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).

Notice: Only Gecko based browsers prior to FF4 support the multipart/mixed "server push" method used by this log reader to auto-update. Since you do not appear to use such a browser, this page will simply show the current log, and not automatically update.

#rockbox log for 2005-06-17

00:00:07stripwax_who knows Tomasz Malesinski? (is that t0mas?)
00:00:16t0masit's not
00:00:20*t0mas is Tomas Salfischberger
00:00:24amiconnYes... rather, it's the cygwin h1x0 simulator bug
00:00:33amiconn(also happens for cygwin x11 builds)
00:00:34stripwax_t0mas :-) Tomasz reckons he's decrypted the iFP firmware :-D
00:00:44stripwax_amiconn - hm, weirdness.
00:00:58 Quit mborus ("CGI:IRC")
00:01:04preglowstripwax_: w00t
00:01:08stripwax_quite!
00:01:36stripwax_He'll be emailing me his source, it's very large (about 60kb, most of which is a lookup table..) - i'm sure it could be simplified..
00:01:48 Quit webguest82 ("CGI:IRC")
00:07:10amiconnThere you go - it's indeed a cygwin bug
00:08:01stripwax_btw - i'm currently building tools for iriver :-)
00:08:04amiconnas soon as this array gets larger than (4MB - something, with something possibly the module size), it tries to execute it as a function
00:08:47amiconnThis wasn't there with earlier versions :/
00:08:55stripwax_amiconn- wow. when does it try to execute it? (not that that's so important, really..)
00:09:11amiconnAt startup, even before calling main()
00:09:41stripwax_insane! nice find. is that known to the cygwin boys+girls?
00:09:54amiconnDunno... it's definitely new
00:10:56 Quit muesli- (Read error: 60 (Operation timed out))
00:13:14hubbelcygwin sucks, use vmware instead
00:13:38amiconnHow do I use the win32 sim then?
00:13:42stripwax_I had a go with colinux but couldn't get too far..
00:13:50amiconnBtw, I do use VMware for occasional linux builds
00:17:19preglowstripwax_: we got any docs on the cpu it uses anyway?
00:17:59 Quit hubbel ()
00:18:45stripwax_preglow - hrm, not sure.
00:18:54stripwax_make
00:19:01stripwax_:-p arf
00:21:08 Quit vmx (Read error: 110 (Connection timed out))
00:21:19linuxstbamiconn: The win32 sim runs perfectly under Linux/Wine
00:21:47amiconnMaybe... but linux is just too cumbersome
00:21:52 Join vmx [0] (oma@p549B7CB0.dip.t-dialin.net)
00:22:29linuxstbI'm not going to start that argument - I'm just saying you can compile and run the win32 sim using it!
00:22:37amiconnI know
00:23:26amiconnBut why should I run a different os (that is much more time consuming to set up and configure) and then run windows software using an emulation layer,
00:23:37amiconnwhen I can just run it on windows directly?
00:24:02stripwax_hmph. ok, so I can't even compile binutils-2.16 on cygwin. so, um, how do I build any tools again?
00:24:16linuxstbI don't know - you asked the question "How do I use the win32 sim then?"
00:25:28amiconnyeah, ok
00:25:40amiconnstripwax_: What's the problem?
00:25:56amiconnCompiling binutils 2.16 was no problem here
00:26:20stripwax_hmm.. "in function fibheap_delete_node error 'LONG_MIN' undeclared"
00:27:03amiconnHmm, maybe cygwin's upgrade to gcc 3.4.4 broke more than just the h1x0 sims
00:27:10stripwax_wait.. maybe I've gone insane.. two tics...
00:27:13amiconnI did that when I still had gcc 3.3.3
00:27:36 Quit Shagnar (Remote closed the connection)
00:28:43stripwax_amiconn, maybe.. i might give that a go
00:29:06amiconnI'll just downgrade to gcc 3.3.3 again, and see what happens to the sims
00:30:10amiconnAny opinions on the sleep timer/ idle poweroff? Should it always power off completely on iriver, even with the charger connected?
00:31:11stripwax_personally, i'd say it should stay on while charger is connected
00:31:24stripwax_no dice. i'm downgrading to 3.3.3. too
00:31:48amiconnI think the same for idle poweroff, however, for sleep timer it might be better to poweroff completely
00:32:20stripwax_makes sense too
00:34:04amiconnIt's just a new config variable... needing a proper name
00:34:15amiconnHAVE_POWEROFF_WHILE_CHARGING perhaps?
00:34:35amiconnirivers have that, archoses do not
00:36:21stripwax_mm, why not archos too?
00:36:37amiconnarchos can't poweroff while the charger is connected
00:36:44amiconn(forecd on by hardware)
00:37:14amiconn*forced
00:37:16stripwax_interesting! that config var name is pretty much spot on then
00:38:12amiconnInteresting detail - the only archos where this actually makes sense is the recorder v1, as it has software controlled charging
00:38:54stripwax_binutils build seems much happier on 3.3.3
00:39:00amiconnAll other archoses with charging (player, recorder v2, fm) have hardware controlled charging, and this behaviour is unneeded
00:39:37amiconnNo way for the sim though, must be some other change there
00:39:37stripwax_amiconn wow! (does that mean there's a hardware mod that could be made to prevent that behaviour?)
00:39:54amiconnI don't know, never checked
00:45:14 Join Meddler [0] (~56817823@labb.contactor.se)
00:45:49preglowstripwax_: if rockbox gets going on a flash based device, i might easily get one
00:47:02preglowoooh, arm core :>
00:47:13amiconnpreglow: It already does, so get an Ondio
00:47:26stripwax_preglow - :-) if solid state devices get to around 20GB maybe i will too :-)
00:47:27 Quit Meddler (Client Quit)
00:47:38preglowdon't care about that, i want a smaller placer as well
00:47:52preglowlugging around the great big h120 while trekking in the woods isn't too good an idea
00:48:12preglowi wonder why it's got a dsp core as well
00:48:18preglowthe arm should be able to decode most things
00:49:42 Quit Sucka ("a bird in the bush is worth two in your house")
00:57:33 Join Blair_42 [0] (cbc@c-24-13-3-178.hsd1.il.comcast.net)
00:58:14Blair_42I need a feature added. Who do I talk to?
00:59:03bill20r3heh
00:59:27Blair_42was that funny?
00:59:32stripwax_Blair_42 - what's the feature
01:00
01:00:11Blair_42I have an almost blind friend who wants to use the ondio to record meetings
01:00:42Blair_42you can get into the recording menu, but you can't tell if it is actually in record
01:01:12Blair_42what I propose is that you get a beep in your headphones when you go into record and a double beep when you stop
01:01:24bill20r3can you play and record at the same time?
01:01:35bill20r3or you could delay the start of recording I guess
01:01:38Blair_42I actually made a hardware device that added this feature to a radio station console for another friend
01:01:45amiconnIf the voice UI is on, there is a beep (or rather, burp) that acknowledges start of recording
01:02:02amiconnWe can't do more due to hardware limitations
01:02:21Blair_42I tried it today and didn't hear any beep (or blurp)
01:02:43 Quit Aison ("( www.nnscript.de :: NoNameScript 3.72 :: www.XLhost.de )")
01:02:46Blair_42I have voice menus enabled
01:02:56amiconnHmm, should be there. I'll check
01:03:31amiconnIt's not very loud, but it should be noticeable both with and without input signal
01:03:34 Quit Harpy (Read error: 145 (Connection timed out))
01:04:42Blair_42I don't hear anything but my button press
01:05:22Blair_42this is an Ondio with voice menu enabled file names not spoken recording from line in with no prebuffer
01:05:33Blair_42sorry, recording from MIC
01:05:43amiconnRecording source doesn't matter
01:06:46Blair_42just tried that
01:06:59amiconnIf you enter the recording screen and the start the recording, you should hear that blurp. Unpuasing a paused recording gives a shorter blurp
01:07:00Blair_42this is ver 2.4
01:07:09amiconnAh, that might be it
01:07:21amiconnGet a daily build, they're way better anyway
01:07:22Blair_42this has been added in more recent builds?
01:07:27amiconn2.4 is half a year old
01:07:33Blair_42ah
01:07:38amiconnlemme check
01:08:17amiconnHah, it was added *just* after 2.4 release
01:08:27stripwax_:-)
01:08:31Blair_42ah
01:08:41Blair_42ok, looking for daily build ...
01:08:43preglowdidn't the guys talk about a new release any moment now?
01:10:02 Quit vmx ("Have fun!")
01:10:19preglowstripwax_: custom dsp core doesn't exactly bode well for a speedy rockbox port
01:11:18stripwax_preglow - bah. 'how' custom?
01:11:30preglowi've sure never heard of it
01:11:33preglowit's not THAT custom
01:11:39preglowbut i bet it's not got a gcc port
01:11:43stripwax_:-)
01:11:48stripwax_well, :-( really
01:11:51preglowgeneral purpose code also seldom runs well on a dsp
01:11:52Blair_42ok, I have the zip. My 2.4 is in the boot rom. How do I upgrade?
01:12:01preglowso codecs will probably need to be slightly rewritten :/
01:12:31stripwax_w00t - first self-built iriver rockbox running on my player. now to get hacking into the code
01:13:50stripwax_amiconn/preglow - thx for all the help! yep, lassauge doesn't work, cygwin gcc 3.4.x doesn't work either, so looks like the only way is to use gcc 3.3.3 and build m68k 3.4.x yourself. but yes it works. THANKSQ!
01:15:00preglowi didn't do anything ;)
01:15:02 Quit mrlala (Read error: 110 (Connection timed out))
01:15:15preglowstripwax_: perhaps the wiki needs updating?
01:15:23stripwax_preglow - aye, i'll do it
01:15:29preglowgoodie
01:15:59preglowperhaps we should put up our own cross compiler binaries too
01:16:11amiconnBlair_42: Unzip it to the root dir of the Ondio, the disconnect, start the Ondio, RoLo into the new build ('play' ajbrec.ajz), the flash the new build by playing rockbox.ucl (found in /.rockbox) and following the instructions that the flash plugin tells you
01:16:16preglowlast time i tried to compile gcc in cygwin, i escaped screaming to linux
01:18:03amiconnBlair_42: The procedure is exactly the same as with the release versions
01:18:30Blair_42thanks. Just been so long since i loaded it I forgot exactly how the extra step was done
01:18:32*amiconn is testing the sleep timer fix
01:19:08Blair_42I wish I could get into writting code for this stuff. I do embedded system programming
01:19:28amiconnRockbox is about embedded systems programming
01:19:32preglowindeed
01:19:37Blair_42yup
01:19:53Blair_42I've done a lot of work with PICs
01:20:13preglowand we've got more resources than those, usually
01:20:33amiconn...even on archos
01:20:34preglowat least you wont be scared of low-level stuff ;)
01:22:49stripwax_woohoo! so my lcd driver scroll fns nearly work first time :-)
01:23:10Blair_42wow, I see lots of differences, but still dont hear any noises
01:23:19Blair_42being called for dinner
01:26:43 Quit hicks ("Too lazy to change my quit message")
01:27:30 Part rotator
01:28:05stripwax_amiconn - hm, gcc3.3.3 still makes broken rockboxui.exe (or did you realise that already?)
01:28:20amiconnyeps
01:28:39amiconnI guess it's either binutils or cygwin.dll
01:29:02amiconnChecking the first one now (with gcc 3.4.4 again)
01:30:31amiconnstripwax_: [00:39:38] <amiconn> No way for the sim though, must be some other change there
01:32:22stripwax_hrm
01:33:28amiconnHmm, using older binutils doesn't help either :/
01:35:48amiconnSame with older cygwin.dll
01:36:01amiconnGuess a bug report to the cygwin guys is due
01:38:29 Quit pbvas ("Leaving")
01:39:38***Saving seen data "./dancer.seen"
01:43:33preglowgotta go bed
01:43:34preglowlater
01:43:36 Quit preglow ("leaving")
01:43:55 Quit gromit` (Read error: 104 (Connection reset by peer))
01:46:54yngwigood night, im gonna sleep too
01:48:43Moosgood night
01:49:04 Quit yngwi ("Chatzilla 0.9.68a [Firefox 1.0.4/20050511]")
02:00
02:16:08 Part Moos
02:25:25 Quit cYmen (Read error: 110 (Connection timed out))
02:28:01Blair_42back from dinner. Never did get any bleep bloops or blorps on record
02:28:33Blair_42found some interesting firmware changes though (and several mistakes)
02:29:13Blair_42and even one or two "why the heck did they do thats?"
02:43:05HClhrm...
02:43:07 Quit thegeek (Read error: 131 (Connection reset by peer))
02:43:21 Join thegeek [0] (na@ti521110a080-4222.bb.online.no)
02:43:33*HCl woke up from something ruffling in his room..
02:49:53 Join mrlala [0] (~mistame@cpe-66-75-129-164.san.res.rr.com)
02:51:37 Quit Stryke` (Read error: 60 (Operation timed out))
02:55:16Blair_42there are no voice tags for the last 2 items on the recording settings menu
03:00
03:01:12 Quit stripwax_ (Read error: 110 (Connection timed out))
03:18:15 Quit Blair_42 ("Error Quitting: Permission Denied <SRAP>")
03:26:24 Quit mrlala (Read error: 145 (Connection timed out))
03:34:50 Join amiconn_ [0] (~jens@p54BD563A.dip.t-dialin.net)
03:39:42***Saving seen data "./dancer.seen"
03:53:20 Quit amiconn (Read error: 110 (Connection timed out))
03:53:20 Nick amiconn_ is now known as amiconn (~jens@p54BD563A.dip.t-dialin.net)
05:00
05:01:21 Join mrlala [0] (~mistame@cpe-66-75-129-164.san.res.rr.com)
05:14:06 Quit DMJC-L (Read error: 110 (Connection timed out))
05:27:20 Join soom [0] (saa@ool-44c67f68.dyn.optonline.net)
05:30:37HClugh.
05:30:43HClsickness drives me insane X.x.
05:31:24 Quit mrlala (Read error: 110 (Connection timed out))
05:39:43***Saving seen data "./dancer.seen"
06:00
06:57:32zeyeah being sick blows
06:58:01zeyou got a cold or something? zinc lozenges are supposed to cut the intensity and durration of colds roughly in half if you start taking them at the first signs
06:58:13zebut thats just for colds i guess
07:00
07:01:31 Join mrlala [0] (~mistame@cpe-66-75-129-164.san.res.rr.com)
07:09:34 Quit crwl (brown.freenode.net irc.freenode.net)
07:09:34NSplitbrown.freenode.net irc.freenode.net
07:11:12NHealbrown.freenode.net irc.freenode.net
07:11:12NJoincrwl [0] (~crawlie@dsl-83.148.225-157-dynip.ssp.fi)
07:15:21 Join lostlogic [0] (~lostlogic@node-4024215a.mdw.onnet.us.uu.net)
07:24:58 Join B4gder [0] (~dast@static-213-115-255-230.sme.bredbandsbolaget.se)
07:25:08B4gdermorning
07:27:21Bgermorning, B4gder:)
07:28:12 Quit mrlala (Read error: 110 (Connection timed out))
07:33:34amiconnmorning
07:39:45***Saving seen data "./dancer.seen"
07:56:03Bgeramiconn did you saw the chip i asked you yesterday ?
07:56:18Bgers/saw/see
07:56:40Bgers/asked you/asked you about ...
08:00
08:00:09 Join ehntoo [0] (~noclue2@24-177-147-34.dhcp.mrqt.mi.charter.com)
08:01:19 Quit einhirn ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
08:02:45amiconnBger: What chip do you mean? If it's the small unknown chip on H3x0, I have no idea
08:03:31Bgeryeah, that' what i mean
08:03:39Bger10x:)
08:03:52Bgeri supposed so, but it never hurts to ask...
08:04:22amiconnHmm, the rockbox->uda1380 volume mapping is non-linear.
08:06:40amiconnThere's a break in the curve. 0..208 work in 0.25 dB steps, 209..255 in 2 dB steps (uda values)
08:07:12amiconnI'll have a look, changing that should be easy (and
08:07:44amiconnperhaps it leads to a slightly better balance handling on archos player)
08:10:06amiconnThe audiobuffer() problem is not a pure cygwin problem after all. Something must be wrong with the H1x0 sims, as it doesn't happen with an archos recorder sim even when I increase the buffer to iriver size
08:11:29B4gderyou tried enabling logf?
08:11:39B4gderI mean, does the problem remain the same then
08:11:58amiconnDidn't try yet
08:12:33B4gderI suspect it'll change then
08:12:59amiconnbuilding...
08:23:18amiconnDoesn't change anything :( (The crashing address changes a bit, but still within the audiobuffer funtion)
08:26:34B4gderit still works fine on linux :-/
08:29:53*B4gder fights the evil demons of gdb configure scripts
08:38:28amiconnHmm, does a current win32 cross-compiled h1x0 sim work on real windows?
08:39:58B4gderhang on
08:40:26dwihnoIf it breaks, you get to keep the pieces? :)
08:42:23B4gderhttp://daniel.haxx.se/rockbox/rockbox.zip
08:43:47 Join Zoom2 [0] (~41081727@labb.contactor.se)
08:45:14dwihnonothing really happens
08:45:17 Quit DomZ ("CGI:IRC")
08:47:50amiconnB4gder: This has the same effect as a cygwin build. It crashes in audiobuffer()
08:48:27B4gderat least it makes the problem consistent
08:48:55amiconnTells a tiny bit more though - a source line within mainCRTStartup
08:48:57amiconnProgram received signal SIGSEGV, Segmentation fault.
08:48:57amiconn0x020dafc0 in audiobuffer ()
08:48:57amiconn(gdb) bt
08:48:57DBUGEnqueued KICK amiconn
08:48:57amiconn#0 0x020dafc0 in audiobuffer ()
08:48:57amiconn#1 0x004011de in mainCRTStartup ()
08:48:58***Alert Mode level 1
08:48:58amiconn at /home/ron/devel/debian/mingw32-runtime/mingw32-runtime-3.7/build_dir/src/mingw-runtime-3.7/cr
08:48:58 Join oxygen77 [0] (~c1c28427@labb.contactor.se)
08:49:00amiconnt1.c:225
08:49:02amiconn(gdb)
08:49:40 Quit oxygen77 (Client Quit)
08:50:30B4gderyes, it seems to use that instead of main
08:52:23B4gderbut the win sim has no main
08:52:31amiconnIf all else fails, I'll do a brute-force binary backward search which commit broke it
08:53:35linuxstbI don't want to distract anyone, but I've found an easy-to-fix win32 simulator problem - the UI-recorderv2.bmp file is not in CVS (but there is a file called UI-v2.bmp)
08:54:29B4gderamiconn: try adding a main function and see what happens
08:54:59B4gdernot that I can understand what possible broke this
08:55:03B4gderpossibly
08:58:54 Quit tvelocity ("Leaving")
08:58:59***Alert Mode OFF
09:00
09:03:55 Quit Zoom2 ("CGI:IRC (EOF)")
09:04:09B4gdercross-compile hell
09:04:25B4gderthere's where I am
09:05:23 Join einhirn [0] (Miranda@bsod.rz.tu-clausthal.de)
09:05:24 Quit einhirn (Read error: 131 (Connection reset by peer))
09:05:29 Join einhirn [0] (Miranda@bsod.rz.tu-clausthal.de)
09:19:47 Join tvelocity [0] (~tony@ipa0.2.tellas.gr)
09:22:54Bgeramiconn: offt: are the pictures of H3x0 you've got good enough
09:26:55 Join Harpy [0] (tbeliE5bCm@dsl-hkigw7wbb.dial.inet.fi)
09:30:40 Join bobTHC [0] (~bobthc@l03v-34-162.d1.club-internet.fr)
09:31:46bobTHChi folks !
09:33:53 Join Sucka [0] (~NNSCRIPT@host81-156-209-158.range81-156.btcentralplus.com)
09:38:13 Join Touns [0] (~plop@lns-p19-9-idf-82-65-82-203.adsl.proxad.net)
09:38:32 Part Touns
09:39:48***Saving seen data "./dancer.seen"
09:45:35 Join gromit` [0] (~gromit`@ras75-5-82-234-244-69.fbx.proxad.net)
09:46:14 Quit gromit` (Remote closed the connection)
09:46:18 Join gromit` [0] (~gromit`@ras75-5-82-234-244-69.fbx.proxad.net)
09:46:32 Quit gromit` (Remote closed the connection)
09:47:56 Join gromit` [0] (~gromit`@ras75-5-82-234-244-69.fbx.proxad.net)
09:51:45 Quit Bger (""In the other world, in paradise, the beauty of women surpassed even the beauty of Bulgarian women" Adaloloddin Mohammed Balhi")
09:55:20west-acrehey there
09:55:36B4gderhi
09:58:46west-acreits my birthday YAY
09:58:56west-acrebbl
10:00
10:07:05 Join Chamois [0] (HydraIRC@champigny-5-82-226-182-23.fbx.proxad.net)
10:07:16CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
10:07:16*B4gder added two red builds to the cvs build table
10:08:28linuxstbB4gder: That will get people excited.
10:08:32 Part tvelocity ("Leaving")
10:08:51B4gderperhaps. I just want the ifdefs sorted out
10:08:59B4gderand this helps doing that
10:10:52linuxstbShould the file uisimulator/win32/UI-v2.bmp be called UI-recorderv2.bmp ? The recorderv2 win32 sim doesn't build (it's not included in your build table)
10:12:06B4gderyes I believe it should
10:12:57linuxstbDo you want me to fix it? (cvs remove ; cvs add ?)
10:13:06B4gderplease do
10:13:13*amiconn checks
10:13:46B4gderyou think I should add such a build to the table?
10:14:08linuxstbWe have enough room now - there's no reason to exclude it (apart from build time).
10:15:32amiconnLooks like it was me who was silly, unless someone changed the ARCHOS value for v2 later on
10:16:12 Join Lynx_ [0] (~lynx@tina-10-4.genetik.uni-koeln.de)
10:16:59linuxstbOK, I've just committed it.
10:20:48amiconnApart from the .bmp name, recorderv2 is exactly identical to fmrecorder, so imho it makes no sense to build a separate sim
10:23:30 Join Philip_0729 [0] (~Philip_j_@user-7155.lns6-c10.dsl.pol.co.uk)
10:23:45 Join Bger [0] (~Bager@83.222.160.88)
10:28:13 Join yngwi [0] (~chatzilla@chello080109107064.1.15.vie.surfer.at)
10:39:33 Join hicks [0] (nobody@blackhat4life.org)
10:52:08 Join Aison [0] (~hans@zux166-181.adsl.green.ch)
10:59:55 Join cYmen [0] (~cymen@nat-ph3-wh.rz.uni-karlsruhe.de)
10:59:57west-acrehey, just wondering, i flashed to rockbox yesterday at about 5pm. how do i know when the daily build changes, and if so what the changes are?
11:00
11:00:19B4gderdaily builds are built every day
11:00:27west-acreand they change everyday?
11:00:30B4gderthere's a changelog on that page
11:00:31Bgeryep
11:00:33west-acrekk
11:00:44B4gderbut it is fairly techy
11:00:46Bgerwest-acre: the bleeding edge builds are the last
11:01:00west-acrekk
11:01:05Bgerand they are built every 20mins, iirc
11:01:13west-acreso there the mosts recent the bleeding edge, o 20mins wow
11:01:14west-acrekewl
11:01:21Bgeryes
11:01:36B4gderthey're actually built within one minute after a change
11:02:00B4gderand the build takes roughly 10 minutes
11:02:43west-acreok. and so i just extract rockbox.zip into the root, no need to change the .hex?
11:03:08B4gderthe hex is the bootloader, you won't need to update that very often
11:03:10Bger22 targets :)
11:03:19west-acreok. ;)
11:06:55west-acreso, do i need to upgrade the firmware again then, from the iriver firmware?
11:07:12west-acreafter unzipping to the root
11:07:16west-acre:S
11:07:38B4gderno
11:07:43west-acreOK
11:07:46B4gderunzip and you're done
11:07:51west-acreschweeet
11:08:06west-acreso, wot will be the next big step in rockbox for the h1XX?
11:08:08Bgerwest-acre: if you are with bootloader V2, just unzip
11:08:20west-acrejpegs?
11:08:31B4gderwhat people decide to do
11:08:40B4gdereveryone does what they think is fun
11:08:45west-acrei just patched the origonal firmware and then put that in the root with the bleegin edge build
11:08:53west-acrethat rite yer?
11:09:16west-acreahhh, i loose all my rockbox settings :S
11:09:21Bgerwest-acre: with firmware patcher ?
11:09:28ghode|afkyou dont need to patch the original firmware weverytime you update rockbox
11:09:30west-acreyup
11:09:32B4gderwest-acre: yes, the config struct has changed
11:09:37west-acreok
11:09:41B4gdersave your config as a config file
11:09:49west-acreo well, ill just keep the default for the thyme being
11:09:53B4gderuseful to reload in situations like this
11:10:00west-acreyeh good thinking
11:10:10west-acreill do that, and they can be made from the DAP yer?
11:12:59BgerB4gder one small notice ... (don't get me as petty) according to iriver themselves, it's iriver, not iRiver
11:13:13B4gderaha
11:13:42B4gdertheir logo is more like iRiVeR :-)
11:14:08Bgerheh:) maybe
11:14:16Bgerbut not the new logo
11:14:48B4gderso they use all lowercase always?
11:14:52Bgeryep
11:15:03Bgeryou can see it on www.iriver.com ...
11:15:16Bgersee the (c) notice on the bottom of the page
11:15:21linuxstbA google for "iriver" shows it as well - all the official sites use lower-case.
11:15:33Bgerall other letters are upper case
11:15:36B4gderremade the pics
11:15:40Bgerhehehe :)
11:16:18Bgerwow, you're FAST
11:16:19B4gderand yes, I'm happy I made the script that generates them :-)
11:17:29Bgertext->jpg ?
11:39:52***Saving seen data "./dancer.seen"
11:40:46 Quit Philip_0729 ()
11:52:36 Quit linuxstb ("Leaving")
11:58:18amiconnBagder: 2005-06-06 works, 2005-06-07 crashes
12:00
12:01:46B4gderand 2005-06-07 has only totally unrelated fixes
12:08:12HClmarkun: in case you're reading this, i'm hardly getting better... i'm most likely gonna go to my parents soon.. i don't know when i'll be healthy again... sorry...
12:13:31BgerHCl: r u sick :(
12:19:30amiconnB4gder: When I do cvs checkout/update using -D, what status do I get? That day 00:00?
12:20:37B4gderI believe so
12:20:49B4gderyou can specify time as well
12:21:05amiconnI'll try that next
12:21:12amiconn...narrowing down the time
12:21:20HClBger: yea, absolute killer headache..
12:22:21Bgerbad :(
12:22:46HClnot to mention feeling cold and overheated at the same time and other crap..
12:22:54Bgermaybe you don't sleep enough ?
12:23:05HClprobably
12:23:13HCli'm not sleeping well due to either being too hot or too cold
12:23:55Bgeri suggest you to go to a doctor...
12:24:04HClmy dad is a doctor
12:24:42HClfortunately..
12:24:51Bger:)
12:25:22yngwiget well soon
12:25:51Bgerasap :)
12:26:48 Quit yngwi ("Chatzilla 0.9.68a [Firefox 1.0.4/20050511]")
12:28:17 Join Moos [0] (moos012@m214.net81-66-158.noos.fr)
12:28:26MoosHi folks
12:29:12HClthanks..
12:29:54 Join webguest46 [0] (~5087b476@labb.contactor.se)
12:31:15Bgerdon't think i'm sayng this for other reasons except for completing runtime DB soon (and rockboy opts.) :D
12:31:19Bgerjust kidding ..
12:31:35 Part webguest46
12:35:39Bgerreally, get healthy asap
12:39:20 Join Jim__ [0] (Owner@spr2-cosh3-5-0-cust23.cosh.broadband.ntl.com)
12:39:47 Nick Jim__ is now known as Jleagle (Owner@spr2-cosh3-5-0-cust23.cosh.broadband.ntl.com)
12:51:23 Join leftright [0] (~5087b476@labb.contactor.se)
12:54:36 Join ripnetuk [0] (~george@82-70-100-230.dsl.in-addr.zen.co.uk)
12:55:03leftrightthere's still a pop when pause is selected
12:56:00leftrightusing todays Bleededge 11:55
12:56:41leftrightI think I'm going to trash my UE10pro's
12:56:55leftright:)
12:59:51Jleagleyou man that fuzzy noise?
13:00
13:00:03Jleaglemean*
13:00:42leftrightno, a pop when pause is selected and its fuzzy with play resume before normal vol
13:01:37Jleagleyer
13:01:43leftrightthe fuzzyness (more like scratching vinyl) is for a few millisec's but its ok,
13:02:42Jleaglei get really bad noises when its on charge and not playing
13:03:20*leftright goes to try
13:03:59Jleagleplay a song, then press stop
13:04:25leftrightyeo I've tried that twice and its normal with me
13:04:52leftrightnope quiet as a mouse in church
13:06:09leftrightare you using the standard power suplly unit
13:06:28Jleagleyup, standard everything, iriver 120
13:06:37leftright140 here
13:07:03Jleaglemaybe thats why, i have mine going through my computer speakers... ill try headphones
13:07:34leftrightso thats with line out then ?
13:08:45Jleaglei have the remote going streight to the blue hole in my sound card
13:09:02Jleagleooh, when i use headphones i dont get the noise :|
13:12:32 Quit edx (Read error: 110 (Connection timed out))
13:14:29Jleagledoesnt that mean when its on charge its somehow affecting the sound output?
13:15:59leftrightsomeone else needs to replicate that,
13:19:32Jleagleya, ill post it
13:23:18ripnetuk#b$/quit
13:23:19 Quit ripnetuk ("Leaving")
13:23:29leftrightthe 'upside down' screen mode is a neat feature
13:24:51 Join ashridah [0] (ashridah@220-253-122-117.VIC.netspace.net.au)
13:25:48Jleaglewas it requested or just added?
13:26:16leftrightscreen mode ?
13:26:24Jleagleya
13:26:39Jleagleseems sort of pointless to be able to have the screen upside down
13:26:45leftrightits always been there afaik
13:26:50leftrightI use it
13:26:56Jleaglereally? why?
13:27:21leftrightis its on my bekt, I dont have to unclip it to read the screen
13:27:29leftrightbekt=belt
13:27:54Jleaglehmm, never thoght about that, nice
13:30:33Jleaglewhat does the car mode do?
13:31:21leftrightit sync the unit to your cars ignition switch if you're using car dc power
13:32:44leftrightor something like that, I briefly glanced through it in the odf manual
13:33:13Jleaglesame here, but missed it
13:33:22Jleagleill go ehck it out
13:33:27Jleaglecheck*
13:38:19 Join LinusN [0] (~linus@labb.contactor.se)
13:39:13Bgerhello, Linus
13:39:14amiconnhi LinusN
13:39:27LinusNhey ho
13:39:52LinusNmy nokia phone has a really shaky gprs
13:39:54***Saving seen data "./dancer.seen"
13:40:31 Quit einhirn ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
13:41:10Bgerhow is the holiday
13:41:19leftrighthows the fishing LinusN
13:41:57LinusNleftright: haven't got around to fishing yet
13:42:36leftright*chuckles, thought it was a fishing trip
13:46:00amiconnLinusN: How difficult would it be to connect the logic analyser to the archos recorder again, and capturing the recording?
13:46:10amiconn(once you're back of course)
13:46:42amiconnI investigated a bit further; it seems the bit shifting does only happen when running rombox
13:47:07amiconnI'd like to compare rombox and ram execution timings...
13:49:02LinusNamiconn: it's not difficult, it just takes some time and effort
13:49:20 Quit Jleagle ("Chatzilla 0.9.68.5 [Firefox 1.0.4/20050511]")
13:49:41 Join preglow [0] (thomj@s183a.studby.ntnu.no)
13:49:45 Quit [zmaj] (Read error: 104 (Connection reset by peer))
13:50:22preglowLinusN: did anything ever come of the phillips correspondance?
13:51:01LinusNpreglow: no
13:54:50 Join bipak_ [0] (~bip@p50884D5B.dip.t-dialin.net)
13:58:56preglowoh my god, how i loathe the nokia theme
14:00
14:00:57LinusNi think i have solved the uda1380 init problem
14:01:03LinusNcheck my commit
14:03:22 Quit bipak (Read error: 60 (Operation timed out))
14:04:32preglowyou mean the problems some are having with getting sound?
14:05:20LinusNyes
14:05:40LinusNthe solved-by-booting-the-original problem
14:08:04preglowthat's a really strange one
14:08:16preglowi never had the problem, so can't test it
14:09:01 Part leftright
14:09:15preglowi'm sure pedro will test it as soon as he can
14:09:32LinusNgotta go cu later
14:09:35 Part LinusN
14:27:45 Join thegeek_ [0] (na@ti521110a080-2751.bb.online.no)
14:30:07amiconnB4gder: I narrowed it down as much as possible.
14:31:02amiconn2005-06-06 01:02 runs perfectly, 01:10 doesn't compile, and 01:30 crashes
14:31:35amiconncvs uses local time (cest) by default, so the real times are 2005-06-05 23:02, 23:10 and 23:30
14:31:41amiconnGMT
14:32:05amiconn==> It's Slasheri's playback patch that breaks the h1x0 sims on cygwin
14:34:41 Quit thegeek (Read error: 145 (Connection timed out))
14:36:44 Join linuxstb [0] (dave@dsl-212-23-31-215.zen.co.uk)
14:39:34Bgeramiconn anything unusual in this patch ?
14:40:04 Join thegeek [0] (na@ti521110a080-9143.bb.online.no)
14:43:27 Join ghostiger [0] (~ghostiger@658507467e999e71.session.tor)
14:43:30 Quit ehntoo (Remote closed the connection)
14:44:52Slasheriamiconn: Hmm, what is the problem with simulator?
14:47:23 Quit thegeek_ (Read error: 145 (Connection timed out))
14:48:01preglowit's not all that easy to tell, i think, heh
14:48:19preglowsomething in your playback code triggers a very weird bug
14:48:35SlasheriWhat will happen?
14:49:33amiconnAll cygwin H1x0 simulators (win32 _and_ x11) are crashing in function audiobuffer()
14:49:55amiconnHowever, audiobuffer is not a function, it's a char array...
14:50:08 Quit thegeek (Read error: 145 (Connection timed out))
14:50:34SlasheriHmm, weird.. I can't make linux version of simulator to crash
14:50:53preglowi suspect this is a compiler bug of some sort
14:50:58preglow'cause my sim too works perfectly
14:51:05linuxstbamiconn: I don't know if Bagder has found the same thing, but I tried cross-compiling the win32 sim from Linux (mingw 3.4.2) and the sim crashes immediately.
14:51:13amiconnyeps
14:51:23Slasherilinuxstb: can you backtrace the crash?
14:51:35 Join edx [0] (edx@p54A8CCCE.dip.t-dialin.net)
14:51:37amiconnI tried Bagder's build (cross compiled on linux) on my machine
14:51:49amiconnSlasheri: A backtrace doesn't tell much
14:51:52linuxstbSlasheri: I'm not sure how to debug win32 under Linux...
14:51:58Slasheriok.. :/
14:52:01amiconnIt happens on startup, before main()
14:52:14Slasheribefore main?
14:52:20amiconnSee today's log for gdb output
14:52:35Slasherithen that _should_ have nothing to do with playback code :)
14:53:08 Quit Aison ("( www.nnscript.de :: NoNameScript 3.72 :: www.XLhost.de )")
14:53:28amiconnNot directly, but it definitely causes the crash
14:54:09amiconnA cvs build from immediately before initial checkin of playback code is running fine
14:54:17SlasheriHmm, maybe there is some stack/buffer overflow then..
14:55:50SlasheriIf the code is not called anywhere before the main, there must be something other wrong before the main..
14:57:07 Join thegeek [0] (na@ti521110a080-1004.bb.online.no)
14:57:56 Join einhirn [0] (Miranda@carlsberg.heim2.tu-clausthal.de)
14:58:28 Quit einhirn (Client Quit)
14:58:49 Quit ghostiger (Read error: 145 (Connection timed out))
14:59:22amiconnSlasheri: I admit this is a weird bug
14:59:37amiconn...but it's got nothing to do with cygwin itself
15:00
15:00:57 Join ghostiger [0] (~ghostiger@d24dd899bb2bd2ad.session.tor)
15:02:34amiconnlinuxstb: Hmm, interesting. I thought it might have to do with the mingw version, but that's also not the case. Your 3.4.2 is an older version, and I have 3.7.? on cygwin
15:06:19amiconn3.7.1
15:10:32B4gderbuild table got all silly... :-)
15:12:02B4gderdue to my changed case
15:12:17 Quit lostlogic ("Going to the moon")
15:17:28preglowhahaa
15:17:31preglow"all silly" indeed
15:17:53B4gderespecially since the different case isn't visible in the table
15:17:58preglowoh, you've added h300 as well?
15:18:05B4gderyeps
15:18:13preglowgoodie
15:18:25*Bger screams "The most stupid scripting language on the earth is definitely M$'s batch lang"...
15:18:38preglowi wouldn't exactly call it a language...
15:19:14*Bger just realised that there *is* a way to extract different parts of filename with it ... through *FOR* command .....
15:19:34ashridahBger: good thing they're making a new one that might not end up in longhorn eh? :)
15:19:46Bgerno idea ....
15:19:59B4gderthe only sane way to do scripting in windows is to install a script language
15:19:59preglowit's not even comparable
15:20:09preglowvbscript......
15:20:23Bgeryep...
15:20:41preglowperl > *
15:20:50Bgerbut if VBScript is language, then Batch scripting is ... nothing ...
15:22:17preglowonly thing that might beat perl is perl 6
15:29:57west-acregot a question. the rockbox manual, is there one for the h1XX or is it the "THE ROCKBOX MANUAL"??
15:30:08preglowno, none yet
15:30:13BgerTHE ...
15:30:14Bgeronly one
15:30:19preglowmost of the existing manual should apply
15:30:28B4gderRockbox is a software made for many players
15:31:31west-acreyeh yeh i know, but. obv some features in the manual wont exist yet on the iriver
15:31:37B4gdertrue
15:31:46B4gderbut we haven't released anything for iriver
15:31:59B4gderyou're using an early dev version
15:31:59west-acrekk
15:32:01west-acrecheers
15:32:09B4gderand besides
15:32:19B4gderwe appreciate all help in adding docs
15:32:26west-acre;)
15:32:30west-acreok br0
15:32:59west-acreim printing off the manual...i may regret it!
15:38:43 Join bumi [0] (~blabla@62.219.122.195)
15:38:46bumihey everyone
15:38:58bumianything new with rockbox for iriver in the last 3 days?
15:39:29crwlbass/treble/balance settings, maybe
15:39:31bobTHCtake a look on website
15:39:43bumiwhere in the site?
15:39:50bumii have a version from 3 days ago
15:39:56***Saving seen data "./dancer.seen"
15:39:58bumibut it has tons of bugs
15:40:09bumiso i thoght maybe some of them have been resolved...
15:40:14linuxstbThinking about the documentation, is the plan to have seperate manuals for the different types of hardware, or stay with one universal manual?
15:41:11bobTHCmost versatile as possible is ez to maintain ...
15:41:22B4gderwe've managed with a single one so far
15:43:18preglowwhy are there so many holes in the build chart?
15:43:28 Join LinusN [0] (~linus@labb.contactor.se)
15:43:31B4gderI changed the name of the iriver builds
15:43:40B4gderso they appear as differnt ones
15:43:55preglowit just looks so... irregular
15:44:07B4gderwell, I added three builds earlier
15:44:32bumihttp://forums.rockbox.org/index.php?topic=828.0
15:44:40bobTHCclean up ure desk, the boss is here ;)
15:45:11amiconnlinuxstb: Afair Christi plans to implement a system for creating per-model manuals, probably using tex instead of openoffice and doing some preprocessing
15:45:12bumican anyone comment on that bug?
15:45:45B4gderhaving such a system would of course be very cool
15:45:53linuxstbamiconn: That could become a #ifdef hell.
15:47:01LinusNbumi: it works for me
15:47:13 Quit bumi (Read error: 131 (Connection reset by peer))
15:48:17preglowlinuxstb: nah, not very much, i believe
15:48:44preglowhaving a simple shell script that generates a root file that just \includes all the different parts needed for a specific model manual should work ok-ish
15:48:48 Join DMJC-L [0] (~DMJC-L@60-240-163-90.tpgi.com.au)
15:49:53linuxstbHas DocBook been discussed rejected?
15:50:08linuxstb^discussed and rejected?
15:50:24B4gderno
15:50:34preglowno, but i believe christi has used latex, and right now she seems to be the one writing the manual
15:50:39linuxstbNot discussed or not rejected?
15:50:49preglowunless anyone else is interested in helping her, i say leave her to it
15:51:01linuxstbpreglow: Yes, of course - we can't force someone to use a particular tool.
15:51:13preglowand besides, latex should be nice for this
15:51:22B4gdernot discussed and therefor not rejected
15:54:17 Join tucoz [0] (~81b1111b@labb.contactor.se)
15:54:34tucozHi, any old-timers here? Just to check wps-behaviour
15:54:50linuxstbI've used docbook a little. It's nice, but I think you need a Docbook expert on the team to customise the formatting scripts. If we don't have anyone to do that, then I would say stick to latex.
15:55:19tucozIf I have 10 songs in a directory, and 7 of them are songs (the other 3 text-files, or somthing else)
15:55:21B4gderI remember the old days... :-)
15:55:34tucozthe wps screen shows like track 7/10
15:56:01tucozno, i have 10 files, not songs
15:56:28preglowlinuxstb: formating scripts? ouch, then i definitely say let's use latex, unless we have any serious typographers around here ;)
15:56:41tucozIs that how it's supposed to be? Shouldn't the wps only count songs in the directory as part of the playlist?
15:57:58linuxstbpreglow: I'm probably making it sound worse than it is, it's just that if you want anything other than the standard output (even things like a custom front-page), then it's a steep learning curve to either change the DSSSL (if we use SGML Docbook) or XSLT (if we use XML).
15:58:10LinusNtucoz: works fine for me?
15:58:17 Join Philip_0729 [0] (~Philip_j_@user-7155.lns6-c10.dsl.pol.co.uk)
16:00
16:00:19tucozLinusN: Ok, I can try something. I had show all files on when I noticed this. Changed to supported, and still the same. Still, it was no reboot in between
16:00:26linuxstbtucoz: What exactly are the other files? What filename extensions?
16:00:37tucozletme check
16:01:24linuxstbe.g. I have 17 FLAC files, and 2 text files in a directory. Pressing play on the first track shows 1/17 in the WPS. The two text files appear last alphabetically.
16:01:40preglowanywho, i'm entirely for latex, but that's of course a consequence of me being very familiar with it ;)
16:02:19 Quit DMJC (Read error: 110 (Connection timed out))
16:02:19tucoz1 m3u, 1 nfo, sfv
16:02:25tucoz1 sfv
16:02:42tucozHmm, forgive me
16:03:19linuxstbI spent four years at University writing Latex, so I'm not against it either. It would be nice to have it in CVS as well.
16:03:30tucozI'm just plain stupid. Had an exam today so my brain is not with me. I was certain that the record had only 7 song.
16:03:31tucozs
16:03:43tucozWhen it in fact had 10.
16:04:14ChamoisLinusN : your UDA driver modification seems to work
16:04:38tucozLinusN: That uda commit, do you think that will take care of the occasional buzz on startup as well?
16:05:59tucozIf I've had my player off for a few hours, and sometimes, very seldom, it's start buzzing until the uda is initialized.
16:08:23tucozIt's hard to check what is causing this to happen, as it is _very_ seldom. Like two times a week perhaps
16:14:18 Join RotAtoR [0] (~e@dhcp54-47.calvin.edu)
16:18:21west-acrehey, just wondering, do iriver know about rockbox?
16:19:37preglowprobably
16:19:43bobTHCsurely
16:19:46bobTHC:)
16:20:28HClwhy is iriver h3x0 and iriver h1x0 listed twice... on the daily builds..
16:21:24tucozit's not. Your'e hallucinating ;)
16:21:28HClshould h300 even be listed cause we don't have a bootloader yet.
16:21:42HClnope, its listed twice...
16:21:47tucozJust kidding
16:22:07*HCl has too much of a headache for jokes at the moment ;/
16:22:10tucozI read that you're not feeling too well.
16:22:21tucozHope you get better :)
16:22:26HClthanks
16:22:30preglowHCl: don't you read mailing list?
16:22:33HClgonna go to my parents place in a while..
16:22:47HCli do, but i haven't seen anything about that topic on it
16:23:00preglowbagder added it to show how badly it crashes because of bad defines
16:23:12SlasheriHmm, yes it's listed twice and the whole page is quite confusing now..
16:23:13preglowso it'll hopefully be fixed
16:23:27SlasheriHCl: You will have the event thing today if i have time to code it :)
16:23:44HClnice
16:23:53HClpreglow: um. right...
16:24:06HClpersonally, i'd say it would create way too much confusion for h3x0 users
16:24:23HClcause it suddenly seems as if we have an h3x0 build
16:24:26preglowwell
16:24:29preglowno one has come asking yet
16:24:48HClright o.O
16:24:53preglowit has to be done sooner or later anyway
16:25:00HCli'll just refer them to bagder when they do then..
16:25:14preglowwhat were we supposed to do, then? add the build when the port is finished?
16:25:27HClwhen the bootloader works?
16:25:36preglowyou're forgetting the lcd driver
16:25:42HCland that.
16:26:03HCli'm just saying that to me it looks misleading, but okay, nevermind
16:26:08west-acrehey, ATM in rockBLOX can you rotate the pieceS?
16:26:09B4gderthat table is for development
16:26:15B4gdernot for people to download
16:26:54tucozBut like someone said. Rockbox should be able to work with the remote as display, before lcd is enabled
16:27:02tucozon h3x0 that is
16:27:08B4gderyes
16:27:18west-acrehey, just wondering, do iriver know about rockbox???
16:27:26west-acreooops SORRY
16:27:31west-acrehey, ATM in rockBLOX can you rotate the pieceS?
16:27:33west-acre????????
16:27:44HClcheck your ? key
16:27:46HCli think its stuck.
16:27:56west-acresorry
16:28:06west-acreno wait, it really is...
16:28:21west-acresorry hydrochloric acid
16:28:34tucozIt's working for me. The rotating
16:28:42Bgerit's not necessary to be "acid" :)
16:29:01Bgerhe can be gas
16:29:38west-acrewot key rotates ?
16:29:41tucozBut it's a pain to play right now
16:29:45tucozjoystick
16:29:46west-acre:-o
16:29:49west-acrestill acidic
16:30:21west-acrewot direction rotates? or pressing into the DAP?
16:30:23preglowonly if it's dissolved!
16:30:45tucozthat would be right if you hold your player in the 'normal' way
16:30:49Philip_0729only acidic when gives out H+ to water.... making H30 ;)
16:30:55tucozjoystick right
16:30:58west-acrekk
16:31:03west-acreahh yes tnx br)
16:31:06west-acreBR()
16:31:07west-acre;)
16:31:13west-acrethats much bettter ;)
16:31:18west-acreotherwise kinda crap!
16:31:52Philip_0729??
16:32:10tucozI tried a patch a while ago with the screen not tilted. Much better. Wonder why it never reached cvs
16:32:27LinusNit's on its way
16:32:38tucozOk, goodie
16:32:52west-acrethe game would be a bit crap if you couldn't rotate, but you can so it's gr8!!!
16:32:52Bgerbye, guys, have a nice weekend
16:32:55west-acrecya
16:32:59west-acretnx for ur help
16:33:13tucozThat scrolling patch also looked nice. With the key-repeat.
16:34:46west-acrehey, i think there's a bug in SNAKE. the new apple doesn't appear for around a second after you ate the last...
16:35:15tucoz..or was that just a suggestion. Anyway I read about it some time ago and it seemed like a good idea.
16:35:23west-acreok
16:35:36west-acrei dunno i spose im comparing it to snake 2 ;)
16:39:59 Part LinusN
16:40:13 Part tucoz
16:47:34 Part preglow
16:48:21west-acrehey, is the stopwatch accurate, has it been tested?
16:53:12bobTHCtest it and u tell us
16:55:18west-acrewill doo
16:55:24west-acre:0
17:00
17:02:55 Quit DMJC-L (Success)
17:06:17 Join dionoea [0] (~dionoea@muscipula152.via.ecp.fr)
17:06:31dionoeahello
17:06:37B4gderhowdy
17:06:43dionoeafile permissions seem all fuckedup in the CVS
17:06:56 Join DMJC-L [0] (~DMJC-L@60-240-163-90.tpgi.com.au)
17:06:57dionoeafind . -iname "*.h" -perm +111|xargs chmod a-x
17:07:01dionoeafind . -iname "*.c" -perm +111|xargs chmod a-x
17:07:18dionoeai can't submit a patch since permission changes don't show in a diff
17:07:30dionoea(some doc files are also executable)
17:08:34B4gderthey're also not possible to change with cvs
17:08:48ashridahanyone tested the firmware in its most recent incarnation?
17:09:13Chamoisyes
17:09:17dionoeaarg ... cvs suxxs
17:10:59dionoeayou should move to svn ...
17:11:20dionoea-you +sf
17:11:23HCldon't fix it if its not broken
17:11:31B4gderyes, we've been talking about that for a long time ;-)
17:11:55B4gderwe have a fair amount of scripts and stuff using cvs
17:12:02B4gderso it isn't a small task to move
17:12:19B4gderdionoea: we don't use sf for cvs
17:12:26dionoeamigration is pretty easy ... just change all the cvs up to svn up
17:12:32B4gderyes
17:12:36dionoeaand remove the extra options
17:12:37B4gderfor that simple case indeed
17:12:54B4gderI know how svn works
17:14:37*B4gder has commit-acccess in the Subversion project
17:14:46dionoeanice :)
17:15:26dionoeawe've been using it for about a year now and its getting better
17:15:54dionoeathe old database format was horrible ... but now its ok (with the one file per commit db)
17:16:22dionoeaare there any cool changes planned ?
17:17:13B4gderno, we only do boring things :-)
17:17:28dionoeahehe
17:18:14 Quit Philip_0729 ()
17:21:57dionoeabtw, i'd like to thank the whole rockbox dev team for this great project
17:29:01 Join djek [0] (~djek@djek.xs4all.nl)
17:34:03 Quit B4gder ("Lämnar")
17:39:10 Quit linuxstb ("Leaving")
17:40:00***Saving seen data "./dancer.seen"
17:46:39 Join niobos [0] (~niels@33.196-200-80.adsl.skynet.be)
17:51:10west-acrehi, is it possible to take screenshots of what is displayed on the iriver? Like in the manual?
17:51:31t0masyes
17:51:50t0masnormally you take screenshots of the simulator... but you can do it on iriver too
17:51:51t0masiirc
17:52:07west-acreo rite, can i get the simulator?
17:52:17west-acrehow is it done on the iriver?
17:52:20ashridahwest-acre: yeah, there's a screenshot option in menu->info->debug->screendump
17:52:25ashridahturn it on, then plug in the usb cable
17:52:38ashridahit won't activate the usb-storage stuff, it'll write a dump of the current screen contents to disk
17:52:52ashridahthen go back to that option, and disable it when you're done
17:53:05west-acrerite. it says stay out of the debug menu :S
17:53:08west-acreis this wise?
17:53:28ashridahyou won't break things by turning on the screendump feature
17:53:52west-acrekk
17:54:17west-acreitll write to the root of c:?
17:54:35ashridahit'll write to the root as a .bmp file
17:54:36ashridahyeah
17:54:51west-acrewot, every time i change screen or wot?
17:55:06ashridahno, every time you plug in the usb cable (provided the other endo f the usb cable is attached to a pc)
17:55:13west-acreok
17:55:43*niobos is learning dvorak, so be patient
17:56:18niobosI'm going on vacation in 3 weeks, for a month
17:56:42west-acrerite. ive enabled screen dump, then displayed on the iriver wot i want to be a bmp. then plugged the usb in. is it sposed to then save a .bmp to c://???coz it hasnt :S
17:56:52niobosis it wise to start playing with rockbox now?
17:57:00west-acreyes. yes it is
17:57:09west-acreashridah
17:57:10niobosor might I break my iRiver
17:57:15west-acreno u wont
17:57:21west-acreand it's iriver not iRiver
17:57:27ashridahwest-acre: it will only show up when the root of the drive is rescanned
17:57:29west-acreits relatively stable now
17:57:45ashridahniobos: there's still a minor amoutn of risk involved with installing the firmware
17:57:48ashridahbut it plays music
17:57:52ashridahwhich is a major step forward
17:58:32niobosI'd like to have the playlist-creation, so
17:58:49niobosashridah: minor = ?
18:00
18:00:39 Join Audiophil [0] (~c1d85c22@labb.contactor.se)
18:00:56west-acreashridah i gottit now ;)
18:01:22west-acrethat is an immensely cool feature!
18:01:50bobTHCniobos > take a look@ http://www.rockbox.org/twiki/bin/view/Main/IriverPortForNoobs
18:02:00Slasheriwow
18:02:01ashridahniobos: playlists have always worked.
18:02:13ashridahniobos: so yes, it supports playing music inside rockbox from playlists
18:02:15SlasheriI just tested that dimming backlight on iriver is possible with simple pwm pulse
18:02:19Slasherivery nice effect
18:03:09 Join linuxstb [0] (dave@dsl-212-23-31-215.zen.co.uk)
18:03:19bill20r3neat
18:03:29bill20r3does it save any measurable amount of power?
18:03:42bill20r3can you make it pulse w/ the music?
18:03:51Slasherii don't know if it's possible to use continuous dimming (that requires cpu power)
18:03:59bill20r3ahh.
18:04:02Slasheribut nice effect could be added when backlight goes on/off
18:04:05bill20r3can you dim the one on the remote?
18:04:12Slasherii think so
18:04:12bill20r3oh yeah, like the 4g ipods do
18:04:29Slasherii will test that more soon
18:05:20Audiophilwhat's the current work tasks for the RB team?
18:06:24 Join leftright [0] (~5087b476@labb.contactor.se)
18:07:10linuxstbAudiophil: That's a good question - I've started a page here http://www.rockbox.org/twiki/bin/view/Main/IriverStatus where we can add the tasks that we think need to be done before Rockbox is ready for the first iriver release.
18:07:56linuxstbSo if any devs can think of things, please add them there.
18:08:49leftrightSlasheri: did you see my post earlier with regards to the pops when pause is selected
18:09:57 Join Kat [0] (~katerina@c-67-176-83-251.hsd1.co.comcast.net)
18:10:47niobosis there any reason NOT to use the daily build, but the latest release?
18:12:00thegeekyes
18:12:02thegeekit's the newest
18:12:13thegeekand the daily builds are generally not any better than the latest
18:13:38niobosare they worse?
18:14:47Slasherileftright: Hmm, the problem that there will be a short loud piece of the current song after pause has completely faded out?
18:16:00leftrightno, there's a pop when pause is pressed
18:16:26SlasheriHmm.. are you using the headphone out connector?
18:16:26leftrightthe music fades nicely then, pop
18:16:31leftrightyes
18:16:50SlasheriIs that pop just some random buzz or real audio?
18:17:12leftrightdefinate short sound, not music
18:17:32Slasheriok, interesting..
18:17:55Slasherii know there is currently a bug that will cause a short piece of music to be played after fade out is complete
18:18:16leftrightI dont hear that, just a pop
18:18:22west-acreyeh sanem,
18:18:24west-acresam
18:18:25west-acree
18:18:27west-acrei hear a p0p
18:18:37bobTHC[ 18:13:27 ] [ @vessla ] Cypheria on efnet?
18:18:37bobTHC[ 18:13:55 ] [ Cypheria ] efnet IRC network
18:18:37bobTHC[ 18:14:05 ] [ Cypheria ] where we are right now
18:18:37DBUGEnqueued KICK bobTHC
18:18:37bobTHC[ 18:15:03 ] Joins : lyndz [ ccf46e07@c
18:18:59bobTHCsorry, keyboard erro
18:19:17leftrightthe resume play from pause is a bit messy too
18:20:09west-acrehey, wot else is worth using in debug ?
18:20:15west-acrescreendump is kewl
18:20:21SlasheriAh, yes. I hear also a very weak pop aftere fade out is complete
18:20:33leftrightthats it
18:20:48west-acreis the keep out because its unstable or just risky because things there cud do damage?
18:21:20Slasheriwest-acre: the debug menu could at least crash your player
18:21:32west-acrerite.
18:21:37Slasherii don't know if it can do any serious to your player
18:21:39west-acrelike battery test?
18:21:44west-acrepartition?
18:21:49west-acredisk info?
18:21:56west-acreaudio tests?
18:21:56leftrightwith In Ear Monitors, the pop is quite loud at medium volume
18:21:59Slasherithey all should work. But audio tests might crash
18:22:07west-acreok
18:23:37leftrightmp3 playback sound really good :)
18:25:17 Join BTKDaImMaikata [0] (~Bager@83.222.160.88)
18:25:54west-acreyeh. the runtime info, it doesn't restart when the player restarts? or is that how long ive been runnin grockbox?
18:26:00west-acrerockbox*
18:34:27 Join Stryke` [0] (~Chairman8@cpe-24-168-110-99.si.res.rr.com)
18:34:52 Quit Stryke` (Remote closed the connection)
18:37:49 Quit bobTHC ("Smoke Weed Every Day !")
18:42:44niobosis it normal for the shuffel mode to only shuffel through the current dir?
18:44:24 Join mrlala [0] (~mistame@cpe-66-75-129-164.san.res.rr.com)
18:45:38ashridahniobos: yes, since starting playback in the current directory only makes a playlist out of the current directory
18:46:08ashridahplayback doesn't operate the same way as the iriver's default firmware, that always treat the entire disk as a giant braindead playlist
18:46:12niobosis it possible to alter that behaviour?
18:46:13ashridahrockbox doesn't do that
18:46:34ashridahniobos: you can make a playlist of the entire disk, but it won't stay up to date if you add more music. and no.
18:46:49ashridahyou can add other stuff to the current playlist, and save the playlist
18:46:56 Join _aLF [0] (Alexandre@mut38-2-82-67-66-128.fbx.proxad.net)
18:47:36niobosthat giant playlist is an idea, since I always update with a script
18:48:02 Quit Bger (Read error: 110 (Connection timed out))
18:48:19ashridahwell, the playlist is created with rockbox itself, i don't know if there's a third party tool to do it.
18:49:00 Quit Chamois (" Try HydraIRC -> http://www.hydrairc.com <-")
18:50:31SlasheriHmm, now the backlight works quite well. The dimming time is very short because kernel is not pre-emptive and yields are too long so it cannot block cpu for long periods
18:51:28ashridahdimming? as in having the backlight fade out?
18:51:33Slasheriyes
18:51:39Slasherifade in and out
18:52:04Slasheribut that requires still more testing and i should ask somebody before committing that
18:52:04 Join Yokalosh [0] (~3efc0007@labb.contactor.se)
18:52:12Audiophilslasheri, when will this be available in the cvs?
18:52:20Moosashridah: (the rockboxs playlists can converted with the plugin iverify if you want use it with rb)
18:52:29SlasheriAudiophil: maybe never if it was a bad idea :)
18:52:31YokaloshHELP! ARGHHH
18:52:40Yokaloshi think it died!
18:52:45SlasheriAudiophil: But perhaps some menu option for that could be added
18:52:50ashridahMoos: that converts it for use in iriver's firmware. i fail to see how that solves anything i mentioned
18:53:16ashridahYokalosh: what died?
18:53:21Audiophilslasheri, my wish exactly:)
18:54:14Yokalosh:(
18:54:54mrlalai wouldn't mind seeing the dimming feature either ;)
18:55:35ashridahYokalosh: what died?
18:55:51YokaloshGuys what do you do when all you get when you try to open your archos jbr 10 and all u get is a hardware I/O error?
18:56:19Yokaloshbrb
18:56:25Yokaloshback
18:56:37 Quit Yokalosh (Client Quit)
18:57:26 Join Yokalosh [0] (~3efc0007@labb.contactor.se)
18:57:38YokaloshAnyway, like i was saying before my internet cut out
18:58:24 Quit Yokalosh (Client Quit)
18:59:25nioboshmm, it's probabely me, but i can't find how to create/edit/save playlists
19:00
19:00:42 Join preglow [0] (thomj@s183a.studby.ntnu.no)
19:00:47preglowSlasheri: hey, that fade thing sounds cool
19:01:08preglowSlasheri: is the pwm high frequency enough that you can't see it flicker?
19:01:27Slasheripreglow: nice, the only problem is that it will totally block the cpu while it fades the backlight (but short periods won't affect playback)
19:01:35Slasheriyes, there is no flicker at all
19:02:30preglowyeah, i kinda figured you couldn't yield without having it noticable
19:02:54Slasheriyep.. it has annoying flickering if i yield..
19:03:08preglowhow long is the fade, then?
19:03:19thegeekimho I want all fades short
19:03:20Slasheriless than a second
19:03:30preglowhmm
19:03:42Slasheribut it is still a nice effect
19:03:49 Join Yokalosh [0] (~3efc0007@labb.contactor.se)
19:03:53preglowyeah, i know
19:04:06 Quit Yokalosh (Client Quit)
19:04:15preglowyou sure the coldfire hasn't got pwm functionality on that gpio port? :>
19:04:27Slasherihmm, i haven't checked that :)
19:04:37Slasherithat would be really cool if it has
19:04:43preglowyeah, no need to block then
19:04:48preglowi think it's got some pwm outputs
19:05:00Slasherioh
19:05:14Slasheriwe should check that :)
19:05:55 Join Yokalosh [0] (~Yokalosh@cpc1-cbly2-4-0-cust103.glfd.cable.ntl.com)
19:06:01YokaloshHey, sorry bout that
19:06:01Slasheriif it's possible to put hardware pwm to the display then it could be continuosly be dimmed
19:06:06Yokaloshmy internet kicked me off
19:06:15Yokaloshyeh as i was saying before i went
19:06:40Yokaloshwhen i try to view my archos jbr 10 on my pc all i get it a hardware I/O error
19:06:43Yokaloshanyone know why?
19:07:23bill20r3is the drive good?
19:07:27bill20r3does it spin up?
19:07:36Yokaloshyes, i can use it as an mp3 player
19:08:00Yokaloshjust once i plug it into my pc i can't view it, it says it is there and my pc knows it is ther and can identify the drive
19:08:12Yokaloshbut if i try to view its files i get an I/O
19:08:53preglowSlasheri: doesn't look good, but you could use a timer
19:09:02bill20r3odd.
19:09:14Slasheripreglow: ah, timer could be nice too :)
19:10:06Slasheripreglow: how many timers there are in coldfire? Can i have own timer for that baclight fading?
19:10:39Yokaloshwht do you think i should do?
19:10:53SlasheriHmm, it seems to have 2 timers :)
19:10:56Yokaloshif i right click on it 9 out of 10 times my computer crashes so that is useless
19:11:01 Quit hicks (Read error: 145 (Connection timed out))
19:11:10SlasheriAnd second one seems to be not in use :)
19:11:22preglowthere are two, yeah, and rockbox uses one
19:11:27Slasherigreat :)
19:11:44 Join hicks [0] (~hicks@26266b3b40418031.session.tor)
19:11:53preglowif you can implement pwm with a high enough frequency using that, it just might be commitable
19:12:12preglowhogging the cpu during an entire fade is not very nice
19:12:14Slasheripreglow: yes, i will try it
19:12:56preglowexcellent, i love fading backlights ;)
19:13:29Slasheriis it ok if i initialize the timer in backlight.c?
19:13:32Slasheri:D
19:14:36preglowgo one and try
19:14:43Slasheriok =)
19:14:44preglowi have no idea, but just hack it in to see if it even works
19:15:51preglowthere's even a toggle output pin mode that's perfect for this
19:16:05preglowall you need to do is set up the correct timer periods
19:16:08Slasheriah, good
19:16:29Slasherithat sounds really nice, it could even work
19:16:42preglowit SHOULD work, can't see a reason why it shouldn't
19:16:47Slasheri:)
19:16:59preglowbrb, shop
19:20:46west-acre<14preglow> brb, shop
19:26:57 Join sockerteze [0] (~sockertez@pool-162-83-112-105.norf.east.verizon.net)
19:27:05 Quit sockerteze (Client Quit)
19:28:16Slasheripreglow: Hmm, do you know where i can get coldfire cpu datasheet?
19:29:31preglowSlasheri: well, yeah, the rockbox datasheet page has it
19:29:44preglowhttp://www.rockbox.org/twiki/pub/Main/DataSheets/MCF5249UM.pdf
19:29:46Slasherithe link is dead :/
19:29:48preglowthat's specific to our model
19:29:49preglowoh
19:29:52preglowthen gimme a sec
19:29:56Slasherioh, that works
19:30:02Slasheribut the wiki link does not
19:30:37preglowoh
19:30:52Slasherihttp://www.rockbox.org/twiki/bin/view/Main/ColdFire
19:30:57Slasheriat least the link on that page
19:34:31preglowhmm, i wonder _which_ output pin the timer toggles, heh
19:34:35preglowprobably best to do it yourself
19:34:39Slasheri:D
19:37:40SlasheriHmm, it can toggle some TOUT pin. I hope that pin can be selected :)
19:40:03***Saving seen data "./dancer.seen"
19:42:51amiconnRockbox on H1x0 has a "negative gap" for mp3 :(
19:44:09amiconnThat is, it even happens without lame tags (files encoded with lame −−nogap and playing perfectly gapless on archos)
19:45:57preglowamiconn: i knowe
19:46:00preglow-e
19:46:23preglowamiconn: that's my doing, probably, it seems like gapless encoders subtract a encoder delay from the length of the first frame
19:46:43preglowgapless _decoders_
19:46:56amiconnOf course the ear isn't very precise, but it seems the time varies
19:47:36preglowmight be, the mp3 decoder doesn't get it accurately right now
19:47:50amiconnI wonder why...
19:47:50preglowthat is, the time SHOULD be accurate, but for some reason, i can't squeeze all the data i want out of libmad
19:48:08amiconnWithout lame tags, the track change is no special case for the mp3 decoder
19:48:30preglowi calculate the exact duration of an mp3 in samples accurately according to foobar2000, which is widely regarded to do things correctly
19:48:38preglowbut i don't know why it doesn't work perfectly right now
19:49:08preglowbut yes, even for files with no lame info, i currently remove the first 500 odd samples from the start of the first frame in each mp3
19:49:23preglowi have no idea why this is done, but it seems to be the correct thing to do
19:50:12amiconnIt's certainly not...
19:50:27pregloware you sure that's the bug, though?
19:50:30preglowtry removing it and see
19:50:53amiconnOn archos, we simply feed the mas with all data after stripping id3 and xing/vbri tags
19:51:14amiconnThe mas won't notice a track change at all unless mpeg version, layer or sample frequency changes
19:51:36preglowi know
19:51:52preglowstart_skip is assigned only twice, the first assigment is when it's got lame data
19:51:56preglowthe second is without
19:51:59amiconn...and it's perfectly gapless this way when encoded properly
19:52:01preglowtry setting it to zero in the second case
19:52:07preglowand see if that fixes it
19:52:41preglowyeah, i agree it sounds weird, but foobar does this, and somehow manages to play gaplessly
19:53:11preglowthere isn't exactly an abundance of code that does this around
19:53:54 Nick Lynx_ is now known as Lynx_awy (~lynx@tina-10-4.genetik.uni-koeln.de)
19:54:19preglowand be aware that seeking currently destroys possibility of gapless
19:54:58amiconnHmm, I don't find this variable?
19:55:36preglowin codecmpa.c ?
19:55:51preglowi can find start_skip plenty of places
19:56:07preglowline 272 is where you'll want to change it
19:56:07amiconnHmm, I looked in playback.c and metadata.c
19:56:15preglowno, it's a codec specific thing
19:56:21preglowwe don't want this done for any other codecs than mp3
19:56:44amiconnHmm, codec*.c in apps/plugins is highly illogical
19:57:08preglowi know
19:57:11preglowi pet they'll be moved
19:57:12preglowbet
19:57:20preglowwhen linus finishes the codec plugin api
19:57:36Audiophilslasheri: status on the dimming?
20:00
20:02:20SlasheriAudiophil: still investigating but with timers it should be possible
20:02:44Slasherihowever, if we want to dim both displays at same time, it's not possible to use the hardware pin toggling
20:02:47crashdBagder: thanks for adding that link to the rockbox wiki
20:03:48west-acrehttp://www.marevalo.net/iRipDB/iRivDB_structure no longer exists in the iriver wiki links
20:03:56preglowSlasheri: the displays don't necessarily dim out at the same time
20:04:03preglowSlasheri: so you can't anyway
20:04:17Slasheripreglow: yes, but i think if it's possible to dim continuously the displays.. :)
20:04:18preglowyou'd need a separate timer for each display
20:04:31amiconnSlasheri: Fading the backlight with timers is the only possible way
20:04:39amiconnAnything else would block too long
20:04:48Slasheriamiconn: ok, that's good to kno
20:04:49Slasheriw
20:05:07amiconnUsing timers shouldn't be hard, and if we're lucky, the backlight control pin is also hardware controllable by a timer
20:05:15Slasheribut i can't use two timers for both displays because kernel uses one timer
20:05:25amiconn...then it wouldn't need any cpu power
20:05:33Slasheriyep, that's what i mean
20:05:58Slasheribut i think it's not possible to toggle two pins with hardware at a same time?
20:06:34amiconnI don't know, didn't read much about the pin configurability of the cf yet
20:06:46preglowit isn't
20:06:49preglowi'm fairly certain
20:07:08preglowbut the cpu needed to toggle a pin is negligable
20:07:11amiconnAnyway, I think it's a bit early for adding fancy features
20:07:23Slasheriok, maybe the only way is to use interrupt then. If we can make the dim continuous without too much reducing performance
20:07:33preglowyou have to use an interrupt anyway
20:07:56preglowyou need to change the timer period every time the timer triggers
20:07:58amiconnBug-free playback is much more important imho, as well as some completing the other rockbox features
20:08:12Slasheriah, then it shouldn't eat much more cpu whetever the pin toggling is done by pure hardware or software
20:08:26preglowwell, no, not any measureable amount
20:08:27amiconnApart from that, I did experiment with backlight dimming on archos some time ago
20:09:01preglowbut no, i'm fairly certain you'll need one timer per backlight, unless they're supposed to always dim at the same time
20:09:16amiconnOn player, it's quite simple, just use a timer. The sh1 timers are very well suited for that, since they can fire 2 interrupts at 2 different values
20:09:41amiconnUnfortunately the recorders don't allow this, since the backlight isn't connected to a cpu pin
20:10:00amiconn...but to a pin of the rtc instead, which can only be set via i2c
20:10:21amiconn...and i2c within an interrupt is a no-no
20:10:48amiconnSo there's no way for a back-light-organ on archos
20:11:17preglowwhy no i2c in an interrupt?
20:11:21preglowright, you have to do it manually
20:11:24amiconnThe recorder MAS would deliver peak readings, but controlling the backlight from interrupt is impossible
20:11:49amiconnThe player allows easy backlight dimming, but its mas doesn't deliver peak readings
20:12:06amiconni2c from interrupt is a no-no because of 2 reasons
20:12:11amiconn(1) it's too slow
20:13:08amiconn(2) There might already an i2c transfer going on in non-interrupt code. Breaking into this from interrupt will make a major mess
20:13:33t0masSlasheri: what dimming do you want? constant dimmed backlight display? or fade out backlight display insetad of just turning it off?
20:13:47t0mas*instead
20:13:49amiconnI tried it, it even works as long as no music is playing. As soon as some music is playing, the archos will lock up (the mas also uses i2c communication)
20:13:54Slasherit0mas: fading at least is possible
20:14:00Slasheribut i will try both
20:14:15amiconnSlasheri: Light organ as soon as we have peak readings :)
20:14:29Slasheri:)
20:14:36preglowhahah
20:14:46preglowa fade will do for now :>
20:14:55preglowpeak readings, yeah
20:14:57t0mashm? disco flashing stuff? :P
20:15:15preglowhrmph, we might end up needing to do quite a bit of processing on the audio data
20:15:17t0mashow should we do peak readings? hardware? of should the pcm code do that?
20:15:22preglowpcm code needs to do it
20:15:26t0masshit..
20:15:30preglowwhy?
20:15:32preglowshould be fast
20:15:46t0masthat'll cost some cpu for visualization stuff
20:15:58preglowsuch is life
20:16:38amiconnPlain peak readings should be simple
20:17:12amiconnJust calculate max(abs(samples)) across a number of samples
20:17:54t0mashm... and something like a peakmeter splitted in tone levels?
20:17:57t0masimpossible?
20:17:59t0masreally hard...
20:18:04t0masof trivial? :)
20:18:08t0mas*or
20:18:33preglowwhat?
20:18:38preglowyou mean like a spectrum analyzer?
20:19:20t0masjup
20:19:29preglowdepends on how many bands you want
20:19:34t0masjust 3
20:19:43t0masso we can flash the lcd backlight on the bass :D
20:19:47preglowshouldn't be too hard
20:20:13preglowif you only need the bass, we don't need to split it in three
20:20:26preglowwe'll just lowpass the audio and run the peakmeter on it
20:20:33t0mashm... some weighted average... would be nicer...
20:20:46preglowah weighted average _is_ a lowpass
20:21:08t0maserm? I ment taking 100 as a nice value for disco bass
20:21:19t0masand the lower bass weigthed lower...
20:21:29preglowahh
20:21:43preglowif you need that kind of spectral resolution, you can pretty much forget it
20:21:53preglowit'll be too slow
20:21:57t0masthen I don't need it ;)
20:22:08t0masjust cut off at 120... and use the peakmeter isn't hard?
20:22:12preglowno
20:22:21preglowbut you were starting to talk about weighting and...
20:22:22preglow;)
20:23:14t0mashm... and the peak readings... sounds not that difficult either?
20:23:15leftrightany news on mp3 gapless
20:23:17CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
20:23:17*amiconn is now listening to the second part of the mix, with start_skip == 0
20:23:24t0maswrite a function returning the max value since the last read?
20:24:22preglowwe'll see
20:24:25preglowi don't much care about it
20:24:31preglowso someone will probably beat me to coding it anyway
20:25:43 Join webguest32 [0] (~d49f4cf2@labb.contactor.se)
20:26:30leftrightwas listening to the beatles white album, yuck the gaps sucked
20:30:17thegeekwas it horrible?
20:30:23leftrightyes
20:30:46preglowamiconn: so, is it fixed?
20:30:55amiconnpreglow: It's still glitching... It seems rockbox might loose whole frames
20:31:48preglowamiconn: that is exactly what i was struggling against when i tried to make gapless playback work
20:32:04Slasheriamiconn: Hmm, when i initialize timer should i write anything to ICR0 register?
20:32:06preglowit lost entire frames at the end of a file
20:32:11mrlalaleftright: with the latest rockbox build?
20:32:32amiconnSlasheri: I have no idea as I didn't look at cf timers yet
20:32:33leftrightyes todays bedge@11:55
20:32:39Slasheriamiconn: ah, ok
20:32:49preglowSlasheri: i don't think so
20:33:05preglowSlasheri: there is timer init example code in the manual
20:33:08HClhello
20:33:11Slasheriat least kernel writes this when it starts it ticker: ICR0 = (ICR0 & 0xff00ffff) | 0x008c0000; /* Interrupt on level 3.0 */
20:33:14amiconnDidn't need them... but I think looking at the timer tick should be helpful
20:33:24Slasheripreglow: ok, maybe i will check that :)
20:33:48preglowpage 11-7
20:33:49amiconnpreglow: I wonder why it loses frames...
20:33:55preglowamiconn: you're not alone in that
20:34:02mrlalai have yet to experience a gap more that a tiny blip on some very few mp3s
20:34:15preglowamiconn: my problem was that libmad thought the file was ended before it truly was
20:34:31preglowmrlala: the gaps aren't large, but they'rethere
20:34:33amiconnWhat does libmad have to do with the file itself?
20:34:53mrlalaso far so good for me though :)
20:34:57amiconnI thought rockbox just feeds it a compressed data stream, and reads back the uncompressed pcm stream
20:35:09preglowcodecmpa feeds libmad
20:35:18t0masSlasheri? codecs call audiobuffer_insert() and samples are added to the audio buffer... but when are they read out?
20:35:25t0masfrom where?
20:35:32t0masso the peakmeter thing is realtime...
20:35:33leftrightmrlala: which album ?
20:35:42preglowt0mas: ooh, that's a good point
20:35:52preglowwe could make a peak meter queue, though
20:35:59preglowso the data can wait until it's supposed to be used
20:36:21preglowbut we might want to do the peak meter elsewhere, yes
20:37:17leftrightmrlaIa: I was listening the The Beatles White album Disc2
20:37:25preglowamiconn: when i tried to get things working, codecmpa, as it is unmodified, decodes exactly like foobar2000 does for a lame gapless file, up until the very end of the file
20:37:26t0maspreglow: I was thinking of just keeping track of the max value we've read...
20:37:30t0maseh
20:37:30t0massend
20:37:39t0masand when you read out the max value... reset it to 0
20:37:44t0masso the next read out is the next value
20:38:00amiconnpreglow: Forget about lame gapless for now, my glithcing happens without that
20:38:16amiconnt0mas: Yes, that's exactly what the mas does
20:38:21preglowamiconn: it skips the right amount of samples at the start, and it's calculated a track duration sample accurate, but libmad stops giving me data before it should
20:38:29t0masok, then that's what I'm trying in software...
20:38:35preglowamiconn: the code should work fine for those files as well
20:38:42mrlalaoh, i havent tried the album you're talking about, just talking about in general i haven't experienced any signifigant gaps, at least nothing horrible
20:38:53Slasherit0mas: they are read out by DMA0 interrupt
20:39:00t0masbut then Slasheri has to tell me wheren to keep track of the max value.. in a way that it's almost realtime
20:39:06amiconnFor track changes where it doesn't looose frames, start_skip==0 sounds perfect
20:39:12t0masSlasheri; ok, and where is that defined
20:39:23Slasheripcm_playback.c :)
20:39:26t0mask
20:39:47amiconnt0mas: We could do a little better than the mas if we are able to keep track of actual clipping
20:40:21amiconnIirc most codecs deliver >16 bits so the data needs to be scaled down.
20:41:17preglowthat's being dealth with by the codecs themselves, at the moment
20:41:17t0mashm? it is already scaled down when in the DMA functions
20:41:23amiconnClip detection on the mas uses some trickery. It considers 2 consective readings of MAX_PEAK as a clipping event
20:41:43t0mashm.. the codec should report clipping then?
20:41:56amiconnHmm, that might get really tricky
20:42:26leftrightmrlala: Dark Side Of The Moon, also a good album to use to test for gapless.
20:42:47amiconnIf the codec handles it, the time delay would need to be taken care of
20:43:24t0masSlasheri: in the dma interrupt... I can't to much thing right?
20:43:31t0masbecause it would block to long?
20:43:48Slasheriin dma interrupt you should do nothing..
20:43:58SlasheriHmm
20:43:59preglowabsolutely nothing
20:44:24t0masok, where should I read out the data then?
20:44:29amiconnWe might end up with a peakmeter thread
20:45:24t0mashm
20:45:29HClSlasheri, any progress on that event thing?
20:45:37Slasherit0mas: Hmm, maybe some function could be written to get a pointer to fresh data that is going to be written out
20:45:51SlasheriHCl: not yet, i will look it soon :D
20:46:05SlasheriHCl: or there is some progress, not much to do
20:46:19HClkay
20:46:49Slasherit0mas: or then you could just buffer the peak meter data
20:46:50amiconna peakmeter thread listening to 'new dma block' events perhaps. The dma interrupt would then send these events to the thread
20:47:21t0masaudiobuffer_pos is that the thing we're now playing?
20:47:29t0masor where the codec is writing?
20:48:03amiconnWhat's the current dma granularity, in terms of smaples?
20:48:08amiconn*samples
20:48:28t0masoh... yes... that's something to look at too...
20:48:43t0masas if that's a lot... reading it this way would be stupid
20:48:51amiconnpreglow: The amount of missing frames seems to be different every time, and btw, these files are 192 kbps cbr
20:50:07preglowyup, i know, for some files i get perfect gapless playback
20:50:11preglowfor some i miss as much as four frames
20:50:22YokaloshAMICONN!
20:50:31YokaloshIgnorin me!
20:50:32 Join webguest50 [0] (~8b4c8047@labb.contactor.se)
20:50:48preglowamiconn: line 324 is where codecmpa.c exits prematurely for me
20:51:00preglowamiconn: that error signifies libmad thinks it's reached the end of the file
20:51:18pregloweither that, or it's found to many corrupt frames, but that's not the case for me
20:53:16amiconnI think there are no corrupt frames in these files, otherwise the mas would glitch too
20:53:28preglowno, i'm certain my files weren't corrupt
20:53:34amiconnline 324 is commented out here?!
20:53:35preglowi've ripped them myself
20:53:41amiconnsame here
20:53:57preglow332, sorry
20:54:08preglow if (file_end == 30)
20:55:38 Quit Audiophil ("CGI:IRC (EOF)")
20:56:07amiconnHmm, that means if the file_end variable reaches 30 it will bail out.
20:56:38amiconnThat leaves the uestion what things might trigger one of the 2 stream errors which make it counting up
20:56:42amiconn*question
20:57:19amiconnApart from that, the if on line 348 will never be reached
20:57:41amiconnThis is looking fishy
20:58:07preglowfor me, all the file_end increments happened simultaneously
20:58:09preglowat the very end
20:58:16preglowwhich has me concluding that libmad thinks it can find no more data
20:58:45 Part djek ("Leaving")
20:59:52amiconnI think you need to feed it the next frames then
21:00
21:00:01amiconnMight be an effect of the bit reservoir
21:00:31 Nick StrathAFK is now known as Strath (~mike@dgvlwinas01pool0-a247.wi.tds.net)
21:02:44amiconnpreglow: In fact it is necessary to feed the data contiuous, even across track boundaries
21:03:02amiconnOtherwise some frames are undecodable
21:05:35 Quit webguest50 ("CGI:IRC (Ping timeout)")
21:07:58preglowamiconn: hmmm
21:08:18preglowamiconn: but these are vbr files, i didn't think the encoder used the bit reservoir for those
21:08:32preglowamiconn: but it's definitely worth a shot
21:08:40amiconnEncoders do definitely use the bit reservoir even for vbr
21:08:45preglowyes
21:08:47amiconnThe mas does, it's documented
21:08:51preglowbut this is the best explanation i've heard so far
21:08:57amiconn(and even selectable)
21:09:09preglowbut i don't dare try to implement it, i'll end up spending the entire night on rockbox again
21:09:13amiconn...and even selectable _in rockbox_ :)
21:10:04amiconnUsing the bit reservoir gives more fined grained control since the frames sizes are fixed
21:10:14amiconn...so the files end up smaller
21:10:15preglowyeah, it is logical
21:10:29preglowyou're able to emulate a more fine grained selection of bitrates then
21:11:06HClbored...
21:11:11preglowHCl: code
21:11:17HClwaiting for preglow :/
21:11:19 Join aegray [0] (~aegray@h69-21-201-246.69-21.unk.tds.net)
21:11:20HClplus i'm sick.
21:11:26preglowHCl: you're waiting for me?
21:11:26HCli can't think extensively or it hurts :/
21:11:30HCli mean
21:11:31HClslasheri
21:11:33preglowheh
21:11:36HClsee, i'm not 100% ;/
21:11:43preglowthen go read a book or something
21:11:47HClmmm...
21:11:50HClno books to read :/
21:11:54t0masHCl: watch a movie...
21:12:02t0masand fall asleep half way though
21:12:06zethere's always books to read
21:12:08HCli should. but i STILL haven't found the new hitchhikersguide movie on the internet
21:12:09t0maswake up tomorrow... and feel better
21:12:11HCl :/
21:12:15zedown and out in the magic kingdom was entertainig
21:12:22zeand its free, creative commons license
21:12:48zeHCl: there's cam's
21:12:56zeHCl: but who wants that crap
21:13:11HCli haven't found the cams.
21:13:17HClsome cams can be pretty decent, really.
21:13:22preglowamiconn: but is really the bit reservoir kept intact across track boundaries?
21:13:23zepiratebay.org
21:13:27zetorrent tracker
21:13:34HClkay, yea, i know the site..
21:13:41preglowamiconn: i'd think that'd give lower quality playback if you don't play files sequentially
21:13:52zeoh gno
21:13:59zenow i'm on LOGS
21:14:06amiconnpreglow: Yes it's necessary, because that is exactly what happens with lame −−nogap
21:14:06zeor as i call 'em, deadtrees
21:14:28preglowamiconn: yeah, but that doesn't explain my vbr files
21:14:51preglowi don't think keeping the reservoir intact across track changes is standard practice
21:14:56preglowbut again, it's most definitely worth a shot
21:14:59 Quit t0mas ("brb")
21:15:03*HCl shrugs. its legal here to download things..
21:15:43HClsomeone make a timemachine and go buy all the bleach dvds in the future and bring them back :/
21:15:59zei guess there's a new law here where posting/obtaining PRE-RELEASE copyrighted material can be prosecuted as a criminal offense
21:16:35zenot sure if that counts stuff thats out in theatres or not
21:16:36zeheh
21:17:16zebut i didn't post, nor have i downloaded hhgttg
21:17:25*ze is waiting for it to come out on dvd...quality
21:17:26ze:p
21:18:04HCli'm waiting for it to be put online decently, any quality
21:18:16preglowi'm just waiting for it to show up in the cinemas
21:18:21preglowwhich apparently will never bloody happen
21:18:31zei saw a cam or ts or whatever that looked pretty decent
21:18:44zeit doesn't say so, its just like generic looking
21:19:48ze"The Hitchiker's Guide to the Galaxy(2005)1-2.MPG" and etc 2-2.MPG
21:20:05 Join t0mas [0] (~Tomas@ip503c08d1.speed.planet.nl)
21:20:10zebut i closed and rm'd it as soon as i saw it was a cam
21:20:23zei already watched it in the theatre so i don't need some crappy quality version of it
21:21:12zei guess the bbc are/have catching/caught up the radio show to the books
21:21:25zewith new phases of the series with as many of the original actors as they could get
21:21:48zethey even used the part douglas adams recorded for the character he wanted to play
21:21:52zebefore he died
21:21:54amiconnHmm, flipped lcd mode is off by 4 pixels
21:22:20 Quit niobos ("leaving")
21:22:29 Quit t0mas (Client Quit)
21:22:48preglowit would be good to see actually which error libmad returns as well
21:24:19preglowMAD_FLAG_INCOMPLETE? that value should never be in Stream.error
21:24:23 Join t0mas|laptop [0] (~tomas@ip503c08d1.speed.planet.nl)
21:24:43preglowSlasheri: can you remember why that's there?
21:24:57*HCl reads about the struggles of various pathetic companies against the freedom of other people.
21:25:18preglowmoney > freedom
21:25:25HClhah.
21:25:27HClmaybe for them.
21:25:29HClnot for me.
21:25:38preglowof course, but they don't want to see it from your point
21:25:59HClthey're money greedy bastards.
21:26:15preglowwell, yes, they have to be by law
21:26:21preglowat least a lot of places
21:26:26HClfortunately, heh.
21:26:36Slasheripreglow: Hmm, i think MAD_FLAG_INCOMPLETE situation occurs if buffers has not enough data
21:26:37HClwhich is why they're resorting to empty threatening
21:26:44Slasheribut i am not sure about that :)
21:26:49preglowSlasheri: but it's not an error code, it's never returned from a function
21:26:51HCland sueing internet providers for private information about their customers
21:26:53HClthey should fuck off :/
21:26:54preglowSlasheri: it's an internal flag
21:27:03Slasheriok, interesting.. i think you can remove it :)
21:27:06preglowyou can't really blame the companies
21:27:14preglowblame the government for allowing it
21:27:28preglowcompanies will always do whatever they can to secure their own interests before anything else
21:27:31HClcompanies buy the government
21:28:07preglowbuy and buy
21:28:12preglowthey're needed to make a country tick
21:28:19preglowso they do of course have some influence
21:28:58 Quit webguest32 ("CGI:IRC (EOF)")
21:29:30HClmhm.
21:29:31preglowsuch is the world today
21:29:45HClwell. the day that they force tcpa onto everyone is the day i move to a small country without much copyright laws
21:30:00preglowi dont like it very much either, but our way of life depends on it being like that
21:30:09preglowhaha
21:30:09Slasherii think i found the problem. I had to set ICR and IMR right :)
21:30:15preglowi'll probably just ditch computing
21:30:20HClyea, or that.
21:30:24HClgo back to typewriters
21:30:25HClwith ink.
21:30:27HClohh. scary :3
21:30:28preglowstart making my own stuff
21:30:29HCl :p
21:31:00t0mas|laptoppreglow: I have an intresting project for you
21:31:04preglowfree software's gotten far enough that you can build a custom system and have it completely operational
21:31:09preglowt0mas|laptop: oh?
21:31:17t0mas|laptopbut it's dutch...
21:31:19t0mas|laptopshit...
21:31:25HClwhat project?
21:31:33t0mas|laptopwe've build a very simple cpu in school
21:32:00t0mas|laptopHCl: translation of "profielwerkstuk" ?
21:32:14t0mas|laptopclosing essay for highschool?
21:32:16t0mas|laptopor something like it?
21:32:20preglowamiconn: btw, i don't reinit libmad, so the reservoir should intact even after it forces me to change track
21:32:28HClyup, sounds fine
21:32:53t0mas|laptopwell... some people in my class wrote it about building their own cpu
21:33:08t0mas|laptopthen me and a friend started a project for study points out of it... to actually build it
21:35:43t0mas|laptopbut ofcourse we haven't gotten any further than a simple calculator
21:36:07HClwe did that.
21:36:13HClalso for our profielwerkstuk
21:36:17HClnot a cpu though
21:36:22HCljust an add unit and a minus unit
21:36:37HClwe built the actual add unit..
21:36:42HCldo you know those "systeemborden" ?
21:36:47t0mas|laptopthat big things
21:36:49t0mas|laptopyes
21:36:50HClyea
21:36:51HCl :P
21:36:52HClwe like.
21:36:55HClrigged 18 of them
21:36:56t0mas|laptopbut we did it with the simple chips approach
21:36:57HCltogether
21:36:57 Part aegray ("Leaving")
21:37:01t0mas|laptopLOL
21:37:03t0mas|laptopwe did that too
21:37:05HClit was fun.
21:37:08HClit actually worked too.
21:37:12t0mas|laptopand after that we switched to normal cheap chips
21:37:17HCl*nods*
21:37:21t0mas|laptop8 ands in 1 chip is a lot smaller... and cheaper
21:37:24HClyea
21:37:26HClwe had that
21:37:34HClthough we never got to getting the chips
21:37:38preglowmaking systems out of logic chips = great fun
21:37:40HCljust planning it out *shrugs*
21:37:44HClanyways.
21:37:46HClhighschool.
21:37:47t0mas|laptopwe actually did it
21:38:03HClhighschool = lots of people forcing you to learn crap you needn't know.
21:38:12t0mas|laptoplolo
21:38:17*t0mas|laptop does it better than
21:38:21preglowi don't really know what high school corresponds to here
21:38:31t0mas|laptopwe just start projects... and get studiepoints for it
21:38:33preglowbut i did three years of learning stuff i didn't want to know, just so i could go to uni
21:38:48t0mas|laptoppreglow: lucky... it's 6 years here
21:39:18preglowi regret it now, of course, but i was young and even stupider then
21:39:41t0mas|laptophm? what else would you've done then?
21:39:43 Quit Yokalosh ()
21:39:55preglowsomething else
21:40:01preglowi've learnt all i know on my own anyway
21:40:04***Saving seen data "./dancer.seen"
21:41:21t0mas|laptophm... what study have you done on uni?
21:41:31t0mas|laptop(or is it in uni?)
21:42:51 Quit mrlala (Read error: 110 (Connection timed out))
21:43:10 Quit ashridah ("Leaving")
21:43:17preglowelectronics
21:43:29preglowelectronics, programming and some signal processing
21:43:53preglowi might have learnt a bit of electronics, but the rest i already knew
21:43:58t0mas|laptophm, and it was't intresting?
21:44:16preglowinteresting, sure, but when you already know what they're teaching you, it's hard to keep interest going for long
21:45:43 Join muesli- [0] (muesli_tv@hmln-d9b8ef5d.pool.mediaWays.net)
21:45:56muesli-g'day mates
21:46:40t0mas|laptophi
21:49:54muesli-how's goin t0mas|laptop
21:53:39 Quit HCl ("Lost terminal")
21:54:13 Join DomZ [0] (~52426222@labb.contactor.se)
21:54:19 Join hcl [0] (hcl@titania.student.utwente.nl)
21:54:23*hcl smacks screen
21:54:25 Nick hcl is now known as HCl (hcl@titania.student.utwente.nl)
21:55:10*muesli- yawns :D
21:56:08 Join thegeek_ [0] (na@ti521110a080-1004.bb.online.no)
21:56:40 Quit thegeek (Read error: 131 (Connection reset by peer))
22:00
22:00:47amiconnYay! Graphical glitch with flipped display is fixed, about to commit :)
22:01:33 Quit RotAtoR ()
22:02:43HClwhat was the glitch?
22:02:54t0mas|laptop4 pixels off
22:02:58HClmk.
22:03:11amiconn4 pixels off, and gibberish in the 4 top pixel rows
22:03:28preglowamiconn: did you see my last message to you?
22:06:47amiconnHmm, yes, scrolled back now
22:07:07preglowi didn't think of that, but libmad is never reinitialised
22:07:29preglowseems i thought of this when i tried coding gapless, heh
22:07:50amiconnHowever, it may happen that you drop out to early from transferring data of the old file
22:08:59 Join Philip_0729 [0] (~Philip_j_@user-7155.lns6-c10.dsl.pol.co.uk)
22:09:28amiconnI don't know the libmad internals; in theory you'll never need a "future" frame to decode the current one; only the current itself and some bits from previous frames
22:09:48amiconnMaybe libmad delays decoding...
22:09:54preglowi doubt it
22:10:03preglowit hasn't got enough memory in its structs for that
22:10:20preglowand there's no reason it shoudl
22:10:56amiconnYes, right, however, the missing frames are observed behaviour...
22:11:10amiconn...and there must be a reason...
22:14:20preglowanywho, i've compared its output to other players
22:14:30preglowit's sample-synced until it starts dropping data at the end
22:14:59preglowso it doesn't delay anything
22:15:00amiconnHmm, how do you determine the start index?
22:15:28preglowwhat start index?
22:15:28amiconnIf I had a device capable of s/pdif in, I could try to compare to MAS...
22:15:50amiconnYou said it's sample synced. How do you determine the start, i.e. sample 0
22:15:51amiconn?
22:15:54preglowahh
22:16:01preglowthere's code in codecmpa that dumps to disk
22:16:08preglowi just assumed it knows what it's doing
22:16:28preglowit just sets a flag when it detects a track change, i believe
22:17:26 Quit leftright ("CGI:IRC")
22:17:32 Join leftright [0] (~5087b476@labb.contactor.se)
22:17:43preglowno, it's true when the codec is initialised
22:17:56preglowso it starts dumping from the start, and that's what it should do
22:18:05amiconnThe interesting point is realtime behaviour here, not file data. I.e. which samples do you get back when you feed frame 1 etc
22:18:17preglowwell, that's the data it's dumping
22:18:26preglowthe data it dumps to disk is the exact samples it also feeds to the dac
22:18:45amiconnYes, but you don't know *when* this data is generated
22:18:45linuxstbHave you tried comparing your output with "madplay -o wav:" on the same file?
22:19:10preglowlinuxstb: no, i just compared it with foobars wavwriter output
22:19:20preglowlinuxstb: nice idea, why didn't i think of that
22:19:43preglowamiconn: i don't see why it matters
22:20:06amiconnIt might, in case decoding is delayed somehow
22:20:47amiconnPerhaps it doesn't even give back anything when feeding the first frame
22:21:08amiconn...but starts to deliver pcm with the next frame(s)
22:21:21preglowthat's highly improbably, but should also be quick enough to find out if is true
22:22:46preglowsomeone make logf for plugins and i'll check it out now :>
22:24:09preglowbtw
22:24:15 Join asdsd____ [0] (~asdsd@h-67-100-24-230.miatflad.dynamic.covad.net)
22:24:15preglowit isn't possible for libmad not to generate data
22:24:18 Part asdsd____
22:24:19preglowit HAS to generate data
22:24:33preglowso it isn't delaying anything, i'd be seeing zeros in the output then
22:29:24preglowand are you completely sure −−nogap assumes the bit reservoirs aren't emptied over track boundaries? i can't think of a single reason why it needs to assume this
22:31:55 Join RotAtoR [0] (~e@c-67-162-206-66.hsd1.mi.comcast.net)
22:35:18amiconn−−nogap is simply one single big mp3 split at the nearest frame boundary
22:35:24preglowi know
22:35:39preglowthat doesn't tell me anything about any special reservoir needs
22:35:54preglowit's just a redistribution of the initial wav data
22:36:02 Join sneakums [0] (~sneakums@dsl092-013-170.sfo1.dsl.speakeasy.net)
22:36:11amiconn..and a frame that uses the bit reservoir is not decodable if the previous frame data is unavailable
22:36:46preglowi doubt lame has the last frame of an mp3 but anything at all in the reservoir, it's more likely it tries to end up with zero bits left
22:36:50preglowbut = put
22:37:02sneakumsiriver rockbox works great, but how do i get at the current-platlist menu?
22:37:15sneakumsto queue up a bunch of assorted files, say
22:37:26t0mas|laptopsneakums: hold select
22:37:30t0mas|laptop(the joystick)
22:37:36preglowsneakums: you select the file, keep the joystick button pressed, and select insert from the menu
22:37:41amiconnTrue for the last frame, but the last frame of the first file at a −−nogap track split *isn't* the last frame of the stream
22:37:42sneakumsah
22:37:44sneakumsrockin'
22:37:54 Quit ghostiger (Ping timeout: 14400 seconds)
22:38:23preglowamiconn: sure, but it doesn't HAVE to put anything in the reservoir for the first frame of the next file to use
22:38:57preglowactually, if it did, it would break playback of the tracks alone
22:38:58amiconnIt doesn't have to, but it might do so
22:39:02preglowsince they wouldn't have the data they need
22:39:05preglowno, i really doubt this i done
22:39:12preglowfor the abovementioned reason
22:39:30amiconnIt doesn't break playback of the track, just playback of the first frame(s)
22:39:41*HCl watches the hitchhikers guide to the galaxy
22:39:53preglownothing says the first frame has to empty the reservoir either
22:40:01preglowit might break playback of a great number of frames
22:40:04amiconnAfaiu lame −−nogap is exactly equivalent to ripping the whole cd as one single wav at first, then encoding that, then splitting at frame boundaries roughly equvalent to the track changes on cd
22:40:19preglowand anyhow, do you know for a certainty the first frames ARE skipped when you play the files back alone?
22:40:46preglowahh, no, i that's BASICALLY what i think happens
22:40:53sneakumshmm, i think it got confused, i did insert next on two tracks, it started playing the first one i inserted, but the little note icon is next to the second one i inserted (which is above the first one i inserted in the playlist)
22:40:59amiconnI know that playback of tracks using bit reservoir does work even if you cut frames from the beginning
22:41:07preglowit pads out the ends of the tracks with data from the next track, but the resulting mp3 streams are independent
22:41:11sneakumsstarted playing it after the then-playing track finished that is
22:42:03preglowanywho
22:42:06preglowthis really doesn't matter
22:42:16preglowif reservoir was the problem, it should be working
22:42:18amiconnI don't think the streams are independent, in fact they can't be
22:42:24preglowwhy?
22:42:42HClbah.
22:42:45amiconnIiuc there is an overlap of 1/2 frame length
22:42:49HClmy dad insists on watching it on television tomorrow -.-
22:42:56preglowindeed
22:43:02preglowthat's intrinsic to the mdct
22:43:14amiconnSo a truly independent track starts with 1/2 volume for half a frame, and ends the same way
22:43:30preglownope, doesn't work like that
22:43:49preglowit just assumes the data outside the wav limits are zero when it does the overlapping
22:43:50amiconnThere's a simple experiment:
22:43:54preglowthe end result isn't a volume reduction
22:43:58preglowit's ringing in the waveform
22:44:35amiconnMake a longer wav. Split it in 2 with an audio editor. Encode both the single wav alone and the 2 part-wavs with −−nogap.
22:44:49amiconnStrip all xing vbri etc stuff from all mp3
22:44:50amiconns
22:45:10amiconnGlue the 2 part-mp3s together
22:45:32amiconnWhen I am right, the resulting mp3 should be identical to the single-wav encoding result
22:45:45 Quit muesli- (Read error: 110 (Connection timed out))
22:45:48preglowi tried that
22:45:49preglowhttp://glow.m0f0.net/gapless_sine.png
22:45:56preglowand there's the result
22:45:59preglownot exactly equal
22:46:02 Join StrathAFK [0] (~mike@dgvlwinas01pool0-a247.wi.tds.net)
22:46:03amiconnThat's from decoding, right?
22:46:06preglowyup
22:46:16amiconnI mean the encoding part with lame
22:46:28preglowthere's ringing in the transition thanks to lame STILL assuming zeros outside the wav
22:46:35preglowoutside the track boundaries
22:46:45preglowwhich goes to show the tracks are still functional stand alone
22:46:48amiconnDid you use −−nogap?
22:46:52preglowyep
22:47:00preglowi tried it with nogap and lame gapless
22:47:06amiconnHmm.
22:47:21amiconnI might do that experiment, but not right now.
22:47:25preglowfeel free to try it again to be completely sure
22:47:35preglowi haven't got any decent audio editors in linux
22:47:47preglowthat screendump is from the lame gapless file
22:47:54preglowbut the nogap mp3 looked much the same
22:47:58amiconnI have audacity
22:48:11preglowme too
22:48:15preglowi hate it with a passion
22:49:01preglowif −−nogap worked like you say, then indeed the reservoir should not be emptied across boundaries
22:49:06preglowbut i don't believe it works like that
22:50:00amiconnTrack splitting on archos recorder definitely does work like that, plus people might use tools like mp3directcut which also do the same
22:50:14amiconnMaybe lame is a bit more clever though, let's see..
22:50:20preglowtrue enough
22:52:00 Quit t0mas|laptop ("brb again")
22:52:26 Quit sneakums ("taster, you're the automatic saint")
22:53:09 Join belgarath [0] (~acd4fbae@labb.contactor.se)
22:53:16belgarathhello again
22:53:25preglowthe fact that some tools does it is a good reason for us not reiniting libmad for each track, if nothing else
22:53:26 Join pbvas [0] (~pbv@res04-pbv.res.st-and.ac.uk)
22:53:29preglowpbvas: yo
22:53:33preglowpbvas: is your bug fixed?
22:53:37pbvashi!
22:53:41pbvasyes it is!!!!
22:53:44preglowwoot!
22:53:49pbvasbig thanks to Linus!
22:53:49preglowi'm sure linus will be happy
22:53:59SlasheriHmm, i have set TIMER1 interrupt handler but i think it's not called when timer1 generates an interrupt..
22:54:18Slasheriinstead rockbox will _almost_ completely freeze
22:54:19belgarathI was reading the 2.4 manual and read this: "The splits are seamless (frame accurate), no audio is lost at the split point.
22:54:43preglowSlasheri: well, it sure is calling something ;)
22:54:56Slasheripreglow: yeah, something else :)
22:55:11Slasherii have set ICR and IMR this way:
22:55:13Slasheri| ICR0 = (ICR0 & 0xffff00ff) | 0x00008d00; /* Interrupt on level 3.0 */
22:55:13Slasheri| IMR &= ~(1<<10);
22:55:13belgarath"The break between recordings is only the time required to stop and restart the recording, on the order of 2-4 seconds"
22:55:24belgarathwill it do that on iriver?
22:55:30SlasheriAh, the level is 3.1, not 3.0
22:55:41preglowbelgarath: what are you talking about? the split plugin?
22:55:59preglowor recording?
22:56:07belgarathpreglow: when you split tracks during recording
22:56:11preglowof course it will
22:56:16belgarathok
22:56:18preglowwe wont deliberately introduce inaccuracies, hehe
22:56:30preglowremember that we can control most things ourselves on iriver
22:56:30 Quit DomZ ("CGI:IRC (EOF)")
22:56:51belgarathi was just wondering seeing as the md players can split tracks seamlessly
22:56:53preglowbelgarath: but don't hold your breath for mp3 recording
22:56:59belgarathok
22:57:02preglowit's still a long way off
22:57:04belgarathwhat will it be just wav?
22:57:14preglowi think wav and wavpack will be the first codecs
22:57:14belgarathwhich is still great
22:57:22preglowpossibly flac if someone gets it working fast enough
22:57:27belgarathcool
22:57:31preglowmp3 might be next
22:58:07belgarathcan the hardware support sdif?
22:58:24preglowyes
22:58:27belgarathwhere the player automatically creates new tracks when recording optically?
22:58:30preglowyes
22:58:36belgarathoh good
22:58:49belgarathwhy didn't iriver firmware implement that
22:58:55belgarathit seems strange
22:58:59amiconnBah, the remote lcd driver is messy :/
22:59:01preglowbusy doing other firmware, i guess
22:59:15preglowamiconn: what do you expect, it was ricks first low-level driver attempt ;)
22:59:37 Join t0mas [0] (~Tomas@ip503c08d1.speed.planet.nl)
22:59:47preglowi just had a look at the shine encoder
22:59:49preglowit's floating point :/
23:00
23:01:37 Quit t0mas (Client Quit)
23:05:10preglowamiconn: read this: http://minnie.tuhs.org/pipermail/mp3encoder/2001-July/003387.html
23:05:11 Quit Strath (Connection timed out)
23:05:15SlasheriHehe, i found the problem :)
23:05:31preglowamiconn: seems we both were right
23:05:53SlasheriI have to assing some free irq level to base vector table for Timer1 :)
23:06:06preglowSlasheri: so, does it fade now?
23:06:18preglowSlasheri: if so, how soon can you send me something to try out? :)
23:06:19Slasheripreglow: it should work very soon ;)
23:06:29Slasherii don't know, testing :D
23:08:19amiconnpreglow: Ok, so it resets the bit reservoir. However, the part about decoding to .wav not being gapless is interesting
23:08:43HClbah.
23:08:51HClthe tv series are tons better than the movie so far...
23:08:58amiconnImho it means that you should just ignore the track boundary and keep decoding, otherwise it won't be gapless, probably due to some delay
23:09:22preglowi remove this delay in codecmpa.c
23:10:01preglowthat's probably where the encoder delay thing comes in
23:10:03 Join mrlala [0] (~mistame@cpe-66-75-129-164.san.res.rr.com)
23:10:15amiconnYes, you remove the delay from one end, but you can't invent the missing samples on the other end
23:10:41Slasherinow it's crashing no more, i got timer1 operational :)
23:10:45amiconnFirst thing is that nothing should be removed unless there is a lame header
23:10:58preglowamiconn: foobar seems to get this right by removing samples in the start, and just ignoring the end
23:11:00amiconnSecond thing is that we need to find out why it's dropping frames
23:11:05HClbleh.
23:11:15HClthe new hhgttg was sure disappointing :/
23:11:27preglowHCl: mixed opinions
23:11:33amiconnpreglow: There must be something non-obvious in the foobar sources then
23:11:59preglowamiconn: perhaps, you should try it out to see if it gets your −−nogap mp3s correct
23:12:19preglowif it does, then my source in codecmpa.c should be correct
23:13:54Slasheriwow, rockbox playing music with display dimmed to 50%! :)
23:14:13HClmaybe
23:14:16preglowgimmegikmmeigmme
23:14:17HClpreglow: have you seen it?
23:14:19Slasherithere will be brightness settings too for backlight ;)
23:15:10preglowHCl: no
23:15:15HClmk...
23:15:44HCli've seen the original tv series and so far the new one is disappointing.. up to such a bad level that i started to skip forward...
23:15:51crashdbah
23:15:59crashdim getting absolutely nowhere with this h10
23:16:07HClportalplayer :/
23:16:10crashdyer
23:16:13preglowcrashd: what're you trying now?
23:16:20crashdjust trying to get my head round it still really
23:16:27crashdi just want to run some damned code on it _ hehe
23:16:39crashdstill got the problem of having arm and thumb instruction set
23:17:28crashdand it's just taunting me :\
23:18:21preglowyeah, couldn't figure that out myself
23:18:29crashdjust need another dev or two
23:18:34crashdthen we might get somewhere :]
23:18:51preglowprobably just need someone good with arms
23:18:57crashdhehe
23:18:58crashdyeah :\
23:20:53amiconnYay, display flip for remote lcd is working :)
23:23:52belgarathwill the h100 series ever have grayscale?
23:23:57preglowof courase
23:23:58preglow-a
23:23:59amiconnyup
23:24:08belgarathcool
23:24:16amiconnI intend to start working on that next week
23:24:30preglowamiconn: like finishing markuns patch?
23:24:58belgarathnice
23:25:01belgaraththanks
23:25:10amiconnImplementing the new gfx api, taking some bits from markuns patch and finishing it
23:25:21belgarathand thx a lot to the person who fixed the sleep timer to work with charger
23:25:30amiconnbelgarath: Thanks :))
23:28:13 Join Audiophil [0] (~c1d85c22@labb.contactor.se)
23:28:29Slasheripreglow: now it fades very nicely :)
23:29:18Slasherislow and smooth fade in/out
23:32:07Slasheripreglow: what do you think about committing it?
23:33:16 Quit Audiophil ("CGI:IRC (EOF)")
23:36:38preglowdunno
23:36:43preglowi wanna try :PP
23:36:50Slasheripreglow: i think it _should_ not break anything
23:36:59Slasheriso maybe i could commit it? :)
23:36:59preglowjust send me rockbox.iriver if you can
23:37:05Slasheriah, ok
23:37:07Slasherijust a moment
23:38:11Slasheridcc?
23:38:34preglowsure
23:38:38Slasherigreat :)
23:40:08***Saving seen data "./dancer.seen"
23:40:30 Quit belgarath ("CGI:IRC (Ping timeout)")
23:40:55preglowhmm
23:40:57preglowglitch
23:41:05Slasherihmm?
23:41:23Slasheridoes it work?
23:41:23preglowright after it's faded in
23:41:25preglowit blinks
23:41:33preglowand sometimes right after it has faded out as well
23:41:35preglowsure, works fine
23:41:39preglowi'd prefer a faster fade, though
23:41:42SlasheriHmm
23:41:50Slasherithat is caused if cpu frequency changes
23:42:07Slasheribut it should not blink when frequency is stable
23:42:28preglowwell, you can't guarantee that during playback :/
23:42:33preglowcpu changes all the time then
23:42:37Slasherithat's true :/
23:42:43Slasheribut it works quite well
23:42:56preglowit works great
23:42:59preglowbut not during playback
23:43:01preglowit blinks a lot then
23:43:07Slasherieh?
23:43:09Slasherireally?
23:43:12preglowyes
23:43:16preglowi'm playing an mp3
23:43:21preglowand it blinks while fading
23:43:21Slasheriit doesn't blink at all on my player
23:43:23SlasheriHmm
23:43:41preglowit just blinks sometimes
23:44:02Slasheriah, ok
23:44:22preglowbut no flicker at all
23:44:23preglowlooks great
23:44:39Slasheri:)
23:44:52Slasherihmm, should i commit it?
23:45:05preglowwell, the blinking should be fixed if possible
23:45:10Slasherilater some brightness settings (max brightness) could be added
23:45:25Slasheriyes, i don't know if that's possible..
23:45:50Slasheribecause cpu doesn't work at all(?) when frequency is being switched
23:45:52preglowhmm, no
23:46:17amiconnThe cpu does work, but at 11 MHz until the pll relocks
23:46:18preglowand the timer will probably be very unstable when it's happening anyway
23:46:29amiconn(for up to 10 ms)
23:46:41preglowamiconn: you sure about it working at 11mhz?
23:46:47Slasheriah, that can be fixed
23:46:49preglowmakes sense
23:46:52 Quit Harpy (Read error: 145 (Connection timed out))
23:47:01Slashericpu should be boosted while fading
23:47:19preglowmno, i don't think so
23:47:26preglowand still it doesn't fix it
23:47:36Slasherihmm..
23:47:49Slasheribut i will fix it tomorrow, nights :)
23:47:55preglowok, gnight
23:48:39preglowyou need to reset the timer for 11mhz before a cpu_boost, and for correct frequency right after,i guess
23:48:48preglowstill a glitch might be noticable, but the brightness wont change
23:49:04amiconnIncrementing the boost counter while fading might fix the blinking, but then it's boosted always if display brightness is <100% and >0%
23:49:13amiconn...not exactly what we want imho
23:49:52amiconnThe brightness shouldn't change much if the cycle is short enough
23:50:06amiconnSlasheri: What cycle frequency do you use?
23:50:42preglowit works fine like it is, but i don't think we can use it with glitching like this
23:50:45Slasheriamiconn: 100 pules / pwm cycle
23:50:47preglowit looks like the player is broken or something
23:50:53amiconn(It's the duty cycle that determines brightness, not the frequency, and duty cycle will be off for a very short time only)
23:50:53preglowat least that's what i think people will conclude
23:51:01Slasheriso 0-100% brightness
23:51:18Slasheriah frequency
23:51:35Slasheri500s intervall
23:51:38amiconnYou using 100 interupts per cycle???
23:51:47Slasheriyes
23:51:50amiconnI'd say 2 are enough
23:52:01SlasheriHmm
23:52:33amiconnOf course the isr would then have to reprogram the timer
23:52:42Slasheriyep
23:52:47amiconn...unless there is a more sophisticated way to use the cpu
23:52:56amiconn...like the sh1 timers allow to do
23:54:04preglowanyway, i think i'd like the fade to be twice as quick as it is now
23:54:22preglowfeels like it uses a couple of seconds now
23:54:35Slasheripreglow: it's possible.. i will look it tomorrow ;) ->
23:54:45amiconnSlasheri: 500 s * 100 means 50 ms, well above the pll relock time
23:54:59amiconnNo wonder it glitches when changing frequency
23:55:30amiconnIt would have to go much faster, but that practically requires going for 2 ints/cycle

Previous day | Next day