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 2004-04-20

00:00:03solarisdaily build?
00:00:11 Join [1]c0utta [0] (
00:00:55[IDC]Dragonsolaris: in your post, your said the PR polarity was changed after flashing?
00:01:16[IDC]DragonThis would indicate the h/w mask was altered.
00:01:49solarisyes, from positive to negative
00:01:54 Quit lImbus (" Like's GUI? Then try HydraIRC -> <-")
00:01:57LinusNsolaris: badness
00:02:25[IDC]Dragonthis should not happen.
00:02:40solarisstop scaring me
00:02:45[IDC]Dragonthe value is kept while flashing
00:03:03 Quit c0utta (Read error: 60 (Operation timed out))
00:03:07[IDC]Dragondo you still have your original flash dump?
00:03:28[IDC]Dragoncan you zip&email it to me?
00:03:31solarisi think i have b4 and after
00:04:01 Join c0utta [0] (
00:04:26 Quit Unknown (Read error: 60 (Operation timed out))
00:05:18solarise-mail or dcc??
00:06:01[IDC]Dragondcc worked, for a change.
00:06:28[IDC]DragonMy firewall seems to have taken the rest of the day off.
00:06:35solarisxinternal is after the 2.1 and internal is b4, i think
00:06:49solarisand xinternal was working, i think
00:08:29[IDC]Dragonboth are identical
00:09:04solaristhan it must be back up b4 the flash
00:09:41amansingerSorry to disturb once again, IDC, but do you happen to recall what you did to the username_ta.txt file to get Textaloud to read it for your German voice? Did you just open te text file, and what options did you select, if you remember?
00:10:12[IDC]Dragonsolaris: the content is no news, I have that variant.
00:10:46amiconn[IDC]Dragon: I did have another look at mp3_play_pause() and the SH hw manual. Now I think I know what may cause the problem.
00:10:48[IDC]Dragonamansinger: you need to place it int the TA program dir.
00:11:17[IDC]Dragonbut name it like the existing file you find there.
00:11:47amansingerYou've already taken care of that. Your cmd file does that quite nicely.
00:11:58LinusNamiconn: i'm all ears
00:12:09[IDC]Dragonnot really, it doesn't know the username.
00:12:43amansingerIt uses your user name and your program files dir, yes. I've edited it to mine.
00:12:48 Quit [1]c0utta (Read error: 60 (Operation timed out))
00:13:04LinusNsolaris: any luck booting ajbrec.ajz
00:13:25[IDC]Dragonamansinger: then you need to set TA to mulpti-article, multi-file wav output.
00:13:51solariswell, i created/formated fat32 and copied the daily build and it's acting same thing
00:13:57amansingerPerfect, multiarticle, multi file.
00:14:07LinusNsolaris: it doesn't boot?
00:14:26[IDC]Dragonamansinger: not user_TAExport.txt ?
00:14:27solarisit boots only with F1
00:14:35solarisand only in archo's firmware
00:15:22solarisif i press ON, the green led comes on and stays for about 15-20 sec
00:15:24amansingerYes, I was doing it from memory. I just copied the path and username.
00:15:26*Laurent_ prays the god of the fiscal administration for mercy
00:15:26solarisand shuts off
00:15:35LinusNdid you try with the fm build?
00:15:49solarisyes, i tried with every build
00:16:07[IDC]Dragonalso with Archos original .ajz?
00:16:12solarisso i'm convinced something went wrong when flashing??
00:16:22LinusNsolaris: i am too
00:16:28solarisoriginal .ajz?
00:16:34solarisi haven't
00:16:40LinusNbut if we can make it boot the ajbrec.ajz, we won't have to do the boot mod
00:16:43[IDC]Dragonto be downloaded from Archos website
00:17:34solarisi think i have back up
00:17:37solarissome where
00:17:50[IDC]Dragonit won't do a difference except taking longer to boot, but would prove only Rockbox is unable to load
00:18:00[IDC]Dragonor nothing is able to load at all
00:18:34[IDC]Dragoncould this be a side effect of a wrong h/w mask?
00:19:02solarisit's same with ON
00:19:15LinusNyou need F1+ON
00:19:19solarislet's see.... F1 +ON....
00:19:32[IDC]Dragonanother thing: does it freeze with F3+On ?
00:19:50solarisafter it starts?
00:19:53 Join ricflair [0] (
00:20:05solarisF1+ON and than f1+F3??
00:20:06[IDC]Dragonpower it off, then do F3+On
00:20:13ricflairhi all. i plugged my jukebox in and it doesnt charge :( nothing happens
00:20:22LinusNricflair: how do you know?
00:20:33solarisok, shutting down...... got it takes forever... archos firmware sucks
00:20:53ricflairlinus, well i plug in the adapter and nothing happens
00:21:08ricflairif i press the on buton it flashes green for half a second or less
00:21:09solarisf3+on seems to at as just ON button
00:21:20LinusNricflair: model type?
00:21:28ricflairjukebox recorder 20
00:21:44solarishmmm, it's not shutting off automatically
00:21:46[IDC]Dragonsolaris: then you have the older bootloader?
00:22:37LinusNricflair: and it used to work?
00:22:40solarisi don't know what u talking about. older bootloader?
00:22:47ricflairlinus, i dont know, ive never ever used one before
00:23:11amiconn[IDC]Dragon, LinusN: My previous though was bullshit, but now I wonder if the MAS doesn't like to be interrupted while being halfway through a 30 byte burst?
00:23:20LinusNricflair: sounds like it's broken
00:23:32[IDC]Dragonsolaris: Once I urged people to reflash, to avoid a flat battery problem (which ricflair may have).
00:23:41LinusNamiconn: perhaps
00:23:46ricflairis there no way around it?
00:23:55ricflaircan i use normal aa batterys in it to test?
00:23:56solariswith F3+ON from off, the green light comes on, i hear disk spinning but no red light indicating any reading. It does not auto. shut off.
00:24:42[IDC]Dragonsolaris: then you may have the newer bootloader, which has my "Minimon" on F3+On.
00:25:02LinusNricflair: i am not sure, the normal AA's have higher voltage
00:25:28solarisin other words???
00:25:29amiconnSo perhaps mp3_play_pause() shouldn't stop the transfer immediately by disabling TIE, but should stop the demand irq instead.
00:25:34LinusN[IDC]Dragon: but the drive shouldn't spin
00:25:47ricflairis there anything i can try at the moment
00:26:13LinusNricflair: plugging in the charger should start the recorder
00:26:30 Quit c0utta (Read error: 60 (Operation timed out))
00:26:40ricflairdo you think that mabey i have the charger in the wrong polaroty and thats why nothing happens?
00:26:43[IDC]Dragonsolaris: this means you don't have to dothe UART boot mod, only the serial mod.
00:26:50LinusNricflair: could be
00:27:15LinusNhow dod you verify the polaroty?
00:27:17ricflairYEAH BABY YEAH
00:27:21LinusNricflair: oh?
00:27:23solarisgreat!!, now point me to the right direction
00:27:24ricflairworks, well starts up anyway
00:27:38ricflairHD register error
00:27:45[IDC]DragonMinimon is sitting there and waiting for commands from the serial. This is why it seems to be frozen.
00:28:16LinusNricflair: ok, so it is broke, or you need to replace the batteries
00:28:36[IDC]DragonTheoretically, you could even go without opening it, by just the remote pin.
00:28:37 Join jkerman [0] (
00:28:44ricflairbut turns on, i just get a hd register error, cant that be fixed?
00:28:48solarisoh, yeah??
00:28:55LinusNricflair: that can be fixed
00:29:04[IDC]DragonBut this would mean "flashing in the dark", without feedback.
00:29:14ricflairGreat, can i fix it now or do i need my usb cable(that i dont have yet)
00:29:25LinusNricflair: try to squeeze the bumpers together
00:29:44LinusNin the direction of the batteries
00:29:57 Quit AciD (Remote closed the connection)
00:30:29ricflairwell that done somthing
00:30:42ricflairnow says archos jukebox recorder - shutting down ....
00:31:45LinusNricflair: ok, you need to open it up and resolder a few solder joints
00:31:57LinusNvery common problem
00:32:14LinusNhappens when you drop it, or if you change batteries often
00:32:19ricflaircool, as long as it can be fixed. How small are the solder joints
00:32:20LinusNa design flaw
00:32:22[IDC]Dragonamiconn: what was that with the 30 byte burst?
00:32:29ricflairjust restarted it and it says battery charging
00:32:59ricflairshould i just leave it charging?
00:33:00amansingerOK, IDC, I've made the text file. The issue is that it mentions the mary voice in all its articles. That is, it says "LANG_SOUND_SETTINGS|Mary|1
00:33:00amansingerSound Settings
00:33:22amansinger" Will this effect the voice and, if so, what's the format to change the voice's name?
00:33:34LinusNricflair: the green ones
00:33:44ricflairlooks easy enough
00:33:47ricflairthanks for the help man
00:33:54amiconn[IDC]Dragon: The mas data transfers (both reading while recording and writing while playback) take place in 30 byte bursts, because that is the size of the dma buffers within the mas.
00:33:55LinusNyou're welcome
00:34:03[IDC]Dragonamansinger: TA can do that, it even suggest this is you change the voice
00:34:31[IDC]Dragonamiconn; I know
00:34:35ricflairshould i leave it as it is, plugged in and with the state "battery recharging" with 0:04 (just turned to 0:05) above the f5 and nothing else on the screen exept for the archos jukebox writing
00:34:48ricflairoh and a little battery in the top corner
00:34:54LinusNricflair: do so
00:34:58ricflairas long as it is a common fault im not to upset
00:35:02solarisso, how can i get start on "flashing in the dark'???
00:35:02amansingerDo you do that by a search and replace in the txt file or by the way through the TTA UI itself?
00:35:04ricflairhow long should i leave it on charge for.
00:35:16LinusNricflair: takes several hours
00:35:20LinusNleave it over night
00:35:51[IDC]Dragonsolaris: I don't recommend it, and it would require canging my uart_boot program.
00:35:53LinusNi suggest you buy a set of NiMH batteries
00:35:59amiconn[IDC]Dragon: mp3_play_pause stops the transfer immediately (by disabling the transfer end interrupt of the serial i/f), so this may happen half-way through a 30 byte burst. Maybe the mas doesn't like that.
00:36:26[IDC]Dragonamiconn: this would be very picky.
00:36:38LinusNamiconn: i think that sounds like a good theory
00:36:45ricflairlinus i will, its just that i go on holiday this week and was hoping i could take it with me. Going to pick up a usb cable tomorrow from a store
00:36:54ricflairagain, thanks very much for you genius advice :)
00:37:01LinusNbut it wouldn't solve the glitch problem
00:37:11[IDC]Dragonhow could we know the 30 byte are over?
00:37:12ricflairwhat glitch?
00:37:28LinusNricflair: sorry, was talking to [IDC]Dragon and amiconn
00:37:36ricflairlol, sorry
00:37:39solarisok, so.... can i return it to amazon? I don't have the original box
00:37:45ricflairim just really panicy. i hope can use this on holiday
00:37:45 Quit silencer ("leaving")
00:37:47[IDC]DragonI suggest we try the DMA pause/play first
00:37:57 Join silencer [0] (
00:38:00amiconnSo, mp3_play_pause() should disable the demand interrupt only. The code then has to wait until the burst is over.
00:38:39LinusNamiconn: it won't solve the audio glitch prob easily
00:41:04LinusNbecause you need to find the next frame boundary and queue the next clip very fast
00:41:47amiconnMaybe this is the point: If we try to do that within a 30 byte burst, we are too slow.
00:42:23[IDC]Dragonthe DMA buffer can't be the only buffer the MAS has
00:43:39amiconnNo, of course it has to buffer at least one complete frame. well as decoded values from previous frames.
00:44:40[IDC]DragonI just swapped CHCR3 and SCR0 usage
00:44:54[IDC]Dragondoesn't recover from pause
00:45:03ricflairdoes the lcd light stay on whilst its on charge?
00:45:05LinusNfigured so
00:45:13LinusNricflair: yes
00:50:12[IDC]Dragonthe voice clipswitching however works, but no better than before.
00:50:37LinusNgotta sleep now
00:50:45LinusNcu around
00:50:52[IDC]Dragonprobably it works because it start a new DMA
00:50:53LinusNricflair: good luck
00:51:07[IDC]Dragonversus the play code doesn't
00:51:17 Part LinusN
00:54:23[IDC]Dragonmattzz: how is your grayscale mandelbrot?
00:58:02mattzzI had a gray set already but than I changed code and now I dont get the buffer allocated any more. IllInstr ....
00:58:03amansingerIDC, what happens once I select the multi article multi file modes? I now have this text file and these modes selected. How do I make TTA speak these articles, if I may ask?
01:02:04[IDC]Dragonamansinger: there is a button to speak it all into files.
01:02:17 Join c0utta [0] (
01:03:22[IDC]Dragonunder file options, you can specify the destination directory, set the speed to maximum, select wav.
01:03:25amansingerOk, these will be wav files, correct? Do I then run wavtrim on them and then voicefont.exe?
01:03:52amansingeror rather, lame encode them and then run voicefont
01:04:21[IDC]Dragonmany wav files, yes, My cmd file then does wavtrim them all, then Lame encode, finally voicefont.
01:04:22amiconn[IDC]Dragon: Try to insert the following line in mp3_play_pause as first line into the block after "else if (!paused && !play)":
01:04:33amiconnwhile(!(*((volatile unsigned char *)PBDR_ADDR) & 0x40)) ;
01:05:04amiconnthis makes it work a little better imo, but please compare yourself.
01:05:30[IDC]Dragonhave you tried?
01:05:43amansingerRight. Thanks. Anyhow, I'm off. It's been a very interesting two hours, watching the project take shape, so to speak. I have no doubt you'll be solving the swallowing problem with the clips.
01:06:06amiconnYes, I have, but I cannot tell for sure if it helps. The problem doesn't disappear completely.
01:06:30 Quit amansinger ()
01:07:27amiconnperhaps the search for the next frame is simply too slow.
01:08:25amiconn(Although I think there should be enough time, since one frame is several milliseconds long.
01:08:37Laurent_does anyone know about token packing here ?
01:08:50Laurent_google was no help
01:09:22Laurent_(almost no help)
01:09:41[IDC]Dragonamiconn: this works very well for me, no hickup yet
01:10:37amiconnI got some hiccups while testing. Basically this loops waits until the mas deactivates the EOD pin before stopping.
01:10:41Laurent_no one ? ok, then time to bed ;)
01:10:55Laurent_goodbye everyone
01:11:12 Quit Laurent_ ("Download Gaim:")
01:12:46[IDC]Dragonamiconn: it seems to have the side effect that clips get barely aborted
01:13:08amiconn[IDC]Dragon: This loop is taken from the recording code; it should definitely be checked if it works on the player.
01:14:14[IDC]DragonI should try a voice file with bit reservoir now.
01:14:28amiconn? You get aborted clips where you shouldn't?
01:14:49[IDC]DragonRather the opposite
01:15:13[IDC]DragonMost of the entries are now completely spoken, not aborted.
01:15:52amiconnMy main "test case" is to alternate between "Sound settings" and "General settings". These get aborted if you switch over, although not always immediately.
01:20:01[IDC]Dragonwe're on the right track, I'd say
01:20:52mattzzok, we have a graubrot.
01:21:21[IDC]Dragonor Grnbrot?
01:22:03mattzzhow do you pronounce that? [g:m-bro:t]?
01:22:25solarisi just noticed, when i press either F3+on or ON, LCD light does not come on, only the green light and the disk spins but no red light indicating any reading
01:22:45solarismy room is getting dark with sun going down
01:22:53[IDC]Dragonsolaris: that's good, yes
01:23:09amiconn[IDC]Dragon: Can you test on the player?
01:23:17[IDC]Dragonamiconn: no
01:25:05[IDC]Dragonthe port pin should be the same according to the docs
01:25:14amiconnI guess it should work on the player since PB14 is EOD as well, but I would not count on it without testing.
01:25:45amiconnToo bad, Linus could test on the player...
01:26:02[IDC]Dragonhe will, no hurry
01:26:14solarisis there anyway to force the archos to read ajz files?
01:27:03[IDC]Dragonsolaris: you tried the Archos one now?
01:27:09amiconnmattzz: How fast is Graubrot?
01:27:15solarisyes, but it's same thing
01:27:36[IDC]Dragondoes it take longer to boot?
01:27:58solarisi thought it was about the same
01:28:01mattzzamiconn: pretty fast if you take into acount that every pixel is drawn now
01:28:02[IDC]Dragon(indicating the second one gets read from disk)
01:28:36mattzzamiconn: there is no stopwatch function included any more
01:28:40[IDC]Dragonmattzz: can we try it?
01:28:57mattzzuh, this is bad, bad code ;-)
01:29:06mattzzbut - sure, yes
01:29:18[IDC]Dragonno review today, I 'm too tired">
01:30:51mattzzthis is my first shot - have mercy ;-)
01:31:55mattzzI haven't found a really nice colorset yet
01:32:25[IDC]DragonI'd prefer green
01:32:39mattzzI _knew_ you would optimize it ;-)
01:34:03[IDC]Dragonlooks interesting, yes
01:34:14mattzzstill way too dark
01:35:14mattzzI want a screenshot of this in the manual!!!!
01:35:47[IDC]Dragongotta go now
01:35:58amiconncu Jrg.
01:36:00 Quit [IDC]Dragon ()
01:39:15amiconnmattzz: graubrot looks really nice.
01:39:55mattzzamiconn: thanks for your grayscale framework!
01:40:48amiconnJust some things I noticed:
01:41:57amiconn(1) Either alter or take out the comment at lines 724..725. The values are wrong.
01:42:31amiconn(2) You can take out gray_black_display() and gray_deferred_update() as well, since you don't use it.
01:43:00mattzzyup, agreed on both.
01:43:24amiconn(3) I would suggest that you try the following for choosing the brightness:
01:44:08amiconnif(n_iter > max_iter)
01:44:16amiconnbrightness = 0;
01:45:31amiconnbrightness = 32 + 31 * (n_iter % 8);
01:47:20mattzzmuch better
01:47:33amiconn(just trying myself)
01:47:58mattzzthough some contrast is missing where n_iter==max_iter
01:48:04 Quit mecraw__ ("Trillian (")
01:50:18mattzzbrightness = 256 - (31 * (n_iter % 8));
01:50:50amiconnshould use 255...
01:51:37amiconn...since 256 does return... but it works here, because the pixel is white anyway.
01:52:10***Saving seen data "./dancer.seen"
01:56:04solarisnow, i'm definetly sure it's not reading .ajz from F1+ON, i formatted the HD and i didn't put anything in. Still archos firmware
01:57:30scott666i thought the point was to find out whether or not it would read it when there was actually something to read
01:58:19amiconnmattzz: Apparently using 28 bits for the fraction causes overflows in a few spots that make it look weird. I changed it locally to use 24 bits for the fraction and a baildout value of 9 (instead of 4). This looks really nice!
01:58:57 Quit jkerman (Read error: 104 (Connection reset by peer))
01:58:58mattzzwhy the bailout value of 9?
01:59:25solarishmmm.... I'm just totally lost here...
01:59:53solarisin anycase, with original archos, it took 14 sec
02:00:05solarisand with rockbox it also took 14 sec
02:00:22solarisand without any .ajz, takes 14 sec
02:01:42amiconnmattzz: You have Xmin = -2.5. The first squaring operation already yields 6.25, so if you use 4, you are above it in the very first iteration. And, although using 9 is a bit slower, it is nearer to the default value other fractal generators use (16 iirc).
02:02:07amiconnJust try it, you will see what I mean.
02:06:27mattzzamiconn: you are definitively right with the overflow. stupid me.
02:08:26mattzznow, _that_ looks much nicer. I will try the other bailout value now.
02:08:52amiconnIf you had a free button left, it would be possible to use it as a "status toggle" to display rendering time and max_iter at the top.
02:10:39mattzzyup, we need more keys!
02:11:24amiconnFirst keypress: place the grayscale overlay down one line (with gray_position_display() ), put out the data with snprintf() / lcd_puts() to line 0, then do gray_deferred_update()
02:11:47amiconnSecond keypress: replace the grayscale overlay to (0, 0)
02:12:13amiconnThe last line is not lost in the process, only hidden.
02:12:26mattzzwho needs F3 anyway? we could use that one
02:12:53*mattzz thinks about a 24bit patch for mandelbrot.c
02:14:41mattzzwe should stay with r_max=2 (i.e. bailout 4)
02:15:08mattzzWhere did you see 16?
02:15:24solariswell, i'll catch you'ol tomo, good nite and thx
02:15:38mattzznight solaris
02:15:54 Quit solaris ("I-n-v-i-s-i-o-n 2.0 Build 3515")
02:20:38mattzzamiconn: do you think I should put graubrot in the pathtracker?
02:21:06amiconnmattzz: Most fractal generator programs I used had a default bailout of 16.
02:21:19amiconn(Just lkooked up some documentation)
02:22:34mattzzDo you see a difference in the resulting image? I wasn't quite sure.
02:26:02amiconnHmm. It looks quite different to me, even the first image.
02:27:22mattzzwith 24 bit fraction we can use 16 as a bailout
02:30:26amiconnJust tried it - this gives overflow artefacts again. I guess that if we use a bailout of r, we have to reserve an integer part capable of holding up to R*R (signed) without overflow. 8 bit integer part is not enough is bailout = 16.
02:32:42amiconnThis may be because if in one iteration it reaches a value just below the bailout, it becomes squared in the next, and this should not produce an overflow, otherwise you cannot reliably detect that it is above the bailout.
02:33:33amiconnIf bailout = 9, integer values up to 81 should be possible, this just fits into 8 bit (+127 / -128)
02:35:13mattzzyou convinced me ;-)
02:38:04amiconnJust tested bailout = 4 and a 6/26 split for integer/fraction, this is also artefact-free, which almost proves my theory.
02:38:30amiconnbailout = 4 looks not as good in the outer area of the starting image though
02:38:36zeyou can't prove a theory, only gain support for it or disprove it
02:39:13mattzzamiconn: ok, that should be it
02:39:34amiconnWhat should be what?
02:39:54mattzz6/26 and 9
02:40:25amiconnNo, either 8/24 and 9, or 6/26 and 4, otherwise -> overflow artefacts.
02:40:38mattzzargh - ignore me ;-)
02:40:45ricflairhi all. just getting this going, is it normal that the jukebox recorder makes a rattle stylenoise from the hd when it plays a song or when it browses?
02:40:49mattzz2 tired
02:43:46amiconnmattzz: btw, I learned something from your code - it is sufficient to comment out a prototype to prevent the function from being compiled in if it is not used. I didn't know that.
02:47:21amiconnIf someone has generated a nice looking mandelbrot picture, we would need a way to save it to disk. Perhaps it could generate a .rvf video from a mandelbrot zoom sequence ;-)
02:48:35mattzzyep, there's a lot of work left to do...
02:49:05mattzzguess, I will have to sleep a bit...
02:49:09mattzzc u amiconn
02:49:25amiconncu and good night.
02:49:34 Quit mattzz ("Client exiting")
02:50:00 Part amiconn
03:16:19 Join diddystar5 [0] (
03:16:29scott666hey diddy
03:23:12 Join jkerman [0] (
03:26:28 Quit diddystar5 ("Leaving")
03:30:44 Join diddystar5 [0] (
03:52:12***Saving seen data "./dancer.seen"
04:27:20 Join monkey [0] (
04:28:58monkeycan anyone here tell me a n00b definition of what the "mandelbrot" and "grey scale" projects are about?
04:29:29scott666have you seen the video plugin?
04:29:44scott666.rvfs and all that
04:29:54monkeyi used it
04:31:07monkeyoh i just yah i've used rvf but how does it relaste to mandelbrot and greyscale
04:31:15monkeyi just meant*
04:32:01 Join lostuser [0] (
04:33:04lostuserdoes any1 kno anything about converting avi to rvf?
04:34:11scott666monkey: theyre working on a way to use the same greyscale as in the video plugin in regular rockbox
04:34:23lostuserno the thing is that when i convert my avi's the video is crap, it is only a bunch of lines
04:34:37scott666or other plugins−−like games
04:35:00scott666do you re-size them to 112x64?
04:35:17lostuserof course
04:35:33scott666then i dunno what to tell you
04:35:44monkeyscott666: and mandelbrot?
04:35:52lostuserdo you know anyone who might be able to help me?
04:35:52scott666i have no idea
04:36:00scott666its something on the same topic, but damned if i know what
04:36:08scott666and ive sat in on half the discussions about it
04:36:24scott666lostuser: what do you make videos with? the UI?
04:36:33monkeyyah i've lurked while its on but dont know what its for
04:37:06scott666i get the feeling its something german, as i know at least 2 of them are
04:37:12 Quit facted (Read error: 54 (Connection reset by peer))
04:37:13scott666but im not sure at all
04:37:56lostuseri use the GUI
04:38:08scott666i dunno
04:38:21scott666ask on the mailing list? read the mailing list archives of the past month or so?
04:38:58scott666trasnlator doesnt like mandlebrot, there goes that idea
04:39:08monkeyscott666: ever heard anything about "color" images from b&w images?
04:39:43scott666we have green...
04:39:49diddystar5i have only used the command line
04:40:05monkeyyou mean green from the LEDs?
04:40:12diddystar5i have to go
04:40:15diddystar5see you guys
04:40:51 Quit diddystar5 ("Leaving")
04:43:00monkeynah i was thinking it might be possible to convey color by manipulating b&w images
04:46:14scott666well...if anyone can do it, jrg can
04:48:23monkeyi will study it some more and then maybe i will bring to him
04:48:34monkeybring it up*
04:56:09lostusernow i won't be botherin y'all. i'll see ya later. thanks fo all the help again...kinda. btw my problem was the FPS rate, it was too high, so iowered it to 10... until we meet again. err um or until i get another problem. later
04:56:39lostuserframes per sec
04:56:56scott666FPS rates should be rate of the video youre feeding it, and ~70 output
04:57:27lostusererr well i changed that thingy and my videos worked.
04:57:51scott666no audio?
05:00:41 Join Unknown [0] (
05:01:10lostuserno i have audio
05:01:43scott666and it's snyced right?
05:02:02 Quit c0utta (Read error: 60 (Operation timed out))
05:02:03 Nick Unknown is now known as c0utta (
05:52:16***Saving seen data "./dancer.seen"
05:54:54 Quit scott666 ("i'll be back...eventually...")
05:54:54 Quit Nibbler (Read error: 54 (Connection reset by peer))
05:56:55 Quit lostuser (Read error: 60 (Operation timed out))
06:23:50 Quit monkey ()
06:29:59 Join monkey [0] (
06:43:13 Quit monkey ()
07:14:49 Join AciD [0] (
07:18:13 Join elinenbe [0] (
07:23:52 Join Nibbler [0] (
07:52:19***Saving seen data "./dancer.seen"
08:04:22 Join LinusN [200] (
08:27:15 Join amiconn [0] (
08:33:35 Join [IDC]Dragon [0] (
08:33:55[IDC]DragonHey, everybody is up so early!
08:50:17LinusNany clues about the mp3 playback issues?
08:50:48[IDC]Dragonhave you read the log?
08:51:08[IDC]Dragonno more than that, so far
08:51:48[IDC]Dragonthe wait loop fixes it pretty well, but makes most clips run completely
08:54:58amiconn[IDC]Dragon: The loop may not catch every gap where EOD is inactive, because it is relatively slow. I will try an asm version that will react faster.
08:55:26 Join mattzz [0] (
08:58:35amiconn(To be more preice, the loop may not be too slow, but _if_ the loop detects a gap, the serial has to be stopped very fast)
09:00:04*dwihno sighs... No stores in town are carrying MDR EX71's :(
09:00:10LinusNamiconn: couldn't it be done in the irq?
09:05:24LinusNwith a pause_pending flag
09:05:33amiconnGood point - this would be a much cleaner implementation.
09:07:23amiconnI think it had to be placed in the stop demand irq, but I wonder if this is possible. I know that the 3587 uses irqs for both start and stop, but the 3507 only for one of these actions.
09:07:48LinusNhowever, the start irq knows if the dma is stopped
09:08:08LinusNif pause_pending and dma is stopped, then never start it
09:08:46amiconnAnother thing has to be taken care of - if mp3_play_pause only sets the flag, it becomes an asynchronous operation.
09:09:12LinusNyes, so the wait loop has to stay, in a different shape
09:09:26LinusNlike waiting for pause_pending to be false again
09:09:38LinusNand the irq sets it to false
09:11:27amiconnYes, and it has to be done in the stop demand irq, so we have more time to have the new data ready when the mas starts demanding again.
09:12:38 Quit AciD (Read error: 54 (Connection reset by peer))
09:13:15LinusNamiconn: good point
09:14:29mattzzcygwin compile of the W32 simulator woes (good morning, folx)
09:14:40LinusNmattzz: morn
09:14:42amiconnOne thing I would like to know: Does the start/stop demand irq occur once for every 30 byte burst, or once for every frame?
09:15:04LinusNamiconn: don't know, i guess i'll have to measure
09:15:46amiconn(I suppose it is the former)
09:15:53LinusNme too
09:16:55mattzz../../firmware/include/dir.h:77: error: conflicting types for `rmdir'
09:16:55mattzz/usr/i686-pc-mingw32/include/io.h:153: error: previous declaration of `rmdir'
09:17:16LinusNmattzz: are you using the latest cvs?
09:17:26mattzzlatest tarball
09:17:32LinusNah, sorry
09:17:44LinusNyes, the win32 sim is unable to compile atm
09:18:19mattzzok, no prob. the normal build is the important one ;-)
09:18:21LinusNi have added the rmdir() function, and i'm waiting for someone to fix the win32 sim version
09:19:16LinusNmattzz: that person may be you :-)
09:19:22mattzzLinusN: yeah, rmdir is part of io.h, too
09:19:29*mattzz is _no_ w32 wizzard
09:20:04LinusNactually, since it's cygwin, rmdir() is posix iirc
09:20:20LinusNso it should be relatively simple to fix
09:20:21*mattzz is _no_ wizzard at all
09:20:44mattzzI will give it a try
09:24:15 Nick amiconn is now known as amiconn|away (
09:24:16dwihnoAre there any standard ways to truncate a file?
09:24:22dwihnospeaking of standards... :)
09:24:50LinusNopen("name", O_WRONLY|O_TRUNC);
09:25:27dwihnoSorry. I ment truncating a file, let's say I want to cut a file from 700 bytes to 500.
09:28:06LinusNprogrammatically or using a ready-made tool?
09:28:33LinusNin linux/unix, dd is your friend
09:31:26dwihnoprogrammatically :) dd is everybodys friend \o/
09:33:50 Join amiconn [0] (
09:39:14LinusNdwihno: truncate() or ftruncate() will do what you want
09:39:44dwihnoit's standard?
09:39:57LinusNposix methinks
09:40:43LinusNftruncate is the safest bet
09:40:59*dwihno checks the mingw include stuff
09:41:26dwihnonot there.
09:41:50dwihnoI guess I have to do a windows hack as well.
09:43:17dwihnoYeah. Booh!
09:44:10dwihnodjgpp has ftruncate though
09:44:17LinusNyou could try #define ftruncate chsize
09:45:12dwihnouses file descriptors
09:45:20dwihnochsize (int, long );
09:45:27LinusNthe cygwin docs claims that ftruncate is supported
09:46:31dwihnowell, mingw and cygwin are not friends then ;)
09:46:41LinusN #include <unistd.h>
09:51:50 Quit amiconn|away (Read error: 110 (Connection timed out))
09:51:50 Nick amiconn is now known as amiconn|away (
09:52:20***Saving seen data "./dancer.seen"
09:56:35 Nick mattzz is now known as mattzz|meeting (
09:57:35 Quit [IDC]Dragon ()
10:28:21 Nick amiconn|away is now known as amiconn|work (
10:49:45 Join AEnertia [0] (~aenertia@
10:50:39AEnertiaHi people, I have read the faq, but I'm going to ask a question about the av320 (which I know is unsupported).
10:51:45*dwihno holds his breath
10:51:45 Nick AEnertia is now known as `Nert (~aenertia@
10:52:29 Nick `Nert is now known as AEnertia (~aenertia@
10:53:17AEnertiaare there any projects to alter/add functionality to it? (the av3x0 series)
10:54:47AEnertiaI have googled but come up blank.
10:54:55BCI'm lost with all the different units now - what is AVOS(.sourceofrge) for?
10:56:09AEnertiathanx... that's just what I was looking for
10:56:18AEnertiaIt's not ranked on any of the google searches I tried?
10:56:42AEnertiaperhaps someone should add that to the faq on the rockbox site
10:59:12c0uttaAEnertia: /join avos
11:12:37 Join AciD [0] (
11:27:25ricflairhi all
11:28:03ricflairis it normal that the jukebox recorder makes a ticking or rattle style noise evry now and again (the unit itself, not sound throug the eaphones)
11:28:08BChi 2u2
11:28:28BCit's the hdd - mine does it a fair amount
11:28:43ricflaircool, i was worried mine is broken
11:28:58BCpanic over :)
11:29:04ricflairim getting my usb cable today, is it easy to transfer music accross and instal rockbox?
11:29:15LinusNvery easy
11:29:17dwihnosimple as pie \o/
11:29:18BCyes drag'n'drop
11:30:18ricflaircant wait
11:30:51ricflairi had it on charge since 12 last night and my wife unplugged it :( about an hour ago, do you think it will have charged ok?
11:31:16BCyeah :)
11:31:41BCbest bet is pop the batteries out CAREFULLY and give them two or three cycles in a proper recharger
11:33:04ricflairi dont have a propper recharger, should i buy one? how much. and is it not voiding my warranty if i take em out
11:34:03BCyou got a new unit with no usb cable?
11:34:29ricflairno, its not new , but its still under warranty
11:34:36BCalso thinks ...what unit JBR v1, v2, fmr???
11:34:53ricflairjukebox recorder 20
11:35:14ricflairwith firmware 1.26 on
11:35:25BCv1 or v2?
11:36:12LinusN1.26 sounds like a v1 to me
11:36:21BCgood point
11:36:35BCthen you're safe with replacing the batteries :)
11:36:52ricflairok, cool.
11:36:53BCjust BE GENTLE because of a design fault
11:37:00dwihnoThe v1 is so good.
11:37:48ricflairwhats the design fault.
11:37:53ricflairshould i have not baught it?
11:38:10BCv1's are by far and away the best units
11:38:49amiconn|workI agree.
11:38:49BCmy only complaint is that I struck unlucky and cannot flash
11:39:01ricflairoooh, i hope i can
11:39:12ricflairwhat is the deisgn fault
11:39:23BCthe battery spring is held on with solder
11:39:39dwihnoBC: the boot time with a flashed unit is so goddamn fast! :D
11:40:02ricflairbc, is that the problem that you can fix easilly?
11:40:15ricflairi think linus told me about tha tyesterday
11:40:23BCif the need ever arises Linus did a page on it
11:40:49ricflairwell evry now and again, when i plug it in, it says shutting down
11:41:23BCremember you CANNOT run the unit from the mains - it ONLY runs from batteries
11:41:52BCgod knows why they did that!
11:41:59ricflairyeah, i mean when i plugged it in to charge it up
11:43:13BCbut even when plugged it is still dependant on the battery connections
11:43:26ricflairright, so i could have that problem then
11:43:57BCthere is no 'usual' "when plugged the batteries are bypassed" thing going on
11:44:55BCalso note that when the batteries are totally dead, it take about 20-30mins of charging before the unit will have enough power to fire up the hdd
11:52:24***Saving seen data "./dancer.seen"
11:56:44ricflaircool, thank sofr the info
11:57:00ricflairare they ok to use as usb harddisks aswell
12:00:14ricflairi hope i can install rockbox later on tonight when i get my cable
12:01:50LinusNyes, it's essentially a usb hard disk with mp3 playback capability
12:04:48ricflairso i can store anything on the drive? thats cool.
12:11:23 Quit c0utta (Read error: 60 (Operation timed out))
12:14:31 Join monkey [0] (
12:16:37 Join c0utta [0] (
12:29:52 Join Laurent1 [0] (
12:33:34 Join pfavr [0] (
12:34:08 Quit Laurent1 ("Download Gaim:")
12:34:29 Join LaurentG [0] (
12:34:34dwihnoAnyone tested the koss "plug" earbuds?
12:34:55BCI must stop typing "ho" - one day someone is going to miss the reference
12:35:05BCkoss headphones are nice
12:35:10c0uttadwihno: yes, i've got a set
12:35:48c0uttathey're good
12:36:37LaurentGBC: I missed the reference indeed ;)
12:37:11BCas opposed to american slang
12:37:30BCas in "geddin' it on wid mi ho's"
12:38:06LaurentGBC: "getting it on with my ??"
12:39:52dwihnoc0utta: Just _how_ good are they then?: )
12:40:12BCI feel a monty python joke approaching...
12:40:29BC"how much do you hate the romans?"
12:40:33LaurentGso, let's get back on technical tracks
12:40:42c0uttaa lot
12:40:44LaurentGdoes anyone here know about token packing ?
12:40:47c0uttaright, you're in
12:40:48BCyou're in then
12:41:02c0uttathe only people we ate more than the romans..
12:41:19c0uttadwihno: they've got good base response
12:41:24*BC shuts up in fear of picking the wrong group
12:41:51BCLaurentG: I know of tokens and of packing but never heard the two together before
12:42:40c0uttadwihno: i like the fact the "the plug" blocks out surrounding noise
12:45:17LinusNLaurentG: i haven't heard of anything called "token packing" before
12:46:12BCsurely the point of a token is that it is so small it does not need packing
12:48:22dwihnoc0utta: I'm in the need of new plugs. I'm considering the Sony EX71 (had the EX70 earlier, but lost them) and perhaps the Koss plug might be a viable choice.
12:49:18c0uttathe koss are cheaper than the sony's, correct ?
12:55:48ricflairis the archos multimedia player better or worse than the jukebox recorder
12:56:04pfavrdwihno, I have a pair of closed headphones (not plugs), the Sennheiser HD-25 - although not that comfortable (hot ears) I'll recommend them because of the exceptional sound, but be sure to get the 70ohms version (they have a semi-prof version at half the price but it sounds bad).
12:57:16BCric: based on what - it is more "different" than better or worse
12:57:40ricflairi just read that the battery life isnt that good on it#
12:58:13BCif that is your point of comparison, then you kinda have your answer :)
12:58:36BCbut you're not going to be playing mpeg-4 video in colour on your jukebox
12:59:07LaurentGLinusN: i've found two pages about it on google but they contain only source code, no explanation of any kind... I wonder where Bjorn got this term
13:00:08BCas Enlish is not his first language maybe the term was descriptive rather than technical?
13:00:54LaurentGBC: that's possible
13:01:14 Join [IDC]Dragon [0] (
13:01:19BCWhat is it in reference to?
13:02:08LaurentGBC : about the multimedia models, he said that he found the firmware was compressed with "token packing"
13:02:52LaurentGBC: it seems a simple mechanism, since most plain text in the decrypted firmware appears nearly undisturbed
13:03:09LaurentGwhich Bjorn confirms in his post about the multimedia models firmware
13:03:48BC"most" ...what IS "disturbed"?
13:03:58BCrecurrent words?
13:05:50LaurentGBC: wait til I get a sample
13:06:21BC" a custom token packing algorithm"
13:06:42BCi think that is descriptive
13:08:28mattzz|meetinghm, anybody here who can explain firmware/include/dir.h ?
13:08:35 Nick mattzz|meeting is now known as mattzz (
13:08:35LinusNi can
13:09:36mattzzargh, plugin compile with cygwin woes also...
13:09:55mattzzduucc.s: Assembler messages:
13:09:55mattzzduucc.s:2: Warning: Line numbers must be positive; line number 0 rejected.
13:10:26mattzzusing the toolchain from
13:10:40mattzzbut that seems to be another prob
13:10:45BCMatzz, what OS do you have?
13:11:02BCI have a more updated dev-kit for that os
13:11:27mattzzworth a new release?
13:11:53BCmmmm, not an "official" one because there is still an issue with 98 that I can't solve :(
13:11:56LaurentGdamn, I have no decrypted firmware anymore ?
13:13:41[IDC]DragonLinusN: I made an experiment for the mp3 clips: I don't search for frame boundary, just throw the whol rest back in.
13:14:00LaurentGok, can build one
13:14:11[IDC]DragonSo the data should be undisturbed.
13:14:30[IDC]Dragon(just to test is we have a performance problem or else)
13:15:22[IDC]Dragonare you with me?
13:15:28LaurentGHere's a sample of the "packed" data :
13:15:32LinusNthrow back in? performance problem?
13:15:41LaurentGff20 5570 6772 6164 65ff 2050 726f 7465 . Upgrade. Prote
13:15:50LaurentG6374 ff65 6420 4a75 6b65 42ff 6f78 2020 ct.ed JukeB.ox
13:16:03[IDC]Dragon(seems not)
13:16:11LaurentGas you see, the text bits are almost "plain"
13:16:31LaurentGbut every few bytes, there's a code byte probably
13:16:32BCwhat a weiirrdd text format
13:16:44 Quit [IDC]Dragon ("no fate but what we make (EOF)")
13:17:15BC8bytes+FF, repeat
13:17:41BCa NINE byte cycle!!??
13:18:07BCthat has GOT to be a clue - who would want their data constantly moving off-boundry
13:18:27BCmaybe that FF is a PREfix
13:18:35LaurentGBC: it's not text, it's the text parts of the firmware code
13:19:08BClol - yes
13:19:26BCI've a pretty solid background in hex-dump world :)
13:22:41BCMaybe I do not get what it is that you are trying to ascertain?
13:26:02 Quit monkey ()
13:26:03LaurentGhuh ?
13:26:18BCwhat are you trying to do?
13:26:50LaurentGBC: i'm looking forward to understand the packing algorithm for the firmware ;)
13:27:19LaurentGBC: and I do not doubt your skills ;)
13:27:37BCmy first guess is that it reads 9 bytes at a time, the first byte will indicate the packet type, the latter 8 will be the data
13:27:42BCff means, raw data
13:28:04LaurentGI think it's a good guess
13:28:14*BC overstates the GUESSwork nature of his words
13:29:09LaurentGe43c 03ff 87e6 3031 3233 3435 ff36 3738 .<....012345.678
13:29:19LaurentG3941 4243 44ff 4546 4e6f 2063 6865 df63 9ABCD.EFNo che.c
13:29:26LaurentG6b73 756d a813 2e2e ff2e 0000 5352 4c00 ksum........SRL.
13:29:54LaurentGthe token + 8 data bytes seems to fit that part too
13:30:00 Quit AciD (Read error: 104 (Connection reset by peer))
13:30:10LaurentGinsert "pattern" after "bytes" ;)
13:30:54BCnow to start following it back
13:31:00BCcan you get a RAM dump?
13:31:33LaurentGRAM ? straight from the gmini ? i'm affraid not
13:33:20LaurentGbut there's a lot of text within the firmware, so this should help guessing all possible "packing" codes
13:34:24BCare there any garbage packets in amongst the text?
13:34:29LaurentGat the end of the 2nd line above, you can see a "df" instead of a "ff"
13:35:13LaurentGBC: do you mean "large parts of non text" ?
13:35:18BCwhats the next line?
13:35:40BCsmall bits of data in amongst the text
13:36:00LaurentGok, that's what I meant ;) there are a lot of them
13:36:08LaurentGlet me flood this channel and show you ;)
13:36:37LaurentG00001bb0: d123 3a20 7374 6172 749f 2073 6f66 748a .#: start. soft.
13:36:37LaurentG00001bc0: 232c 3564 ff69 7370 6c61 7920 6fcb 7074 #,5d.isplay
13:36:37LaurentG00001bd0: ad20 7385 2244 3675 6dff 7020 6d65 6d6f . s."D6um.p memo
13:36:37DBUGEnqueued KICK LaurentG
13:36:37LaurentG00001be0: 7279 af00 4572 7211 2055 c411 69ff 6e67 ry..Err.
13:36:37LaurentG00001bf0: 0000 4461 7461 bf20 4f4b 0066 64a8 106b ..Data. OK.fd..k
13:36:54BCfeel free to dcc or im me if it is appropriate
13:37:06LaurentG(BTW I can send the file to anyone wants it)
13:37:22CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
13:37:22*LaurentG wonders if gaim supports dcc ?
13:38:30 Join [IDC]Dragon [0] (
13:38:50LaurentGBC: are you receiving anything ?
13:39:02[IDC]DragonLinusN: sorry
13:39:31LaurentGBC : now ?
13:39:53LaurentGsuccess rate : 0/2 -> plain success
13:40:03LaurentGok, I'll put this on the web instead
13:40:12[IDC]Dragonin my shutup() function, I pause and search forward for the frame end. The missing part is re-scheduled.
13:40:50[IDC]DragonFor a test, I did not search, just re-scheduled the whole rest, to test if it's seamless.
13:41:10[IDC]DragonTurns out it isn't.
13:41:31[IDC]DragonSo I need to revise that part first.
13:43:59 Quit [IDC]Dragon ("no fate but what we make (EOF)")
13:44:14BC[12:39] *** Malformed DCC SEND request from laurentg, ignoring.
13:44:24BCsorry missed that
13:44:33LaurentGBC : no prob
13:44:50LaurentGIf I can remember what my home page URL is you'll get it on the web soon ;)
13:46:26LaurentGhere it is :
13:49:48BCoffset a9a0
13:50:22LaurentGhuh ?
13:50:45 Quit AEnertia (Remote closed the connection)
13:50:51BCmaybe it is my font
13:51:18BCthe third character on my screen is a mathematical "n"
13:51:52BCjust made me laugh
13:51:52LaurentGangel's are trying to communicate with you ;)
13:52:26***Saving seen data "./dancer.seen"
13:52:36BCARGH the apostrophe police
13:53:39BCthat's interesting
13:54:34LaurentGok, the 8 bytes is not a constant
13:54:41LaurentGor rather, it's not always 8 bytes
13:56:27LaurentG U p g r a d e .
13:56:27LaurentG P r o t e c t .
13:56:27LaurentGe d J u k e B .
13:56:27LaurentGo x B a c .
13:56:27LaurentGk d o o r c l .
13:56:28***Alert Mode level 1
13:56:28LaurentGo s . . . . . . .
13:56:30LaurentG. . . . & . a c t
13:56:32LaurentGi v . . . * . . B
13:56:57LaurentGafter dumping with a width of 9 bytes -> the "code" byte is not always at the end, it shifts
13:57:53LaurentGas you said, when it's ff it's 8 plain bytes, and otherwise the length of data is variable
13:58:22BCso FF is "8 uncompressed bytes"
13:59:35 Nick mattzz is now known as mattzz|work (
13:59:46BCso for other codes is the length IN or AFTER the token
13:59:57BCi suggest that it is 1+8 NOT 8+1
14:02:08BCIs there a reason (other than "cos its there") that we are repeating this work?
14:02:43 Quit pfavr ("ChatZilla 0.9.52B [Mozilla rv:1.6/1]")
14:04:24mattzz|workBC: how do we procede with the not-working w2k toolchain? Can you provide documentation or a link for what you fixed for w2k?
14:05:20BCI have recenlt ybeen offered web space, but have been too busy to upload it yet, I wish to apply the update patches and rebuild the installer with them included
14:05:46BCif you want my dev kit, just say how and I will send it your way
14:05:57BC"my" - it is NOT all "my" work
14:06:11BCroland did most of the donkey work
14:06:29***Alert Mode OFF
14:06:49BC(that was meant to be a compliment)
14:11:03LaurentGback (was on the phone)
14:11:22LaurentGBC : ok with your deductions
14:11:44LaurentGBC : i'm working on the gmini's firmware
14:12:10LaurentGthat's why I'm repeating this work ;)
14:12:11BCbut why not ask Bjo"rn
14:12:19LaurentGI did, I'm waiting for his reply
14:12:20mattzz|workBC: I will get back to you when I am @ home. I will provide you with an ftp-account for my server @ home.
14:12:40BCemail if u like?
14:13:16mattzz|workBC: how big is it?
14:13:34BC7.5MB before mime - i also have it split in 2 with rar
14:14:20mattzz|workBC: that should be ok. send it to "mattzz at"
14:14:52mattzz|workBC: smaller chunks might be safer, dunno what exim accepts as max size at the moment....
14:15:11mattzz|workgotta go... meeting.
14:15:18 Nick mattzz|work is now known as mattzz|meeting (
14:28:05 Part LinusN
14:32:46 Join deathsc0ut|Delir [0] (
14:34:29DBUGEnqueued KICK deathsc0ut|Delir
14:34:29deathsc0ut|DelirHello could somebody tell me some hard disks which work good with the AJBR. Or could somebody tell me whats most important for chosing a hard disk for the jukebox?
14:42:26BCpower consumption
14:42:29BCslow speed
14:42:40BC*low* power consumption
14:42:56deathsc0ut|DelirOne question .. if the speed isnt that good do i access my data slower
14:43:16deathsc0ut|Deliras if it would be faster..? Or does it depend on the buffer of the ram in the jb?
14:43:22BCyou need to read up to 320Kb per second to play the worst case MP3
14:43:36BCon drive cache is not used
14:44:06BCthat's 40KB/S
14:44:25deathsc0ut|DelirOkay Could you give me a hint of a good Happy Day because I dont know which one I should take
14:44:35deathsc0ut|Delirhard disk not happy day
14:45:12BCshould I do it with embedded commands
14:45:24***Alert Mode level 1
14:45:24deathsc0ut|DelirHe He He
14:46:20 Nick mattzz|meeting is now known as mattzz|work (
14:46:32BCif you think that my advice is valuable you will review the power consumption ratings
14:47:14deathsc0ut|DelirDo they say them in the hard disk information?
14:51:12***Alert Mode level 2
14:51:12deathsc0ut|DelirAh Okay I found a site where I can find most of this stuf
15:01:13***Alert Mode OFF
15:06:21 Join [IDC]Dragon [0] (
15:11:22LaurentGoops got an answer from Bjorn
15:11:38LaurentGbasically it says : "if it's the same you should find it" ;)
15:13:09LaurentGok, first a tea, then let's try ;)
15:13:20BCyes, good idea
15:13:24BC(kettle) brb
15:20:09 Quit deathsc0ut|Delir (" Info-[v9.4.22]- Released-[August 19, 2002]- Channel-[#Excurs")
15:23:54LaurentGbk too
15:24:20LaurentGDarjeeling anyone ? ;)
15:24:46BC"brown" today
15:25:33BC"Tesco Premium" a choice, not a default
15:25:35LaurentGactually, I let it infuse all along while I drink, so it's light at the beginning, and brown in the end ;)
15:25:48BClikewise :)
15:26:08LaurentGall great minds do the same I guess
15:26:38BCgotta get every milliounce of caffeine out if it :)
15:26:47LaurentGteine ;)
15:29:50 Join AciD [0] (
15:36:40 Quit [IDC]Dragon ("no fate but what we make (EOF)")
15:36:57 Join deathsc0ut|Delir [0] (
15:39:12DBUGEnqueued KICK deathsc0ut|Delir
15:39:12deathsc0ut|DelirHello another question just came into my mind before I will order a new hard disk.. I want to be sure that I can format the new hard disk while it is connected via usb to the pc - is that correct?
15:41:46LaurentGhuh ?
15:42:07LaurentGshould'nt it be ok as long as the HD is connected to the Archos ?
15:42:57BCi guess death meant via the usb in the jokebox
15:43:40LaurentGsure, that makes sens
15:44:35BCI'm glad we have served you well in a way tailored to what is right for you :)
15:44:39BC(be happy)
15:44:44deathsc0ut|Delirjep you have
15:45:18LaurentGnow, please consider making a donation to our headquarters ;)
15:45:22LaurentG(joking !)
15:45:36***Alert Mode level 1
15:45:36deathsc0ut|Delir{{ gHg t }}
15:47:07deathsc0ut|DelirAlright ... I think I dunno another question at the moment and if i will need your help again i will be back :p
15:47:09***Alert Mode level 2
15:47:09deathsc0ut|Delirthx and See Ya!
15:47:24 Quit deathsc0ut|Delir (" Info-[v9.4.22]- Released-[August 19, 2002]- Channel-[#Excurs")
15:49:45 Join mecraw__ [0] (~mecraw@
15:52:29***Saving seen data "./dancer.seen"
15:52:57mattzz|workBC: toolchain on its way?
15:53:06BCshould be there
15:53:46BCwent 48mins ago
15:54:27mattzz|worknot yet5
15:57:10***Alert Mode OFF
15:58:38mattzz|workBC: it is there
15:59:06mattzz|workBC: mailfilter was smarter than me - again....
16:00:28LaurentGGOTCHA !
16:00:48BCdone it
16:00:59LaurentGI got the unpacking stuff
16:01:04LaurentG(part of it actually)
16:01:10BCwell done
16:02:30LaurentGhum, no
16:02:34LaurentGdumbass me
16:02:52LaurentGyou said it
16:03:13LaurentGI need to learn that "turn your tongue ..." stuff
16:04:08LaurentGactually it would have been a pretty dumb scheme if it was that
16:04:45BCare these tokens pointers to data blocks?
16:04:58LaurentGI had : "FF" -> 8 ue, 0 ce ; "FE" -> 1 ce, 7 ue ; "7F" -> 7 ue, 1 ce ; "9F" -> 6 ue, 1 ce
16:05:12LaurentGue = uncompressed elt, ce = compressed elt
16:05:56LaurentGyes, it looks like every "non ff" token means "there are two bytes pointers to data blocks within the next 8 elements
16:06:14LaurentGelt = element, because pointers seem to be two bytes
16:07:19 Nick BC is now known as bc|brb (
16:07:19DBUGEnqueued KICK bc|brb
16:07:19LaurentGI need to get more examples to figure the coding, but as Bjorn sais, it's probably not very difficult
16:38:39 Nick bc|brb is now known as BC (
16:38:39DBUGEnqueued KICK BC
16:39:12BCif you can find a pointer in the MIDDLE of a block of text you can find the missing text elsewhere in ROM and match it up with the point?
16:39:56LaurentGI'm not sure I understood, what do you mean by "ROM" ?
16:40:15BCscrub that -the ROMs encrypted
16:40:28BCAFTER encryption
16:40:50CtcpIgnored 4 channel CTCP requests in 2 hours and 4 minutes at the last flood
16:40:50*BC has a brain fart
16:41:16LaurentGok, but I haven't access to the ROM, my gimi is hardware opaque to me
16:41:32BCwhat is this rom dump then?
16:42:00LaurentGit's the archos firmware after unxoring
16:42:24LaurentGgmini firmware sorry
16:43:03BCokay ROM==gmini firmware dump
16:43:10LaurentGhum, ok I just understood what you said
16:43:24LaurentGneurons take some time to heat ;)
16:43:44BCmore tea?
16:44:49LaurentGI haven't searched yet, I was focusing on identifying all "codes" and their relation with packed/unpacked elements following them
16:45:16LaurentGbut you're right, this might explain how pointers work
16:45:37LaurentGhum, wait I already have such an example, see :
16:45:55LaurentG00001a90: 54ff 6f20 7570 6461 7465 ff2c 2070 6c65 T.o update., ple
16:45:55LaurentG00001aa0: 6173 65fe ca10 7567 2044 4320 49ff 6e00 DC I.n.
16:45:55LaurentG00001ab0: 0023 5345 5249 7f41 4c23 0000 2023 e911 .#SERI.AL#.. #..
16:46:57LaurentGone can assume that the "<space>pl" from ",<space> please" is used to produce "<space>plug DC"
16:47:55LaurentGso the code here is "FE" which means 1 ce (pointer to " pl"), then 7 ue
16:48:13LaurentGand the pointer to " pl" is $ca10
16:49:28BCi would need to see some onscreen messages to be convinced by that
16:49:33LaurentGI think the pointer is in fact one byte long only otherwise the text would be less readable
16:49:58BCis it a pointer - or is it a token - if it is a token then there are only 255 of them!?
16:50:17LaurentGBC : that's an easy one, it's the message displayed when updating the firmware, I did it thrice and it says to just that
16:50:20BCif it is a pointer then it can only address 255bytes of memory space
16:50:55LaurentG"says just that" (where did this "to" come from ??)
16:52:37BCWhen the gmini is running what is the full message "To update, please ********ug DC In
16:52:56BC"JukeBox Backdoor closed" ?????
16:53:13LaurentGfrom memory, it's "To update, please plug DC in"
16:53:22LaurentGBC : interesting one no ? ;)
16:53:45LaurentGI noticed it too, wonder how to use it ;)
16:54:03BCsomething will print it - trace it back like that
16:54:27LaurentGBC : yes, but that means we can disassemble it, it's not the case yet :(
16:54:38*LaurentG curses Samsung
17:04:21LaurentGgood, there's a new version of the firmware so I'll be able to check if the message is correct
17:04:39LaurentGlots of bugfixes, good thing ;)
17:07:49 Join Ka__ [0] (~tkirk@
17:08:25 Join methangas [0] (
17:13:16mattzz|workcu l8er
17:13:42 Quit mattzz|work ("CGI:IRC")
17:26:35LaurentGOk, here's the update string : "Update file found Please plug DC-IN to proceed"
17:27:00LaurentGwhich demonstrates one thing -> I have unxored the wrong firmware :)
17:27:29LaurentGwhich means ?
17:27:38BClaughinh my arse off
17:28:22LaurentGhum, confirmation : i'm working on the 1.20 while my last one was the 1.50
17:28:23BCIn trillian (and others) you get a little smiley with tears rolling down his face
17:28:44LaurentGI think I disabled smileys in gaim
17:28:59BCoften wise when sending code
17:29:23LaurentGsure :)
17:29:31LaurentGbut I'm not that wise !
17:29:47LaurentGso, let's see how LMAO appears
17:29:52LaurentGplain text :)
17:30:15LaurentGnow I know why I disabled smileys :)
17:52:31***Saving seen data "./dancer.seen"
17:53:41 Quit Nibbler (Read error: 54 (Connection reset by peer))
17:58:06 Nick amiconn|work is now known as amiconn|away (
17:58:25 Part BC
18:03:38 Quit AciD (Read error: 54 (Connection reset by peer))
18:05:19 Join BC [0] (
18:26:08 Join pfavr [0] (
18:56:37 Nick amiconn|away is now known as amiconn (
18:58:22 Quit pfavr ("ChatZilla 0.9.52B [Mozilla rv:1.6/1]")
19:07:33 Join Ka___ [0] (~tkirk@
19:10:45 Nick BC is now known as BC|bbl (
19:13:24 Join CpuMan2001 [0] (~jirc@
19:15:37 Quit CpuMan2001 (Client Quit)
19:16:05 Quit Ka__ (Read error: 60 (Operation timed out))
19:17:54 Join mattzz [0] (
19:18:45amiconnHi mattzz
19:18:51mattzzhi amiconn
19:19:55 Join matsl [0] (
19:22:39amiconnmattzz: Just an idea for a minor speedup: Instead of (n_iter % 8) you can use (n_iter & 7)
19:23:28LaurentGi'd call that a major speedup ;)
19:23:50amiconnAnother one: for implementing the idea "incremental recalculation when moving" the gray_scroll_xxx() functions are suited perfectly.
19:24:34 Join AciD [0] (
19:24:38mattzzamiconn: I will check. Perhaps using only 6 rows for graphics would be also a good idea?
19:25:22amiconnLaurentG: It is not so minor if called often, because the % operator uses a subroutine, while the & is included directly as a machine instruction.
19:26:02mattzzamiconn: a major drawback is that I can't simulate the grayscale stuff (yet?!).... The usb-plug becomes kind of worn out....
19:26:27amiconnmattzz: You could do that, but I think it looks better with 8 rows.
19:26:28LaurentGamiconn: youre 100% right
19:26:36mattzzLaurentG: dont you mess with Mr. 3% ;-)
19:26:49mattzzLaurentG: or he will speed you up
19:26:50LaurentGmattzz: ;)
19:27:20amiconnI did also have to plug the usb in & out a number of times; meanwhile I do it at the PC end when developing.
19:29:00amiconnThe problem with simulation grayscale - how do you simulate something that is only implemented as a plugin and hence not part of the core?
19:30:03mattzzamiconn: regarding the r_max discussion (4 vs. 9): Maybe the 2 bits increased accuracy for the fractional part are worth a r_max of 4? (Sorry for me starting that topic again;-)
19:30:37mattzzamiconn: 26 bit with r_max of 4 looks fine @ 17 levels of gray
19:30:57mattzzamiconn: and is _waaaay_ faster ;-)
19:32:15mattzzargh, wrong window
19:33:14amiconnAccuracy - maybe getting 2 extra bits is good; I did hit the precision barrier already while zooming in (looks very blocky then).
19:33:55amiconnmattzz: Is it something like 3% faster? ;-)
19:34:23mattzzamiconn: definitively
19:35:25amiconnTo make the default image look a bit better when using r_max = 4, you should crop it a little, especially at the left (maybe using Xmin = -2.2)
19:37:43 Join Nibbler [0] (
19:37:46amiconnBtw: While tweaking the X and Y boundaries, a check of the aspect ratio would be good - the default mandelbrot looks a bit distorted. Maybe you didn't take into account that the pixels are not square.
19:39:30mattzzx_min = -2 is fine (I want integers in init_mandelbrot_set() ...)
19:39:51mattzzwhat is the aspect ratio of a pixel?
19:41:05amiconnThe pixels are taller than wide; width/height = 0.8
19:41:51amiconnmattzz: Do you want an extra 3% speedup? ;)
19:43:26amiconnI could send you my 3% accelerated (average) pixel setting routine (that I didn't post), but beware that the asm looks really strange.
19:43:47amiconn(The thing that caused you to call me Mr. 3%)
19:43:50mattzzasm always looks strange (to me)
19:44:39mattzzamiconn: btw, remember that email you sent me the other day? t-online kept it for nearly 7 hours... finally I got it.
19:45:33amiconnHmm, most of the time the t-online servers aren't _that_ lazy...
19:49:52 Join nelliep [0] (jirc@
19:50:19BC|bbl% is horribly expensive in comparison to &
19:50:25 Nick BC|bbl is now known as bc (
19:52:32***Saving seen data "./dancer.seen"
19:53:45amiconnBC: yes, because it uses a lib call which even does division (slow on SH).
19:54:16 Nick bc is now known as BC (
19:56:31 Join deadite66 [0] (
19:59:55 Quit nelliep ("Leaving")
20:00:50LaurentGbe back
20:00:52 Quit LaurentG ("Download Gaim:")
20:03:04mattzzamiconn: I made a screenshot of the grayscale mandelbrot
20:03:19mattzzamiconn: I did it the old school way...
20:04:25BCcool - how long to generate?
20:05:05mattzzBC: about 4 secs
20:05:45mattzzBC: how long to zoom? depends on how much "black" there is.
20:05:54BCit does zoom though?
20:06:21mattzzBC: sure
20:06:53mattzzboy, this plugin is so much pointless ;-)
20:07:22BCaren't they all?
20:07:42BCtesting my card game framework by writing solitarie atm
20:08:35mattzzBC: I was addicted to solitaire on my windoze 3.11 notebook once. vegas points, 3 cards - a nightmare
20:11:26mattzzBC: btw: the devkit installer rules! nice package. Unfortunately we have some policies installed on our company's PCs that allows me to run it as localadmin only.
20:11:57mattzzBC: regedit is blocked
20:12:12mattzzBC: when logged into the domain
20:12:59BChmmmmmm - not sure what the workaround for that is
20:13:06BCthanks for the compliment :)
20:31:07 Nick BC is now known as bc|busy (
20:42:07 Nick mattzz is now known as mattzz|away (
20:49:25 Quit deadite66 (Read error: 110 (Connection timed out))
20:51:14 Quit AciD (Connection timed out)
20:59:22 Join pfavr [0] (
21:06:51 Join LaurentG [0] (
21:10:00 Quit pfavr ("ChatZilla 0.9.52B [Mozilla rv:1.6/1]")
21:17:19 Join ricflaira [0] (
21:17:26ricflairahello all
21:18:14ricflairaall is fine with my archos recorder jukebox now aside from one problem, if i plug it in to charge, somtimes if i slightly touch the cable of the adapter or the archos itself it says "shutting down"
21:19:58LaurentGI see two possibilities : either a bad contact which causes the firmware to think you asked him to shutdown, either a bug in the firmware. Are you using the original firmware or Rockbox ?
21:21:26ricflairajust installed rockbox now, but in flash is original firmware. Should i flash with rockbox?
21:21:41LaurentGno, wait until it suits you before flashing
21:22:05LaurentGit's hard to get back in case of mistake so better take care first ;)
21:22:31LaurentGso does the shutdown occurs when you are using rockbox ?
21:23:28ricflairano, only when it is turned off and i plug in the charger and either touch the charger wire or knock the jukebox(charger is fine as i have tried 2 diff ones)
21:25:01LaurentGok, i'm no rockbox expert, but I guess it's the original firmware which handles this part. but the cause definitely looks like there's a bad electrical contact within your recorder's case.
21:25:47ricflairahmm, i hope not. i cant afford to post it to usa to get it fixed
21:25:49LaurentGI suggest either mailing to the list or waiting for an expert ;)
21:26:03ricflairahow do i know when it is fully charged
21:26:04LaurentGdid you open the case ?
21:26:22LaurentGis it new ?
21:26:32ricflaira2nd hand
21:27:21LaurentGok, so you're on your own. as for battery charge, I think there's an indicator of it within rockbox (a percentage I think)
21:27:41ricflairawhere abouts is that? because the battery in the top corner has a ? in it
21:28:22LaurentGonce again, i'm not sure of it since I do not have any rockbox able player/recorder, but I seem to recall that it shows on the top menu
21:29:03LaurentGa "?" ? well, you should definitely see with someone else than me ;)
21:29:16LaurentGI've reached the limits of my rockbox knowledge ;)
21:30:18ricflaira:) thanks anyway
21:30:59LaurentGis it the first time you charged it ?
21:31:26amiconnricflair: See, last question.
21:33:43ricflairajust wondering if when it gets to 100 percent it will say full or somthing
21:36:34LaurentGno idea
21:42:43amiconnricflaira: It doesn't say "full", but if you have rockbox running you can look under Info->Rockbox Info. The third text line tells you the battery state.
21:43:05ricflairaok, thanks, ill check that
21:44:12ricflairajust checked that and all it tells me is the free hd space and also buffer
21:45:00amiconnThe battery state is given below the Buffer size
21:45:14ricflairasays (na)
21:45:38ricflairanow says 100 percent
21:45:42amiconnYes, you have to wait at least 1 minute, see battery faq.
21:45:44ricflairabut ive only had it on charge for about 3 hours
21:46:44 Join scott666 [0] (
21:46:58amiconnSo maybe either the batteries weren't completely empty when you started charging, or the batteries have lost some capacity.
21:47:38amiconn(Or the batteries have a lower capacity rating than the original ones)
21:48:55*LaurentG misses Visual C++'s IDE, and notices that he never thought he'd say such a thing about a Microsoft product....
21:49:19ricflairathey are the original ones.
21:49:30ricflairaits now down to 97 percent after just about 2 or 3 mins of play
21:51:48amiconnThis is perfectly ok as the battery state is (and only can be) a rough estimate. I recommend you to have a look at the manual ( as well as the faqs.
21:52:19ricflairaok thank you.
21:52:31ricflairafrom empty to full, how long od yuou reckon it will take
21:52:35***Saving seen data "./dancer.seen"
21:53:59amiconnIf the batteries are completely empty, the original one (1500 mAh) and the original charger is used (assuming you have a recorder v1), it should take about 6 hours.
21:54:24ricflairacool, i have the recorder 20 v1 yeah. and a new charger
21:57:30amiconnIf you use a different charger, the charging time depends on the output voltage of the charger, the maximum current it can deliver and whether the voltage is regulated or not.
21:59:11ricflairaright. Ive got a feeling my batteries might be diying , ill buy some new ones soon
22:00:10amiconnHowever, you cannot charge faster than with the original faster, and if the charging voltage goes above 12V the charging circuit within the archos is at risk.
22:00:24amiconns/original faster/original charger/
22:01:19ricflairaim charging at 9v.
22:01:38ricflairaill get some new batteries and use an external charger, is it easy to take the batteries out?
22:03:36amiconnIt is not too complicated (it is described in the original manual), but you should do it with a little caution, because the construction of the battery contacts is a bit fragile.
22:03:54ricflairawould you reccomend getting new batteries then??
22:04:11ricflairaand if i get some, and charge at 9v, will the archos charge the new batteries ok?
22:06:14ricflairamy battery life has gone from 97 percent to 83 percent within about 6 songs
22:06:14 Quit mecraw__ (Read error: 104 (Connection reset by peer))
22:06:23 Join mecraw__ [0] (~mecraw@
22:07:04amiconnIf your batteries are worn out (that is, you don't get the runtime you could normally expect even after a full charge), I would definitely recommend new ones.
22:08:06LaurentGbut you should try the deep discharge mode first, to see if they can recover
22:08:13amiconnIf the batteries are worn out chances are they may start to leak (just discovered that with mine).
22:09:51amiconn(LaurentG), ricflaira: Yes, first try to run down the batteries until the unit stops, then recharge until they are full. Repeat this several times to get rid of any possible memory effect.
22:10:34ricflairawill do, thanks.
22:10:37amiconnricflaira: If your 9V charger is regulated (stabilized) it may take more than 6 hours for a full recharge.
22:11:00ricflairawould formatting the drive effect it?
22:11:47ricflairai guess not
22:14:00LaurentGricflaira: if you want to drain your batteries as fast as possible, then connect the unit to a PC and keep reading from the disk, this is the highest power drain you can get
22:14:27LaurentG(beware not to write to the disk, as a power failure during such an operation might cause harm to it)
22:14:37 Join pfavr [0] (
22:14:57LaurentGbut there's no reason to hurry, normal usage is fine too ;)
22:16:15ricflairaok, cool. Ill drain it right down and then recharge and see how that goes. Failing that ill go and buy some new ones from maplin
22:16:46LaurentGricflaira : do it several times to be sure
22:17:34ricflairais this a common faukt
22:20:19ricflairawould these be ok
22:21:04scott666VBscript runtime error...
22:21:08 Nick mattzz|away is now known as mattzz (
22:22:16mattzzamiconn: I reverted x_min to -2.5 again. I need an aspect ration of 1.75, otherwise there will be stretching while zooming
22:23:28mattzzamiconn: And I put some nicer screenshots on">
22:24:22scott666crazy looking
22:28:54 Quit Ka___ ("Leaving")
22:29:16 Join lurwas [0] (
22:29:23LaurentGnice grayscale levels
22:29:41amiconnThe sections are looking really nice. Btw, I did also get various nice pictures. Too bad there is no save function yet ;-)
22:30:09lurwasDoes anyone know if it's possible to upgrade the Archos 6000 with a 20GB HD instead?
22:30:58amiconnmattzz: Concerning the aspect ratio: Either you can change the aspect ration as well if you change the initial aspect, or better still, you could use separate delta values for x and y.
22:31:04amiconnlurwas: yes
22:31:46lurwasamiconn: Ok, great :) Have you done it yourself? Anything special you have to do to get it open etc?
22:33:03 Join AciD [0] (
22:33:09mattzzamiconn: that would make -3% I guess
22:33:20mattzzamiconn: using a delta
22:33:36amiconnNo, I haven't done it myself, since I happen to have a recorder (and didn't upgrade the hd yet). But there is a link at the rockbox docs page:
22:34:13lurwasamiconn: Thanks man :)
22:34:59amiconnmattzz: Why? You use a delta atm and would replace it with, say deltax and deltay. These are only recalculated when the zoom factor changes.
22:35:49lurwasDoes anyone know of any case mods for the Archos Jukebox? I find it to be rather ugly, and want ti make a casemod for it....
22:35:59lurwasto even
22:36:04amiconnmattzz: It would ensure that the initial aspect will be kept regardless of the starting value.
22:36:46mattzzamiconn: you are right
22:37:27 Quit ricflaira (Read error: 60 (Operation timed out))
22:40:18mattzzI am too tired to do this tonight (last night was short enough). I will put mandelbrot.c and graubrot.c into the patchtracker for now - maybe some of the allmighty CVS gods have mercy with me ;-)
22:40:41LaurentGlurwas: all case mods are listed on the web site
22:41:09LaurentGbut I don't recall anyone replacing the case
22:41:11lurwasLaurentG: Ok, I need to surf around a bit more there then, thanks :)
22:41:41 Quit methangas (" HydraIRC -> <- irc client ownage!")
22:42:00LaurentGCVS god are nice ones, the ugliest ones are those of the fiscal administration ;)
22:42:46amiconnmattzz: Do you put it up with or without deltax-deltay?
22:43:04mattzzamiconn: without, should I wait?
22:43:41amiconnHmm, I really don't know, but I _could_ be temted to commit it...
22:44:04amiconn*tempted even
22:44:51mattzzamiconn: feel free to do so. If you want, I put both .rocks in the usual place and you can modify and commit them to the patchtracker
22:48:04 Quit pfavr ("ChatZilla 0.9.52B [Mozilla rv:1.6/1]")
23:01:03mattzzamiconn: coding? ;-)
23:07:44 Join hardeep [0] (1098@
23:08:01 Quit matsl (Remote closed the connection)
23:50:27mattzznight folx
23:50:53 Quit mattzz ("Client exiting")
23:52:38***Saving seen data "./dancer.seen"

Previous day | Next day