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

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

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

#rockbox log for 2005-06-22

00:00:09preglowi'm coding resampling now
00:00:41preglowsince rockbox can'thandle 48khz files
00:01:09eightbitWhat I also would like to do some time is turn my thing into a little sampler
00:01:51 Join MrStaticVoid [0] (
00:02:19 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
00:02:26eightbitFor a joke I made a loop with the A-B function then speed shifted it until it matched the speed of another song
00:02:40eightbitplayed to my other ear from my friends ipod
00:02:40preglowi'm hoping plugins will have recording functionality some time
00:03:00eightbitThat would be needed for an fx box
00:03:01 Join cYmen [0] (
00:03:05preglowof course
00:03:09preglowthat's why i'm hoping it ;)
00:03:28preglowcan't see a reason why they shouldn't if they can play back sound
00:03:39 Quit Febs_away ("Chatzilla [Firefox 1.0.4/20050511]")
00:03:48eightbitI agree. It would lead to some amazing plugins
00:03:54eightbitIs anyone going to implant a better a-b function
00:04:01eightbitOne with out a pause
00:04:22preglowa-b function?
00:04:32eightbitthe looper
00:04:40preglowthere is a looper? :)
00:04:55eightbityeah, press the a-b in the ordinary firmware
00:05:23preglowif someone wants it and is able to code it, it'll be there
00:05:48eightbityeah, its just I think it would need some flexibility in the playback code
00:06:28eightbitif you make a long enough loop rockbox would have to rebuffer the first part of it
00:07:27preglowmight be tricky
00:07:49Bagderisn't our size_t unsigned?
00:10:07preglowsize_t should alays be unsigned
00:10:39Bagderthat's what I think too
00:10:52preglowit's not a matter of what you think, it's a c spec thing :)
00:11:01preglowif ours isn't, it needs to be fixed, heh
00:11:13Bagderimagine size_t size;..
00:11:19Bagdersize -= copy_n;
00:11:19Bagder if (size > 0) {
00:11:27Bagderthat is a weird check
00:11:30preglowoh yes
00:12:17 Join matsl [0] (
00:12:34BagderI don't think it explains my patch's failures, but it sure isn't right either
00:12:57preglowoh no, you need to check if copy_n is larger than size before you subtract
00:13:13preglowmight not fit snugly in with the code, but
00:13:53BagderI'll read the code again to verify
00:14:28Bagdercopy_n is never bigger than size, so it works
00:16:57*HCl slumps over cause of boredom :/
00:19:33preglowi really don't understand people who complain about boredom
00:19:36preglowthere's plenty to do
00:19:50preglowtoo little time, though
00:20:14Suckathe worst kind of boredom is boredom when you know you have things you should be doing :)
00:20:20HCl :P
00:20:29tvelocityboredom is so boring
00:20:32HClSlasheri: event done yet?
00:20:36HCli made runtime database code
00:20:38tvelocityand you are too bored to do anything
00:20:40HClbut i can't test till tomorrow.
00:20:42preglowwhat event?
00:20:51Suckaruntime database code? what sorcery is this!?
00:20:56Bagderthe audio buffer we use atm is... _huge_
00:21:03HCltrack end event thing.
00:21:08preglowhacking in an event callback in the playback code should be trivial
00:21:18HCli have a hack like that
00:21:24HClbut slasheri was gonna make it a not hack
00:21:26preglowBagder: yes, indeed
00:21:29HClwhich is better.
00:21:39Bagder32166376 bytes with my patch applied
00:21:49HClhow many bytes do we have in total?
00:21:56preglowBagder: ahh, that buffer, that's a good thing, yes?
00:22:03Bagderthat's goodie
00:22:12Bagder32MB is 33554432
00:22:17HCl :)
00:22:36 Join n0bby [0] (
00:22:38Bagderso 95.8% is for the buffer
00:22:41HCl :P
00:22:53HClwe might want to start thinking about other things that would be nice in ram
00:23:02HCli can certainly think of some regarding the database
00:23:31HClwith a red black tree we could most likely even save memory space.
00:23:47HClhrm. i'm not too good at red black trees though :/
00:23:51HClcomplicated buggers
00:24:06HClfast though
00:24:19HClguarantees log n
00:24:39HCland adding and removing is fast too.
00:24:41preglowi don't have 48khz mp3s
00:24:47HClme neither..
00:24:48HClmake one?
00:25:09preglowoh, i do have some other 48khz files
00:25:43 Quit matsl (Remote closed the connection)
00:25:58preglowwell, if you want to use trees, you pretty much need to have the entire database in memory at all times
00:25:59n0bbyout of interest, do you guys use rockbox on your irivers when youre out and about?
00:26:05preglown0bby: oh sure
00:26:21n0bbycarry a pin in case of crashage? :P
00:26:22Bagdersince long
00:26:29Suckai have a 48khz track if you want it
00:26:36Suckaseveral in fact
00:26:39preglowSucka: nah, i'm alright
00:27:13n0bbydoes anyone actually find that 48khz sounds better?
00:27:49Suckaon rockbox you mean, or in general?
00:27:59preglowon rockbox it sounds worse, heh
00:28:09Suckai never really noticed, i only have some 48khz tracks because i left it too high by mistake
00:28:24Suckathe first few times they played back on rockboy it gave me quite a fright
00:28:25HClpreglow: i don't think it would cost too much memory...
00:28:35HClwhen you use trees, you can also get rid of the allocate the max length of memory
00:28:38Suckathought the heat was playing up my speakers or something, then i realised i wasnt listening to cassette tapes anymore
00:28:40 Nick n0bby is now known as n0bby-brb (
00:28:50HCland then its what, 3000 times the full path of a file
00:28:56HClsay 120 on average
00:29:06HCl360000 bytes
00:29:12HClplus some overhead for the tree structure
00:29:23HClprobably 12 bytes left right and parent
00:29:38HCland the hash and the rundb entry and tagdb entry
00:29:57HCli dunno
00:30:05HCllets see how it performs with plain binary search on disk
00:30:11HClit might very well not be needed
00:30:32HClit would be nice if we could get rid of a database generator completely though.
00:30:38HCland have it build itself on the fly
00:31:14HClprobably only possible on iriver
00:31:19HClsince archos simply does not have enough ram..
00:31:22 Quit dwihno (Read error: 110 (Connection timed out))
00:31:29preglowunified code for all platforms would be nice..
00:31:34 Nick n0bby-brb is now known as n0bby (
00:31:48HClyou could do the database generation with mergesort on disk
00:31:49preglowanywho, i don't much care, i don't think i'll start using the database now
00:31:49HCleven on archos
00:32:00HClmmm, i want to get the runtime database in
00:32:07HCladd my changes to the searchengine
00:32:16HCland then start for a menu with preselectable searches
00:32:26HClduplicating rio karma functionality
00:32:35n0bbyhcl, i like this idea :)
00:32:37 Join Febs [0] (
00:32:43 Quit Febs (Client Quit)
00:32:49HClwhich? :P
00:32:59n0bbythe predefined searches :)
00:33:13HClit will most likely be something like a directory with search files
00:33:16HClthat you can just select one from
00:33:26n0bbynot very slick
00:33:41HClwell, if done well
00:33:44HClyou wouldn't see the difference
00:33:48Bagderstill crashes ;-(
00:33:50HClonly that you can adjust them 100%
00:33:56preglowresampling kinda works
00:34:14HCland add your own predefined searches
00:34:22HClthats why i've been keeping databox and searchengine split up
00:34:31n0bbyi have a feature request, but im not sure if its possible. different icons for different audio filetypes?
00:34:33 Join Febs [0] (
00:34:42HClit is.
00:34:42n0bbyso i can tell my oggs from my mp3s
00:34:43HClvery easily.
00:34:49HCljust modify viewers.config
00:34:54HClthe last bytes are the icon
00:35:13HClfeel free to improve and submit a patch :)
00:35:13n0bbyi once wasted a whole battery with a single album in wav format :S
00:35:20n0bbyon an old iriver firmware
00:35:38HCl :X
00:35:50HClwhy does wav waste battery on iriver?
00:35:57n0bbybad buffering
00:36:05HClafkish bit.
00:36:07n0bbynear constant HDD access
00:36:18preglowi resample
00:36:57BagderSlasheri: if you read the logs, my latest (still not working) patch is now
00:37:06n0bbyi thought the archos had to have icons in the font, and it had a 5 icon limit or something?
00:37:48preglowcan anyone think of how to convert two sampling rates to a 16.16 fixed point delta?
00:38:38 Quit Stryke` ("Friends don't let friends listen to Anti-Flag")
00:38:45 Join Stryke` [0] (
00:42:32 Join webguest49 [0] (
00:42:39 Join TCK [0] (
00:42:58 Part webguest49
00:43:30 Part Febs
00:46:13preglowthis might be a problem
00:47:33 Quit Aison ("( :: NoNameScript 3.72 :: )")
00:48:58 Join muesli- [0] (
00:49:54n0bbycan bluechip's dev enviroment complile DUMB properly, or does it choke the same as a home-made cygwin complier?
00:51:57preglowchoke, i guess
00:52:42 Quit n0bby ()
00:56:12preglowhelp :/
00:59:04preglowsomething's wrong and i don't know what, heh
01:00:44preglowdecent resampling :P
01:00:55Christi-SYou won?
01:01:00preglowbig time
01:01:03 Quit Stryke` (Read error: 110 (Connection timed out))
01:02:19preglowsounds ok ish
01:02:22preglowanyone wanna try?
01:02:31preglow48khz mp3s required
01:02:45Christi-SI would but I only have 44khz files.
01:03:06preglowhmm, nasty clicking
01:08:13 Join muesli__ [0] (
01:14:51 Join jwalk [0] (~jwalk_edm@
01:15:46 Quit Harpy (Read error: 145 (Connection timed out))
01:16:15preglowTHERE, goddamnit, nice resampling
01:16:45preglowi'm off to bed, later all
01:16:58 Quit preglow ("leaving")
01:17:54jwalkhi all great work. no brick yet :)
01:18:26jwalkAm I correct in assuming that the rockbox volume on iRiver is set as loud as it can go?
01:19:18jwalkThat is, can the maximum be made louder?
01:19:51jwalkAt 100% with Shure e2C's, I still feel I'd like more.
01:19:58 Quit Christi-S ("If I were actually witty, this quitline would be funny.")
01:23:53 Quit muesli__ (Read error: 60 (Operation timed out))
01:24:20thegeekthen get a small amp
01:24:23thegeeka cmoy f.ex
01:25:55jwalkhehe. to much to carry already
01:33:24 Quit muesli- (Read error: 110 (Connection timed out))
01:34:44HCljwalk: we could probably do some software scaling, me thinks?
01:34:58HClbut that would be a seperate volume thing, planned with the runtime database
01:35:11HClit supports a custom volume setting per song
01:35:22HClyou can also try getting other headphones
01:35:23jwalkas in replaygain?
01:35:26HClthey seem to make a huge difference
01:35:34HClwell, we're gonna have replaygain too
01:35:39HClbut this one will be user settable
01:36:09jwalkI just can't believe that in-ear phones are too quiet for me. maybe I'm going deaf?
01:36:27HClthere's a huge difference in headphones
01:36:42HClmine are too quiet too, but amiconn has ones that can function as loudspeakers on max
01:36:46jwalkany suggestions?
01:36:56HCleither patience or new headphones
01:37:10jwalkas in headphone suggestions :)
01:37:11HClor help on coding replaygain :p if you know how
01:37:13HCl :p
01:37:16HClno, sorry
01:37:18HClno clue
01:37:21HCli need new ones myself
01:37:26jwalkmy coding is limited to VB
01:42:07 Quit Kohlrabi (Read error: 131 (Connection reset by peer))
01:42:21***Saving seen data "./dancer.seen"
01:48:57 Quit cYmen ("zZz")
01:52:20 Quit Sucka ("a bird in the bush is worth two in your house")
01:57:21 Part eightbit
01:57:51 Quit jwalk ("Chatzilla [Firefox 1.0.4/20050509]")
02:06:37 Join StrathAFK [0] (
02:10:19 Quit hicks (Remote closed the connection)
02:21:53 Join xen` [0] (
02:25:21 Quit Strath (Read error: 110 (Connection timed out))
03:10:28 Join DMJC [0] (
03:17:39 Join lostlogic_ [0] (
03:17:50 Quit lostlogic (Read error: 110 (Connection timed out))
03:26:32 Quit kenshin (Read error: 110 (Connection timed out))
03:42:22***Saving seen data "./dancer.seen"
03:50:42 Join LinusN [0] (
03:51:32 Join amiconn_ [0] (
04:05:43 Join QT_ [0] (as@area51.users.madwifi)
04:09:05 Quit QT (Read error: 145 (Connection timed out))
04:09:08 Quit amiconn (Read error: 110 (Connection timed out))
04:09:08 Nick amiconn_ is now known as amiconn (
04:29:19 Join wacky_ [0] (
04:30:04wacky_hey guys, would it be possible to compile the rockbox fort the iRiver.. but with the Remote display only
04:30:26wacky_like if we used only the remote (it's about the size of the archos), and not the main LCD
04:30:36*wacky_ just broke his main unit's screen! :(
04:31:02wacky_Rockbox would be my saviour if it could be totally controlled with the remote, just like the Archos's screens
04:32:06LinusNit will be, but it doesn't yet work
04:32:18LinusNwhat happened to your lcd?
04:32:21thegeekdo you get up this early?
04:32:22wacky_total control ??! :)
04:32:28wacky_someone step on it
04:32:40LinusNhaven't slept yet :-)
04:32:46wacky_yeah.. that's the sound it made :)
04:33:11LinusNwacky_: no chance of rma?
04:33:18wacky_iRiver America doesn't even repair, even if I pay
04:33:22wacky_warranty void
04:33:31LinusNhow nice of them
04:33:31wacky_I mean, expired
04:33:36thegeekthey are good with the warranty in the 1 year
04:33:39thegeekbut after that
04:33:42thegeekyou are shit out of luck
04:34:21LinusNwacky_: there are a few broken H1xx around on misticriver
04:34:39LinusNyou could buy one for the lcd
04:34:50wacky_heh :) but how to repair it ?
04:35:06thegeekLinusN : did you do all the detail scans?
04:35:06LinusNthe lcd isn't that hard to replace
04:35:07thegeekif so
04:35:08wacky_and I don't want these H3xx fuzzy colored craps
04:35:11thegeekis the lcd hard to remove?
04:35:24thegeekI have to resolder the headphone jacks
04:35:33thegeekand I want to take of the lcd so I dont scratch/bork it
04:35:44wacky_how much would they sell such a thing on misticriver ?
04:36:10LinusNi dunno
04:36:19thegeekif it's just the harddrive broken on one of them
04:36:23wacky_anyway, I'm going on a trip in august.. and I would have liked to bring the machine.. but now the screen is broken.. and I don't think I will be able to repair it until then
04:36:26thegeekyou could just take your hd and put it in too
04:36:38 Join kenshin [0] (
04:36:43wacky_oh :) maybe :)
04:36:52thegeekwould certainly be easier
04:37:00thegeekbut perhaps
04:37:09thegeekyou can get someone to sell you just the pcb/lcd
04:37:14wacky_LinusN - what do you guys plan for the remote ?% to have a dual display or you would toggle between the two ?
04:37:30wacky_In my case, I really wouldn't care a toggled display :)
04:37:49wacky_oh yeah ? does this require a big rewrite of the display funcs ?
04:38:00wacky_how will you deal with that ?
04:38:12wacky_any design planned ?
04:38:17LinusNwell, we do the rewrite
04:40:27LinusNsomeone is working on it, but i don't remember who
04:40:50wacky_no wiki page for discussion ?
04:41:51LinusNdon't think so
04:42:38wacky_ok.. thanks
04:43:22wacky_hey, btw I'd like to thank you guys, you just did an awsome job, freely like that, with all that passion we can see that result in some so harmonious code
04:44:07wacky_a job that turns electronics into a delight
04:44:14wacky_a big thanks
04:44:44*LinusN blushes
04:45:27wacky_good night guys
04:45:33 Quit wacky_ ("rrrrrock box")
04:46:03 Nick StrathAFK is now known as Strath (
04:49:51*LinusN just committed daniel's patch that allows plugins while playing music
04:59:46 Part LinusN
05:03:30 Join courtc_ [0] (
05:04:10 Quit courtc (Read error: 110 (Connection timed out))
05:04:12 Nick courtc_ is now known as courtc (
05:25:21 Join elinenbe_ [0] (
05:25:46elinenbe_the latest bleeding edge doesn't even boot on my device! (h120)
05:26:44 Nick elinenbe_ is now known as elinenbe|sleepin (
05:33:54 Quit kenshin (Read error: 60 (Operation timed out))
05:35:18 Join ashridah [0] (
05:42:24***Saving seen data "./dancer.seen"
06:02:51 Quit RotAtoR ()
06:04:27 Join CheeseBurgerMan [0] (
06:34:24 Quit CheeseBurgerMan ("CGI:IRC (EOF)")
07:04:15 Join kenshin [0] (
07:13:13 Quit ashridah ("Leaving")
07:14:08 Join Chamois [0] (
07:22:03 Join fuzzie [0] (
07:26:23amiconnGood morning
07:28:33*HCl woke up choking in his vitamin drink, what a way to wake up
07:42:27***Saving seen data "./dancer.seen"
07:55:14 Join [IDC]Dragon [0] (
08:00:05amiconnHi Jörg :)
08:00:50[IDC]Dragon"bootbox" is working now
08:01:16*HCl got 2 million gold from another player on his mmorpg
08:01:22HClthats certainly a way to wake up o.o.
08:01:24HClhow goes?
08:01:37amiconn[IDC]Dragon: Hmm, btw, I have a question. In main.c, we don't need the if(!usb_detect()) check around the autorock execution since we have early usb detection, do we?
08:01:57*[IDC]Dragon looks
08:02:49amiconn...and the comment in line 295 is clearly wrong
08:03:18[IDC]Dragonthe autorock check is legacy
08:04:24[IDC]Dragonwhat's wrong with the comment?
08:07:33amiconnIt says 'no complaint if it doesn't exit' but imho it should read 'no complaint if it doesn't open'
08:08:23[IDC]DragonI meant "exists", it's a typo
08:08:39amiconnSo the check could go away, but the braces have to stay (unless we want another #ifdef AUTOROCK at the top of the function)
08:08:58amiconnAh, exist makes sense too
08:09:08*[IDC]Dragon fixes
08:12:22 Quit kenshin (Read error: 110 (Connection timed out))
08:12:52amiconn[IDC]Dragon: Bootbox is different per model?
08:13:28[IDC]Dragonamiconn: do you have a clue why I had to make the *(.rodata*) so cumbersome in
08:13:39[IDC]Dragonyes, built per model
08:13:56amiconnHmm. Did you check whether rolo actually works on Ondio?
08:14:14amiconnI still have the problem that rolo'ed archos firmware won't play music
08:14:22[IDC]Dragonnot yet, I only tested the recorder so far
08:14:38 Join t0mas [0] (
08:14:39[IDC]Dragonrolo in general, or bootbox?
08:14:53amiconnRolo in rockbox, didn't try bootbox so far
08:15:22amiconnMaybe it'll work in bootbox, due to some init not happening
08:15:24[IDC]Dragondoes it fail if you rolo the same firmware again?
08:15:54amiconnRolo does work for all firmwares, but archos ondio firmware doesn't like to be roloed
08:16:09amiconnObviously I can't rolo another firmware from archos
08:16:10[IDC]Dragonah, Archos
08:16:24[IDC]DragonI thought rolo in general, overlooked that
08:16:43amiconnDoes bootbox include lcd drivers?
08:17:34amiconnJust builkding, to see how small it is
08:17:53[IDC]Dragonabout 23 K
08:18:04amiconnOooh, the ajz is 37K
08:18:08[IDC]Dragonas .ucl
08:18:26[IDC]Dragonsounds ok
08:18:26amiconnTried different models? Like player?
08:18:32[IDC]Dragonnot yet
08:18:48 Join oxygen77 [0] (
08:19:10amiconnPlayer will be fun. I doubt that we can exclude rocklatin handling, although simple ascii would be sufficient
08:19:29[IDC]Dragonthe size is ok, we'd have >200k left in flash
08:19:39amiconnShouldn't be much of a problem though, there's plenty of space
08:20:16[IDC]Dragonany idea about the *(.rodata*) ?
08:20:37[IDC]Dragonthe other sections were ok with wildcards
08:21:19[IDC]Dragonperhaps because of the rodata.str1.1 and rodata.str1.4 following
08:21:27[IDC]Dragonbtw, what is that?
08:21:57amiconn.rodata.str1.* is for strings, .rodata without suffix for other constants
08:22:08amiconn.4 means long aligned, .1 unaligned
08:22:24[IDC]Dragonwhy is that separate?
08:22:26amiconnsh-elf-gcc always uses .rodata.str1.4 for strings
08:23:06amiconnBootbox for Ondio doesn't link
08:23:20amiconnundefined reference to `_mmc_remove_request'
08:23:39[IDC]Dragonwithout the original *(.rodata) I got different images than before, with no wildcard
08:23:57[IDC]Dragonso I made it this way, to play safe
08:24:00amiconnDid you check the .map ?
08:24:17[IDC]Dragonis shuffled some sections around
08:24:34amiconnWith wildcard, I guess the linker intermixes .rodata and the .rodata.str1.* stuff
08:24:38amiconnShouldn't hurt
08:24:50[IDC]Dragonprobably not harmful, but I didn't want to break the regular builds
08:24:59amiconnYou could do
08:25:16[IDC]Dragonwe can simplify if we feel more assured
08:25:16amiconnShould be safe
08:25:31amiconn(Note the second dot in the wildcard)
08:25:44[IDC]Dragonor just *(.rodata*) alone
08:26:01amiconnNo, that will mix the sections as you observed
08:26:11[IDC]Dragonis that harmful?
08:26:18amiconnI guess it's not.
08:26:22amiconnWe could try
08:26:32[IDC]Dragona question for LinusN
08:27:01amiconnWe could test 5 targets...
08:27:03[IDC]Dragonhe should know the intention of this construction
08:27:41[IDC]Dragonthe recorder works, but I made only a quick check
08:27:51amiconnOf what construction? The compiler using those different sections?
08:28:16[IDC]Dragonnaming .str1 explicitely
08:28:51amiconnI guess this is just because he didn't want to use wildcards (or didn't know about them)
08:29:09 Join dwihno [0] (~dw@
08:29:47amiconnIirc I did my dead code experiments using *(.rodata*) only, and it worked
08:30:02amiconn...although not yet tested on iriver
08:30:39[IDC]Dragonthe iriver bootloader should do sectioned linking, too
08:30:45[IDC]Dragonwould be much smaller
08:31:00[IDC]Dragonbut I didn't dare to blindly change that
08:31:32[IDC]Dragonsee the new "if" in the configure script
08:31:45amiconnI'm thinking about sectioned compiling and linking for the plugin lib
08:32:22amiconnThat would make maintaining some things like the grayscale lib much easier
08:33:20[IDC]Dragonso the ondio needs a stripped down mmc_remove_request()
08:34:11amiconnThe mmc_remove_request is nothing fancy
08:34:55amiconnIt's in screens.c iirc
08:35:10[IDC]Dragonyes, I'm facing it
08:35:36[IDC]Dragonfor bootbox, it would suffice to just tackle the internal MMC, but that would be wore work
08:35:47[IDC]Dragon...more work
08:36:30amiconnI don't think it would be too difficult
08:36:30[IDC]Dragonyou can't change the card, since bootbox has no event loop
08:36:45amiconnJust don't enable multivolume & hotswap
08:36:54[IDC]Dragonso mmc_remove_request is a dummy
08:37:19amiconn..and the mmc detection in the init should be changed to always access the internal
08:37:32amiconn-> no remove request necessary
08:37:54amiconnSame goes for usb thread of course
08:38:23[IDC]Dragonall this is useless anyway
08:39:29amiconnAh ok
08:41:40[IDC]Dragonactually, the USB screen does loop
08:42:03amiconnThis way bootbox won't touch the external at all, like the archos firmware during boot
08:43:12amiconnIf bootbox stays below ~50KB (compressed) for all platforms, we won't have a problem with rombox at all
08:43:40[IDC]Dragonit will
08:44:04[IDC]Dragoneven with some more bells&whistles
08:45:08amiconnThe #ifdef HAVE_MMC in both usb thread and usb_tick need to be changed #if defined(HAVE_MMC) && !defined(BOOTLOADER)
08:46:02amiconnTo remove the code for handling insertions/extractions
08:46:20[IDC]DragonI'd try to modify as little of the codebase as possible
08:46:31[IDC]Dragon(currently, no change)
08:46:52amiconnThere are some #ifdef BOOTLOADER for iriver
08:47:02 Join Aison [0] (
08:47:05amiconn...and the hotswap code should be disabled, imho
08:47:23amiconnI think it would improve reliability
08:48:13[IDC]Dragoncan be done in the config-ondio*.h
08:48:47amiconnI don't think so. HAVE_MMC probably needs to stay defined for other places
08:49:22[IDC]Dragonyes, but I meant the hotswap switched
08:49:31[IDC]Dragon... switches
08:50:04amiconnThis switches off hotswap and multivolume for rockbox, but the usb hotswap is hardcoded
08:51:04amiconn...and so is the hotswap in the debug menu, but that isn't included in bootbox
08:53:27[IDC]Dragon24K for Ondio
08:54:37amiconnHow does bootbox rolo? Is there a minimalistic browser, or does it simply try to rolo ajbrec.ajz?
08:54:57[IDC]Dragondumb question: does lcd_puts() clear the restof the line?
08:55:26amiconnAfaik it does
08:55:39[IDC]Dragonamiconn: no browser, just hard coded /ajbrec.ajz resp. /archos.mod
08:55:47amiconnHmm, okay
08:56:23amiconnBtw, I think ~37KB of code is still quite much. Archos managed to put the first player firmwares in ~40KB
08:56:46[IDC]Dragonyes, we carry some ballast
08:56:55amiconnDo you include the language stuff?
08:57:38[IDC]Dragonit just occured to me that mmc_remove_request() is still a dummy, we won't have accessed the MMC before USB
08:58:33amiconnThat's correct
08:58:59amiconnI'd prefer not to access the mmc at all, but that requires some (slight) code change in ata_mmc.c
08:59:21 Quit Chamois (Read error: 131 (Connection reset by peer))
08:59:23amiconn...then mmc_remove_request can go away
08:59:27 Join Chamois [0] (
09:00:27amiconnWhen compiling without hotswap and an mmc is plugged, current code would access the mmc exclusively instead of internal flash
09:01:38[IDC]Dragonor, it has nothing to do with hotswap
09:02:14amiconnIt has - indirectly. Hotswap won't work without multivolume
09:02:46[IDC]DragonI meant the swap in USB mode
09:02:56*[IDC]Dragon might be confused
09:03:16amiconnI meant rockbox, not usb mode
09:03:46amiconnYou're right about usb mode
09:04:07[IDC]Dragoncd ..
09:08:12amiconnHmm, how can I use parameters in a shell script?
09:08:50Bagder$1 $2 etc
09:09:04Bagderor $@ for all of them
09:09:11amiconnThanks :)
09:09:30*Bagder bows
09:09:51amiconnI want to put the objdump disassembler command for m68k in a script
09:15:08*amiconn just found an obvious optimisation in memset_a.S...
09:16:36 Quit Strath (
09:17:01[IDC]Dragonanother 2.4 speedup? ;-)
09:17:09amiconnHaha, no
09:17:36amiconnSaving two words
09:17:57NJoinStrath [0] (
09:17:57amiconn...might execute a bit faster if not yet cached
09:17:58[IDC]Dragonbootbox works on Ondio
09:18:13amiconnNice :)
09:18:27amiconnSo it should work on rec fm/v2 and Ondio SP too
09:18:40amiconnNow there's the player...
09:18:48[IDC]Dragondid you try?
09:20:58amiconnBagder: I had a quick look at your codec separation. Do you agree that it would be even better to have a separate api as well?
09:21:51[IDC]Dragoncompiling bootbox for the Studio gives a heapload of errors :-(
09:22:04[IDC]Dragonlinking, that is...
09:23:28 Quit [IDC]Dragon ()
09:28:00 Join bobTHC [0] (
09:29:41 Join Harpy [0] (
09:30:28bobTHChi folks !
09:37:08amiconnWow, multiplies are slooow on coldfire
09:39:50 Join kenshin [0] (
09:42:31***Saving seen data "./dancer.seen"
09:43:25 Join yngwi [0] (
09:48:18 Quit tvelocity ("Leaving")
10:08:20 Join B4gder [0] (
10:12:20 Nick Lynx_awy is now known as Lynx_ (
10:20:04 Nick Strath is now known as StrathAFK (
10:22:46 Quit oxygen77 ("CGI:IRC")
10:22:52 Join markun [0] (
10:24:24 Quit kenshin (Read error: 110 (Connection timed out))
10:24:36markunI've recalculated the table for the iriver battery. I put 3.08V as 0%, should I put a higher value there?
10:30:19 Join webguest01 [0] (
10:30:52 Quit Andrew179 ("Leaving")
10:30:57 Quit webguest01 (Client Quit)
10:34:35*amiconn accidentally dropped his H140 to the floor :-/
10:35:30HCl :X
10:35:49HClmarkun: with original battery?
10:36:11markunHCl: Yes, calculated it from the measurements in the wiki
10:36:21 Quit DMJC (Read error: 110 (Connection timed out))
10:36:43amiconnHD is still working, but it's a bit louder than before :(
10:36:59yngwiamiconn: :-(
10:37:47amiconnThat never happened with archos, probably thanks to the bumpers...
10:38:29B4gdermaybe you should get a few spare bumpers from Jörg's junkyard and glue them onto your iriver? :-P
10:38:52amiconnBah utf-8 looking fugly here
10:39:01B4gderuh, right
10:39:27amiconnI should place a feature request for my irc client...
10:39:58amiconnHydra is quite nice, but it doesn't support utf-8
10:40:19B4gderI changed to iso-8859-15 for next time
10:40:56amiconnI doubt any current windows irc client supports utf-8. Hydra could be the first...
10:41:05B4gderxchat does :-)
10:41:14amiconnHmm, it does?
10:41:24amiconnXchat on windows is just plain ugly
10:41:24B4gderthis is xchat on windows
10:41:36amiconn...and shaky
10:41:48amiconnIt crashes a lot
10:41:52B4gdernot for me
10:42:03amiconnHydra never crashed so far, in spite of being versioned <1
10:42:14B4gderI've never seen xchat crash actually
10:42:33markunamiconn: Do you think I should put a higher value than 3.08V as 0% battery to prevent people from draining the batteries so far?
10:42:39amiconnI had it crash multiple times when I tried it, within just a few days
10:43:13amiconnmarkun: We should test how low the battery can go with the hd still able to spin
10:44:42markunamiconn: I based the values on the values in
10:45:06markunHe let the test run until the drive was unable to spin up the drive.
10:45:28markunEh.. until the iriver..
10:47:00markunThere is a bug in the old values anyway because it goed from 373 to 370 to 386. That can't be right, can it?
10:51:32markunamiconn: would you like to try the values I calculated?
10:51:58Chamoisi made a test a long time ago
10:52:06Chamoismaybe can serve
10:52:30Chamoisi hope
10:53:18markunChamois: ok, let me check if I don't go under C8 then..
10:55:04Chamoismaybe Linus had to put in the bootloader a battery check
10:55:25Chamoisif it is under a certain value iHP doesn't start
10:55:42B4gderthen there's need to check :-)
10:55:47B4gderno need
10:56:24Chamoismy iriver stays sometimes blocked after i started it but hard disk couldn't spin
10:56:47B4gderthen it did start
10:57:14ChamoisB4gder: ??
10:58:02B4gderstart == program runs
10:58:18Chamoisi explain
10:58:25Chamoisi push the play button
10:58:34Chamoisbootloader STARTED
10:58:44B4gderright, and that's my point
10:58:44Chamoisbut hard disk couldn't spin
10:58:47B4gderit _started_
10:59:10Chamoisso iHP was blocked and i had to plug the charger and RESET
10:59:10B4gderbut sure, you need a certain amount of juice to spin up the disk
10:59:39B4gderthe bootloader should time-out the spinup attempt and shut down I guess
10:59:54Chamoisor mesure the battery voltage
11:00:03B4gderI prefer time-out
11:00:08Chamoisi prefer the other
11:00:12B4gdermeasuring is always very rough
11:00:25B4gderbattery levels is a very inexact method
11:01:12Chamoisyes but i don't think to make spin the hard disk and stop it in a anormal postion is not good
11:01:15B4gderwhat if someone replaces his HD with a very low power consuming one?
11:01:39B4gderor the other way around
11:02:19 Quit Bger (Read error: 131 (Connection reset by peer))
11:05:31amiconnChamois: If the hd can't spin up it will not go into an anormal position
11:06:21Chamoisbut sometimes my hd start to spin up
11:06:46Chamoisbut can't read anything and stop to spin
11:06:52Chamoisis it good ?
11:07:29B4gderthe question is not if its good
11:07:36B4gderthe question is how rockbox should deal with it
11:07:56 Quit yngwi (Read error: 104 (Connection reset by peer))
11:10:03 Join DMJC [0] (
11:10:17 Join preglow [0] (
11:10:49preglowso, ok, who wants to adopt some resampling code?
11:11:00amiconnHi preglow
11:11:38 Join yngwi [0] (
11:12:27preglowi have linear interpolation working, but i can't spent the time to integrate it into whatever
11:12:31preglowfor at least a week
11:13:06 Join Bger [0] (~Bager@
11:15:24 Join Sucka [0] (
11:17:01 Join einhirn [0] (
11:24:21preglowat least someone should try it to see if it sounds good
11:25:12*B4gder works on a new language system suggestion
11:26:25 Join LinusN [0] (
11:26:41B4gderhowdy LinusN
11:26:44LinusNB4gder: oops, forgot to add CODECS :-)
11:26:53preglowgimme a sec and i'll put up some code
11:26:59LinusNgood work btw, with the codec separation
11:27:16B4gderLinusN: so can you run a plugin while listening now?
11:27:29LinusNit rocks, so to speak
11:27:30B4gderok, good
11:27:38amiconnI'll repeat myself: Would it be wise to separate the api as well?
11:27:46B4gderI must've erred on my last test yday
11:27:48LinusNamiconn: nah
11:28:14preglowwhy not?
11:28:20B4gderthe current approach is fairly nice code re-use wise
11:28:27amiconnPlus codecs should get a different extension, to avoid running them as plugins
11:28:43B4gderthey should be put in a separate location
11:28:46dwihnorockdec! :D
11:28:51amiconn...and the codec source files should be moved to apps/codecs/ imho
11:29:12B4gderamiconn: they re-use big chunks of the plugib build setup
11:29:18B4gderthat's why I didn't move them
11:30:20preglowstill think they should be moved somehow
11:30:29preglowthey don't have proper dependecies at the moment either
11:30:47preglowwhich kind of annoys me while coding
11:30:52B4gderI agree
11:30:58preglowi need to touch codecmpa every time i change libmad
11:31:25B4gderso we move them into apps/codecs
11:31:32amiconnB4gder: I think a separate api would be a cleaner solution; currently codecs could call every function plugins can, and vice versa
11:31:39B4gderchange extension and fix the deps
11:31:54amiconnFor some functions this isn't desirable
11:32:11B4gderyes, let's make a separate API
11:32:17B4gderit'll be cleaner
11:32:40amiconncodec.[ch] in apps/ the same way as plugin.[ch] ?
11:32:55B4gdersounds good
11:33:02preglowand please drop the codec prefix
11:33:15amiconnpreglow: ?
11:33:26amiconnDo you mean sources, codecs, or both?
11:33:27preglowcodec in front of mpa, for example
11:33:35amiconnmpa.c / mpa.codec
11:33:35preglowwhen they're placed in codecs/, it redundant
11:34:14amiconnHmm, I think I should be working on gfx api...
11:35:27amiconnThat'll cause a lot of changes...
11:36:56B4gderok, I can do the codec move then
11:37:19preglowbut ok
11:37:23preglowanyone here got 48khz mp3s?
11:39:57LinusNi have
11:40:03amiconnI have my mp2/mp3 test set
11:40:13amiconnContains everyting from 8 to 48 kHz
11:40:40preglowi'll toss up a rock
11:41:11amiconnDoes the resampling code handle all cases, or just 48 kHz?
11:41:25preglowjust 48khz, i don't know how to calculate a delta without using floating point
11:41:42amiconnHow do you mean?
11:41:57amiconn48 kHz won't use a delta?
11:42:00preglowlike, when i need to step through the samples
11:42:00LinusNpreglow: src please
11:42:04preglowLinusN: a sec
11:42:23LinusNpreglow: 16.16 fixed?
11:42:26preglowLinusN: same place
11:42:28preglowLinusN: indeed
11:42:35***Saving seen data "./dancer.seen"
11:42:42preglowLinusN: i figure we'll never have a buffer larger than 65kb
11:42:48preglowamiconn: oh, and btw, it's mono at the moment
11:43:40preglowwhat i need to do is mp3_sample_rate/target_sample_rate*65536, then cast that to an int
11:44:28amiconnmp3_sample_rate*65536/target_sample_rate should work with unsigned int
11:44:36amiconnunsigned long to make sure
11:45:35amiconnThat doesn't support sample rates > 65535 Hz, but fixing that is simple, losing 1 bit of precision
11:45:36preglowwell, yeah
11:45:41preglowfor sample rates below 65khz
11:45:49preglowthat's true
11:45:53preglowthe delta doesn't need to be hyper accurate
11:46:40amiconnDo we have any format that supports sample rates >65 kHz?
11:47:10preglowi'll admit we wont see that often
11:47:11markunyes, wavpack
11:47:24markunsampling rates from 6 to 192 kHz
11:48:28amiconnWe could auto-scale the calculation
11:48:42Sucka192khz O_O
11:48:54 Join rasher [0] (
11:49:08Suckai thought the maximum worth sampling at was 44khz
11:49:10preglowi don't think we'll be able to handle 192khz...
11:49:11amiconnI wonder who needs 192 kHz sample rate. Even 96 kHz seems to be overkill...
11:49:12LinusNamiconn: autoscale...sounds awfully close to floating point :-)
11:49:15Suckaas the human ear cant tell the difference
11:49:32preglowseriously, 96khz is a nice maximum
11:49:40preglowfor 192khz, linear interpolation will not do
11:49:41LinusNSucka: but we need to account for the superhumans as well :-)
11:49:45preglowwe'll need proper filters for that
11:50:00rasherSo.. it seems that fwpatcher doesn't work on windows 95/98
11:50:18preglowrasher: not the non-unicode one either?
11:51:04rasherThe final move fails, giving them the "is the file write-protected?" error
11:51:14 Join kenshin [0] (
11:51:45Suckathose crazy kids
11:52:00preglowwell, i have no windows 98 machine to test on
11:52:23amiconnI do have a Win98 VM...
11:53:48amiconnRunning plugins while playing music is nice :-) Thanks B4gder
11:54:12markunYes it's nice, I was just playing mine sweeper :)
11:54:25amiconnNow I need a 4-grey solid cube ;-)
11:54:55amiconnOr perhaps someone implements peak reading, and I could adapt oscillo(graph|scope)?
11:56:27amiconnpreglow: Regarding your yesterday question: No wavpack track is skipping, but a total of 4 FLAC tracks from that ABBA album are skipping
11:56:40preglowno surprise
11:56:50preglowi wont be fondling flac anymore
11:56:57preglowthe source code gives me coughing fits
11:58:26preglowamiconn: tested the resampler?
11:58:40 Part LinusN
11:58:40preglowchamois says it doesn't work for vbr files, but i can't imagine my code broke that
11:59:29rasheramiconn: could you test fwpatcher and fwpatchernu on win98?
12:00:16preglowfwpathcer won't work, that's for sure
12:00:36rasherProbably not (I think you can after-install some unicode support on windows98 at least?)
12:00:41Chamoispreglow : i'm pretty sure of me
12:00:57preglowi'll try it afterwards
12:01:00preglowfirst i'll shower
12:04:09amiconnrasher: Are the versions in the wiki up to date?
12:05:12rashershould be.. although when they fail, they're very uninformative
12:06:56rasherNow let me look at msdn and try to extract the actual errormessage
12:07:29amiconnpreglow: It seems your codecmpa.rock always resamples... so 44.1 kHz sounds weird now :-/
12:09:17rasherWindows Me/98/95: The MoveFileEx function is not supported. To rename or delete a file at restart, use the following procedure.
12:09:20rasherwell there we go
12:09:22*rasher sighs
12:10:06rasherthat may just be for MOVEFILE_DELAY_UNTIL_REBOOT
12:11:09preglowamiconn: i know
12:11:16preglowamiconn: it's not a complete solution, just a proof of concept
12:11:24preglowamiconn: it don't even want the resample code to be in the codec
12:11:35amiconnMakes sense.
12:11:36rasherMoveFileEx is not supported on windows me/95/98
12:11:59rasherpreglow: Why is MoveFileEx used at all?
12:13:07rasherto be able to overwrite
12:13:54rasherwell that's annoying
12:15:20rasherMoveFile is pretty braindead
12:15:37preglowwell, then just do a delete and copy
12:15:43 Join B4gd3r [0] (
12:15:50rashermore points of error :-\
12:15:56rasherbagder overflow
12:17:25preglowthe more bagder, the merrier
12:17:40amiconnrasher: DeleteFile(), MoveFile() ?
12:18:34rasherif (MoveFileEx(name3, fn, MOVEFILE_COPY_ALLOWED | MOVEFILE_REPLACE_EXISTING)) becomes if (DeleteFile(fn) && MoveFile(name3, fn))
12:18:55rasherI'll compile this and let you have it
12:18:57preglowChamois: seems you're right
12:19:18amiconnpreglow: Your resampling sounds about right.
12:19:35amiconnI wonder if we should do like the iriver firmware and resample everything to 44.1 kHz
12:19:40preglowamiconn: that'll only be the case for small resampling ratios, mind you
12:19:45preglowamiconn: i can see no reason why we should do that
12:19:56preglowamiconn: we should switch to 22050 and 11025 whenever we can
12:19:58amiconnHow would you do that?
12:20:05preglowhow how?
12:20:08preglowshould it be a problem?
12:20:18amiconnResampling everything to 44.1 kHz has at least 2 advantages
12:20:26preglowthe codec api sees it can use a lower sampling rate, and does so
12:20:28preglowreinits the uda
12:20:32preglowreinits the optical output
12:20:33markunpreglow: switching will give problems with crossfading
12:20:38preglowmarkun: well, duh
12:20:40amiconn(1) It makes mixing in the buffer easier (thinking crossfade, although I don't like it)
12:20:57amiconn(2) S/PDIF doesn't allow <32 kHz
12:20:59preglowmoot point for me
12:21:05preglowamiconn: i forgot about that
12:21:10preglowthen i say we stick to 44.1khz
12:21:29preglowHCl: moot, a good point, but unimportant
12:21:30rasheramiconn:"> and - could you test these?
12:21:39HClnever heard of that word, okay o.o.
12:21:45preglowwell, now you have
12:22:04rasheractually, the fix should work for wine-emulating-98
12:22:07 Join hicks [0] (debian-tor@6e97f016a4beba05.session.tor)
12:22:07rasherlet me test..
12:22:27rasherexcept my keyboard's key-less
12:22:33 Join LinusN [0] (
12:23:03preglowwhat the hell is going on
12:23:07preglowmy vbr files don't work now
12:23:15preglowamiconn: can you play vbr files?
12:23:20preglowperhaps something else is wrong with my setup
12:23:33amiconnI don't have vbr test files
12:23:40preglowplay any vbr file
12:23:50preglowdont matter, it'll try to resample anyway
12:24:29amiconn^Hmm, doesn't werk
12:24:40preglownow, why the hell would that be the case
12:24:48preglowi can think of no changes that destroys vbr playbac
12:26:28amiconnHmm, vbr definitely doesn't work
12:26:41preglowyou compiled your own copy, or?
12:26:58amiconnIt starts to prebuffer, but the time counter stays at 0:00 and there is no sound
12:27:07preglowbut did you compile codecmpa from the source?
12:27:10preglowor use my rock
12:27:11amiconnI didn't compile myself
12:27:18preglowperhaps something else is broken here
12:27:57preglowi think i know now
12:29:26 Part LinusN
12:29:42preglowgot it fixed
12:29:55preglowstart_skip had me fooled
12:29:57 Join hicks_ [0] (debian-tor@f774ff41f1b833f6.session.tor)
12:30:13preglowlame header is only read if it's a vbr file at the moment
12:30:48preglowthis resampler can also be used quite easily for tempo adjustments
12:30:52preglowpitch shifting
12:32:32 Quit B4gder (Read error: 110 (Connection timed out))
12:32:33amiconnWe do that on archos... using faked clock values for the MAS
12:33:01amiconnWhile pitching up means overclocking the MAS, we can pitch up to 200%... stable
12:33:51preglowdamn! this thing _canm
12:33:54preglowdamn! this thing _can_ do vorbis!
12:34:01HClwhat can?
12:34:03preglowsomeone code a decoder!
12:34:05preglowthe mas
12:34:23preglowa guess memory is more of a requirement
12:35:04amiconnrasher: I'm trying to test, but VMware doesn't like me atm
12:35:35 Join LinusN [0] (
12:36:52preglowLinusN: still on a boat? :)
12:38:20LinusNon an island
12:39:21pregloweven better
12:39:23rasherokay, fwpatcher works in wine now
12:39:52rasherI'm not entirely sure it did previously
12:40:02 Join Coldtoast [0] (
12:40:11amiconnrasher: "Firmware patched successfully" :-)
12:40:11preglowit didn't
12:40:17preglowseveral people have complained about that
12:40:24Coldtoastwas walking back from the gym listening to music and 2 things occurred to me
12:40:25 Quit Chamois (Read error: 104 (Connection reset by peer))
12:40:28 Join Chamois [0] (
12:40:29amiconn(the unicode variant doesn't start)
12:40:36preglowit shouldn't
12:40:42Coldtoast1) in the Playlist menu, you need a Clear Playlist
12:40:44preglowwin98 doesn't have a unicode api
12:40:57rasherI'll commit my fix then
12:40:59amiconnIt can be added iirc
12:41:14rasherYes, I'm pretty sure you can install some stuff to get the unicode
12:41:23Coldtoast2) when you get to doing th eEQ, it'd be GREAT if you could assign saved EQs to Artist or Genre
12:41:24amiconnColdtoast: Why would you need 'clear playlist'?
12:41:46Coldtoastamiconn: if I'm listening to, say, Tool and I decide I want to listen to a bunch of other stuff
12:42:05amiconnThen simply browse to the other stuff and start playing
12:42:06Coldtoastlike, dump what I'm listening to
12:42:26amiconnThe old playlist will be replaced by the new one
12:42:28Coldtoastwhen I change dirs, it builds a playlist of all the tracks in the dir
12:42:35LinusNthe playlist is automatically cleared when you play a new file in the browser
12:42:41Coldtoastbut I may not want the whole dir added automatically
12:42:45Bger <= that's the unicode support in win98
12:43:35Coldtoastif I select somethign else, then view the playlist, it's automatically added all files in the dir to the playlist
12:44:08 Join DaKi][er [0] (
12:44:21Coldtoasthow about the EQ idea tho? sound silly?
12:44:36Coldtoastso I could set up a bunch of EQs and assign them to either genre or artist
12:44:38LinusNColdtoast: yes :-)
12:44:45Coldtoastreally? hmmm. ok
12:44:48B4gd3rthat idea is old too
12:44:49 Part LinusN
12:45:01B4gd3rmost ideas are already present in the feature-request tracker
12:45:24 Quit hicks (Remote closed the connection)
12:45:39ColdtoastI don't see tha tidea in the list
12:45:40B4gd3rbut we more often talk about config files for specific songs or dirs
12:45:54B4gd3rsince "eq" would be a config file
12:46:03pregloweq assigned to artist or genre sounds more like a runtime db issue
12:46:23Coldtoastpersoanlly, I don't see how that's a silly idea at all
12:46:27B4gd3rthe tricky part is the preloading of songs
12:46:33B4gd3rand applying when the song changes
12:47:11 Nick B4gd3r is now known as B4gder (
12:47:13preglowyou'll get a nice pop for sure ;)
12:47:35preglowunless you crossfade
12:48:51Coldtoastoh well
12:50:04preglowbut ok
12:50:07preglowthe resampling sounds ok?
12:50:46 Quit DMJC ("Leaving")
12:51:32 Join DMJC [0] (
12:51:50amiconnpreglow: Yes it does
12:52:02amiconnWhat would be the problem with bigger ratios btw?
12:54:06preglowdistortion, i guess
12:54:13preglowi've made a version with varying delta
12:54:14preglowworks just fine
12:55:15amiconnWhy should linear interpolation produce distortion? I'd guess it would just sound a bit dull
12:55:22amiconn...for lower sample frequencies
12:55:38amiconnCubic interpolation is out of question?
12:55:49preglowall interpolation produces distortion, higher order interpolations are better at attenuating them, that's all
12:56:04preglowwhen you resample from 48khz -> 44.1khz, all the distortion lies in the inaudible area
12:56:42preglowso linear interpolation to resample from 96khz -> 44.1khz will have audible distortion all over the audio band
12:57:20amiconnI'm thinking more about upsampling, not downsampling
12:57:51preglowupsampling should sound okish
12:58:10preglowcurrent code only handles downsampling well
12:58:31preglowlistening to music with the resampling ratio varying is... weird
12:58:39amiconnFor mp3, the only downsampling case is 48 kHz
12:59:05amiconnAll others need to be upsampled (32/24/22.05/16/12/11.025/8)
13:01:03rasherI like how we may end up deleting the user's firmware with fwpatcher now
13:01:13rasherif DeleteFile is successful, but MoveFile is not
13:01:26rashercan't see how that should happen though
13:02:32preglowcode handles upsampling as well
13:03:05preglowwell, if movefile fails, just tell a joke or something
13:03:16preglowthat'll put their attention elsewhere!
13:03:57rasherThere should be no way it could happen, unless the temp file gets deleted behind our back
13:04:01preglowi get some weird clicking with upsampling
13:04:06amiconnrasher: Rename old file first, then rename new file. Only if both renames succeeded, delete the old file
13:04:21amiconn...the *renamed* old file
13:04:21rashergood point
13:04:34rasherexcept the rename of the old file might fail
13:04:36*rasher cries
13:05:03preglowahh, the current code doesn't handle upsampling well at all, no
13:05:12preglowit assumes the delta is always above 1
13:06:25amiconnrasher: Where is the problem?
13:07:54amiconnBtw, does fwpatcher protect from accidentally patching twice?
13:08:06rasherYou can't
13:08:10rashernothing happens in that case
13:08:21rasherit just patches on top of it
13:08:24preglowit tries doing the patching, sees the md5 fails
13:08:27preglowthen deletes the result
13:08:34rasherI don't think so
13:08:39rasherI'm pretty sure it succeeds
13:08:46preglowof course it'll succeed
13:08:46rasheras do the command-line tools
13:09:40preglowamiconn: the first stage of the resampler, the one that actually uses the last sample of the previous block, has to go continue for longer in the case of upsampling
13:10:07rasherso where do I move the old file?
13:10:08preglowamiconn: so we'll need another for loop there, we might want to have separate routines for up and down sampling
13:10:15rasherone of the old temp files?
13:10:17preglowamiconn: especially seeing as how the downsampling can be done in place
13:12:51 Quit edx (Read error: 110 (Connection timed out))
13:15:59rashercube.rock needs beat-detecting-zoom
13:16:25preglowbeat detection isn't exactly trivial
13:16:32preglowbut you can make something that's not too accurate
13:16:34B4gderpeople will start requesting that
13:16:44B4gdernow when the plugins work with music
13:16:54preglowi know of a kick ass transient detection algorithm, but i think it's secret, and it requires floating point ;)
13:17:21preglowamiconn: i've got codecmpa upsampling by a ratio of over 2 here, want to have a listen?
13:17:44preglowthe music goes around veeeery slowly
13:18:05amiconnNo adaption to actual sample rate?
13:18:14amiconnI could test with my testset.
13:18:18preglowno, i just wanted to hear how badly linear interpolation works
13:18:23rasherLinux doesn't spin down the harddisk either :-\
13:18:28amiconnWhat's the exact ratio?
13:18:55preglow0.485 or something
13:19:01preglowi just use 30000 as delta
13:19:04amiconnI think I know one of the 'unknown' chips of iAudio M3
13:19:12 Join cYmen [0] (
13:19:13preglowbut i can cook up something that should adapt to sampling rate
13:19:56B4gdernow, should codecs link with the pluginlib?
13:20:03B4gderor should we make a separate codeclib
13:21:04amiconnIn fact I'm pretty sure about the chip. It's our infamous LT3440....
13:21:46Bgeramiconn: which chip
13:21:57amiconnI'll edit the wiki
13:23:14preglowwhat does to pluginlib do?
13:23:17rasherwow, the id3 information sure is messed up for a bit just as you start playing
13:23:18preglowthere already is a codeclib
13:23:30B4gdercodeclib.c is part of the pluginlib atm
13:23:38B4gderand I'm moving the codecs
13:24:23B4gderso I'm considering if just using the pluginlib for codecs is suitable
13:24:33B4gderor if I should introduce a separate codeclib for codecs
13:24:58B4gderpluginlib is meant for general useful functions for plugins
13:25:57amiconnBger: Info added.
13:26:24 Join Febs_away [0] (
13:27:06rasherHrm.. having crossfading enabled breaks "next song info" in the wps?
13:27:22linuxstbB4gder: I'm in favour of a seperate codeclib and not giving the codecs access to the plugin api. Seems the logical thing to do.
13:27:36preglowamiconn: test
13:27:36B4gderI agree
13:27:42B4gdera separate codeclib it will be
13:27:46preglowamiconn: it should only resample when needed, it should also both up and downsample
13:28:25linuxstbB4gder: Only problem is that there is some debugging code (in codecmpa) that uses rb->open/write/close to write debugging data.
13:28:40preglowthen toss in read and write for now ;)
13:28:54preglowbasic debugging possibilities should still be possible
13:29:02preglowcodecs should have logf, etc
13:29:30B4gderlinuxstb: that's not a problem right now, since the codec API will be pretty much cloned from the plugin API to start with
13:30:06linuxstbBut I don't think there's very much needed from the plugin API - just the "string.h" functions and yield()
13:30:23linuxstbThey shouldn't need to access the LCD or buttons for example.
13:30:25B4gderperhaps, but I think it is better to get the move fixed first
13:30:57B4gderit is quite an impact to the whole build setup already
13:31:13linuxstbI can imagine.
13:40:00markunI would like to commit new values for the volt to percent conversion. I put the old and new values in this graph:
13:42:37***Saving seen data "./dancer.seen"
13:42:44amiconnpreglow: Distortions are audible for 8/11.025/12/16/22.05/24 kHz. Seems you're right about the ratio problem...
13:42:53amiconnI compared with archos, btw
13:43:34 Join ashridah [0] (
13:44:41rasher \o/
13:45:13B4gdergood work!
13:45:32amiconniriver fw doesn't have that problem. I wonder what interpolation it uses...
13:45:57amiconnBah, file browsing in iriver fw is cumbersome
13:46:59preglowamiconn: what kind of distortion?
13:47:21amiconnSounds like harmonics
13:47:25preglowit is harmonics
13:47:42amiconnApart from that, interpolation is working fine
13:47:53preglowas long as there is no clicking, i'm ok
13:48:11preglowbut linear interpolation, and indeed, all interpolation generates harmonics
13:48:31amiconnAgain, I wonder what interpolation iriver uses
13:48:49preglownot in itself, but when coupled with the upsampling
13:49:05preglowhave you tested these files on the iriver?
13:49:21preglowthey might not be interpolating
13:49:24preglowthey might be filtering
13:51:15preglowbut ok, code is at
13:51:29preglowdownsampling can and should be done in place, i guess
13:52:21B4gdercodec_api is already in use...
13:52:28preglowrip it out
13:52:33B4gderyes, me rename
13:53:24 Join LinusN [0] (
13:53:25linuxstbB4gder: I assumed you would be adding the functions to the existing codec_api - not creating a new one.
13:53:33B4gderannoyingly enough, codec.h is already "occupied" too
13:53:35preglowamiconn: downsampling from 96khz to 44.1khz will also sounds bad, all the harmonics you hear in the upsampling will be folded down the spectrum again, so they won't even be harmonics anymore
13:53:49B4gderI'm making a new one
13:53:57B4gderto replace the plugin_apu
13:54:18linuxstbBut as you say, there is already a codec api.
13:54:20B4gderI'm doing the move
13:54:29B4gderlets fiddle after the move
13:54:40B4gderlinuxstb: I have no idea how they relate
13:55:58B4gderlinuxstb: can you explain?
13:56:05linuxstbCurrently, the codecs are initialised with both a pointer to the Codec API, and a pointer to the Plugin API. All that's needed is to move a few functions from the plugin API to the codec API, and then we can drop the Plugin API.
13:56:24linuxstb(that's my understanding - it's Slasheri's code)
13:56:49 Join RotAtoR [0] (
13:57:20preglowthe codec api pointer is passed in the other argument to plugin_start
13:58:04B4gdermore work
13:59:17amiconncodec*.c don't conform to rockbox standard
13:59:31linuxstbIn what way?
13:59:46B4gderI'm taking care of some of that now
13:59:47amiconnThey're indented by 2 spaces, not 4
14:00:07preglowi've never read about a rockbox standard indent
14:00:15preglowit just says 'dont use tabs'
14:00:29preglowbut a two space indent is a sin against mankind, yes
14:00:51linuxstbWhat's wrong with a two space indent?
14:01:04amiconn "Always indent your code with four spaces."
14:01:18preglowi just think four spaces makes the source more readable
14:01:21zeindent's should be 3 spaces!!
14:02:43preglowbut yes, for more drastic downsampling, we should implement some filters to use before decimation
14:02:48preglowfor upsampling, i have no idea
14:03:06 Join Andrew179 [0] (
14:03:12preglowperhaps some better interpolation
14:04:13 Quit lostlogic_ (Client Quit)
14:04:37 Join lostlogic [0] (
14:04:56preglowwhat's clever for cleaning lcd screens?
14:05:21LinusNpreglow: toothpaste
14:05:40Andrew179I noticed that one of the unknown M3 chips had been identified as a dc-dc converter. Could the chip labeled 3441 also be a dc-dc converter?
14:06:09Andrew179like this one?,C1,C1003,C1042,C1116,P2149
14:06:21preglowLinusN: no, no scratches, just general cleaning
14:06:35preglowi don't think toothpaste is too good at removing smudges ;)
14:06:43LinusNi use the same cloth as i use for cleaning my glasses
14:06:58preglowi don't have glasses :/&
14:07:22amiconnAndrew179: Could be.
14:07:34amiconnThe datasheet says part marking is 3441, and it has 12 pins
14:07:35preglowany special cleaning solution or anything?
14:08:28 Quit Febs_away (Read error: 145 (Connection timed out))
14:08:30LinusNpreglow: my breath
14:08:36amiconnAndrew179: I'm sure about the 3440, we know this infamous chip from the Ondio
14:10:12amiconnAndrew179: To confirmation for the 3441, you should check which pin is connected to what parts. There should be a coil somewhere...
14:10:19amiconn*To get
14:11:17Andrew179I'm not too knowledgeable about the hardware, but I'll see what I can do
14:11:33Andrew179at the very least I'll put it into the wiki as a possibility
14:13:26preglowif there's a coil connected to it, you should be able to spot it quite easily, heh
14:13:49 Part LinusN
14:15:53*B4gder just saw a "OBJCOPY vorbis.codec"
14:16:34B4gderthe concept holds at least ;-)
14:19:17Andrew179something like this is a coil, right? :)
14:19:20 Join t0mas_ [0] (
14:20:33t0masoh lol...
14:20:36t0masmissed 1 client...
14:20:40 Quit t0mas_ (Client Quit)
14:21:38preglowit has an ungodly amount of legs to be just a simple coil...
14:21:49preglowlooks more like a mini-transformer, heh
14:23:28amiconnAndrew179: Look at this
14:23:58amiconnIn the first row under 'view inside' look at any version you want
14:24:15t0mashm.. has Bagder fixed the plugin + codec = crash issue?
14:24:26preglowone can't exactly call the early archos players beauties
14:24:33preglowt0mas: yes
14:24:40Bger hm, guys, what about
14:24:53amiconnNear the bow-shaped recessing of the board at the lower right corner, you'll see our famous LTC3440 (marked LTNP)
14:25:10amiconnLeft from that there is a part with a round head marked "220"
14:25:22amiconnThis is the coil belonging to the LTC3440
14:25:25 Quit DMJC ("Leaving")
14:25:58Bgerregarding the last unknown chip
14:26:50Andrew179amiconn: ok, it looks like the M3 has a coil on the backside of the PCB for the M3
14:29:13rasherBger: certainly looks like the same type of marking.. guess asking TI would be a good plan
14:29:53Bgerrasher: also
14:29:59amiconnAndrew179: If the 3441 is an LTC3441, there should be another coil near it
14:31:20Andrew179amiconn: yes, there's two coils on the back in that area
14:33:25preglowstill no idea what the shielded thing is?
14:33:36 Quit B4gder ("Lämnar")
14:33:47preglowif m3 does radio, it's probably the receiver
14:34:19Andrew179preglow: it does do radio
14:35:44rasherIs there any reason why IriverHardwareComponents isn't IriverH1XXHardwareComponents ?
14:36:19rasherI guess that goes for most of the Iriver* pages though :-\
14:36:52amiconnpreglow: Radio receiver is probably a good guess
14:37:11amiconnThere's the LM052 near it, and the backside is well shielded
14:37:16 Join DMJC [0] (
14:38:28 Quit rasher ("CGI:IRC")
14:42:57amiconnAndrew179: I'm convinced that this 3441 chip is really an LTC3440, and I believe the 'AZS' chip is the LiIon charger
14:43:10amiconn...TI BQ24020 in this case
14:43:25amiconnErm, I mean LTC3441 of course
14:44:16amiconnThere's another Linear chip, marked LTZW
14:44:27amiconn6 Pins.
14:44:41amiconnProbably another voltage converter or regulator
14:45:20Andrew179where are you seeing the LTZW chip? I must have missed that one
14:45:47amiconn above the LTNP
14:47:38 Join hicks [0] (~hicks@4cb2e00e3afd58a1.session.tor)
14:47:40 Quit lostlogic ("Going to the moon")
14:47:51Andrew179yes, looks like a regulator:,C1,C1003,C1042,C1032,C1063,P2183
14:48:30 Join LinusN [0] (
14:48:43amiconnAndrew179: Part marking: LTZW. Bingo
14:49:48LinusNpreglow: resampling sounds pretty decent to me
14:50:49preglowLinusN: yeah, me too
14:50:57preglowbut it wont at higher downsampling ratios
14:51:36amiconnThe upsampling also doesn't sound decent for higher ratios
14:51:48preglowi can live with harmonic distortion
14:51:52preglownonharmonic distortion i can't live with
14:52:03preglowand that's what we'll have at 2x downsampling, for example
14:52:44amiconnWhat about cubic interpolation for upsampling? Too demanding?
14:52:49preglownot really
14:54:05preglowin the case of upsampling we'll have pretty much processing power to spare anyway
14:54:21preglowsince the codec will only we working at around half throttle
14:54:28 Join edx [0] (
14:55:59amiconnAh, yes. 96 kHz will probably present us with a cpu performance problem at least for some codecs...
14:58:17LinusNbtw, we could probably use 140MHz as the boost freq
14:58:29LinusNas long as the disk isn't spinning full time
14:59:53LinusNbut it may be too close to the edge...
15:00:00preglowi think we should avoid it
15:00:12preglowif the codecs can't do 96khz, too bad
15:00:44LinusNpreglow: the resampling code should be in the codec lib
15:01:01preglowhmm, i'd rather it was in the playback system
15:01:18LinusNah, yes of course
15:01:26preglowbut that will present us with some design questions
15:01:42preglowin what format should the codecs pass along the data?
15:02:11preglowLinusN: as it is, codecs speak directly to the pcm buffering system, i'd like that to stop, so the playback system can apply some processing, like perhaps eqing and replaygain, etc
15:02:32LinusNhmmm, it surely would be handy if all codecs delivered 16-bit 44.1kHz pcm data
15:02:39preglowfor processing, the higher resolution data we have, the better, so we might also want the playback system to scale everything down to 16 bits
15:03:09LinusNso 24 bits/44.1kHz might be nice
15:03:34LinusNleavs 8 bits headroom for mixing
15:03:57preglowbut then we'll need to scale the data twice, once in the codec and once in the playback system, what we might want to do is just pass everything along as 32 bit ints, and tell the playback system how much each sample needs to be shifted
15:04:19LinusNpreglow: ah, yes i see
15:05:24LinusNdo the codecs differ a lot today, regarding sample formats?
15:05:37preglowthis way, dithering can easily be implemented in the playback system as well
15:05:41preglowyes, they do
15:05:52preglowlibmad uses non-interleaved 32 bit ints
15:05:58preglowtremor uses interleaved 16 bit ints
15:06:04preglowbut i believe tremor gives you whatever you want
15:06:22preglowwavpack and flac delivers 32 bit ints as well, i believe
15:06:45preglowit surely needs to be investigated further, but i think what i'm saying is quite sound
15:06:54LinusNme too
15:07:10LinusNthe less intermediate processing the better
15:07:55LinusNi would like a lot less decoding delay
15:08:01preglowme too
15:08:07LinusNswitching tracks takes too long imho
15:08:16preglowbut i dunno, how much will that affect decoding cpu usage?
15:08:25preglowswitching takes way too long
15:08:32preglowand we'll also have a humongous latency for eqs and the like
15:09:44LinusNshorter decoding bursts will affect performance a little
15:10:04LinusNsince we lose ~10ms for each cpu_boost()
15:10:32preglowhmm, yes, but i believe a shorter latency really is needed
15:10:36preglowtrack switching takes aeons
15:12:21 Join ]RowaN[ [0] (a2b0y@
15:12:39]RowaN[where can i find some info on how to add images to .wps?
15:13:23preglowi don't believe you can, yet, t0mas is dealing with this
15:13:33LinusN]RowaN[: in the rockbox forum
15:13:52LinusNthere is a patch somewhere, but i don't know how well it works
15:14:18]RowaN["Image support for the WPS"
15:14:21]RowaN[it says in cvs now
15:15:39preglowoh well, then he needs to document it as well
15:15:49crashdit's hot today
15:16:41preglowno, not really
15:16:42preglowthe weather is shit, just like any other random day in trondheim
15:17:15crashdit's scorching hot in london
15:17:38preglownice, i hope it stays that way until i get there in about three weeks time
15:18:09crashdwhat you doing in londo n?
15:18:40preglowi'm not staying in london, really, but i'll probably be going there a couple of times
15:18:46preglowi'm visiting some family
15:18:51crashdin the uk ?
15:19:00preglowi'm half british ;)
15:19:08crashdshould be good weather for the next few weeks, at least ;)
15:19:36preglowit's always been pretty nice when i've been over
15:19:38preglowi suspect i bring it with me
15:19:44linuxstbpreglow: If you're needing free Wifi in central London, stand outside the Apple Store on Regents Street.
15:19:45crashdyou must miss the winter months ;)
15:19:55preglowlinuxstb: no laptop :/
15:20:01crashdthere's loads of free wifi in london
15:20:05crashdjust gotta know where to go ;)
15:20:15linuxstbpreglow: In which case, go inside and use one of the Macs.
15:20:27crashdfree internet
15:20:30crashdon nice machines
15:20:33crashdpity about the fucken price
15:20:41preglowcan't do that, i'll run the risk of wanting to buy one
15:20:48linuxstbThe Apple store is full of tourists checking their Yahoo mail...
15:23:12t0mas]RowaN[: in the forum there is a topic about theme modding
15:23:17t0masthere is the explaination
15:23:23t0masand maybe someone should add it to the wiki
15:23:27preglowLinusN: there's also the question of whether the codecs should hand over data in interleaved or deinterleaved form
15:23:38crashdright, time to brave the underground sweatbox
15:23:39crashdlater guys
15:23:56LinusNpreglow: interleaved feels better in some respects
15:24:09LinusNbut not good for bursted filtering loops
15:24:32preglowfor processing, deinterleaved is way better
15:25:20LinusNso let's go for that
15:25:51preglowwe might want to take advantage of those libs that already have data in deinterleaved form
15:25:51preglowbut we also don't want to deinterleave interleaved data
15:25:52preglowwe might end up with several versions of all the processing loops :/
15:26:33preglowof course, it's just a question of variable array stride, but that'll hinder use of movem, etc
15:28:39 Join GodEater [0] (
15:29:34preglowtremor too has data available deinterleaved internally
15:30:39 Join Rori [0] (
15:31:06preglowwe need to handle tremor a different way, the current way doesn't let us dither at lal
15:33:39 Join Febs_away [0] (
15:36:33 Quit bobTHC ("Smoke Weed Every Day !")
15:38:17t0masWPS images added to the wiki
15:41:08t0masnow time to finally get the wps to the remote
15:41:25preglowi think there should be a wps data directory
15:41:31preglowkeeping bmps in the .rockbox root isn't very elegant
15:41:41t0masok, can you or LinusN add that?
15:41:50preglowcan't you? :P
15:41:52t0masI don't know how to add dirs and stuff like that to the build system
15:42:10preglowoh, but the images are added by users, no?
15:42:16preglowthey just make the directory themselves
15:42:40***Saving seen data "./dancer.seen"
15:42:56t0mashm? so make the users add an empty dir?
15:43:03t0masand let them put the images + wps files there?
15:43:28LinusNwhy not have the path in the .wps file?
15:43:41pregloweven better!
15:43:50t0mas*pling* nice idea
15:43:53*t0mas changes
15:43:56preglow.rockbox relative
15:44:11preglowor relative to .rockbox/wps
15:44:31preglowso you can have several sets of wps imagines lying around at once
15:44:31t0masit is already .rockbox relative..
15:45:01t0masI remember what the problem with that is
15:45:21preglowyeah, i just don't know how nice a practice it is to have bunch of bmps lying directly in .rockbox
15:45:21t0maswe need to parse the string then... and check where the filename ends
15:45:21preglowwill get very untidy very fast
15:45:39preglowwhy not just require a .bmp ending?
15:45:45t0masbut moving all to /.rockbox/wps/ is possible
15:45:50preglowif you see .bmp, that's the end of the filename
15:46:11t0masthat still requires parsing it...
15:46:42preglowso? that's not hard
15:46:57preglowshouldn't be, at least
15:46:57t0masok, maybe just change the format?
15:47:08t0masand split it on |
15:47:12preglowthat too would suffice lovely
15:47:33t0masthere is a define for max filename lenght right?
15:47:33preglowi know nothing of wps, but it looks good
15:47:59t0maswoops... stop...
15:48:01t0masproblem :)
15:48:19t0masI just skip the fixed lenght tag (like all others) now... for player
15:48:41t0masshould I do parsing there too?
15:48:42t0masso skip to the next ' '?
15:49:13*preglow has no idea about player related issues
15:49:29t0masI'll just try it :)
15:52:08preglowso, what to do with this resampling code? just have it lying around until we can use it in the playback system?
15:52:34LinusNeither commit it as-is, or #ifdef:ed
15:52:38LinusNyour call
15:52:59LinusNor do you think it sounds too bad?
15:53:57LinusNpreglow: make it compile in the sim first
15:54:57preglowi think it sounds just fine
15:55:11preglowok, i'll just use long longs liberally
15:56:44preglowi have no way of testing it in the sim, though, unless playback's started working there
15:56:59LinusNjust #ifdef SIMULATOR
15:57:09LinusNto make it compile
15:57:57preglowahh, yes, of course
15:58:06preglowi added a general MUL routine as well
15:59:21 Join Maxime`Mrn [0] (
15:59:44 Nick Maxime`Mrn is now known as Maxime (
16:00:01 Nick Maxime is now known as Maxime` (
16:01:19Maxime`first thing, thanks for rockbox on iriver ! :p
16:01:45Maxime`even if it's not over it's great already
16:02:16Maxime`and I'm trying to make a plugin for iRiver, (a small stepmania-like game) lol
16:03:25preglowit'll never be over ;)
16:03:36Maxime`preglow: you have an iRiver?
16:04:14Maxime`I can send you the -beginning- ? :p
16:04:35preglowLinusN: but what should i do with the code? just keep it in codecmpa.c so it can be used when someone gets around to bettering the playback system?
16:04:43preglowMaxime`: sorry, no, i gotta start doing other work now
16:04:53preglowi'll commit now
16:05:45preglowi'd much prefer it if the downsampling was in-place, though, but no time now
16:05:53Maxime`but, how to handle a loop like "every seconds.." "every millisecond.."?
16:06:11LinusNMaxime`: button_get_w_tmo()
16:06:39Maxime`hm, i'll have a look
16:08:29preglowah, i forgot it resamples even if it's not needed now
16:09:10LinusNpreglow: check out my new audio thread debug screen
16:09:15preglowi will
16:09:16LinusNupdate cvs
16:09:22preglowjust built i new rockbox
16:09:26LinusNdebug->view audio thread
16:09:42preglowoooh, greatness!
16:10:05preglowit panicked :-)
16:10:15LinusNupdate cvs again
16:10:23preglowi thought i got that last one
16:10:53LinusNi always have an "oops" commit right after my regular commits :-)
16:11:20Maxime`hm, why sometimes when the iriver is plugged to the power supply, it sends nasty sounds? :s
16:11:30LinusNit does?
16:11:50preglowvert buce
16:11:52preglowvery nice <-
16:12:02Maxime`I dunno if its my iriver who does this or if it's an rockbox-issue :x
16:12:20 Quit Seed (Read error: 60 (Operation timed out))
16:12:27preglowcodecmpa is around 50% boost ratio now
16:12:28LinusNi wonder why the pcm buffer usage fluctuates like that
16:12:50preglowlet's hope slasheri knows, heh
16:12:53 Nick Febs_away is now known as Febs (
16:12:54Maxime`you have the same problem LinusN?
16:13:22LinusNMaxime`: no
16:13:39*amiconn wonders whether he should tackle memcpy.
16:13:42LinusNgotta go now
16:13:54LinusNcu around
16:13:57 Part LinusN
16:13:58Maxime`cu LinusN
16:14:00Maxime`too late
16:15:18FebsMaxime': are you noticing the "nasty sounds" through headphones or when plugged into something else (like computer speakers, for example)?
16:15:22elinenbet0mas: how does the WPS image support work? :-)
16:15:35Maxime`Febs: when plugged to the computer, i'll try with headphones
16:16:19FebsI've noticed that with the original firmware and I've seen it reported on Mistic River. I think that there is a grounding issue.
16:16:24markunMaxime`: Do you hear it only when the iriver is not playing music? I think you can only hear the sound when the headphone amp is off.
16:16:41elinenbemy bad... I see it in the wiki.
16:16:53amiconnFor connecting the iriver to audio equipment, I'd rather use the line out
16:17:03Maxime`I only hear it with another thing than headphones
16:17:04amiconnThen there shouldn't be any grounding problems
16:17:08Maxime`and when listening music too markun
16:17:27markunMaxime`: Hm, yes, I also hear it now.
16:17:42 Join bobTHC [0] (
16:17:54Maxime`it's strange lol
16:18:11markunWhen I touch the iriver the sound is gone.
16:18:18FebsGround loop?
16:18:47Maxime`maybe the ground of the power supply touch the hearphone plug
16:19:00Maxime`it must be an iriver-issue lol
16:19:08amiconnpreglow: I wonder why your codecmpa.rock is 20 KB larger than mine
16:19:12preglowwill anyone cry if i leave codecmpa always resampling? it wont affect the sound
16:19:18amiconnDoes dithering add so much code?
16:19:24preglowamiconn because of the static buffers?
16:19:31preglowthey should be bss, though
16:20:08amiconn20 KB buffers for resampling?
16:20:08preglowoh yes
16:20:09preglownow i use more
16:20:10DBUGEnqueued KICK preglow
16:20:10preglowsince i just added stereo
16:20:16linuxstbpreglow: I'm happy with that (I can then play my 48KHz MP2 files). I'm assuming that it's possible to disable it easily?
16:20:23preglowlinuxstb: of course
16:20:27amiconnShouldn't resampling be possible with just buffering a few samples?
16:20:38preglowamiconn: because i need to upsample a frame of 1152 samples to a frame of four times that for the case of 11khz files
16:21:04preglowamiconn: this isn't how it's going to look, it's just proof of concept for when i/someone else puts it properly in playback
16:21:09amiconnI thought it resamples on the fly.
16:21:19preglowit resamples a whole frame at a time
16:21:26amiconnThe frame is already buffered in the codec
16:21:38preglowbut that doesn't help me
16:22:44Maxime`will someone test the pre-pre-pre-beta of my plugin? :s lol
16:22:44preglowideally, we should upsample just a bit of the buffer at a time, so we don't need tons of intermediate storage space
16:22:53preglowhowever, i sorely need to start working on something else right now
16:24:13preglowi'd be more than happy not to commit this, i don't really care
16:24:20amiconnWould it be possible to let the audiobuffer insert function do the resampling?
16:24:26preglowthat's what i want
16:24:34preglowbut no time for it now
16:25:26amiconnThe pcm buffer fluctuation is looking weird
16:25:57preglowwhen i tried running the audio debug screen again, it paniced again
16:26:14preglowmy reset button is destroyed
16:26:41preglowi don't need this
16:26:41t0masclean build:
16:26:41t0mas/home/tomas/dev/rockbox-devel/firmware/export/logf.h:31: error: parse error before "logfwrap"
16:26:41t0mas/home/tomas/dev/rockbox-devel/firmware/export/logf.h:31: warning: type defaults to `int' in declaration of `logfwrap'
16:26:41DBUGEnqueued KICK t0mas
16:26:41t0mas/home/tomas/dev/rockbox-devel/firmware/export/logf.h:31: warning: data definition has no type or storage class
16:26:47t0mas(a logf build)
16:26:53preglowTHANK GOD
16:26:55preglowTHANK YOU
16:27:00preglowi used the wrong screwdriver
16:28:05linuxstbt0mas: I've just tried a logf build, and it worked fine.
16:28:14preglowamiconn: this resampling sure as hell didn't do the cpu usage any favours
16:28:14amiconnI think I do understand the fluctuations now
16:28:38CtcpIgnored 2 channel CTCP requests in 6 minutes and 31 seconds at the last flood
16:28:38*t0mas starts digging
16:29:05linuxstbt0mas: I did an "rm -fr *" in my build directory, then configure, then make.
16:29:51t0maslinuxstb: I created a new dir for logf build
16:29:59preglowamiconn: i can't use the audio thread debug screen twice in a row, it locks up my unit
16:30:44amiconnHmm strange. I can enter & leave it at will
16:30:48preglowlinux just fixed it
16:30:59amiconnHe told so...
16:31:26amiconn[16:09:57] <LinusN> i always have an "oops" commit right after my regular commits :-)
16:31:35 Join DMJC-L [0] (
16:31:52amiconnThis time it was an OOps commit ;)
16:31:54preglowi've got that
16:32:26preglowseems i've got that, it still panics
16:33:01amiconnWhat panic do you get?
16:33:22 Join O112351 [0] (
16:34:08preglowtick_add_task(): out of tasks
16:34:08 Quit hicks_ (Client Quit)
16:34:31O112351this room is very very full, yet nobody will speak to me :(
16:34:41preglowcompletely reproducible, i enter the audio thread debug, press stop, enter it again
16:34:44preglowand it crashes
16:34:49GodEatermost of us are asleep that's why :)
16:34:52amiconnpreglow: Hmm, then you probably didn't compile with Linus' fix, or didn't install. I can't imagine what else would cause this
16:35:11preglowthen there's gotta be a broken dep somewhere
16:35:17ashridahO112351: questions usually get some kind of response too :)
16:35:41GodEaterI have a question - though it might all get a bit metaphysical...
16:35:44preglowi'll try a make clean
16:35:56GodEaterDid I imagine the speed of switching between mp3s getting slower between recent builds ?
16:35:57ashridahmetaphysical? as in ESP?
16:36:00GodEateror was it always that
16:36:12preglowGodEater: no
16:36:31GodEaterthank goodness. Any idea what's done it ?
16:36:31t0masGodEater: crossfade...
16:36:31t0mashave you enabled that?
16:36:34GodEaterdon't think so
16:36:37GodEaterI'll check
16:36:50O112351how can i tell who the moderator people are?
16:36:58GodEaterNope - it's off
16:36:58preglowwhy do you care?
16:37:14O112351am i just using a rubbish irc client?
16:37:25O112351i am using gaim
16:37:30ashridahO112351: there are no ops in here
16:37:44O112351that explains it :)
16:37:45ashridahat least, no ops who're sitting around with ops, there are people who have access to get it
16:37:59GodEaterso would having crossfade turned on help, or do I need to leave it off to keep the switchign responsive ?
16:38:05preglowamiconn: make clean, cvs update, make, complete delete of rockbox then unzip and i STILL get it
16:38:11linuxstbaudio debug is working perfectly for me (and very nice it is too).
16:38:58O112351does anyone else use Gaim?
16:38:58t0masno, Xchat
16:38:59preglowi use gaim, but not for irc
16:38:59t0maslinuxstb: your logf build works?
16:38:59GodEaterI like XChat - wish the firewall here let me use a decent IRC client.
16:38:59t0masmine still doesn't :X
16:38:59amiconnO112351: HydraIRC
16:39:00t0masrm -Rf *
16:39:00t0mascvs update
16:39:00***Alert Mode level 1
16:39:00***Alert Mode level 2
16:39:00t0maserror.... :X
16:39:04linuxstbt0mas: Yes.
16:39:19amiconnpreglow: Strange...
16:39:20O112351why not use Gaim for irc?
16:39:43Maxime`t0mas: -rf not -Rf
16:39:44linuxstbt0mas: Try deleting your apps and firmware directories and checking out fresh copies.
16:39:57O112351oops my name is wrong :P
16:40:14t0maslinuxstb: I'm working...
16:40:19t0masand it chokes on bool
16:40:30t0masextern bool logfwrap;
16:40:37t0maserror: parse error before "logfwrap"
16:40:38amiconnt0mas: wps image code adds about 1 KB on archos
16:40:46O112351what would people say is the best, most simple irc client?
16:40:47t0masamiconn: recorder only?
16:40:59t0masO112351: xhcat
16:41:03elinenbet0mas: you committed the WPS image commit? right?
16:41:08t0masyes I did
16:41:09amiconnt0mas: I'll compile for all archos targets...
16:41:16t0masamiconn: wait a moment
16:41:22t0masI'm improving it now..
16:41:22O112351i will try and find and download it
16:41:23elinenbet0mas: can you increase the image count from 9 to 99?
16:41:30t0maselinenbe: next version :)
16:41:34elinenbethanks :)
16:42:02elinenbealso, is it correct when i assume that image tags can be contained in conditional tags?
16:42:05amiconnelinenbe: That would be giant overkill imho. *1* image should be enough
16:42:15t0maselinenbe: no, they can't
16:42:17preglowamiconn: did it another time, now it works...
16:42:30t0masamiconn: I'm parsing the string anyway...
16:42:40amiconnpreglow: Linux, right? My guess is unclean dismount...
16:42:40t0masso everybody can chose for n or nn
16:42:43O112351bye bye, i'm going to switch to xchat
16:42:54elinenbewell, I am thinking more along the lines of "mp3" image when there is an MP3 playing −− ogg image when OGG is playing, etc...
16:42:54preglowamiconn: might be, can't remember ;)
16:42:59 Part O112351
16:43:14elinenbegiving the effect of multiple fonts, etc..
16:43:21t0mashow is it possible that gcc doesn't know the bool type??
16:44:12preglowamiconn: did you know why the buffer position fluctuates?
16:44:46linuxstbt0mas: Try adding #include <stdbool.h> at the top of logf.h
16:45:11t0masshall I commit that?
16:45:24linuxstbI think so. It doesn't hurt.
16:45:27t0masas using bool without that is kind of dirty... (because we don't use stdlib?)
16:45:36preglowbut ok, i gotta go, should i commit the resampler or not?
16:45:48amiconnMy interpretation: The small up&down fluctuations are caused by decoding in chunks. The overall tendency is downwards until it reaches the low watermark. Then the cpu gets boosted, and the overall tendency will be switched to upwards, until the high watermark is reached, where the cpu will be unboosted etc
16:46:12linuxstbIt probably works most of the time because <stdbool.h> is included in the main .c file, before logf.h is included. You must have found a case where that's not true.
16:46:24linuxstbMaybe Bagder's recent changes.
16:46:52t0maspreglow: commit
16:47:08linuxstbpreglow: commit
16:47:40preglowaight, if people complain why mpeg playback got so much slower, you know what to answer
16:47:43t0maspreglow: do you think it will be integrated in the playback code soon? I have some 48 khz files :)
16:47:55preglowt0mas: i have no idea, i won't have the time
16:48:17amiconnpreglow: You keep saying that....
16:49:01***Alert Mode OFF
16:50:24amiconnt0mas: It seems the code impact is even less than 1 KB
16:50:37amiconn(unless it isn't compiled for Ondio)
16:50:46 Join O11 [0] (
16:51:09O11hello again everyone
16:51:18*GodEater waves
16:51:45t0masamiconn: it's compiled for HAVE_LCD_BITMAP units
16:51:57t0masdon't know if that's ondio?
16:52:01t0masok, then it's < 1 kb :)
16:52:07preglowok, commited
16:52:09amiconnThey have the same display as the recorders
16:52:10preglowfeedback, plz
16:52:17O11why does it say that my old nickname; O112358 is already in use?
16:52:31O11it can't be
16:52:42amiconnIt's just mounted flipped, so the flip code has to work complementary
16:52:42t0masis gaim closed?
16:53:25O11i also deleted my irc setup in gaim
16:53:32preglowamiconn: don't .rocks exclude bss, btw?
16:53:38preglowmakes no sense to include it
16:53:50amiconnThey include .bss
16:53:59amiconnIt could be excluded, yes
16:54:14linuxstbpreglow: I'm assuming there is no space left in IRAM for your resampled_data buffer?
16:54:15O11is there a way to search all users in a server?
16:54:19amiconn...since we clear the plugin area before loading plugins
16:54:46preglowlinuxstb: oh god no, not by far
16:54:58preglowlinuxstb: the most critical case is downsampling, and that can be done in place
16:55:02amiconnNo, we can't do that
16:55:02preglowlinuxstb: i just haven't done it
16:55:19amiconnThis way the plugin loader wouldn't know how much ram the plugin is going to use
16:55:33preglowa header?
16:55:58O11so how can i solve my nick problem, any ideas?
16:56:11amiconnThe current solution is probably the simplest. Hard disk space shouldn't be a problem...
16:56:46 Quit DMJC-L ("Leaving")
16:57:34amiconnpreglow: U're resampling in the 32 bit domain?
16:57:41preglowamiconn: yep
16:57:52preglowamiconn: it pays to do all processing with as much precision as possible
16:58:21preglowremember that we might need to do further processing as well
16:58:21amiconnHmm, but reading 32 bit samples then writing to a buffer then reading again... doesn't help performance
16:58:55preglowi sincerely doubt doing 16 bit multiplies would help by much
16:59:03O11please help me someone :(
16:59:06preglowand we'd have more roundoffnoise
16:59:16amiconnI don't mean the multiplies, I mean the memory access
16:59:33t0masO11: type /whois [oldnick]
16:59:38preglowthat should be done in iram
16:59:38t0masand /whois O11
16:59:40t0masand then compare
16:59:42preglowlike i say, in-place
16:59:44amiconnApart from that, the standard coldfire multiplies are slooow and 32 bit multiplies are slooooooow
16:59:50t0masif the ip's are the same... it's you :P
17:00:00preglowamiconn: they're not THAT slow
17:00:22O11i see
17:00:22amiconn5 clocks for 16x16 and 9 clocks for 9 clocks for 32x32 is slow
17:00:39t0mas<Insomniak`> Stupid fucking Google
17:00:39t0mas<Insomniak`> "The" is a common word, and was not included in your search
17:00:53t0mas<Insomniak`> "Who" is a common word, and was not included in your search
17:01:15preglow5 clock for 32x32, i believe
17:01:16preglowwith two registers
17:01:20Maxime`("The Who" will be greater :p)
17:01:30O11thanks t0mas, gaim is still connecting me even though i have closed it and made it so it does not connect me to irc
17:01:38amiconnThe SH does 16x16 in 3 clocks, and that can be shadowed when doing other insns in between, so it needs 1 clock
17:01:41t0masO11: then kill it
17:02:10preglowanywho, i believe the mul i use now has a 5 cycle latency
17:02:43preglowfeel free to optimise, it needs it
17:03:11preglowwhat it needs most is an in-place downsample, so feel especially free to do that ;)
17:03:11O11what do you mean by kill it t0mas, and also how do you do personal chat things?
17:03:21DaKi][er011 - task manager....
17:03:33amiconnpreglow: It's even worse muls.w and mulu.w take 9 clocks, mulu.l and muls.l 18 clocks (!)
17:03:53amiconnWe're practically obliged to use EMAC
17:03:54preglowamiconn: i believe that's on a coldfire with no emac unit
17:04:13O11 i can't tell which task is the gaim one
17:04:23preglowamiconn: on coldfires with emac units, it's done faster, but i use the emac unit directly wherever i can anyway
17:04:26amiconnHmm, might be. I wonder how to tell which coldfire belongs to which family
17:04:40O11and how do you say something to a specific person?
17:04:47preglowours is a 2M
17:04:59amiconnOkay, then it's 3 and 5 clocks
17:05:09preglowwhere do you read this info, btw=
17:05:30amiconnThere's that nice COLDFIRE2UM.pdf in the datasheets wiki page
17:05:43preglowthere's timing info in COLDFIRE2UM.pdf and MCF5249UM.pdf, and they differ some places
17:07:49preglowamiconn: but yes, with all the processing stages we might end up doing, i think we should keep the precision as long as we can
17:08:06O11how do you send a message to a specific person?
17:08:14preglownot a must, of course, but that most certainly is how it's usually done in signal processing
17:08:14amiconnHmm, indeed there's contradicting information...
17:08:18Maxime`O11: /query nickname
17:08:33amiconnDidn't find the timing info in MCF5249UM.pdf btw
17:08:35O11only way?
17:08:38amiconn...until now
17:08:57 Join DMJC-L [0] (
17:08:57amiconnO11: Try /whois nickname
17:08:59O11can you not click on the person's name or something?
17:09:03preglowbut yes, someone please make downsample in-place, and please put it somewhere in playback
17:09:06preglowi gotta go, later
17:09:11 Quit preglow ("leaving")
17:09:23Maxime`O11: right button on the person, "Query"
17:09:30Maxime`If I remember
17:10:28Maxime`no not Chat
17:10:51Maxime`w8 a sec
17:10:54O11i don't see anything saying query
17:11:05Maxime`I'm launching Xchat w8 lol
17:11:46Maxime`"Open dialog windowé"
17:17:06 Quit GodEater ("CGI:IRC")
17:17:32amiconnt0mas: It's somewhat above 400 bytes
17:17:45t0masamiconn: ok
17:18:04t0maschar *a;
17:18:09t0masconst char* b;
17:18:12t0masb = a;
17:18:16amiconnNot too bad, compared to some other features...
17:18:30t0masthat's impossible right?
17:18:30amiconnNo it's not
17:18:54amiconnconst char* means that the data pointed to is constant, not the pointer itself
17:19:00t0masah ok
17:19:07t0masand I get a warning on that:
17:19:08t0maswarning: assignment discards qualifiers from pointer target type
17:19:10amiconnThat would be char const *
17:19:34amiconnYes, that is because you assign a pointer to constant data to a pointer to non-constant data
17:20:01amiconnCode using a is allowed to alter the data pointed
17:20:09amiconn...while code using b is not
17:20:18t0mashm... and what if I know the pointer a is to some part of b?
17:20:21t0mascast it?
17:20:37amiconnconst char* a;
17:20:55t0masoh... duh..
17:20:59*t0mas wakes up
17:21:10amiconnI know this, I did the constification of rockbox for ROM execution
17:21:14Roriwhere do I find the whole CVS list for iRiver?
17:21:53 Join Dave99 [0] (
17:21:53Rorior rather where do I find a more complete list of recent activity? :)
17:22:11Maxime`O11: like this?
17:22:14BagderRori: 'cvs log'
17:22:42Roriwhere is that?
17:22:45HClfront page?
17:22:50BagderRori: that's a cvs command
17:22:54O11yeah maxime, that is what i want
17:22:56Bagdercheck out the sources
17:22:59Bagderissue the command
17:23:07Bagderthe entire rockbox history
17:23:22amiconnt0mas: You can have fun things like const char const * ....
17:23:45amiconn...meaning a constant pointer pointing to constant data.
17:23:51Maxime`(hum, for a "long" number, it's %l no ?
17:24:01Dave99hi, I've been away for a while - has playback broken on hp120?
17:24:08RoriI must be thick
17:24:09Bagderits %ld for longs
17:24:14Maxime`thx ^^
17:24:14Rorican't figure out how to find it
17:25:35Roriah well too much effort anyhow
17:25:56 Join Seed [0] (
17:27:52Maxime`why does a long divided by an int returns 0? o_O
17:30:45 Join Stryke` [0] (
17:31:55 Quit Dave99 ("CGI:IRC")
17:34:17t0masMaxime`: no compiler warnings?
17:34:47 Join mborus [0] (
17:35:44 Quit mborus (Client Quit)
17:37:10 Quit DaKi][er ()
17:42:27 Join Maxime`Mrn [0] (
17:42:28 Quit Maxime` (Read error: 104 (Connection reset by peer))
17:42:44linuxstbMaxime`: When you say "long", do you mean "long" (32-bit) or "long long" (64-bit) ?
17:42:45***Saving seen data "./dancer.seen"
17:42:51t0maswhy does logfdump dump in reverse order?
17:42:52 Part fuzzie
17:43:28Maxime`Mrnlinuxstb: in fact, I want to do sth like this : result = ((i % 300) / 300);
17:43:28Maxime`Mrn(i is an int)
17:43:39Maxime`Mrnand i'd like to know what type must be "result" to have something.. :x
17:44:06 Join greygnome [0] (~greygnome@
17:44:16linuxstbIt must be a float - i % 300 will always be between 0 and 299. So integer dividing will give 0.
17:44:38 Quit Andrew179 ("Leaving")
17:44:38Maxime`Mrnyes i know
17:44:48Maxime`Mrn(it's for a kind of progress-bar)
17:45:07 Part greygnome
17:45:12 Join greygnome [0] (~greygnome@
17:46:44linuxstbMaxime`Mrn: Can you explain your problem? What exactly is "i"?
17:46:57Maxime`Mrni do but this always returns 0 -_-((int % int) / int)
17:47:07Maxime`Mrni do ((int % int) / int) but this always returns 0 -_-
17:47:21t0masint = ((int % int) / int)
17:47:39Maxime`Mrnno i've tried float = ((int % int) / int)
17:47:44Maxime`Mrnbut always 0
17:48:05t0masand I != 0 ?
17:48:36*amiconn is completely puzzled about some asm code in rolo.c
17:49:06t0mascall linus ;)
17:49:25linuxstbMaxime`Mrn: What exactly are you trying to achieve?
17:49:30amiconnOne instruction seems to be completely superfluous, and some others can be shortened considerably
17:49:55amiconn(not that it hurts performance, but why waste space?)
17:50:15t0masmaybe Linus forgot his *oops* commit there?
17:50:38Maxime`Mrnso, i have an int always increasing, and I want to have a number beetween 0 and 1, using (-myintincreasing- % 300) / 300, to have a number restarting after 300 increases
17:50:39amiconnI also wonder why he doesn't use moveq.l instead of move.l where appropriate
17:50:51Maxime`Mrn(don't know if I'm well explaining)
17:51:11bobTHCamiconn > not a part of the sanitize code ?
17:51:37amiconnbobTHC: Huh? What do you mean?
17:53:09t0masMaxime`Mrn: try (-myintincreasing- % 300) first
17:53:11t0masand logf that
17:53:15linuxstbMaxime`Mrn: result=((float)(i % 300))/300.0; (where i is int, and result is float)
17:53:29t0masand then devide by 300, and logf that too
17:53:30 Quit ashridah ("holy crap! 2am. sleep!")
17:54:05Maxime`Mrnlinuxstb i'll try
17:58:50greygnomeDoes anyone know if there is a way to recover files on the jukebox? Specifically, I was recording a meeting last night and I turned off the jukebox before hitting the stop button for recording. I have the Jukebox Recorder 20 (JBR20101). I have looked through the faq and searched the wiki so I hope this isn't a repeat question.
18:05:18Maxime`Mrnlinuxstb: it works if I add (float) everywhere ^^ lol :x
18:05:19 Quit Chamois (Read error: 131 (Connection reset by peer))
18:06:02 Join Chamois [0] (
18:08:17amiconnMaxime`Mrn: What are you trying to do? If it is for rockbox, forget about float
18:08:43Maxime`Mrnwhat should I use instead? :s
18:08:51linuxstbUse integers....
18:09:30Maxime`Mrnyeah ok.. lol
18:09:48linuxstbIf you need to draw a progress bar, then you need to scale (i % 300) from 0..299 to the width of your progress bar.
18:10:11Maxime`Mrnin fact, i want to scroll sth from right to left
18:11:28Maxime`Mrnhow to explain
18:11:38Maxime`Mrnyou know the game "stepmania" or "DDR" ?
18:12:24Maxime`MrnI want to show arrows coming to a point
18:12:38linuxstbAs amiconn said, if you want it to work with any speed, you need to think about how to implement it using integer maths only.
18:13:15Maxime`Mrni'll do that ^^
18:18:24O11Can anyone explain a possible reason for getting this message when trying to access a server? :
18:18:25O11<O11> −−- Connected. Now logging in..
18:18:25O11<O11> −−- G-Lined: Client joined '#zurna' - Turks have been banned from this network for years
18:18:25DBUGEnqueued KICK O11
18:18:25O11<O11> −−- Closing Link: O112358[] by Buffalo.NY.US.ZUH.Net (K-lined)
18:18:25O11<O11> −−- Disconnected (Remote host closed socket).
18:19:14O11i started getting it when i tried to join a room called something like KIllER TROJAN RooM EnTeR AND DIE!!
18:19:44O11just out of pure curiosity
18:19:53O11did i get banned from the server or something?
18:20:03 Quit DMJC ("Leaving")
18:20:10t0masyes you did O11
18:20:50O11do you think it might have assumed i was some kind of hacker BOT or something?
18:21:11O11well well well
18:21:30O11but i'm not a bot! i'm a real person :'(
18:21:39O11and not a hacker in any way
18:21:49O11it's made a mistake :'(
18:23:50thegeektry another server in that net
18:25:05t0masthegeek: gline is global
18:25:13t0mashe has to email the serveradmin
18:25:30amiconnIs k-line also global?
18:25:58 Part greygnome
18:26:11t0masamiconn: no
18:26:43amiconnO11: Some trojans use those special channels, and you'll get banned for sure when you try to enter them
18:26:44 Quit Stryke` (Read error: 60 (Operation timed out))
18:28:36O11could they not just remove the channel to avoid people like me getting banned?
18:28:37O11should i look around on to try and contact an admin or something?
18:28:58bobTHCu have been klined so only ircop can help u
18:29:48t0masbobTHC: glined
18:30:01 Quit ]RowaN[ ()
18:30:05bobTHCand K
18:30:24bobTHC>> "O112358[] by Buffalo.NY.US.ZUH.Net (K-lined)"
18:30:48t0masoh lol
18:31:02O11what should i do?
18:31:22CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
18:31:22*amiconn summons LinusN
18:31:22O11i have been wrongly k-lined :(
18:31:23 Nick QT_ is now known as QT (as@area51.users.madwifi)
18:31:29O11how can i contact someone?
18:31:52t0mascheck the website
18:31:56bobTHCchange from buffalo ircd, in general it's serv specific
18:32:35 Nick Febs is now known as Febs_away (
18:32:43t0mashm... we use unreal... there it's a setting
18:33:04O11the ircop's don't seem to have any contact information
18:33:53bobTHCchange serv or contact the op
18:35:03O11i want to go to a specific room so i can't really change server can i?
18:35:33bobTHCof course they're linked...
18:35:48bobTHCexcept if a netsplit occur
18:36:19t0masbobTHC: but he still has a Gline
18:36:26t0masso he can't use any other server in that net either
18:36:47t0masamiconn: can you test WPS image stuff again?
18:36:47bobTHCdepend of op rules
18:37:00t0masit takes some more space now... but Linus and preglow liked this format :)
18:40:08linuxstbt0mas: Why do you need both a number and a filename for each image?
18:41:06t0masbecause I can't use the filename as an index in the array
18:41:22t0masand I have to do a simple check for loaded or not...
18:41:55t0masbecause I would waste resources by loading it every time...
18:46:58SlasheriHmm, i will move that resampler code to playback.c so codecs don't need to worry about that
18:48:18O11i have contacted the op now via email and am awaiting a response, I hope he isn't asleep :)
18:48:43linuxstbt0mas: I may be misreading your code, but what happens if you try and display the same image twice? I can't see how it would work.
18:49:20SlasheriI would also suggest an option not to animate the charge status while charging. instead it could display current battery charge level (something like for example current_voltage - 0.05 to make it accurate)
18:50:02bobTHCO11 > Glined is usualy temporary... so wait and see ;)
18:50:30 Join TCK- [0] (
18:50:30t0maslinuxstb: it won't display the second image :)
18:50:33O11well i have sent an email now
18:50:38t0masyou'll have to chose 2 different id's
18:50:40t0masand load the same file twice
18:50:49t0mas(how inefficient)
18:51:22t0masdamn... preglow rocks... my 22 khz song sounds way better than with the iriver firmware :D
18:51:32linuxstbAlso, can you have conditionally displayed images?
18:51:57t0masbut it's on my list
18:52:08t0massomeone wanted it
18:53:20linuxstbSo when would rockbox try to reload an image that is already loaded? I would have thought that either they were all loaded, or none of them were.
18:55:09t0masno... if you want to use 1 image 2 times... then you have to add 2 tags
18:55:13t0mas%x1 and %x2
18:55:17t0maswith the same filename
18:55:26amiconnt0mas: I think the wps formatting codes for bitmaps should be split in 2, one for loading the .bmp and assigning a number, and another for actually displaying the image, giving position and all
18:55:41t0masthat sounds goot to me
18:55:42 Quit thegeek (Read error: 145 (Connection timed out))
18:55:57 Join belgarath [0] (
18:56:00t0masit would solve the 2x one image too
18:56:24amiconnThis way one could display the same bitmap in multiple places and still load it only once
18:57:42linuxstbOr extend the command to accept multiple locations.
18:58:35amiconnThat would be less flexible
18:58:41amiconn..and harder to read
18:59:20amiconnHaving 2 format codes would e.g. allow to put the display codes into different conditionals
18:59:30amiconn(once conditional rendering is done)
19:00:02O11Sorry everyone if I was wasting everyones' precious time with my questions about getting my server bans lifted etc. someone has been telling me off so i thought i should say sorry
19:00:17 Quit DMJC-L (Connection timed out)
19:00:50t0masamiconn: it's a great idea...
19:01:11t0masI can parse the second tag (displaying) somewhere else...
19:01:50t0masand that way make it conditional :)
19:02:08amiconnt0mas: I imagine displaying an album symbol in front of an album tag, but if there is no album tag and the folder name is displayed, it could display a folder symbol instead...
19:02:16t0masyes, that would be cool :)
19:02:36amiconnThat _could_ even look good on archos if the symbols are small enough
19:03:03amiconn8x8 is probably the maximum
19:04:41 Quit bobTHC ("Smoke Weed Every Day !")
19:05:45linuxstbIs it worth thinking about album art - e.g. a special tag for files inside the same directory as the currently played track.
19:06:47amiconnNot with b&w imho. That reminds me...
19:07:28 Join tvelocity [0] (~tony@
19:07:33 Quit TCK (Read error: 110 (Connection timed out))
19:10:18 Join Tangleding [0] (
19:15:00TangledingNice work today again
19:17:19amiconnLinus' audio debug code added 5 warnings to gmini build :-/
19:18:36TangledingHello Amiconn :)
19:19:46 Join daven [0] (
19:19:46Tangledingi also read that the iRiver wave recording GLITCH is fixed by Rbx...
19:19:56TangledingThat's very cool new!
19:20:27Tangleding(i barely understand why iRiver didn't fixed this stupid bug if it was only software...)
19:22:36TangledingOoooh the MisticRiver rockbox board is more and more crappy... :(
19:23:10linuxstbTangleding: I didn't think enough testing had been done to say for certain that the recording glitch is gone.
19:23:33Tangledingah okay
19:23:40Tangledingi read this somewhere
19:23:57Tangledingshould have misunderstood
19:25:57TangledingWas this post in fact:
19:26:15 Join ghostiger [0] (~ghostiger@e2b2959e7b53fe7c.session.tor)
19:28:59davenhave you seen this post regarding the glitch?
19:29:30davenapparently a solution has been found with iriver fw
19:29:57davenwhich gives hope it can be dealt with in rockbox
19:30:25Tangledingi watch it thanks Daveb$
19:32:40TangledingSeems very strange if i understand well...
19:33:01TangledingI guess it means it's really only a stupid soft bug no?
19:33:28 Join thegeek [0] (
19:33:40davenappears so. i guess i river just didn't put much effort into fixing it
19:33:52davenhifi recording is not a priority for most unfortunately
19:35:29TangledingThat's quite incredible
19:35:44Tangledingi remember about big petition for glitch fix
19:36:13Tangledingreally can't believe they did nothing to solve something so silly
19:36:22daventhat was the only thing keeping me from buying an iriver for the longest time
19:36:31TangledingThanks again for the link daven :)
19:36:43davennp :)
19:36:46Tangledingsouds really incredible
19:41:36TangledingI m' really astonished
19:41:56Tangledingi m gonna relay this information to some HA guys who will be very interested
19:42:03Tangleding(JeanLuc for exemple)
19:42:48***Saving seen data "./dancer.seen"
19:45:37 Join muesli- [0] (
19:45:53Tangledinghi HCl
19:46:05muesli-hi tang
19:46:06Maxime`Mrn(hum, 'dont understand what RoLo is? or will be..)? :s
19:46:14muesli-hi HCl
19:46:16Tangledingyou're laughing about this GLITCH story?
19:46:24TangledingHi Muesli
19:49:15HCl :)
19:49:27*HCl goes back to catch up on sleep
19:49:38*muesli- yawns :D
19:50:32TangledingGod night HCl
19:50:32DBUGEnqueued KICK Tangleding
19:50:49Maxime`Mrni've understood ^^ (best way, test! :p) (I can launch original firmware using RoLo?)
19:51:12amiconnNot on iriver, since the original firmware doesn't exist as a file
19:51:25Maxime`Mrnk it's what I thought ^^
19:53:14 Join LinusN [0] (
19:53:46amiconnhi again LinusN
19:53:47LinusN17.48.36 # * amiconn is completely puzzled about some asm code in rolo.c
19:54:14amiconnI'm no longer puzzled as you might have guessed from my latest commit
19:54:34amiconnHowever, I still wonder why crt0.S doesn't use some obvious optimisations
19:54:41Slasherihi using moveq.l where appropriate
19:54:49 Join preglow [0] (
19:54:49amiconnhi Slasheri
19:55:21amiconnLinusN: Btw, there was a bug in rolo_reload() which fortunately didn't hit
19:55:43amiconnIt still used 0x30000000 as the start address (hardcoded) instead of 0x31000000
19:55:58LinusNamiconn: you still wonder why i haven't spent more time optimizing stuff that doesn't need optimizing? :-)
19:56:16Slasheripreglow: i am thinking about moving the resampler code to playback system. But if i understood correctly that requires the resampling to be made before dithering. Maybe i can move dithering stuff to playback system too?
19:56:26amiconnIt's not that it needs speedup, but saving code can never hurt
19:56:37amiconnmoveq.l uses 1 word. move.l uses 3
19:56:56amiconnmove.l #immediate,dest
19:56:58preglowSlasheri: yes, you should, if you ask me, we should do as much processing as we can with the higher precision, then dither it down to 16 bits
19:56:58LinusNoh, 4 bytes! :-)
19:57:07amiconn4 bytes and again 4 bytes and....
19:57:10LinusNamiconn: i'm being sarcastic
19:57:56Slasheripreglow: ok, good. Is the dithering/rescaling code in codecmpa so general that it could be integrated to playback system?
19:58:04preglowSlasheri: no idea
19:58:15Slasherihmm :)
19:58:29preglowi can't hear the difference with dithering, so i don't care either :)
19:58:54Slasherii think that ideally it's only needed in codec to configure(SET_SAMPLERATE, ...) and configure (SET_DITHER_ENABLE)
19:59:11preglowthose should be in settings
19:59:15Slasheripreglow: hehe, but what about the scaling?
19:59:17Bagdercvs update: conflict: removed `apps/plugins/codecmpa.c' was modified by second party
19:59:23SlasheriThat should be general?
19:59:48preglowbagder: sorry :/
19:59:48Bagderno worries
19:59:51preglowslasheri: scaling and clipping needs to be done
20:00:00preglowslasheri: the scaling should always be a simple right shift
20:00:22Slasheripreglow: ok, so that could be integrated to playback code :)
20:00:34preglowclipping depends on the codec
20:00:57preglowbut all transform codecs need clipping
20:01:08preglowclipping is not needed for lossless codecs, nor is dithering
20:01:17Slasherihmm, so clipping should be configurable also
20:01:24 Quit linuxstb (Read error: 145 (Connection timed out))
20:01:24preglowso there should be a codec flag or something, telling if the codec is lossless
20:01:58 Join linuxstb [0] (
20:01:59preglowslasheri: the problem is, we can't afford too many passes over the buffer, we should integrate as much as we can into one loop
20:02:10Slasherii think that every codec can configure that clipping (enable it if necessary)
20:02:19preglow320kbps mp3 is REALLY struggling with resampling enabled
20:02:24preglowSlasheri: good idea
20:02:29Slasheripreglow: yes, i keep that in mind
20:02:40Slasherieverything will be done on one buffer
20:02:51preglowdownsample() should be made to work in place
20:02:57preglowsince that should be possible there
20:03:01preglownot so for upsample()
20:03:07amiconnLinusN: There are 3 places in crt0.S that could use moveq.l. And line 183 could be replaced with clr.l (%a2)
20:03:41LinusNamiconn: do whatever changes you feel necessary
20:04:54amiconnLine 183 is bootloader only, so I can't test
20:04:55preglowof course you can ;)
20:04:58Bagderfirst codec move test coming right up
20:05:16amiconnThis is the bootloader crash recovery, so I'd need to make it crash first
20:06:23belgarath"currently it slows down playback quite a bit" - is that normal 44khz files as well or just if files need sampling?
20:06:29amiconnLinusN: A related question: Why are the m68k assembler parts indented by 8 spaces, unlike all other asm?
20:06:34Bagder inflating: .rockbox/codecs/vorbis.codec
20:07:23preglowbelgarath: ALL files are resampled, currently
20:07:55Slasheripreglow: Hmm, is there any need to resample 44 kHz?
20:07:58CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
20:07:58*t0mas is away: party! :D
20:07:58preglowbelgarath: there's no difference in quality if the file doesn't need it, but the cpu is still being loaded
20:08:09Bagdergood old paperclip
20:08:26preglowSlasheri: 'course not, i just couldn't be bothered to do the code, since it was all going to be reimplemented in playback anyway
20:08:26t0masBagder: saw my question?
20:08:31t0mas(logf reverse dumped)
20:08:34Slasheripreglow: ah, ok :)
20:08:36preglowand i'm writing a report that badly needs writing now
20:08:37t0masis there a reason for that?
20:08:44preglowrockbox is too bloody fun :/
20:08:52Bagdert0mas: you mean when saving to file?
20:08:55t0masor shall I fix it tonight/tomorrow? or do you have time left?
20:09:01t0masBagder: logf dump in the debug menu
20:09:24BagderI was just being lazy when I wrote it and I didn't think it mattered
20:09:48t0masah, do you mind if I change that?
20:09:55Bagderno, go ahead
20:10:01t0masok, I'll take a loot tomorrow
20:10:15Bagdera loot! ;-P
20:10:16LinusNamiconn: i guess i just didn't care when i wrote that asm part, and emacs has 8 spaces by default in asm mode
20:11:49preglowemacs :/
20:12:06preglowgreat! the oven's broken
20:13:28Bagderin case anyone wants to have a look/comment
20:13:33Bagderdoesn't work yet
20:14:37 Join Stryke` [0] (
20:17:09 Part LinusN
20:20:28Bagder... and I see why
20:20:32preglowSlasheri: you should probably also make it so you don't need to resample structs to resample a stereo stream, that should speed things up somewhat, since most of the resample variables wont have to be calculateed twice for each channelø
20:20:42preglowto = two resample structs
20:21:28Slasheriok, good :)
20:25:04 Join webguest87 [0] (
20:25:04 Quit TCK- (Read error: 104 (Connection reset by peer))
20:26:42 Join TCK [0] (
20:28:56 Quit Stryke` (Read error: 60 (Operation timed out))
20:30:21 Quit webguest87 ("CGI:IRC (EOF)")
20:30:31 Join webguest74 [0] (
20:32:05 Quit O11 ("Leaving")
20:32:33 Quit muesli- (Read error: 110 (Connection timed out))
20:41:54 Join webguest58 [0] (
20:41:59 Quit webguest58 (Client Quit)
20:45:32Coldtoastwhat are the dimensions of the iriver screen in pixels?
20:45:47Bagderh1x0 models
20:45:50Coldtoastk. thanx
20:46:20 Join asdsd____ [0] (
20:46:23 Part asdsd____
20:48:45BagderSlasheri: here?
20:49:20Bagderah never mind
20:52:25Maxime`Mrnhm, cygwin .. pure shit :s
20:52:54preglowcygwin is spawn of satan himself
20:53:23Bagderbtw, I'm currently disabling the *2wav plugins
20:53:27Maxime`Mrnyeah :x lol
20:53:28preglowBagder: agreed
20:53:49preglowsomeone needs to add 'dump to wav' to the file menu, though
20:54:05preglowso we've still got the functionality
20:54:25preglowor better yet, dump to aiff!
20:54:30preglowno byte swapping :)
20:56:23*Bagder prepares paperclip
20:56:46Bagderwooo, music!
20:57:06 Quit belgarath ("CGI:IRC")
20:57:06Tangledingpaperclip is our friend
20:57:35Bagdercarmina burana, rotating cube *and* a codec api
20:58:22Maxime`Mrnhm, can we now listen music and rocckboy @ the same time?
20:58:32Bagderyou can try at least
20:58:47Bagderit might result in fun effects
20:58:50 Join TCK- [0] (
20:59:11Maxime`Mrnbefore, it hanged "ilptr"
20:59:21Bagder127KB patch
20:59:39BagderSlasheri: I could use your feedback on this before I commit
20:59:50BagderMaxime`Mrn: we fixed that this morning
21:00:10preglownow there's a patch with a great big head of hair
21:01:37Maxime`Mrna long loading and..
21:02:27 Join Audiophil [0] (
21:02:44Audiophilhi, anyone made a image-wps yet?
21:06:05Tangledingabout the glitch
21:06:17Tangledingi wonder about a strange hypothesis
21:06:44Tangledingwould you think it could have been sorta unfaire watermark from iRiver?
21:08:26preglowtheir firmware isn't exactly high quality
21:08:31preglowi don't know what to expect from them
21:08:59BagderI think it is a plain bug, possibly related to something technically tricky
21:09:06Tangledingyou should be right in fact Prelow
21:10:09preglowi wonder why you can't control playback speed for anything but mp3
21:10:18preglowperhaps their vorbis implementation runs really marginally
21:10:23 Join Susu [0] (
21:10:56Susuanyone around?
21:10:57preglowmore likely they implemented it like i implemented resampling in rockbox ;)
21:10:57preglowsusu: apparently
21:11:16SusuI've got an mp3 player with fm radio that was shipped with the rockbox firmware, but it's not on the rockbox site
21:11:39Bagdercan you show us a pic?
21:11:46SusuI put it into firmware upgrade mode
21:11:55Susuthen plugged it into the usb port and got a file 'firmware.h'
21:11:56Susu$Id: uda1380.h,v 1.5 2005/06/16 20:16:58 preglow Exp $
21:12:04Susuyeah I'll get a picture
21:12:12Coldtoastwhat brand is it?
21:12:20Susuno idea what the brand is
21:12:24Susuit's unbranded
21:12:30SusuI've been trying to find the firmware for it
21:12:34Coldtoastdid you get if off of ebay?
21:12:41Susuwhen I searched for a quote from the manual...
21:13:02Susubut the maxdigital site has no firmware on it
21:13:19SusuColdtoast yes I got it from ebay
21:13:36Bagderis it hard-disk or flash based?
21:13:59 Quit cYmen (Read error: 113 (No route to host))
21:13:59*preglow does the 'wtf'
21:14:12Susuthe box was sealed
21:15:00Bagderquite a mystery!
21:15:17 Quit yngwi (Read error: 60 (Operation timed out))
21:15:45preglowa picture would be very cool
21:16:01Susujust took some on webcam
21:16:08SusuI'll upload them somewhere
21:16:33 Quit TCK (Read error: 110 (Connection timed out))
21:16:40 Quit Tangleding ("Chatzilla 0.9.68a [Firefox 1.0.4/20050511]")
21:17:03tvelocitysounds like it's a harware ripoff of some other player rockbox supports
21:18:29preglowthat would be fun :]
21:19:41Sususorry the quality's not so great
21:19:52preglowyou sure it's running rockbox?
21:20:00SusuI uploaded the 'firmware.h' file I got from it
21:20:20Susuyou put it in firmware upgrade mode and plug it in, then you see this on the disk instead of your mp3s
21:20:33preglowthat's pretty weird
21:20:40preglowthat's the header for the sound chip in the h1x0
21:20:58preglowgod knows why there's a header file lying around
21:21:02SlasheriBagder: nice patch :D just go ahead if it works
21:21:12preglowit's very recent as well
21:21:31preglowsusu: and that's all that you can see on it?
21:21:57Susuthe date is wrong
21:22:15Susuit's not 16th june
21:22:21Susuunless they meant 2004
21:23:09preglowno, that's the exact date i modified that file
21:23:26Susubut that file has been there since march :\
21:23:39tvelocityrockbox code from the future!
21:23:40preglowthat's not possible
21:23:42Suckamaybe they have a time machine!!!
21:23:58Susuthe player has not been updated and it was bought in march :\
21:24:08preglowyou must have placed the file there by accident somehow
21:24:16SlasheriBagder: btw, codecs.c is missing from the patch
21:24:40Susuthere's no way I put the file there by mistake
21:24:55SusuI'd not even heard of rockbox until I saw the file
21:24:58preglowsusu: it's quite impossible that that file's been on the device since march, it contains changes i did on the exact date it is marched
21:25:12 Join ]RowaN[ [0] (
21:25:26]RowaN[time to play with wps images =]
21:25:29preglowsusu: then this is weird indeed
21:25:35tvelocitymaybe there is more future rockbox code hidden there
21:25:41preglowbut still, unless somehow has a time machine, it's quite impossible
21:25:43BagderSlasheri: there's a codecmv-4.patch there now with those
21:25:47Coldtoastdyou said "(Susu): you put it in firmware upgrade mode and plug it in,"
21:25:51Coldtoasthad you done that?
21:25:53SlasheriBagder: great :)
21:26:02*amiconn wonders why his H140 drop on the floor caused the HD to increase its noise level
21:26:04SusuColdtoast yes
21:26:13Susuyou should plug it in and then run the firmware upgrader
21:26:23Susubut I just browsed the 'USB Storage Device'
21:26:26Coldtoastwell then. would that not be how the file got on there?
21:26:27Susuand found that file
21:26:33]RowaN[greyscale isnt implemented on rockbox yet properly is it?
21:26:40preglowrowan: correct
21:26:56preglowthe bmp loader only supports 1 bit files anyway
21:27:02amiconnI dropped it from the table to the floor, about 75 cm. If it stopped again within 2mm on the floor (reasonable since there is some carpet), the shock shouldn't have exceeded 400G
21:27:06]RowaN[1bit porn will be leet
21:27:27amiconnHowever, the Toshiba should survive an 800G shock when not operating...
21:27:59Susuok I just installed the firmware upgrading software
21:28:25tvelocityis it called "cygwin"?:P
21:29:09Susuat least I know it's made by 'wilson co. ltd' now
21:29:10SlasheriBagder: Hmm, it looks good. Just commit :)
21:29:38Bagderbuckle up my friends ;-)
21:30:53 Quit thegeek (Read error: 131 (Connection reset by peer))
21:30:57Susu"This file is in the directory "(install directory)\recover", and its name is $ACTSDK$.FW."
21:31:07Susuis ACTSDK anything to do with rockbox?
21:31:14 Join thegeek [0] (
21:32:25amiconnI think in case the HD breaks, I'll upgrade to the 60 GB model if I can find one. It will be able to take twice the shock...
21:33:22]RowaN[images in wps seem to be inverted.. on purpose?
21:34:09]RowaN[what is the resolution of the iriver screen?
21:34:39]RowaN[thanks max
21:35:00Susuthanks for the link tvelocity
21:35:37Coldtoast"Be careful about upgrading the firmware. One of my first reactions was to upgrade the firmware but I soon found out that there are a lot of MP3 players out there identified by the firmware upgrading software as: "Device Name: S1 Mp3 Player" and "Productor: Wilson Co. Ltd.". These players have little in common and therefore the Firmware is often incompatible leading to non functioning players. For some more info see here"
21:36:11Susuyeah Coldtoast I just flashed it and it's not booting up
21:36:48SusuI just got it into recover mode :D
21:37:37 Join Maxime` [0] (
21:37:38Coldtoastis that the box is came in?
21:37:50 Quit Maxime`Mrn (Read error: 131 (Connection reset by peer))
21:38:20thegeekSusu : you just flashed an iriver?
21:38:24thegeekif it takes a lot of time to start
21:38:26thegeeklet it work
21:38:36thegeekit can take little while on first startup
21:39:15 Join Sususu [0] (
21:39:15Sususuhow queer
21:39:32SususuI ran the recover software, it randomly hung for a bit and then my machine rebooted
21:39:42 Quit Susu (Read error: 104 (Connection reset by peer))
21:39:48Coldtoaston that page I posted. is that the player?
21:39:54]RowaN[only scrolling text overlaps images in wps.. would be cool if all text would overlap.. or none heh
21:40:30ColdtoastI'm just going to stich the rockbox logo on the bottom of teh wps I think
21:41:48 Join Susu [0] (
21:41:53Sususame again :\
21:42:03 Quit Sususu (Read error: 131 (Connection reset by peer))
21:42:50***Saving seen data "./dancer.seen"
21:43:43*preglow dives for cover
21:43:45Susunow I really have no idea what to do
21:43:58Coldtoastwhy were you upgrding the firmware?
21:44:18Susufor some reason the radio has never worked
21:44:32Bagderoh warnings
21:44:33Susuand the guy who sold it said that it's a common problem and that the firmware needs to be upgraded
21:44:43Susubut then he disappeared
21:45:00Coldtoastis that the one?
21:45:17Susuno but the guy who sold it asked if the box had a violin on it
21:46:51Susutrying a different version...
21:46:52SusuI wonder if it will reboot this time
21:47:55Coldtoast might have relevent stuff for you
21:48:01Bagderthis overhaul can use some further cleaning up
21:48:25Coldtoastseems that project is for devices known as "S1 Mp3 Player".
21:49:19 Join Sususu [0] (
21:49:26Sususuyep, it rebooted again >_M
21:49:56 Quit Susu (Read error: 104 (Connection reset by peer))
21:50:10ColdtoastSususu: read this > se some further cleaning up
21:50:10Coldtoast(Coldtoast): seems that projec
21:50:20ColdtoastSususu: read this
21:50:22preglowi sincerely doubt this mp3 player runs rockbox
21:50:26preglowit's got a z80 core!
21:50:38Susususo how the hell did that file get there?
21:50:52preglowi don't know, but i'm 100% certain it wasn't there when you bought it
21:51:13tvelocitymaybe they use parts of rbx code?
21:51:29preglowthe mp3 player doesn't even have the hardware that firmware.h is meant to use
21:51:40]RowaN[what prog can i use to make fonts for rockbox?
21:52:20preglowseems h1x0 uses the same fm chip, though
21:52:24Sususupreglow I really have no idea how that file got there
21:52:25Bagderyour favourite font editor that can save fonts as BDF
21:53:14thegeekI get an error about bootbox when I try to checkout rockbox-devel now
21:55:19]RowaN[i dont have a favourite hehe
21:56:20Sususuoh yay
21:56:23Sususuthanks for the help guys
21:56:26SususuI got it working :D
21:56:44preglowthat's lucky
21:56:45Maxime`(hum, FM support for iRiver is not one priority, that's it?)
21:57:12BagderMaxime`: feel free to send us a patch
21:57:14preglowmaxime: some guy has it working, but it hasn't been put in cvs yet
21:57:45ColdtoastI've had my h140 since Nov last year and I don't think I've used the radio once
21:57:51Audiophilpreglow, seriously? :thumbsup: (:
21:58:55 Join LinusN [0] (
21:59:04preglowaudiophil: yes, austriancoder said he had it working, but then he vanished and reportedly destroyed his mp3 player
21:59:28Audiophilarf, someone donate one to him ;)
22:01:36 Quit RotAtoR ()
22:01:58thegeekprimus compiles # cvs co rockbox-devel
22:01:58thegeekcvs server: cannot find module `bootbox' - ignored
22:01:58thegeekcvs [checkout aborted]: cannot expand modules
22:02:03thegeekanyone else getting the same?
22:02:24Maxime`pcm_playback.c:270: warning: `interrupt_handler' attribute directive ignored < normal?
22:02:33Bagderbootbox is not a module
22:02:39BagderLinusN: you modified?
22:02:48thegeekand when using latest build from the webpage, I ge I0B:Line-F at 32F47770
22:02:54thegeekwhen trying to play music
22:03:04LinusNthegeek: get the latest blleding edge
22:03:06 Join benjimima [0] (
22:03:08thegeekI did
22:03:12thegeekand I get that line-f ;)
22:03:15thegeekwhen playing music
22:03:20thegeekI deleted all my old rockbox stuff
22:03:28thegeekand just unzipped the latest build
22:03:53thegeekthat's the latest right?
22:04:19 Quit benjimima (Client Quit)
22:04:19Coldtoastthat's teh bleeding edge build
22:04:24 Join benjimima [0] (
22:04:24thegeekbut that does not build the codecs
22:04:35Maxime`kernel.c:234: warning: `interrupt_handler' attribute directive ignored < :x
22:04:42thegeekI did a co of just "rockbox"
22:04:44thegeekand got the same
22:04:52thegeekbut I need the codes too in order to play anything)
22:05:28Maxime`hum .. another .. /home/maxou/m68k/lib/gcc-lib/m68k-elf/3.3.4/../../../../m68k-elf/bin/ld: error: no memory region specified for loadable section `.entry'
22:06:18thegeekif I could just co rockbox-devel I'd build it myself;)
22:06:29preglowyou people need to stop using gcc 3.3
22:06:40preglowit's got known problems with some rockbox code
22:06:53Maxime`ah yeh
22:07:01thegeekit's in the wiki
22:07:02Maxime`sry i've downloaded the wrong package
22:07:14*Maxime` −−> [- ]
22:07:53amiconnBagder: Your codec move broke the gmini build, though I don't understand why..-.
22:07:56Bagderthegeek: try again now
22:08:28Bagderamiconn: no proper #ifdef uda1380 protection
22:08:41amiconnOops, my ellipsis looked like morse code...
22:12:34thegeekI dont think the codecs are built
22:12:42amiconn-... .- −−. -.. . .-.
22:12:48Bagderthegeek: because... ?
22:12:53thegeekI mean
22:13:00thegeekwhen I do what I've always done
22:13:05thegeekthe codecs are just not there;)
22:13:08Bagderyou're right
22:13:25Bagdermissed a file in my commit
22:13:52Bagder... or two
22:14:42 Join OnkelJonas [0] (
22:17:57thegeekworks good:)
22:18:48Bagdernow run cube while listening! ;-P
22:19:46benjimimacan any of the pther plugins except cube be ran while listening to music?
22:19:52Bagderall of them
22:20:07LinusNnot rockboy
22:20:13Bagderwhat happens then?
22:20:22Maxime`erm.. compiling gcc on a 333Mhz.. tiring.. lol
22:20:23LinusNit needs the buffer mem
22:20:49Maxime`trying to run rockboy as listening = big loading then... loading..
22:20:56Bagdergot 32 fresh new gmini warnings
22:20:59preglowmarking which plugins can't run while playing would be nice
22:21:08preglowbut i guess that's impossible with the current plugin foirmat
22:21:20Maxime`or add an (if_playing() = return; }
22:21:37Bagderyes, the plugin could check if it can run or not
22:21:39 Join belgarath [0] (
22:21:41Bagderor rather, it should
22:22:03preglowis it possible to suspend a plugin while it runs?
22:22:07thegeekwhen I use images in the wps
22:22:10thegeekand then change the wps
22:22:17thegeekand try to reload it or any other wps using images
22:22:20Bagderpreglow: no
22:22:24thegeekit does not update
22:22:30preglowBagder: then that solution should work wonders
22:22:47Bagderrockboy _does_ do that, but in a weird order
22:23:09Bagderor perhaps I'm stupid
22:26:19preglowbagder: the recent cvs activity table certainly does indicate you're mad
22:26:30preglowif not stupid
22:27:07 Join Moxon [0] (~Moxon@
22:27:13preglowa couple more and you've filled the entire table!
22:27:38Bagderis that a challange? ;-P
22:28:30Moxonheya ppl!
22:28:45MoxonI have got a question. Is there any roadmap for the iriver portin effort?
22:29:12Bagderfull speed ahead!
22:29:28Maxime` < the table here isn't actualised?
22:29:41LinusNMaxime`: sort of
22:29:42benjimimais the bleeding edge build meant to be only 500kb?
22:30:11preglowwhat the hell?
22:30:20benjimimawell the builds are usually around 900kb arnt they?
22:30:24MoxonLinusN: do you need testers who check the daily builds? I could try to do that.
22:30:28preglowi've still got mpa2wav and mpc2wav, and i can't bloody well rid myself of them
22:30:46preglowdeleting them only has them coming back on the next cvs update!
22:31:08Maxime`hum, does anyone has a little idea on how we may handle the FM chip? :x
22:31:12LinusNMoxon: lots of people do that
22:31:23MoxonLinusN: is there any coordinated effort?
22:31:23LinusNjust install, run and enjoy
22:31:28Bagderpreglow: the sources are still there
22:31:28Moxonsome I could join?
22:31:34preglowBagder: oh
22:31:37LinusNMoxon: for testing?
22:31:38MoxonI already do that :)
22:31:42MoxonLinusN: yes.
22:31:47Bagderpreglow: I just removed them from SOURCES so they don't compile anymore
22:31:54LinusNno, there isn't
22:32:01preglowwe should remove the sources soon anywho
22:32:17Moxonhm, do you need some? like once a week checking a list of functions which should/might work?
22:32:36LinusNMoxon: that would surely be nice
22:33:06MoxonLinusN: I would appreciate that, too. Kind of quality measurement. Where can I apply, anyone to contact?
22:33:24LinusNMoxon: just go ahead
22:33:36Coldtoastthen come here
22:33:40LinusNthe Wiki is yours
22:33:53Moxonok, I will try to do that :) Expect regular posts in the forum ... or wiki (have to learn that).
22:34:23LinusNMoxon: personally, i prefer the mailing list, or irc
22:34:32 Quit tvelocity ("Leaving")
22:34:33Maxime`half an hour after, gcc hasn't finished compiling..
22:35:12MoxonLinusN: ic. I will subscribe to the list, too. it will be interessting anyways.
22:35:23Bagderthe cvs build machine builds 22 Rockbox targets in 10 minutes ;-)
22:35:24LinusNMoxon: gr8
22:35:51thegeekwps bug : if you try to change wps/update it wont work if the wps has images in it
22:35:54Maxime`Bagder: I do with what I have.. lol
22:35:59Maxime`only my server on linux.. :x
22:36:12Maxime`(and cygwin.. bleh)
22:38:08Maxime`hum, does anyone has a little idea on how we may handle the FM chip? :x
22:38:12 Quit Chamois (" HydraIRC -> <- IRC has never been so good")
22:38:25LinusNyes i do
22:38:52 Join Moos [0] (
22:38:56MoosHi all
22:39:03Maxime`hm, ahve to use asm tricks or else? or only C ? lol :x
22:39:17LinusNc will do fine
22:39:28LinusNno magic involved
22:39:37LinusNin fact, we have most of the code ready
22:40:52 Join Stryke` [0] (
22:40:57LinusNthe fm chip is the same as in the Archos Ondio
22:41:08Bagderdarn Linus broke my cvs series
22:41:20LinusNstart over again
22:41:44preglowmost of the codecs call the api both rb and ci now :)
22:41:49preglowi suppose that's supposed to be cleaned up?
22:42:01BagderI took some shortcuts
22:42:07preglowok, i need a break anyway
22:42:07amiconnLinusN: I did the mentioned fixes/changes to crt0.S, and got rid of all the TABs...
22:43:19 Join muesli- [0] (
22:43:54muesli-what happpened to the last rbx rls..only 540kb !?
22:44:12muesli-in comparison to 950kb the times before
22:44:26benjimimacodecs changed so size diff i believe
22:44:34Bagderif you want, we can add some random junk to fill it up again?
22:44:51muesli-am just wondering..thats all
22:44:52benjimimawoooo random junk :D
22:44:57Bagderbefore, we had the codecs twice
22:44:59Maxime`(ehe, better with a gcc 3.4 .. lol :x)
22:45:08Bagderand we had a few more plugins that no longer are included
22:45:16muesli-for 1 to 100000000 write "bla" ;
22:45:41Maxime`/home/maxou/m68k/lib/gcc/m68k-elf/3.4.4/../../../../m68k-elf/bin/ld: error: no memory region specified for loadable section `.entry'
22:45:43*Maxime` unlucky
22:48:17benjimimais it possible for rockboy to run at same time as music? or is it impossible due to hardware, becos none of the games like pong never used to run with music until recently so just wondered if its time, or hardware ;) thanks and great job btw
22:48:38amiconnbelgarath: It's a memory problem
22:48:41Maxime`[22:20] <LinusN> not rockboy [22:20] <LinusN> it needs the buffer mem
22:48:47amiconnrockboy uses the audio buffer
22:49:33Bagderbut only the archos version does it nicely
22:49:35amiconnBagder: I wondered why you chose such a weird memory size for the codecs
22:49:40Maxime`/home/maxou/m68k/lib/gcc/m68k-elf/3.4.4/../../../../m68k-elf/bin/ld: error: no memory region specified for loadable section `.entry'
22:49:44Maxime`what does this means?!
22:49:59amiconn0x3C000, I had simply chosen 0x40000
22:50:04Bagderamiconn: I did a binary search
22:50:16amiconn0x40000 would be 256 KB
22:50:54amiconnBagder: What do you mean, only the archos version does it nicely?
22:51:16Bagderamiconn: rockboy uses the audio buffer without asking for it the nice way
22:51:50amiconnIt should use plugin_get_audio_buffer(), which in turn should stop the musc
22:52:01amiconnOn archos my loader does this iirc
22:53:36 Quit Sususu (Read error: 54 (Connection reset by peer))
22:53:37preglowBagder: has deps for the codecs been done?
22:53:57amiconnBagder: mpc.codec probably won't load now. It's 252 KB
22:54:30Bagderbut it links! ;-)
22:54:45preglowamiconn: it loads here, but it sounds terrible
22:55:21preglowi was wondering why it sounded so bad...
22:55:27amiconnHaha, a52.codec is 52 KB
22:55:38preglowa52 is a nice codec
22:55:57amiconnOnly wavpack and wav are smaller
22:56:01preglowwavpack isn't large either
22:56:13preglowwe could probably fit all the code and data for wavpack in iram
22:56:47preglowbut yes
22:57:01preglowso there is good reason for mpc to sound bad? if so, i wont try to debug it :)
22:57:44amiconnBagder, LinusN: I think about moving the bootfile check to the end of the default_event_handler() usb handling
22:58:01amiconnThis way it would (1) always catch a changed bootfile
22:58:16amiconnNow it works only in the browser, and only if you're browsing the root
22:58:52amiconn...and (2) now it checks everytime it browses the root, that's unnecessary
22:58:58Bagdergood enough for me!
23:00:14Bagdernow, back to my upcoming revised language system suggestion...
23:00:29amiconn...unless it should be able to catch a new bootfile moved to the root by rockbox itself, but that won't be necessary imho
23:00:45amiconnIf you are in rockbox, it's easier to RoLo directly
23:03:18preglowanyone else have mpc codec that sounds like crap?
23:03:37amiconnI don't have a single .mpc file so I can't tell
23:04:23benjimimawhat even is .mpc?
23:05:19 Join rasher [0] ("> - if you *really* want to catch the digital flow
23:06:40preglowmy file is taken from rasher
23:07:33amiconnBagder: My idea might creates a problem (or rather, that problem is laready there with current code); probably academic:
23:07:40Maxime`I always wondered what's the utility of flac :x
23:07:54rasherThe audio thread debug screen doesn't take into account the font-height..
23:08:10amiconnIf you change the bootfile and rockbox catches it, but you insert usb again instead, and do that several times, it might overflow the stack
23:08:52Bagderacademic it is
23:08:59 Quit muesli- (Read error: 113 (No route to host))
23:10:43preglowignore it you should
23:14:21Bagderh100 sim crashes
23:14:57preglowso, what's status on mpc?
23:19:12preglowMAN, mp3 is struggling with the resampling code as it is
23:19:43 Join matsl [0] (
23:22:36preglowliba52 is perfectly content at 48mhz
23:22:38preglowit never boosts
23:22:44preglowsame for wavpack
23:22:53 Quit Stryke` (Read error: 60 (Operation timed out))
23:24:28HClwhat compression does wavpack give compared to ogg/mp3
23:24:35preglowwavpack is lossless
23:25:29preglowbut someone please test mpc so i can commit mpc.c knowing i haven't destroyed anything
23:25:38preglow'cuz it sounds like crap
23:25:39rasherit's broken here as well
23:25:54rasherAnyone mind if I commit an addition to the audio thread debug thing to take into account the line-height?
23:26:33rashersuch that you don't get the bars on top of your text
23:26:43rasherit assumes a line-height of 8px as it is
23:28:33preglowi sure as hell don't mind
23:28:47preglowbut so, why is mpc broken? is the codec file too large?
23:30:55rasherI blame Bagder.
23:31:27rasherwhat does flac use for tags? vorbiscomment? Or is that only for ogg-flac?
23:31:37Bagderyes, the file is too large
23:31:41Bagderor rather, the buffer too small
23:31:49preglowwell, that needs to be fixed
23:31:51Bagderand the codec loader doesn't check this properly
23:32:06 Quit rasher ("CGI:IRC 0.5.4 (2004/01/29)")
23:32:10preglowthere's nothing wrong in mpc.c that makes it too large, libmusepack is just a bloated inelegant lib
23:32:31Bagderwe should make it 0x40000
23:32:43preglowi wonder how they managed to make such a small nice codec result in such a big footprint
23:33:22preglowBagder: how/where do i change this?
23:33:32Bagderme fix
23:34:17amiconnRick: The simplest solution is to make the audio thread debug screen set the system default font
23:34:18preglowwe're certainly giving the build box a run for its money
23:34:34LinusNamiconn: agreed
23:35:38amiconn..same goes for archos btw
23:36:24amiconnpreglow: mp3 boost ratio is ~65% here now (with 192 kbps files)
23:37:03preglowit's _REALLY_ struggling here, with iriver.mp3
23:37:10preglowand another vbr file i had
23:37:33preglowit stays boosted the entire time
23:37:50preglowthe resampling accounts for much of this
23:38:17amiconnI should try mp2
23:39:03amiconn384 kbps mp2 playing...
23:39:06preglownow mpc locks my player up
23:41:09 Quit TCK- (Read error: 110 (Connection timed out))
23:42:53***Saving seen data "./dancer.seen"
23:43:05amiconnBoost ratio is ~80%
23:43:05 Quit belgarath ("CGI:IRC (EOF)")
23:43:13preglow /home/thomj/build/rockbox-devel/apps/codecs.h:76:1: warning: "CODEC_BUFFER_SIZE" redefined
23:43:14amiconn(with buffering inbetween
23:43:24preglowanyone else get that?
23:44:25 Quit Coldtoast ("Peace and Protection 4.22")
23:45:09preglowall codecs lock up my player now
23:45:11preglowwhat the hell is this
23:45:36linuxstbpreglow: Yes, I get that CODEC_BUFFER_SIZE warning.
23:46:10preglowanyone else get a constantly crashing rockbox too? :-)
23:46:51linuxstbHaven't got that far yet - just arrived back home...
23:48:14linuxstbpreglow: Yes, it crashed as soon as I tried to play some music (AC3)
23:49:23amiconnSlasheri: I get that premature stopping again...
23:51:26preglowi don't see how bagders latest commit can have caused this
23:51:30preglowunless it overwrites something
23:52:28preglowi bet it's connected to that warning
23:52:28amiconnI do...
23:52:37Moxonquick question: where do I find docs about rockbox specifix twiki variables (and variables lik %Y%) ?
23:52:49amiconnThe size must be changed in plugin.h and in the linker scripts as well
23:53:10preglowyeah,he forget to change
23:53:22 Join psycho_maniac420 [0] (
23:53:29linuxstbI think you can just remove the CODEC_BUFFER_SIZE line from plugin.h
23:53:40LinusNme too
23:53:52amiconnHe changed codec.h and
23:54:03psycho_maniac420hello. i have a question about the daily build and the main site of rockbox
23:54:12amiconnMaybe in plugin.h it can be removed
23:54:17LinusNpsycho_maniac420: shoot should know about it
23:54:23LinusNamiconn: yes
23:54:48psycho_maniac420i saw on the main page that tere was going to be a pictuer in the wps earlier today but then i go to look at the changelog again and it is gone
23:55:07preglowit's not gone
23:55:14preglowa lot has happened today, it's been pushed away
23:55:27psycho_maniac420is it on the main page?
23:55:28thegeekit's there
23:55:45thegeekas preglow said
23:55:51thegeekit has been pushed away
23:55:57thegeekthere have been a lot of changes today
23:55:59psycho_maniac420oph alrighty
23:56:00amiconnLinusN: Me fix, or?
23:56:03thegeekbut the code is still in rockbox
23:56:04LinusNi fix
23:56:06thegeekand it works great
23:56:11thegeekI have images in my wps now;)
23:57:12OnkelJonaswhats the tag for images in the wps?
23:57:48 Nick Febs_away is now known as Febs (
23:58:14OnkelJonasahh thx :)
23:58:37linuxstbYes, changing CODECSIZE in and removing the CODEC_BUFFER_SIZE from plugin.h seems to work.

Previous day | Next day