeven though it was quite possibly _the_ worst sentence ever uttered
22:51:26rasherI can't for the life of me remember what the fix to this problem is
22:52:34preglowouch, long long spotted
22:53:22HClyea, dumb author was talking about a 32x32 -> 64 bit multiplication..
22:53:28preglowbut that doesn't matter
22:53:31preglowi see it now, that's ok
22:53:39preglowjust use emac there
22:53:50HClcan we?
22:53:53preglowwhy not?=
22:54:10HCli thought markun said we couldn't do 32x32 -> upper 32 bits multiplication
22:54:15HClonly the lower 32 bits, which are useless
22:54:17preglowwell, then he's plain wrong
22:54:25HCloh, well, thats good news for tremor then
22:54:27preglowi do those all the time
22:54:38HCli must admit i couldn't find it either
22:54:40preglowi should write a emac users guide or something
22:54:48stevenmpreglow, added guard samples; no difference
22:54:56preglowsince i seem to be the one who knows the most about it
22:55:04preglowstevenm: ok, explain exactly what you've done
22:55:09stevenmpreglow, actually, taht even added additional clicking
22:55:35preglowHCl: as a matter of, the mac unit in fractional mode does exactly what PROD31 in tremor does, and that's one of the more commonly used macros
22:56:03HClkay, hope markun is reading
22:56:13*preglow summons markun
22:56:29HClxD it'd be funny if he'd join instantly
22:56:33preglowhe is on
22:56:48preglowi actually did summon linus like that once
22:56:53HCl :P
22:57:52 Quit Dicko (Remote closed the connection)
22:58:01rashergar.. I need some kernel parameter
22:58:06rasherand I can't remember which
22:58:11HClfor what?
23:00:08rashersomething with acpi interrupt handling
23:00:11rasheror rather, to avoid it
23:00:43HClor maybe it was plain noacpi
23:00:49preglowthat's it
23:02:48rasherit appears that I have thermal zone support
23:02:51rasherand I never knew
23:03:17rasherokay, same error
23:04:11HClwith which?
23:04:15HClpci=noacpi or noacpi?
23:04:29rasherif I did it right..
23:04:35rasheroh silly me
23:04:39rasherrunning lilo might help
23:04:49HClaccording to the kernel docs
23:04:51HClits the first one
23:04:53HClwith pci=
23:05:00rasherdammit, already rebooting
23:05:23HClwith lilo
23:05:28HClyou can just add to the kernel opts...
23:05:36HClas long as you have prompt turned no
23:05:51*rasher tries
23:07:01amiconnGah, x11 is lame
23:07:14stevenmpreglow, I'm back. I had it recompute the second sample used for interpolation in the event that the thing hit the end of a loop and was bumped elsewhere
23:07:24stevenmpreglow, it sounds no differently now that it did.
23:07:41preglowamiconn: oh, we agree once again
23:08:13preglowstevenm: so if the x + 1 or x + 2 samples now go beyond the loop point, you wrap them around to the start loop point?
23:08:24preglowbrb, switch os
23:09:16amiconn(1) I want to disable keyboard autorepeat within the x11 sim only. Since X11 does only allow to do this globally, I need to track whether the sim has focus (FocusIn / FocusOut events) (2) Cygwin x11 sends an initial FocusIn event, x11 on linux doesn't
23:09:23stevenmpreglow, yes.
23:11:08amiconnGrr, cygwin is even undecided what to do :-/
23:13:43 Join bippy [0] (
23:14:33prpplague^2amiconn: hey, the ata_mmc.c stuff is GPL correct?
23:14:46 Quit _Lucretia_ ("Leaving")
23:15:10amiconnprpplague^2: Yes of course, all rockbox code is
23:15:21prpplague^2amiconn: ok, just double checking
23:16:42 Nick prpplague^2 is now known as prpplague (
23:17:36bippyi think ye should treble check
23:19:03rasherexcellent, my cpu is 340K
23:19:47bippyis that good
23:20:03rasherit's roughly 67 degrees celsius
23:20:10bippyThats not good
23:20:23HClthats pushing it..
23:20:31bippyIs this in the iriver ?
23:20:36rashernooo, laptop
23:20:57bippyThats alright then, thought you were making handwarmers and all
23:22:01stevenmpreglow, i'm going to go eat lunch, back a bit later
23:22:15bippyBring some to share around
23:22:33stevenmtrust me, you don't want college food :-P
23:22:52bippytrue, soggy chips...
23:22:59*HCl gets reminded that there's new fat for the deepfrier...
23:23:06*HCl goes to heat it up...
23:23:29bippyI might have a bowl of alpen
23:26:39bippymmm bacon fried egg toast beans
23:26:45*bippy drools
23:27:10bippyHCI = Human Computer Interface?
23:27:34 Join Camilo [0] (
23:27:59bippySame thing
23:29:09bippySo, how is mister iriver coming along
23:29:57 Join langhaarrocker [0] (
23:30:24 Quit langhaarrocker (Client Quit)
23:31:14preglowstevenm: strange, should sound better
23:31:32bippyHe isnt here
23:31:40bippyhe is eating college food
23:31:50preglowi now, i'm counting on him reading the backlog
23:31:53preglowi know, even
23:32:38bippyI wouldnt read the back log, guess thats why im not a Dev
23:32:56preglowi always read the backlog
23:33:01preglowfirst thing i do, heh
23:33:32bippyId eat first or get dressed
23:33:38preglowi don't
23:33:53preglowthen again, i'm not on irc while sleeping
23:34:00preglowstill, there's tons of stuff i do before i do either of those things
23:34:12bippyBrush your teeth ?
23:34:25preglowno, i don't see the point in brushing my teeth before eating
23:34:54bippyThat means you'd never brush your teeth :|
23:34:59*bippy is scared
23:35:06preglowten minutes before eating, then
23:35:21bippy9 minutes is pushing it then?
23:36:05preglowi usually toss myself down on the chair and sit around scratching my nuts and reading up on news, then i get dress, then i eat, then i brush my teeth
23:36:11preglowinsert shower before getting dressed if you're lucky
23:36:39 Join hi [0] (
23:37:14bippyI shower at night, dont have time to shower in the mornin
23:37:27preglowyeah, depends on the schedule
23:37:39 Part hi
23:37:53bippyyou seem pretty normal for a programmer
23:38:15preglowbrb, pack some more
23:38:16 Quit Aison ("( :: NoNameScript 3.72 :: )")
23:39:53bippyRight im off for my shower, keep up the good work
23:39:55 Quit bippy ("CGI:IRC")
23:42:28 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
23:50:50 Join hi [0] (
23:50:59 Part hi
23:56:39 Join amiconn_ [0] (
23:57:02 Quit amiconn (Nick collision from services.)
23:57:03 Nick amiconn_ is now known as amiconn (

Previous day | Next day