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

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

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

#rockbox log for 2005-11-04

00:08:47 Part arkascha ("Konversation terminated!")
00:12:55 Quit Kohlrabi (Read error: 104 (Connection reset by peer))
00:17:27 Join darkskiez [0] (
00:19:17darkskiezdoes anyone know if you can put an ihp140's hdd into an ihp120? are they the same chassis? mine broke, but the hdd is fine
00:19:28darkskiezwondering if i can save money and get a 120 and put my hdd in it
00:19:40preglowit wont fit
00:20:01preglowit's slightly too thick
00:20:04darkskiezis there somewhere i can buy an ihp140 with a broken hdd ?
00:20:32darkskiezis the screen easy to attach/detach? i could take one with a broken screen and hard disk
00:20:38darkskiezi just need the mainboard
00:24:56 Quit cYmen ("w00t")
00:28:45 Join linuxstb [0] (
00:29:10darkskiezwould this work with my drive:
00:30:47markundarkskiez: no, the hitachi connector is different
00:31:34markundarkskiez: this should work:
00:37:05 Quit RotAtoR (Read error: 110 (Connection timed out))
00:43:52 Join webguest20 [0] (
00:44:02webguest20Hello every one
00:44:26webguest20anyone here know a bit the WPS style?
00:46:35webguest20I %s%?it<%?in<%in. I >%it I %fn>
00:47:15webguest20anyone know what I need to remove for remove the track number info?
00:53:17webguest20anyone please?
00:53:41webguest20I'm lost in the WPS language
00:53:41phaedrus961remove %?in<%in. I >
00:54:04webguest20oh ok let me try, thanks
00:54:53webguest20I don't understand well
00:55:07phaedrus961it should look like this: %s%?it<%it I %fn>
00:55:11 Nick ashridah is now known as Lost-ash (
00:55:30phaedrus961but I believe the I should be a vertical bar: |
00:55:43webguest20that will do the same thing without the track info?
00:55:52webguest20yes it is
00:55:58 Join RotAtoR [0] (
00:56:19webguest20ok thanks I'll try
00:58:43 Join bagawk [0] (n=lee@unaffiliated/bagawk)
01:09:28webguest20thanks a lot phaedrus961 that rocks
01:10:46webguest20before to go a ltle question please :)
01:10:58webguest20it's about MPC decoder
01:11:27webguest20the APE tags seems to don't be full suported yet
01:12:57preglowwebguest20: well, no, they should be supported
01:13:05webguest20in my WPS screen when I'm playing MPC file, few infos are missing
01:13:38webguest20like file bitrate
01:14:22webguest20and next song info :(
01:14:26preglowthat's not in the ape tag
01:14:40preglownext song info i don't know about, that really should work
01:14:45preglowi'll fix the bitrate thing afterwards
01:14:55preglowat least hack it to work somewhat
01:15:22webguest20for the next info thing, nope for MPC (all others yes)
01:15:36webguest20thanks for fix bitrate info
01:15:53webguest20currently i've got: ?
01:16:18webguest20and for next info song it's freeze in searching info
01:16:24webguest20for all MPC files
01:16:38preglowi'll see if i can see why
01:16:55webguest20thanks very much, I've got a lot of MPC files
01:17:08preglowmpc is currently the worst codec in rockbox
01:17:09preglowapart from aac
01:17:39webguest20thanks for have it in realtime at least, really impresive works men
01:18:29webguest20the "Multi codec support" in the front of us iriver is very true
01:18:55webguest20I can't imagine any commercial player with all those codecs we have
01:19:18preglowneither can i
01:19:30preglowmusepack should at least be realtime for all bitrates now, yes
01:19:36preglowbut it could easily have been much, much faster
01:19:46preglowbut libmusepack isn't very easy to work with :/
01:19:54webguest20yes and less cpu using :)
01:20:04webguest20I can imagine
01:20:45webguest20just if you can fix please those missed infos, it would be wonderfull
01:20:49preglowi'll try
01:20:57preglowjust need to finish something else i'm working on
01:21:14webguest20ok, thanks much
01:22:38***Saving seen data "./dancer.seen"
01:23:48webguest20I love you all men
01:24:09markunwe are men, men is what we are..
01:24:34webguest20have a good night at all and preglow I hope you'll can fix MPC bugs
01:24:56markungood night
01:24:56webguest20markun: :D
01:25:07markunwebguest20: fight club :)
01:25:07 Quit webguest20 ("CGI:IRC")
01:25:11preglownow lookie here
01:25:17preglowsomeone posted fixed point wma source on the forums
01:25:43markunWhat a nice surprice!
01:26:11linuxstb"partially fixed point" is more accurate.
01:26:20linuxstbBut it's still a useful start.
01:26:55preglowquite a large codec too
01:27:02preglowyou wont see me working on that one for a while
01:27:07preglowi think i've said that before...
01:27:20linuxstbThere's still far too much rubbish from ffmpeg in there. I'm sure a lot more can be stripped out.
01:28:29preglowyeah, me too
01:28:34linuxstbBut I don't have any motivation to do anything either.
01:29:24preglowit's not going to be the first codec i fix, for sure
01:29:28preglowi have one wma file i ever listen to
01:29:33linuxstbI really want to find time to make some ipod progress - the ipod will be a fun platform for rockbox.
01:31:34preglowyes, i think you should do that too :)
01:32:01preglowjust ditch everything else, i'd love ipod support
01:32:17 Quit bagawk ("Leaving")
01:47:01 Nick Lost-ash is now known as ashridah (
01:59:08preglowimplemented the jump table idea in coldfire.S, and now it just emits static bursts for higher orders...
02:00:53 Quit linuxstb ("CGI:IRC")
02:02:15 Nick linuxstb_ is now known as linuxstb (
02:04:21linuxstbThe "fixed point" wma decoder still includes -lm in the makefile and -msoft-float. So it's still going to need a lot of work I think.
02:04:43preglowlibm we haven't got, so that's just ok
02:05:18linuxstbThey obviously _just_ did enough to get it working realtime on their platform.
02:08:24preglowi used to have just one malfunctioning lpc_decode loop, now i have two
02:08:38preglowso much for a little pleasureable diversion from the 64 bit version
02:09:11linuxstbA PhotoBridge has a 300MHz MIPS CPU and 64MB of DDR DRAM. Slightly better than our irivers I feel.
02:09:19preglowyes, slightly...
02:09:25preglowbut ok, i've unrolled loops up to order = 10 now
02:09:36linuxstbAnd WMA still only runs at twice realtime...
02:09:39preglowthe default case should also be slightly faster now
02:09:45linuxstbCool, thanks.
02:09:54preglowbut i just need to get it working again :P
02:10:14preglowlike you said, theres the floating point still
02:10:20preglowand they probably haven't got iram
02:10:33preglowbut i think wma is pretty close to aac in terms of complexity
02:14:21 Nick AliasAle is now known as AliasCoffee (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
02:19:30linuxstbTime for bed now. I should have the results of a new FLAC playback test in the morning - I'm curious to see how much time we gained by putting the code in IRAM.
02:19:45linuxstbI'll then have to do another test when you finish your EMAC routines...
02:21:00 Join preglowII [0] (
02:21:09preglowIIseems my irc box fell on its face
02:21:16preglowIIsounds good, linuxstb
02:21:18preglowIIsee you later
02:22:21MoospreglowII: the bugs related by webguest about MPC files are reproductable :(
02:23:20MoosI just test with my MPC files, and right, bitrate info don't here+scrolling for next track info too
02:26:31preglowi'll fix the bitrate thing now
02:26:45Moosgoodie :)
02:26:51preglowbut as for the next track info, i have no idea why that happens
02:26:52 Quit preglowII ("CGI:IRC (EOF)")
02:28:08Moosparser problems maybe, no? scuse I know nothing about codecs things :)
02:28:58preglowahh, sorry
02:29:05preglowthe bitrate thing isn't as easily fixable as i thought
02:29:18Moosreally? :(
02:29:19preglowit'll have to wait until i can make a proper metadata parser for musepack
02:29:41MoosI suspected this :(
02:29:57preglowmpc doesn't have a proper metadata parser yet
02:30:01preglowjust apev2 parser
02:30:07preglowand bitrate isn't stored in the ape tag
02:30:09Moosyou planed to do it yourself?
02:31:00preglowi can do it right now if you don't mind that only sv7 files will be supported
02:31:39Moos??? sorry :)
02:32:31preglowmusepack files exist in several formats
02:32:34preglowsv7 is the newest one
02:32:43Moosah ok
02:32:57Moosmine are not so young
02:33:19Moosbut if you can start with it, it could be fine
02:33:43MoosI like Musepack :)
02:34:59preglowyes, i like musepack myself
02:35:38preglowi'll try to make a parser now
02:35:59Moosgood to hear
02:51:30preglowthe sv7 bitstream specification is written in german
02:51:38preglowthey persist in making it easy for me
03:07:20 Join yngwi [0] (
03:07:25 Quit yngwi_away (Read error: 104 (Connection reset by peer))
03:07:39Moostime to sleep here
03:08:06Moospreglow: good luck for your MPC parser and good night
03:08:19preglowbitrate isn't stored in the header
03:08:30*preglow gives musepack a round of applause and goes to bed himself
03:09:02preglowso bitrate info in the wps isn't looking good
03:09:10preglowmusepack sure as hell isn't making it easy on us
03:09:19 Part darkskiez
03:09:25preglowbut bed now
03:09:31Moosgood night
03:09:38 Quit Moos ("Glory to Rockbox")
03:22:41***Saving seen data "./dancer.seen"
03:59:27 Quit tvelocity ("Leaving")
04:04:43 Join MN [0] (
04:12:03 Quit Slasher (
04:12:03 Quit markun (
04:12:03 Quit ghode|afk (
04:12:03 Quit dwihno (
04:12:03 Quit linuxstb (
04:12:03 Quit ashridah (
04:12:03 Quit blackax (
04:12:03 Quit amiconn (
04:12:03 Quit solexx_ (
04:12:03 Quit Seed (
04:12:03 Quit Zagor (
04:12:03 Quit wireddd (
04:12:03 Quit merbanan (
04:12:03 Quit pill (
04:12:03 Quit HCl (
04:12:03 Quit crash_ (
04:12:03 Quit yngwi (
04:12:03 Quit RotAtoR (
04:12:03 Quit Strath (
04:12:03 Quit goa (
04:12:03 Quit Bagder (
04:12:03 Quit crwl (
04:12:03 Quit ]Maggie[ (
04:12:03 Quit MN (
04:12:03 Quit _Nilisco (
04:12:03 Quit bluebrother^ (
04:12:03 Quit godzirra (
04:12:03 Quit thegeek_ (
04:12:03 Quit gromit` (
04:12:03 Quit novimon_ (
04:12:03 Quit ze (
04:12:03 Quit phaedrus961 (
04:12:03 Quit preglow (
04:12:03 Quit AliasCoffee (
04:12:03 Quit mbr (
04:12:03 Quit Lynx_ (
04:12:03 Quit Ismo (
04:12:21 Quit Rick (Success)
04:12:25 Quit CoCoLUS (Read error: 104 (Connection reset by peer))
04:13:24 Quit elinenbe (Read error: 104 (Connection reset by peer))
04:13:32 Quit Maxime (Nick collision)
04:13:32 Join CoCoLUS [0] (
04:13:32 Join Maxime [0] (
04:13:32 Join Rick [0] (i=rick@unaffiliated/Rick)
04:13:32 Join elinenbe [0] (
04:13:32NJoinMN [0] (
04:13:32NJoinyngwi [0] (
04:13:32NJoinRotAtoR [0] (
04:13:32NJoinlinuxstb [0] (
04:13:32NJoin_Nilisco [0] (
04:13:32NJoinashridah [0] (
04:13:32NJoinAliasCoffee [0] (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
04:13:32NJoinbluebrother^ [0] (
04:13:32NJoinSlasher [0] (
04:13:32NJoinmarkun [0] (
04:13:32NJoingodzirra [0] (
04:13:32NJointhegeek_ [0] (
04:13:32NJoinblackax [0] (
04:13:32NJoinamiconn [0] (
04:13:32NJoinghode|afk [0] (
04:13:32NJoinsolexx_ [0] (
04:13:32NJoinwireddd [0] (
04:13:32NJoinStrath [0] (
04:13:32NJoinSeed [0] (
04:13:32NJoinZagor [0] (n=bjst@pdpc/supporter/sustaining/Zagor)
04:13:32NJoingoa [0] (
04:13:32NJoindwihno [0] (n=dw@
04:13:32NJoinpreglow [0] (
04:13:32NJoinBagder [0] (
04:13:32NJoingromit` [0] (n=gromit`
04:13:32NJoinmbr [0] (
04:13:32NJoincrwl [0] (
04:13:32NJoinLynx_ [0] (
04:13:32NJoinnovimon_ [0] (
04:13:32NJoin]Maggie[ [0] (
04:13:32NJoinze [0] (
04:13:32NJoinIsmo [0] (
04:13:32NJoinphaedrus961 [0] (
04:13:32NJoinmerbanan [0] (
04:13:32NJoinpill [0] (
04:13:32NJoinHCl [0] (
04:13:32NJoincrash_ [0] (
04:16:36 Quit Slasher (
04:16:36 Quit markun (
04:16:36 Quit dwihno (
04:16:36 Quit ghode|afk (
04:16:36 Quit HCl (
04:16:36 Quit merbanan (
04:16:36 Quit Seed (
04:16:36 Quit amiconn (
04:16:36 Quit blackax (
04:16:36 Quit ashridah (
04:16:36 Quit linuxstb (
04:16:36 Quit elinenbe (
04:16:36 Quit solexx_ (
04:16:36 Quit wireddd (
04:16:36 Quit crash_ (
04:16:36 Quit Zagor (
04:16:36 Quit pill (
04:16:36 Quit ]Maggie[ (
04:16:36 Quit crwl (
04:16:36 Quit RotAtoR (
04:16:36 Quit yngwi (
04:16:36 Quit Strath (
04:16:36 Quit goa (
04:16:36 Quit Bagder (
04:16:36 Quit phaedrus961 (
04:16:36 Quit novimon_ (
04:16:36 Quit bluebrother^ (
04:16:36 Quit _Nilisco (
04:16:36 Quit MN (
04:16:36 Quit Rick (
04:16:36 Quit Maxime (
04:16:36 Quit godzirra (
04:16:36 Quit thegeek_ (
04:16:36 Quit ze (
04:16:36 Quit gromit` (
04:16:36 Quit preglow (
04:16:36 Quit AliasCoffee (
04:16:36 Quit Ismo (
04:16:36 Quit mbr (
04:16:36 Quit CoCoLUS (
04:16:36 Quit Lynx_ (
04:40:43 Quit TiMiD (Read error: 104 (Connection reset by peer))
04:42:24 Quit pabs (Nick collision)
04:42:25 Join Lost-ash [0] (
04:42:25NJoinIsmo [0] (
04:42:25NJoinLynx_ [0] (
04:42:25NJoinmbr [0] (
04:42:25NJoinAliasCoffee [0] (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
04:42:25NJoinCoCoLUS [0] (
04:42:25 Join pabs [0] (
04:42:25NJoinSlasher [0] (
04:42:25NJoinmarkun [0] (
04:42:25NJoinpill [0] (
04:42:25NJoinHCl [0] (
04:42:25NJoincrash_ [0] (
04:42:25NJoinmerbanan [0] (
04:42:25NJoinphaedrus961 [0] (
04:42:25NJoinze [0] (
04:42:25NJoin]Maggie[ [0] (
04:42:25NJoinnovimon_ [0] (
04:42:25NJoincrwl [0] (
04:42:25NJoingromit` [0] (n=gromit`
04:42:25NJoinBagder [0] (
04:42:25NJoindwihno [0] (n=dw@
04:42:25NJoingoa [0] (
04:42:25NJoinZagor [0] (n=bjst@pdpc/supporter/sustaining/Zagor)
04:42:25NJoinSeed [0] (
04:42:25NJoinStrath [0] (
04:42:25NJoinwireddd [0] (
04:42:25NJoinsolexx_ [0] (
04:42:25NJoinghode|afk [0] (
04:42:25NJoinamiconn [0] (
04:42:25NJoinblackax [0] (
04:42:25NJointhegeek_ [0] (
04:42:25NJoingodzirra [0] (
04:42:25NJoinbluebrother^ [0] (
04:42:25NJoinashridah [0] (
04:42:25 Join Nilisco [0] (
04:42:25NJoinlinuxstb [0] (
04:42:25NJoinRotAtoR [0] (
04:42:25NJoinyngwi [0] (
04:42:25NJoinelinenbe [0] (
04:42:25NJoinRick [0] (i=rick@unaffiliated/Rick)
04:42:25NJoinMaxime [0] (
04:42:25NJoinpreglow [0] (
04:43:21 Quit markun (
04:43:40 Join TiMiD [0] (
04:43:44 Join Naked [0] (
04:44:51 Quit RotAtoR ()
04:45:12CtcpVersion from freenode-connect!freenode@freenode/bot/connect
04:47:29NJoinmarkun [0] (
04:48:12 Quit Hadaka (Connection refused)
04:48:13 Nick Naked is now known as Hadaka (
04:52:40 Quit ashridah (Nick collision from services.)
04:53:34DEBUGEOF from server (Success) (snapshot: netstuff.c line 545)
04:53:34***Saving seen data "./dancer.seen"
04:53:34***Started Dancer V4.16
04:53:34DEBUGconnect(2) failed on socket 3 (Connection refused) (snapshot: netstuff.c line 150)
04:53:34DEBUGconnect(2) failed on socket 3 (Connection refused) (snapshot: netstuff.c line 150)
04:53:34DEBUGconnect(2) failed on socket 3 (Connection refused) (snapshot: netstuff.c line 150)
04:53:34***Connected to on port 6667
04:53:34***Logfile for #rockbox started
04:53:41***Server message 501: 'logbot :Unknown MODE flag'
04:53:41Mode"logbot :+i" by logbot
04:53:41***Server message 487: 'logbot :Due to a network split, you can not obtain channel operator status in a new channel at this time'
04:53:41Mode"#rockbox +ns" by
04:53:41Mode"#rockbox +t-s " by
04:53:41 Join markun [0] (
04:53:41 Join TiMiD [0] (
04:53:41 Join Lost-ash [0] (
04:53:41 Join Ismo [0] (
04:53:41 Join Lynx_ [0] (
04:53:41 Join mbr [0] (
04:53:41 Join AliasCoffee [0] (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
04:53:41 Join CoCoLUS [0] (
04:53:41 Join pabs [0] (
04:53:41 Join Slasher [0] (
04:53:41 Join pill [0] (
04:53:41 Join HCl [0] (
04:53:41 Join crash_ [0] (
04:53:41 Join merbanan [0] (
04:53:41 Join phaedrus961 [0] (
04:53:41 Join ze [0] (
04:53:41 Join ]Maggie[ [0] (
04:53:41 Join novimon_ [0] (
04:53:41 Join crwl [0] (
04:53:41 Join gromit` [0] (n=gromit`
04:53:41 Join Bagder [0] (
04:53:41 Join dwihno [0] (n=dw@
04:53:41 Join goa [0] (
04:53:41 Join Zagor [0] (n=bjst@pdpc/supporter/sustaining/Zagor)
04:53:41 Join Seed [0] (
04:53:41 Join Strath [0] (
04:53:41 Join wireddd [0] (
04:53:41 Join solexx_ [0] (
04:53:41 Join ghode|afk [0] (
04:53:41 Join amiconn [0] (
04:53:41 Join blackax [0] (
04:53:41 Join thegeek_ [0] (
04:53:41 Join godzirra [0] (
04:53:41 Join bluebrother^ [0] (
04:53:41 Join Nilisco [0] (
04:53:41 Join linuxstb [0] (
04:53:41 Join yngwi [0] (
04:53:41 Join elinenbe [0] (
04:53:41 Join Rick [0] (i=rick@unaffiliated/Rick)
04:53:41 Join Maxime [0] (
04:53:41 Join preglow [0] (
04:53:41Mode"#rockbox +bbb *!*@* *!*ju_@* *!*icechat*@* " by
04:53:41Mode"#rockbox +bbb *!* *!*jirc*@* *!*Yu-Gi-oH@* " by
04:53:41Mode"#rockbox +bbb *!*user@82.67.68.* *!* *!* " by
04:53:41DEBUGParse error in OnTopic(from = "", line = "#rockbox :Rockbox 2.5 | New developers mailing list!") (snapshot: server.c line 299)
04:53:41 Join logbot [0] (
04:53:43CtcpVersion from freenode-connect!freenode@freenode/bot/connect
04:53:43***Server message 477: 'logbot #rockbox :[freenode-info] if you need to send private messages, please register:'
04:53:49 Join lostlogic [0] (
04:54:25 Nick Lost-ash is now known as ashridah (
04:56:43 Quit AliasCoffee ("Leaving")
04:57:54 Join webguest82 [0] (
04:58:34 Join Naked [0] (
04:58:50 Nick Naked is now known as Hadaka (
05:00:42 Quit webguest82 (Client Quit)
05:43:09 Join XShocK [0] (
06:05:22 Quit ashridah ("phone")
06:12:01 Join ashridah [0] (
06:53:35***Saving seen data "./dancer.seen"
07:17:34 Quit blackax (Read error: 110 (Connection timed out))
08:19:38 Join solexx [0] (
08:31:31 Quit solexx_ (Read error: 110 (Connection timed out))
08:48:00 Join B4gder [0] (
08:48:05 Join _FireFly_ [0] (
08:53:38***Saving seen data "./dancer.seen"
09:01:54 Join einhirn [0] (
09:08:45*B4gder had to use the paperclip today
09:17:38 Join LinusN [0] (
09:17:50LinusNB4gder: congratulations to the paperclip experience
09:18:16B4gderslightly annoying since I had none in my pocket on the train
09:18:34B4gderbut hey, who leaves home without a paperclip?
09:18:47LinusNhow did it freeze?
09:18:55B4gderon NEXT
09:19:03B4gderI skipped forwared a few songs
09:19:13B4gderand then freeeze
09:19:31B4gderbut I'm not running the very lateset
09:19:52B4gderyeah, I'm such a looser
09:19:55LinusNyou're a shame to all the devs
09:20:11*B4gder stands in the corner
09:20:30LinusNi sentence you to 2 weeks with "follow playlist" enabled
09:20:48*amiconn wants a working recovery from I0C: Debug, but that means to make UIE() an asm routine
09:21:03amiconn...calling a C function for display etc
09:21:42amiconnThe pb is that it will break building with gcc4 for the same reason as on archos
09:22:21amiconn(weak) alias to a symbol defined in the asm block and hence invisible to gcc
09:22:41amiconnOtoh, the implementation would be quite a bit cleaner
09:23:08amiconnAtm, UIE() relies on a specific number of registers to be saved on the stack
09:23:33amiconnLet the compiler decide to some register allocation different, and it will break
09:24:02amiconn(talking about system.c lines 436, 437)
09:28:33B4gdernow there's a normal thing to spend your time on
09:28:36 Quit ashridah ("Leaving")
09:30:32B4gderworld record on normal Rubik cube solving seems to be 11.51 seconds
09:30:42B4gderand here we are, wasting our time on Rockbox instead ;-)
09:32:34 Join _FireFly__ [0] (
09:34:47 Join ender` [0] (i=ychat@
09:37:21LinusNwe should extend it
09:40:45ender`make it support .plg files :)
09:44:36 Join linuxstb_ [0] (
09:45:39linuxstb_Morning all. I'm confused - the FLAC decoder managed 12h 38m in my first test. I tested it again last night (after my change to move to all the code into IRAM), and it managed 12h 32m of playback.
09:46:32linuxstb_But the decoder is definitely faster - it can now play back 24-bit FLAC files in realtime. Previously it couldn't.
09:47:43_FireFly__maybe your measurement methode is inaccuracy
09:48:37 Quit _FireFly_ (Read error: 110 (Connection timed out))
09:50:28 Nick _FireFly__ is now known as _FireFly_ (
09:50:50 Join ashridah [0] (
09:51:53linuxstb_My methodology is simple - fully charge the player, play an album in repeat mode and count the hours until the player shuts down.
09:52:39_FireFly_who or what is counting the hours ??
09:52:53linuxstb_Me. The time is accurate.
09:53:11linuxstb_Both times, I've been staring at the player when it shuts down.
09:53:45_FireFly_i bet if you redo the test you will get another result
09:53:47B4gderthe charging might not be that accurate
09:55:27 Quit linuxstb_ ("CGI:IRC (EOF)")
10:04:12 Join linuxstb_ [0] (
10:04:38linuxstb_OK, I'll try some more tests. But is it possible that no power (or negligible power) is saved when the CPU is in idle sleep mode?
10:05:02linuxstb_The FLAC decoder was alredy running at a constant 45MHz.
10:05:19B4gderit'll consume the same while in idle
10:06:10B4gderwe'd have to lower the mhz bar for it to get noticable
10:06:58Slasherlinuxstb_: It will save something like 20 mA when cpu is idle (sleep(1), not yield)
10:07:09linuxstb_The 6 minute difference can be explained by Rockbox simply managing to fill the buffer once (or maybe twice) more time during the first test compared to the second.
10:07:34B4gderlinuxstb_: that, and the fact that you don't know if the battery was exactly as much charged both times
10:08:01B4gdercharging levels being a shaky business
10:08:14linuxstb_B4gder: That's what I'm saying. There was probably slightly more charge the first time.
10:08:51linuxstb_I wouldn't expect identical runtimes in each test, but I was expecting this to be an improvement.
10:09:06B4gderand 6 minutes diff is a mere 0.8%
10:09:49linuxstb_Slasher: Will rockbox be calling sleep() very often during audio playback?
10:10:21linuxstb_Obviously, if the codec is fast enough.
10:10:37Slasherlinuxstb_: it should when the pcm buffer is full. But if any other thread is not calling sleep too, then it wont work
10:12:53LinusNrockbox is pretty busy during playback, with wps and all
10:13:08LinusNso i don't expect it to sleep much
10:14:08linuxstb_Is there anything we could do to make Rockbox less busy? e.g. disable the WPS display?
10:16:03linuxstb_If I put Rockbox in file browser mode (instead of WPS mode), could that potentially increase playback time?
10:17:27LinusNlinuxstb: yes i think so
10:19:36linuxstb_In any case, it seems more tests are needed.
10:19:44_FireFly_but then a test wouldn't be a real test because many users will have wps enabled during playback
10:20:33linuxstb__FireFly_: It's just a theoretical test. If disabling WPS during playback significantly increases playback time, then we could possibly change how Rockbox behaves.
10:23:51LinusNthe wps is a lot more busy nowadays, after we added the bmp support
10:24:10LinusNand i guess it's busy even if no bmp's are loaded
10:24:32linuxstb_I do have the peakmeter enabled on my WPS.
10:25:18_FireFly_i think the peakmeter is a other element which needs some cpu power
10:26:09linuxstb_Looks like more test cases - no WPS, WPS with peakmeter, very simple WPS
10:26:50_FireFly_bmp using wps
10:26:54 Quit linuxstb_ ("CGI:IRC")
10:31:20linuxstbLooks like RealAudio (the "cook" codec) has been re'ed and will soon be in ffmpeg:
10:32:16linuxstbBut it's floating point, so no use to Rockbox, even if someone wanted to include it.
10:52:10 Join DJDD [0] (
10:52:26DJDDLinus you there?
10:53:39***Saving seen data "./dancer.seen"
10:54:57 Quit yngwi ("Chatzilla 0.9.68a [Firefox 1.0.7/20050919]")
10:55:57LinusNDJDD: yup
10:56:47LinusNdjdoomsday, i presume?
10:58:04DJDDCould i grab your undivided attention for a little while to discuss my defunct player?
10:58:59 Join cYmen [0] (
10:59:38DJDDI think your right about the CPU, i managed to boot into the iriver firmware and its pretty screwed. There are artifacts with the graphics everywhere and its EXTREMELY slow. As in you'd call it frozen, but i could navigate the folder tree a few times.
11:00:24LinusNis it hot when this happens?
11:00:42DJDDThough i'm wondering that that may have have been the cause of a lot of heat on the CPU. I had to plug in the charger for it to boot as the battery is dead i think (it doesn't seem to be able to charge very much)
11:00:58DJDDIt only takes 30 seconds for the to get hot
11:01:24LinusNok, a hot cpu probably will act up as you describe it
11:01:42DJDDgrand, so we could call the CPU undamaged, for now
11:01:46LinusNmy suggestion is not to try to run it again
11:02:11DJDDYeah, im not going to. I managed to turn it off via the reset+hold button
11:02:19DJDDI can't do it any other way
11:02:35LinusNthe most common scenario when a 3rd party charger is used is that the regulator breaks down and delivers too high voltage to the cpu
11:02:41DJDDI just read the warrenty info on the iriver website, i doubt i'd get anywhere there
11:03:05LinusNmost players i have examined have had stone dead cpu's
11:03:20DJDDOnly the charger is rated pretty low, maybe he CPU could have survived
11:03:26LinusNyour is still alive, so there is still a chance
11:04:12 Join cYmen_ [0] (
11:05:09DJDDSo what would you advise now? replacing the regulator?
11:05:11ashridahvisual artifacts and whatnot sounds more like screwed memory to me.
11:05:21LinusNmy autopsy of the dead players has also shown that it isn't always the regulator itself that breaks
11:05:23ashridahthat type of thing happened to me when one of my videocards failed.
11:05:31ashridah(of course, that's a pc, completely different hardware)
11:05:37DJDDPerminantly screwed memory or an effect of a lot of heat?
11:06:06ashridahLinusN: i'm mildly surprised they didn't toss in enough circuitry to short it to ground on overvoltage and save the hardware
11:06:08DJDDAsh, yes but that would have been your video cards CPU. It happens when you overclock them and they overheat
11:06:14LinusNi had the same effect when i overheated my iriver when running at 140MHz
11:06:21linuxstbThere was someone in this channel a few days ago with a spare H1?0 mainboard. Can't remember his/her name though.
11:07:03amiconnImho the iriver protection against some mistakes is pretty lame
11:07:19LinusNeven more lame that archos
11:07:42LinusNarchos still has no protection against reversed polarity
11:07:44DJDDLinus, what do you mean? That the entire power supply circuitry on my iriver could be fried? Not just the reg chip?
11:07:45ashridahmy electronics is sketchy after this long, but wouldn't a simple zener diode with a breakdown voltage a few volts above the charger's input done the trick?
11:07:49amiconnI'd never expect that too high/reversed voltage will kill the whole player
11:07:55LinusNamiconn: true
11:07:58ashridahsure, it would have blown a fuse higher up the chain, but hell, who uses actual fuse wire fuses anymore? :)
11:08:03LinusNthe archos only kills the regulator
11:08:17amiconn...especially since they use a standard socket for dc in
11:08:19DJDDLinus, i was thinking just before that reversed polarity is probably what did the damage
11:08:33LinusNnokia charger?
11:08:39DJDDyep :P
11:08:48amiconnThe HD shock protection is another lacking area
11:09:08LinusNi'd say that the iriver drive is pretty protected
11:09:16amiconnI disagree
11:09:23 Join cYmen__ [0] (
11:09:38LinusNthe 1.8" drives themselves can withstand quite a shock, iirc
11:09:39amiconnI dropped my H140 to the floor *once*, while it was *off*
11:09:44DJDDSo where should i start on fixing this thing? If at all possible?
11:09:59amiconnThe HD still works, but became rather loud from this one occasion
11:10:05LinusNamiconn: wow
11:10:27amiconnIt didn't fall that hard, just from the table (75 cm) to the floor (carpet)
11:10:31LinusNDJDD: do you have any experience in electronics?
11:10:33solexxhm. i have dropped my h120 several times now
11:10:39LinusNme too
11:10:46solexxbut not from a great height
11:10:59LinusNamiconn: you have been very unlucky
11:11:14DJDDLinus, i have experience with soldering my own circuit boards and playing the resistors and what not. So yeah, i have a basic knowledge
11:11:32DJDDnever heard of it
11:11:36amiconnI dropped my archos recorder several times (not that often though), also to concrete ground
11:11:40LinusNSurface Mounted Device
11:11:44amiconnThe HD still works fine
11:12:32DJDDLinus, nop, never heard of it.
11:12:46LinusNDJDD: then you won't be able to fix it yourself
11:12:55LinusNthose chips are *tiny*
11:13:15amiconnSince my H140 'event', I try to keep it in the leather case most of the time, for a little more protection
11:13:29LinusNme too
11:13:48DJDDlol. Ah, you mean those chips that have a sort of holder you solder on then plug the chip into?
11:13:50amiconnToo bad you can't reach the reset button without taking it out
11:14:36 Join cYmen___ [0] (
11:14:36***Alert Mode level 1
11:14:36DBUGEnqueued KICK cYmen
11:14:36DBUGEnqueued KICK cYmen_
11:14:36***Alert Mode level 2
11:14:36DBUGEnqueued KICK cYmen__
11:14:36DBUGEnqueued KICK cYmen___
11:14:36***Alert Mode level 3
11:15:07DJDDWell surely i could take it to a professional electrician, or a friend that i think might know how. I think it'd be harder to FIND the replacement chip
11:15:30ashridahDJDD: no, he means tiny resistors and capacitors that are soldered onto the surface of a board, instead of using a pad with a hole in it.
11:15:35ashridahvery delicate to resolder
11:15:53ashridahbut very easyfor a machine to place
11:16:06DJDDI was gonna say, is that even doable by hand?
11:16:12thegeek_oh yes
11:16:28ashridahyou stick them down with non-conductive glue and solder paste, and use things like wave soldering, where you just pass the entire board over a wave of solder in a bath, and everything solders near-instantly. :)
11:16:36ashridahDJDD: yes, it is
11:16:44 Quit cYmen (Connection timed out)
11:16:48LinusNi do it by hand, but with a tiny soldering tip and a magnifying glass
11:17:07DJDDok, so could you guys help me find the chip and i'll find the hands to do it
11:17:32DJDDSome instructions for the unknown hands would probably be grand too
11:17:53LinusNDJDD: digikey, part # LTC3405ES6-ND
11:18:24thegeek_my audio-out is broken
11:18:26DJDDI'm from australia, local would be good :P
11:18:39thegeek_is it a known part?
11:18:55LinusNthegeek_: the connector?
11:19:17DJDDLinus have you ordered from these guys before?
11:19:18thegeek_the metal connectors inside are loose
11:19:24thegeek_digikey is good
11:19:35LinusNthegeek_: the connector looks pretty standard to me
11:19:44thegeek_that is what I thought too
11:19:46 Join cYmen [0] (
11:19:46***Alert Mode level 4
11:19:46***Alert Mode level 5
11:19:46DBUGEnqueued KICK cYmen
11:19:46***Alert Mode level 6
11:19:46***Alert Mode level 7
11:19:46***Alert Mode level 8
11:19:53LinusNDJDD: yes, i bought my regulators from them
11:19:53thegeek_but I could not find anything exactly like it
11:20:05thegeek_I checked both digikey and mouser
11:20:30thegeek_I could find comparable parts but nothing with the same "pin" placement
11:20:46DJDDI dont think DigiKey ships outside the US
11:20:56LinusNthey do, but it's expensive
11:21:00DJDDscrub that
11:21:44LinusNthegeek_: have you opened your player?
11:21:45DJDDexpensive? Its like $25 fuckin bucks. For something that really doesn't wiegh anything
11:21:47DJDDThats insane
11:22:00thegeek_several times;)
11:22:01 Quit cYmen_ (Connection timed out)
11:22:42DJDDOk, you think there'd be any local place that'd stock that sort of part?
11:23:50LinusNDJDD: maybe, do a search
11:24:11DJDDi did a google on the exact part name and nothing
11:24:32DJDDI think i need some good keywords
11:26:57 Quit cYmen__ (Connection timed out)
11:28:28LinusNDJDD: LTC3405 is the official part number
11:28:47DJDDLinus i could this, is the part there? (sorry, i can't make sence of it all):
11:29:47***Alert Mode OFF
11:32:12 Quit cYmen___ (Connection timed out)
11:33:03DJDDmmm, i don't think theres much hope for a local shop. Ok, so i go buy one from bloody digikey. What are the chances that it'll actually work?
11:33:35 Quit _FireFly_ (Read error: 110 (Connection timed out))
11:34:07LinusNDJDD: from my experience, less than 50%
11:35:13DJDDheh, i've never had good luck
11:35:24DJDDi doubt it'll coem good here
11:35:50DJDDAny other alternatives? What about the fellow with the spare mainboard?
11:38:02linuxstbDJDD: He came to IRC anonymously - 15:00 in
11:40:22DJDDway, i think i know who he is. I was reading a thread before with a guy say he had a iriver minus the HDD and the lcd was screwed
11:41:22 Join DangerousDan [0] (
11:42:14 Join ep0ch [0] (n=ep0ch@
11:42:36ep0chhow do you disable the WPS?
11:42:47LinusNep0ch: exit it
11:42:55ep0chohh obviously :)
11:43:38DJDDHas anyone tryed messing with iriver aobut the warranty? You think you could get by without proving when you bourght it?
11:44:08ep0chif the WPS does indeed consume a lot of battery, it would be good if there was a timer to switch it off, and also a timer to switch off the LCD
11:44:18DJDDI've returned a few things in the past and never have i produced proof of purchase
11:44:53ep0chany idea how much power the LCD consumes?
11:48:46linuxstbep0ch: It's only a theory that disabling the WPS will increase playback time. Firstly, it will depend on the codec being used working faster than realtime at 45MHz. If it doesn't, then the WPS will probably not make much difference.
11:49:01DJDDfound the fellow with the iriver parts :D
11:49:13CtcpIgnored 5 channel CTCP requests in 5 minutes and 10 seconds at the last flood
11:49:13*LinusN is quite upset by len0x last post in the mailing list
11:49:27ep0chwell, would be nice to switch off the LCD anyhow
11:50:27linuxstbep0ch: I agree, assuming we save power by doing it, I think a screensaver-like timeout would be a good thing. But I want to do some more runtime tests to see what difference it could make.
11:51:58DJDDLinus: You don't think we could be overlooking the battery do you? What reason do you have that my battery isn't dead?
11:52:57markunLinusN: What a crazy guy..
11:53:27markunMaybe he should start his own fork
11:54:48LinusNDJDD: your battery might be dead as well...
11:55:12DJDDBut could that be the sole cause of it?
11:55:20LinusNi doubt it
11:56:08DJDDAlso, how long has the Hxxx range been discontinued? Im wondering whether there are some still lying around in stores
11:57:18LinusNi gotta run
11:57:27amiconnep0ch: The lcd power consumption should be <1 mA, almost neglectible. (B&w lcd that is. H300 will be an entirely different matter)
11:57:27LinusNcu in an hour or so
11:57:34 Part LinusN
11:58:48ep0chis that all!
11:59:09ep0chnice and effecient
12:07:11linuxstbMy understanding is that the display of the WPS could be preventing the CPU from sleeping. When the CPU is sleeping, we are saving 20mA (according to Slasher - I don't know).
12:08:36Slasherlinuxstb: yes, i measured something like that difference when cpu was sleeping and decoding
12:09:02linuxstbAs a comparison, do you know what power does the CPU use at 45MHz?
12:09:04Slasherbut can't measure now again without opening the unit first
12:09:21Slasherlinuxstb: it should be somewhere in the wiki, just a moment
12:10:33SlasherHmm, it was almost 20 mA (still more than 10)
12:11:06linuxstbThanks. Just what I was looking for.
12:13:32amiconnWe can also lower the default frequency if the majority of codecs is content at 45 MHz
12:13:44amiconnNext lower possible value would be 34 MHz
12:14:30linuxstbHow much work is it to change the frequency? I'm guessing it's not trivial?
12:14:42linuxstbI mean programming time.
12:15:15amiconnPerhaps half an hour
12:15:34amiconnMost of the work is datasheet lookup, and some calculation
12:15:45 Join Moos [0] (
12:15:49linuxstbI know wavpack and FLAC are comfortable at 45MHz. But I don't know about the other codecs.
12:16:34linuxstbALAC definitely isn't - it needs about 40%-50% boost.
12:16:36amiconnIt's necessary to calculate the PLLCR value and the refresh timing
12:16:54amiconn(the latter for both h100 and h120)
12:17:21linuxstbI would just be curious to know how low we can change the CPU and still get 0% boost for FLAC (and other codecs).
12:17:46linuxstbBut I would be very surprised if they didn't boost at 34MHz.
12:36:32 Join XavierGr [0] (
12:36:52XavierGrHi all!
12:42:39preglowamiconn: i dropped my h120 1 meter onto a wooden floor (no protective case), with no ill effects
12:44:56ep0chmine also fell onto a wooden floor while it was playing, the harddisk slowly died a miserable death. :(
12:45:35preglowhas this len0x fellow been aboard long?
12:47:29XavierGrwhen did he get cvs access?
12:47:37B4gdera couple of days ago
12:49:57preglownice couple of first moves, then
12:50:02preglowgetting on everyones good side
12:51:36 Join muesli- [0] (n=muesli_t@
12:52:51XavierGreven in he has some valid points about his commitment, I don't find his way of expressing them a bit harsh...
12:53:22XavierGr^remove the "don't from previous statement
12:53:41***Saving seen data "./dancer.seen"
12:56:57XavierGrI want to think of it as a misunderstanding and not bad intention.
12:57:42 Quit DangerousDan ("Miranda IM! Smaller, Faster, Easier.")
13:03:18preglowplus some slight bad intention
13:05:11preglowtried starting on a musepack metadata parser yesterday
13:05:20preglowthen i gave up when i notioed nothing of importance is stored there...
13:05:33preglowi wonder how foobar etc. finds the bitrate
13:06:31XavierGrIs the playslist scrollbar the same as filetree scrollbar?
13:06:53XavierGrBecause it has a different response from the one in filetree
13:07:31XavierGrWhen I hold up/down the scrolling will not stop on the upper/bottom entry as in filetree scrollbar, is that intended?
13:08:27 Join amiconn_ [0] (
13:10:41 Quit amiconn (Nick collision from services.)
13:10:41 Nick amiconn_ is now known as amiconn (
13:11:23amiconnXavierGr: Afaik, it's not the same atm. The filetree uses TiMiD's new gui list, while the playlist viewer doesn't (yet)
13:11:41amiconnI'm not sure though, as I seldom use the playlist viewer
13:12:03 Join B4gd3r [0] (
13:12:28XavierGrI think that the playlist viewer is quite good, especially when rockbox is playlist oriented.
13:13:32XavierGrAlso I wonder why the playlist viewer is so deep in the settings? I will try to see if I can shortcut it to rec button for my own pleasure.
13:13:47amiconnYes, but usually I know the album that I started
13:13:53 Quit B4gder (Read error: 110 (Connection timed out))
13:14:03amiconnThe playlist viewer isn't buried deep. Check the wps context menu
13:14:32XavierGrthe viewer has more uses when you insert new tracks, or you want to make a new playlist. It is very intuitive IMHO
13:15:58XavierGrwow I am now suprised and impressed!!
13:16:17XavierGrI thought that I had found all possible Rockbox screens....
13:17:38XavierGrI didn't even know the existance of a context menu in wps quite handy I must say, thanks
13:18:58 Nick B4gd3r is now known as B4gder (
13:19:05amiconnI believe that even I don't know all places in rockbox, being a project member for almost 2 years now...
13:20:17linuxstbpreglow: bitrate is normally calculated as ("filesize in bytes" * 8) / "length in ms". Not many codecs store it explicitly in the headers.
13:20:42linuxstbSo to calculate the bitrate you need to determine the length in ms.
13:20:51preglowah, right
13:20:53preglowso i'll do the parser
13:21:17preglowi'll just rip the relevant parts out of libmusepack
13:21:35linuxstbOr you could just set it in mpc.c for now.
13:21:40preglowdoesn't work
13:21:49preglowseems it's only read once
13:21:55preglowforget it
13:22:20preglowrockbox freezes when it tries to access next track data for musepack, they say
13:26:19preglowcould this be related to it not having a proper metadata parser?
13:27:06Moosand seeking doesn't work too :(
13:27:44preglowfor musepack?
13:27:53linuxstbpreglow: I don't see how. It should be no difference between current track and next track.
13:27:57Mooswhen seeking, Rockbox freeze need to stop playback after few seconds of HD freeze
13:28:00preglowmoos: that's not our fault, musepack has horrible seeking support. it works, it just takes three hours
13:28:02Moospreglow: yes
13:28:10Moosah ok
13:28:24linuxstbpreglow: Why is musepack seeking bad?
13:28:36preglowbecause of poor seeking support in the bitstream
13:28:38preglowno seeking table
13:28:43preglowand inter-frame-dependencies
13:28:56linuxstbSo everything to make seeking hard is there then.
13:28:59preglowthey plan on fixing that in sv7.5, it seems
13:29:12preglowSeed: you around?
13:30:18preglowi think musepack has potential to be really great on portables, too bad it's stuck in a rut like this
13:30:33linuxstbA nice thing IMO about .m4a is that every frame is indexed. So seeking is almost instant. There also don't appear to be any inter-frame dependencies in either AAC or ALAC.
13:31:37linuxstbAnd according to the MPEG-4 specs, MPEG-1, Layer III audio is legal inside the .mp4 container. But it's a bit late for that.
13:32:02 Quit XavierGr (Read error: 104 (Connection reset by peer))
13:32:18preglowlinuxstb: well, there is if you use prediction in aac
13:32:27preglowlinuxstb: but i think lc doesn't use that much at all
13:33:00preglowdependencies that just last a frame or two is ok
13:33:11preglowbut there's got to be something insane going on with musepack if seeking is THIS slow
13:33:18linuxstbIs that going to be a problem then? I haven't heard any artifacts when seeking in my aac test files.
13:33:40preglowlinuxstb: artifacts, no, but the frame doesn't sound the same
13:33:49preglowlinuxstb: this is the same with mp3 and its bitreservoir
13:34:02linuxstbBut I guess it's only a few frames, so it will be inaudible to normal people.
13:34:13amiconnlinuxstb: Wanna do an experiment?
13:34:14preglowyes, and the onset frame is a bit special
13:34:19preglownot much context to judge it by
13:34:23ep0chis aac running at realtime?
13:34:27preglowep0ch: far from it
13:34:27linuxstbamiconn: Sure.
13:34:42 Join LinusN [0] (
13:35:14amiconnlinuxstb: (1) In firmware/export/system.h, change line 283 to: #define CPUFREQ_NORMAL_MULT 3
13:35:58amiconn(2) In firmware/system.c, change line 576 to: #define NORMAL_REFRESH_TIMER 15 and line 580 to: #define NORMAL_REFRESH_TIMER 7
13:36:16Zagorwho the heck is this Anton guy?
13:36:36amiconn(3) Change line 611 to: PLLCR = 0x13846005;
13:36:59Zagorwhat did he write before he got commit access?
13:37:00amiconnlinuxstb: This should make the unit run at 34 MHz default frequency
13:37:33LinusNZagor: improvements on the next-dir feature, and some fm screen changes
13:37:37amiconnSmall remark: This is untested, I just calculated the values
13:37:59linuxstbamiconn: Thanks. Compiling now.
13:40:11 Join XavierGr [0] (
13:40:31LinusNand i agree with len0x that the user perspective is important
13:40:34linuxstbIf your calculations were correct, FLAC is happy at 34MHz :)
13:40:51linuxstbMy usual "-8" test album.
13:40:53ep0chnice, lower lower! :)
13:41:00B4gderLinusN: of course it is, no one said anything else
13:41:19ep0chyou tried it at 11mhz via the debug menu?
13:41:33linuxstbYes, but all users - not just ex-users of the iriver firmware.
13:41:33preglowdo i hear "codec dependent cpu frequency"? :PP
13:41:52B4gdera new option! ;-)
13:41:54preglownow don't tell me that wouldn't rock
13:41:58LinusNlinuxstb: that's the issue
13:42:47LinusNhowever, in this particular case (follow playlist), both archos and the iriver firmware behave the same by default
13:42:59LinusNthe original firmwares that is
13:43:06B4gderI wouldn't know
13:43:10linuxstb(lossless) Wavpack is boosting a little at 34MHz though.
13:43:40amiconnlinuxstb: We have two more options (if we don't want to give up the current timer handling): 23 MHz (22.5792 MHz) and 11 MHz ...
13:43:43LinusNi remember, because the follow playlist option was added because old archos users wanted rockbox to behave like the original archos firmware
13:44:00B4gderright, now when you mention it I recall that too
13:44:23LinusNthat was in the good old days, when we listened to our users :-)
13:44:40 Quit Maxime (Read error: 110 (Connection timed out))
13:44:50B4gderLinusN: you mean when "some" of us did? ;-P
13:45:01amiconnLinusN: Imho, follow playlist should be always on, but that's my personal opinion
13:45:10amiconnIt's way more logical behaviour
13:45:19B4gderit is?
13:45:26LinusNit is?
13:45:34amiconnI have it on at all times
13:45:39linuxstbAC3 is also boosting at 34MHz.
13:45:40preglowseems i've always had it on myself
13:45:45B4gderit would confuse me mucho
13:45:48preglowlinuxstb: well, small wonder there
13:45:56amiconnB4gder: How so?
13:45:56preglowlinuxstb: try stuffing bits of ac3 code in iram
13:45:59LinusNi often listen to shuffled playlists
13:46:01preglowlinuxstb: imdct especially
13:46:06XavierGrme too, I always have it on, but this it hink has nothing to do with logic but rather a personal preference
13:46:08linuxstbpreglow: Yes, I will.
13:46:09B4gderamiconn: because I view the browser and wps as separate entities
13:46:18LinusNand i would go nuts if the player tossed me around in the file tree
13:46:19B4gderI want the browser where it was when I left it
13:46:22Moosalways used here too :)
13:46:33linuxstbA discussion :)
13:47:04B4gderI use shuffled playlists almost exclusively
13:47:16linuxstbI have to say I also enable it - but I never use shuffle.
13:47:22preglowi never use them :>
13:47:37amiconnI seldom use shuffle, but even when I do, I think follow playlist is more logical
13:47:42Moostimt to time here
13:48:01B4gderI don't think any of the choices are the more "logical"
13:48:10amiconnI think of the current traclk being the current file, and the browser should be there when I switch to it
13:48:25Zagoramiconn: didn't we discuss your definition of "logical" a while back? ;-)
13:48:32MoosB4gder: Rockbox have his logic XD
13:48:35LinusN(silly example) i wouldn't want my File Explorer in windows follow the Media Player
13:48:36amiconnOtherwise I would (almost) always end up in the root, and that's most of the time not what I want
13:48:45B4gderamiconn: thats
13:48:50B4gderexactly what I want
13:49:12amiconnIf I would end up where I left, that would perhaps be okay, but that it doesn't
13:49:19XavierGrpersonal preference
13:49:41B4gderXavierGr: indeed, not "most logical"
13:50:03B4gderbut then
13:50:04XavierGrbut sometimes logic is 2 sided so... :)
13:50:16XavierGrlike history and so on
13:50:17B4gderI didn't argue all that on the list just about this specific issue
13:50:24B4gderit is more a policy thing
13:50:27XavierGrI know what you mean.
13:51:31preglowso, how to actually reach a decision on something that's only personal preference?
13:51:31ep0choh cool I didn't know about the "follow playlist" option.
13:51:31amiconnLinusN: Your example is indeed silly: The file explorer is not built *into* media player. For an inbuilt file explorer, I would surely expect it to follow
13:51:35preglowjust keep it the way it always was?
13:51:58B4gderpreglow: I wouldn't mind taking heated things to a vote
13:52:04B4gderall committers one vote each
13:52:07B4gdermajority wins
13:52:11amiconnB4gder: I agree 100% on the point that discussion is necessary
13:52:29preglowi, like amiconn, do in fact agree that playlist follow is a tiny bit logical
13:52:38B4gdervoting is done in many open source projects
13:52:47preglowbut then again, i can see why you wouldn't want it as well
13:53:08Zagorpreglow: is that why it's an option? ;)
13:53:30amiconnApart from that, I don't have much of a problem with either default value. Every user can set it the way he likes
13:53:31Moosdemocratic way! :)
13:53:32ep0chfrom a users perspecitive, it should be on by default, as it mimicks the original firmware
13:53:48B4gderwe don't mimic firmwares
13:54:03Zagorep0ch: which original firmware?
13:54:07B4gderI don't know how the original ones work
13:54:21Zagorep0ch: exactly. I don't use the ihp.
13:54:22ep0chhence the words "users perspective"
13:54:28B4gderI am a user
13:54:30preglowZagor: archos as well, they say
13:54:32linuxstbep0ch: So should we insert 2-second gaps between tracks?
13:54:36ep0chno you're a commiter :p
13:54:37XavierGrbut ep0ch rockbox is not intended only for iriver
13:55:10XavierGrIt would be silly to make if statements for default values in every single target.
13:55:18preglowbut hell, i don't care, seems i've had this enabled for yonks
13:55:25preglowB4gder: quick! code a voting mechanism!
13:55:45LinusNin the "follow playlist case", both archos and iriver behave the same
13:55:49B4gderI think we should cast votes by committing to a vote file
13:56:09ep0chi'll send in a patch ;)
13:56:10B4gder(its being done so in the subversion project)
13:56:34linuxstbI would like people to email their default settings files somewhere. We could then analyse what settings people are actually using.
13:57:03B4gderI've never bothered to modify mine until they fit me perfectly ;-)
13:57:04Zagorthe core point is we are making our own firmware, not an imitation.
13:57:10LinusNZagor: amen
13:57:32linuxstbI agree, but defaults should probably reflect what the majority of users want them to be.
13:57:35ep0chlinuxstb: that's a very good idea
13:57:37Zagorand it's a lot less confusing for newcomers to NOT follow playlist if they expect it, than to follow playlist if they don't expect it
13:57:38B4gdermy WPS for example looks like crap ;-)
13:57:54B4gder(iriver default)
13:58:02XavierGrep0ch: Does your nickname resembles the spaceship/timemachine from chrono trigger super nes game?
13:58:05LinusNZagor: both archos and iriver users will expect it
13:58:23B4gder... if they used it with the original that is
13:58:28B4gder... and liked it
13:58:33LinusNwhat else would they be used to?
13:58:34amiconnlinuxstb: How does the ipod stock fiirmware behave?
13:58:42linuxstbThere is no file browser.
13:58:48ep0chXavierGr: hehe, dunno never played it
13:59:01XavierGrlol just a thought
13:59:02amiconnNo, but maybe there is a database browser?
13:59:14B4gderLinusN: lots of people don't remember the original or never used it
13:59:21ep0chXavierGr: I'll see if its on my xbox somewhere...
14:00:03LinusNBagder: perhaps, but then they will be used to rockbox instead
14:00:19LinusNand they won't be confused
14:00:40amiconnLinusN: It depends on whether they are getting the player because they know about rockbox, or whether they discover rockbox after already using the player for a while
14:00:51LinusNamiconn: true
14:00:51linuxstbamiconn: You always enter it from the root. Leaving the "Now playing" screen takes you back to the main menu, from where you navigate down into the tag database.
14:01:29amiconnFor me it was the latter, discovering rockbox after I had my archos for more than a year
14:01:55linuxstbamiconn: Ignore what I said. Leaving the Now Playing screen takes you back to where you were in the database browser.
14:01:56ep0chok time to play chrono trigger :)
14:01:59LinusNi also used the original firmware for a few months before rockbox was usable
14:02:13LinusNon the archos
14:02:19amiconnMy first rockbox version was 2.0 iirc
14:02:28B4gderI don't recall ever using the Archos firmware for more than some initial testing
14:02:56*preglow suddenly wants to play chrono trigger
14:03:17linuxstbThe original firmwares have never been able to play my files - so I'm a rockbox only user as well.
14:04:11LinusN"I see that you're not working in the software industry" :-)
14:05:28amiconnlinuxstb: It takes you back to where you were, or to the currently playing track?
14:05:34XavierGrep0ch: If you are interested in these kinds of games (RPGS) you will enjoy it, it is classic and awesome for its time
14:06:15ep0chXavierGr: how far in until i come across the time machine/spaceship?
14:06:37preglowquite a time
14:06:40XavierGrehmm yeah
14:06:47ep0chforget it then :)
14:06:57XavierGrwait I will attach a screen shot :0
14:07:03ep0chhehe cool
14:08:35preglowmight as well get replaygain support going as well
14:08:46linuxstbamiconn: It's the same thing. The database browser takes from from (e.g.) Artist->Album->list of tracks. or Genre -> list of tracks. It's only these playlists that can be played, so you are always taken back to the same "directory", because it only plays a single "directory" at a time.
14:09:10linuxstbHope I'm making some sense
14:10:32ep0chwho was it that was trying to get rockbox compile on FreeBSD?
14:11:22XavierGrand a little history about it, epoch is a timemachine on the game. The name is not set by luck, in Greek epo(c)hi means season or era.
14:11:55LinusNimho, rockbox should at least remember the browser location when booting
14:12:17B4gderthat would indeed be nice
14:12:31XavierGrwell if you have the option "follow playlist to on"
14:12:37linuxstbSo it behaves more like "sleep" than power-off?
14:12:38XavierGrand then you resume playback
14:12:51LinusNand i personally don't care much about the default value of "follow playlist"
14:13:24XavierGryes I agree it is not very important to argue about.
14:13:49LinusNthe discussion on the mailing lust was not about the option itself
14:14:06XavierGrI know,this just sparkled it...
14:16:25 Quit B4gder ("time to say moo")
14:21:00XavierGron another note for the move to next folder option:
14:21:27XavierGrShouldn't the Next track information be updated when the option is set to on?
14:22:57XavierGror does this works only for playlists?
14:26:59LinusNit should work for follow-dir as well, but unfortunately not when playing the last track in the dir
14:27:18LinusNdue to how it works internally
14:28:55amiconnI see the folder navigation feature as an extra. It doesn't really fit in the normal rockbox operation
14:29:20amiconn(and personally I don't like it)
14:30:08XavierGrso you think it shouldn't be bothered to be fixed to work with this option?
14:30:30XavierGrIt is good sometimes to know what will be after the current folder.
14:31:04ashridahi personally really disliked the way the iriver firmware operated in that manner
14:31:27ashridahi really wanted to be able to repeat a set of nested subdirectories. rockbox makes that much easier with playlists
14:31:39linuxstbpreglow: Adding ICODE to all of liba52 speeds it up a little, but it's still boosting a little (less than 10%) at 34MHz.
14:32:18LinusNXavierGr: the problem is that rockbox is based on playlists
14:32:41LinusNthe next-dir option merely creates a new playlist when the current dir is finished
14:32:42preglowok, so the imdct itself needs speeding up
14:33:04linuxstbI don't understand the need for "move to next folder". Isn't this the same as recursively adding the parent directory to the playlist?
14:33:10LinusNthe next-track information is based on what's next in the current playlist
14:33:26XavierGrGot it.
14:33:27LinusNlinuxstb: i can see a use for it
14:33:27amiconnlinuxstb: Guess you are curious whether flac will boost at 22.5 MHz? ;-)
14:33:34preglowi use move the next folder all the time
14:33:45preglowthe lack actually irritated me before someone added it
14:33:50XavierGrlinuxstb: some times you choose asongle file to play.
14:33:57LinusNlinuxstb: you just start playing a file, and it will automatically continue, without having to add to the playlist
14:33:59amiconnI never use it and don't see the point of it
14:34:10XavierGrthen playback stops
14:34:15ep0chi never use but i can see the point of it
14:34:29preglowi think it's logical behaviour
14:34:30XavierGrTHere is a major point in it.
14:34:45preglowwhen i start a file in a dir, i almost never want that dir to just repeat
14:34:47XavierGrsometimes you don't have the time to bother with playlists
14:35:01LinusNi like the feature
14:35:03amiconnpreglow: You think it's logical that the device continues to play when your (carefully selected) playlist ends? I don't...
14:35:04XavierGrI just select a song and forget it
14:35:12linuxstbBut's it a transparent playlist. You just long-click on a directory and play it
14:35:26preglowamiconn: you think in terms of playlists, i don't
14:35:32LinusNamiconn: playing a directory is not playing a carefully selected playlist
14:35:42XavierGryes sometimes you play files individually
14:36:15LinusNpeople use rockbox in different ways
14:36:33LinusNif you play directories, you will probably like the next-dir feature
14:37:14preglowyes, like i do
14:37:23LinusNamiconn: do you mean that rockbox continues with the next dir even in playlist mode?
14:37:24preglowi almost never bother to use playlists
14:37:31ep0chwith "play next dir" enabled does rockbox keep on playing through every single directory until the battery runs out
14:37:39ep0chor just the next one?
14:37:50LinusNep0ch: yes
14:38:05LinusNit will continue forever
14:38:21linuxstbI understood it to be the next directory at the same level as the current one. Is that not the case? Does it traverse the whole directory tree at all levels?
14:38:45ep0chwould be nice to contain it to certain directory levels
14:39:02ep0chi dont fancy listenting to Abba after listening to 808 state :)
14:39:48preglowhell, why not, abba is better than 808 state :P
14:40:53LinusNit traverses the entire tree, iirc
14:44:00linuxstbamiconn: Yes please if you are offering to calculate the 22.5MHz values :)
14:44:38 Join _FireFly_ [0] (
14:46:38linuxstbLooking at len0x's "folder skip" patch, am I right in saying you can only manually skip to the next directory using a remote?
14:47:00ep0chis it feasible for Rockbox to set the CPU speed per codec?
14:47:08LinusNwe have yet to find out a key combination for the main unit
14:47:34LinusNep0ch: yes, but it requires a fundamental change in the boost paradigm
14:47:49preglowplus some overrides for dsp and crossfading
14:47:53preglowso i don't know if i think it's a good idea
14:47:59LinusNi don't think it is
14:48:42linuxstbBut maybe we could get away with 34MHz instead of 45MHz.
14:49:08LinusNhow about ogg and mpc?
14:49:21linuxstbI've no idea.
14:49:25preglowthey need plenty of boosting
14:49:27preglowif that's what you mean
14:49:38preglowso does mp3, at all but the lowest bitrates
14:49:45preglowso does more or less all the codecs
14:50:06LinusNthen i'd say that 45MHz might be the sweet spot for now
14:50:13preglowso i don't really know if this is the right time to start tweaking clock rate
14:50:49linuxstbIn any case, I'll try and run a playback test for FLAC at 34MHz to see what improvement it could give.
14:51:02preglowlinuxstb: seems i need to skip id3v2 tags for musepack as well, perhaps branch the skip out to a separate routine?
14:51:33LinusNi'm looking into a battery-low shutdown mechanism
14:52:03linuxstbpreglow: Do id3v2 tags always appear at the very start of the file?
14:52:20LinusNdo we want it to shutdown when it's still possible to run the hard drive, or when the battery is at the most critical level?
14:52:25preglowlinuxstb: perhaps :)
14:52:44preglowamiconn: if you've got time, could you just have a look at my unrolled loop now? i'm doing something wrong, but i can't see it:
14:52:59preglowamiconn: i mean the non-unrolled-loop, yes
14:53:00preglowsilly me
14:53:29_FireFly_LinusN: if we whant to save the config-changes on shutdown then we need a battlevel where the disk can spin up
14:53:42LinusNof course
14:53:44***Saving seen data "./dancer.seen"
14:55:04_FireFly_or we had to save the settings every time a setting is changed but i think that isn't a good idea
14:55:30LinusNwe do save the settings when the disk spins up for other reasons
14:55:39preglowi don't really think that's the defining reason for doing a battery level limit
14:55:50LinusNpreglow: no it isn't
14:58:57 Join ]RowaN[ [0] (
14:59:34]RowaN[guys, pressing STOP while in the filetree causes the music to stop suddenly.. no quick fade-down like it does when you stop in the wps
15:00:58LinusN]RowaN[: yes, that's because the fadeout is a lame hack from the days of yore
15:01:11LinusNand it's only in the wps
15:01:53LinusNit should be moved to the audio thread some day
15:06:03 Quit ashridah ("sleep")
15:09:57 Quit lostlogic ("Going to the moon")
15:10:24ep0chis there much of a difference between the two shutdown levels? what would happen if a user tries to access the HD within Rockbox when there isn't enough power to do so?
15:10:26 Join Nibbler [0] (
15:10:28 Quit Nibbler (Remote closed the connection)
15:10:33]RowaN[i see
15:11:17LinusNep0ch: it's a big difference
15:11:32LinusNep0ch: the drive will not spin up and rockbox will hang
15:13:32 Join Nibbler [0] (
15:14:27linuxstbI done two FLAC "play until death" tests, and both times, the iriver has shut down when the battery got too low.
15:14:59LinusNyes, the hardware will shut down when the battery is too low
15:15:31_FireFly_but LinusN will implement a safe shutdown of rb when batter gets low
15:15:36_FireFly_i think
15:15:44preglownice to have software control, yes
15:15:47LinusNthat's what i was thinking
15:16:28LinusNbtw, the ATA code should probably not retry disk operations when the battery is too low
15:16:42XavierGrLinusN: Any luck with those files I sent to you that WPS showed wrong total time?
15:16:56preglowwrong frame count, probably
15:17:06LinusNthis is why rockbox hangs when shutting down with too low battery
15:17:35LinusNXavierGr: no, i haven't looked any further (forgot about it actually)
15:17:54XavierGrok it's nothing important just wondering.
15:18:12XavierGrdo you remember the exact problem?
15:18:24XavierGrwas it ripping problem or rockbox bug?
15:18:31LinusNi seem to remember that is was a rockbox bug
15:19:12XavierGrI wish I could help but playback code is out of the question for my limited programming skills.
15:23:39amiconnLinusN: Imho the ata code should always retry. If it has to retry and battery is low, the low batt is not necessarily the reason for the retry
15:24:29amiconnRockbox doesn't hang indefinitely since the 8-second hard shutdown was added
15:25:19LinusNamiconn: that's only if a shutdown has started
15:29:15linuxstbXavierGr: I'm assuming these are MP3 files giving you problems?
15:30:49XavierGrlinuxstb: There are no playback problems. It is just that sometimes I want to know exact total time.
15:30:50amiconnLinusN: Really? Last time I ran down my battery that far, the box did shit down...
15:31:01XavierGramiconn: lol
15:31:48muesli-somebody should set it as topic ;)
15:31:49preglowi'd give it a nice spanking to teach it a lesson, were i you
15:32:20preglowmuesli-: that would indeed be the new high standard in informative topics
15:33:25LinusNamiconn: maybe you're right
15:33:38XavierGrpreglow: god that was good I had a really good laugh thanks. :D
15:33:51amiconnlinuxstb: Values for 22.5 MHz: PLLCR = 0x1584e005, NORMAL_REFRESH_TIMER is 10 for h100 and 4 for h120. CPUFREQ_NORMAL_MULT is of course 2
15:34:05linuxstbamiconn: Thanks.
15:35:03amiconnXavierGr: Sorry for the tpyo ;)
15:35:57XavierGrI remembered a question I had.
15:36:43XavierGrI want to make a script function in cygwin that will automate the rockbox compiling progress.
15:37:00MoosLinus: me too last one battery left, it auto shuted down
15:37:11XavierGrBut in this script I want to call tools/configure and set the parameters before hand.
15:38:19XavierGre.g I will type build 9 sim and then it will build iriver sim, the problem is that I don't know if I can pass the parameters in my script in the way that tools/configure will know
15:38:44MoosLinusN: bootloader V6 in preparation? :)
15:39:02linuxstbAs expected, the new FLAC is good, but not good. About a 9% boost running at 22.5MHz.
15:39:13preglowlinuxstb: good but not good? ;)
15:39:15XavierGrI've managed to pass the first argument with an echo, but then I can't find a way to pass the 2nd argument (except ofcourse changing tools/configure) any ideas?
15:39:22linuxstbpreglow: I wanted 0% boost :)
15:39:24preglowlinuxstb: that is indeed faster than i'd expected
15:39:28preglowlinuxstb: q8?
15:39:36ep0chXavierGr: use different pre made MakeFiles instead?
15:39:45LinusNXavierGr: echo the CR's too
15:39:48linuxstbpreglow: Yes. But I haven't analysed the LPC orders. I'll do that now.
15:40:04preglowlike i said, i've got a new coldfire.S coming up, with orders up to 10 unrolled
15:40:07preglowi just need to find this final bug
15:40:28LinusNecho "9\n\s\n\n" | tools\configure
15:40:29linuxstbSo we can get it down to 0% @ 22.5MHz then ? :)
15:40:34preglowin the few files i've analyzed, it doesn't look like 9 and 10 is used much at all :/
15:40:37XavierGrLinusN: I think I tried that already? isnt the /r for the cariage return? I will give it a try again.
15:40:47preglowlinuxstb: you can try the file now, if you don't mind the occasional static :)
15:40:51XavierGrah so it is/n and not /r
15:41:20linuxstbYes, my test files have lpc orders all the way up to 12.
15:41:42linuxstbpreglow: OK, I'll unplug the headphones and test it.
15:42:05preglowlinuxstb: only the higher orders are distorted
15:42:06_FireFly_XavierGr: not /n /r -> \n\r
15:42:10preglow11 and up
15:42:14preglowi HOPE that's the case, at least
15:42:38XavierGr_FireFly_: both "/n" and "/r" or just "/n"?
15:43:20_FireFly_under windows a new line is represendet by \r\n
15:43:41_FireFly_a backslash not a slash because they are escape chars
15:43:58_FireFly_under unix is only \n on mac only a \r
15:44:02linuxstbpreglow: It's a tiny bit faster - about 7% instead of 9% boost.
15:44:11preglowlinuxstb: didn't expect much more, no
15:44:19preglowlinuxstb: there's only so much that routine can do
15:44:27]RowaN[guys the iriver has a mic.. would it be possible to have a plugin that could analyse voice input.. i.e. say "play", "stop", "i should be so lucky" etc
15:44:47preglow]RowaN[: of course it's possible
15:44:53preglowit would also be very much work
15:45:03]RowaN[u gotta love possibilities
15:45:08XavierGrand slow I suppose?
15:45:15linuxstbI'm assuming we read the data from the mic at the same time as playback?
15:46:08preglowlinuxstb: that is possible
15:46:10]RowaN[also is the iriver powerful enough to alter song speed without changing pitch?
15:46:17preglow]RowaN[: sure
15:46:21 Join yngwi_away [0] (
15:46:24preglow]RowaN[: just don't expect too good quality
15:46:53preglowwe wont have enough cpu to employ a phase vocoder time stretcher for the majority of the codecs
15:47:22preglowyou'd be amazed at how bad all of the time-stretching methods sound with earphones on anyway
15:48:06]RowaN[what do you mean
15:48:36LinusNtime stretching may distort the audio severely
15:48:39preglowthat all but the very best time-stretching methods sound pretty poor
15:48:49pregloweven the boost do a pretty poor job of some kinds of music
15:48:54preglowpercussive music is very hard, for one
15:49:15preglows/boost/best/ :P
15:49:28]RowaN[i use a plugin called chronotron for winamp (well, when i used winamp), that sounded perfect with everything
15:49:48]RowaN[but i understand you say iriver isnt powerful enough to do best calculations
15:50:02preglownot without substantial time investments
15:50:13preglowin optimising
15:50:32]RowaN[why would percussive music be treated any differently to any other music?
15:51:04preglowbecause most time stretchers use phase vocoders, which are harmonic by nature
15:51:26preglowwhen you stretch beats, it tends to sound very granulated
15:51:39preglowafaik, this isn't something i've spent much time investigating
15:51:49]RowaN[surely thats just the nature of time stretching
15:52:32 Join Kohlrabi [0] (
15:52:54]RowaN[i'd use it to speed up songs anyway, not slow them down heh
15:53:05preglowin which case it's different
15:53:14preglowthat can be done with better quality
15:53:47preglowwe'll see, i've got a billion things on my plate at the moment, this isn't something i'll try to implement very soon, if ever
15:54:06]RowaN[i was just wondering if the iriver is capable thats all
15:54:18LinusNit should be
15:54:56]RowaN[never say never preglow ='[
15:54:57preglowyse, sure, but it might take a lot of optimising
15:55:02preglowit's capable of very much
16:02:06XavierGrLinusN, _FireFly_: Thanks this "echo -ne "9\ns" | mkrock" worked fine. Now I will include variables to it.
16:02:37LinusNXavierGr: what are you trying to accomplish?
16:05:17XavierGrwell I have specific tasks when I compile rockbox code. 1) I have a rockbox-devel folder which I only update. 2)I make a copy of that when I want to fiddle with the code or apply changes. 3) The copied folder is always named with a number and inside I build rockbox in a numbered subfolder. "e.g /1/1" /1 has the source and /1/1/ is the dir which I build.
16:05:56XavierGrso I made a script (more like an alias fn). I type "start" and all these tasks are done without me having to do anything.
16:09:23LinusNXavierGr: sounds like a very cumbersome way of working with cvs
16:11:33XavierGrwell I want to be sure that I always have a safe and unscathed cvs version.
16:11:59XavierGrI update it regularly and then copy it to make my experiments.
16:12:10LinusNyou can have as many repositories as you wish, no need to copy them
16:12:28LinusNjust check out as many as you like
16:13:10preglowi like this len0x person
16:13:12preglownice and cooperative
16:13:22preglowwhat manufacturer does the hds in iriver have, again?
16:13:24LinusNwell, he has the same view of us
16:14:41preglowlinuxstb: sure, but i think he should minimise the stepping on the toes anyway, it doesn't gain anyone anything
16:14:52preglowthat should be
16:15:52ep0chwhat generates the page?
16:16:38ep0chi'd much like it if the page had the logs in reverse order
16:16:45ep0chor even better drop down menus.
16:17:01LinusNep0ch: reverse order sounds like a good idea
16:18:24ep0chhow does one go about changing it?
16:19:42XavierGryeah I agree, with my crappy connection I have to wait before I can read the latest logs.
16:19:44linuxstbpreglow: My H140 has a Toshiba MK4004GAH
16:20:02XavierGr:) mine has a MK6006GAH!!!
16:20:33ep0chare Toshiba the only people churning out these hardisks?
16:20:49preglowbut for long they were the largest
16:20:54preglowstill are, i think
16:22:01linuxstbXavierGr: My ipod's got one of those.
16:22:12linuxstbMaybe I should swap them...
16:24:15XavierGrespecially when you are testing for the ipod port
16:24:41XavierGrbettter to burn a 40GB disk instead of a 60GB.
16:25:02XavierGrknock wood...
16:26:39amiconnI'd rather want a MKx006 disk. They're more shock resistant than the x004 series
16:28:34muesli-linuxstb is it possible to brand a normal thoshiba to an ipod useable?
16:39:54 Join tvelocity [0] (n=tony@
16:41:31amiconnpreglow: In the order3 loop is wrong
16:41:51amiconnAfaiu it must be subq.l #4,%a0 not 8
16:42:11amiconnIt is in the cvs version :)
16:43:23preglowamiconn: yes, you're correct, i pasted all those loops anew and have probably forgot some things here and there
16:43:33preglowamiconn: order3 is used very seldom, so i haven't heard a glitch yet
16:44:18amiconnDidn't you ask about that because there is a glitch somewhere?
16:44:23preglowno, not those
16:44:30preglowi think i corrected myself, i meant non-unrolled loops
16:44:32preglownot unrolled loops
16:44:47preglowbut of course, turned out there was a bug there as well :)
16:45:01preglowit's the default case that glitches
16:47:23linuxstbmuesli-: I don't know. I think I'll try and find out.
16:49:27muesli-guess they did something in its mbr..should be easy to get fixed by copying a mbr from an original drive
16:49:27 Quit DJDD ("Trillian (")
16:53:47***Saving seen data "./dancer.seen"
16:55:19 Quit ep0ch ("brb")
16:56:43linuxstbmuesli-: No, there's nothing special written on the drives. If it physically connects, it should work.
16:57:32muesli-i've read somewhere that non-branded=oem toshibas will not be recognized by ipods fw
16:58:25muesli-but maybe im wrong..dont count on this
16:58:34amiconnpreglow: The thing is I can't find a mistake in the default case...
17:00:13preglowamiconn: me neither, so i'll need to do some more detailed debugging
17:00:15linuxstbmuesli-: If it goes wrong, I can just swap them back. I don't own any torx screwdrivers though, so it will have to wait.
17:00:42preglowamiconn: you're a better german reader than me, that does say the stream is big endian, yes?
17:02:22amiconnJa, erm, yes ;)
17:02:30preglowthen why the hell does streaminfo.c only swap the values in the header #ifndef LITTLE_ENDIAN?
17:03:38amiconnAh, no, the drawing confused me a little
17:03:44preglowyes, me too!
17:04:00preglowplus my german isn't what it used to be
17:05:29amiconnHmpf, I'm not sure at all how to interpret this drawing
17:06:09amiconnIt looks like the bits within a byte are swapped, but that can't be true
17:06:16preglowno it can't
17:06:23preglowlibmusepack wouldn't work, was that the case
17:06:50preglowplus it would be the most impractical design known to modern man
17:06:58preglowand it isn't very practical as it is already
17:07:38preglowbut no, i'm almost certain it is little endian, i'll just continue and see what happens
17:08:55 Quit ]RowaN[ ("CGI:IRC (EOF)")
17:10:05amiconnIt makes sense (somewhat) if I interpret the bit numbers as counting from MSB (1) to LSB (32). Then it's little endian
17:10:10preglowbut rockbox has no 'swap from x endianess to local endianess' functions?
17:10:17amiconnit has
17:10:23amiconnThey're macros
17:10:28preglowas they should be
17:10:33preglowwhat're they called?=
17:10:49amiconnletoh() is for little endian to host order
17:10:58amiconnletoh32() and letoh16()
17:11:07preglowi wonder why metadata.c doesn't use them
17:11:17linuxstbBecause they are new.
17:12:40 Join DangerousDan [0] (
17:35:22 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
17:35:42 Quit Zagor ("Client exiting")
17:44:15 Join Maxime [0] (
17:52:52 Quit muesli- ("ich will Kühe!!!")
18:03:59 Nick Lynx_ is now known as Lynx_awy (
18:10:07 Join Gibbed [0] (
18:10:09 Quit Rick (Nick collision from services.)
18:10:19 Nick Gibbed is now known as Rick (
18:25:37 Join lamed [0] (
18:27:39lamedanyone else not havine his h100 player really build the cache with the last week's builds?
18:27:51lamed(this week's)
18:27:59lamedincl. today
18:28:20preglowsure, me
18:28:27preglowthat is, it works fine here
18:28:29preglowas far as i can see
18:28:36preglowwhat do you mean by 'not really build' ?
18:29:38lamedwell, it seems to be reading a lot at start, but when it's done, the drive still spins up if it was idle/ reads
18:30:07lamedhow do i do a full cache build?
18:30:28lamedmaybe it's because i sometimes use usb boot mode?
18:31:47preglowyou always do a full cache build
18:31:55preglowbut obviously something happens so the cache disables itself
18:34:34lamedwhen the cache was just introduced it was "reading library tree" or something on the boot screen. that what was my idea of full cache build. is this now always in background?
18:37:40lamedpreglow: any clue?
18:39:21preglowthe only time it does that, is the first time you ever use it
18:39:24preglowor when you delete your settings
18:39:28preglowafter that it does it in the background
18:43:16lamedgreat. just freshed installed. lemme c
18:45:35preglowwhy does replaygain contain floating point code? doesn't gcc give us this if we use floats?
18:45:38preglowreplaygain.c, that is
18:51:13amiconnThis is fixed point code afaiu
18:51:46lamedi'm not much into how .cfg's work, but the reason of my problem was an old .cfg file. it had "dircache on" but yet something didn't liked it.
18:53:16preglowright, that they are
18:53:27preglowi was just under the impression that most codecs saved replaygain data as floats
18:53:50***Saving seen data "./dancer.seen"
18:54:45preglowlamed: well, let's just call it quirk, move along, nothing interesting here ;)
18:54:48lamed-> I'm reffering this minibug the second time, Iriver sim uses archos sound settings. (bass & treble -15~15)
18:55:25preglowin that case there is very little info on the fixed point format the replaygain functions use
18:56:05amiconnlamed: A .cfg file shouldn't harm at all as long as you don't load it. They're never loaded automatically
18:57:46lamedpreglow: especially when those minor obstacle are preventing me from really working. :( and the family dinner that's beggining
18:58:29preglowif that's the case, then HAVE_UDA1380 isn't set for the sims
18:58:35lamedamiconn: of course... but that was the file that i was mainly using.
18:59:05lamedpreglow: yes. guess what. that's your job.
18:59:13preglowlamed: i like your attitude
18:59:14lamedyou _are_ a dev right?
19:00:18lamedwhat's your cvs sign in name?
19:00:31preglowsame as my nick, i believe
19:01:37preglowit's neither my job nor my responsibility, it'll be fixed when it's fixed, i don't use the sims
19:02:12lamedwhat i was (trying to) ask, was if you can submit to cvs, but i don't recall i've seen ur nick on the "who" list
19:02:45preglowok, then it's probably thom or something
19:02:53preglowyeah, i'm listed as thom on the main page
19:03:18amiconnThe main page uses a mapping cvs login name −−> realname
19:03:34lamedsims is great. i've only played sims 1 though. (gee long time ago. my pc is getting old) nice to know you thom i've gtg now.. dinner
19:03:46preglowyes, tata
19:13:41 Quit ghode|afk ()
19:19:43 Quit XavierGr (Read error: 110 (Connection timed out))
19:25:48 Part LinusN
19:39:39 Join Mxm`Pas`Bien [0] (
19:39:39 Quit Maxime (Read error: 104 (Connection reset by peer))
19:43:54preglowcan't we allocate a couple of bytes to album_gain_string and track_gain_string in id3.h instead of passing these bloody temporary buffers all around the place
19:44:46linuxstbHow many bytes?
19:45:18linuxstbBut I don't know many copies of that struct are in memory, so I'm not sure of the significance.
19:45:24preglowwell, i have no idea how long those strings are, usually
19:45:31preglowi think about 10 for iriver
19:45:48preglowi don't get how that's possible, all of those point into the id3b2 buffer
19:45:51preglowwhat if we load an mp3 suddenly?
19:47:13preglowi've got the musepack parser working now, but i stopped on replaygain
19:47:50linuxstbWhen are those strings converted to numbers?
19:48:03preglowthey aren't, they're only used for the id3 screen
19:48:18preglowthat is, they're converted to numbers too, but those are stored in album_gain, etc
19:48:27preglowthey keep a text representation for the id3 screen
19:48:39linuxstbI didn't know you could display them on the id3 screen...
19:48:44preglowwell, sure
19:48:44preglowjust try
19:49:00preglowtrack gain and album gain can be displayed
19:49:09preglowalbum and track peak can't be displayed
19:50:07preglowcurrently they only work for vorbiscomment, though, since that format saves the replaygain info as strings
19:50:21preglowmusepack doesn't, so i need to sprintf into the buffers, but then i notice the buffer problĝem
19:50:21 Join Maxime [0] (
19:50:45preglowthey need to point somewhere, and currently i don't understand how we use the id3v2 buffer for that
19:51:50preglowsomehow it does
19:52:09linuxstbNo, I think they should stay in id3v2buf - they have the same status as all the other vorbis/id3v2/ape etc tags.
19:52:57linuxstbMy understanding that there is one id3v2 buffer, and that's used for storing any strings referenced by the char pointers in the main id3 struct.
19:53:14linuxstbI mean one id3v2 buffer per track.
19:53:24preglowmakes sense, i suppose
19:54:17linuxstbI'm not 100% familiar with that code - but no-one else seems to be around.
19:54:41preglowwhat you say makes sense, so i'll pretend it's true for now
19:55:44linuxstbhehe. That's what everyone tells me.
19:58:44preglowas long as you remember it's ultimately your fault, it's ok by me
19:59:12linuxstbOK, you can blame me...
19:59:36 Quit Mxm`Pas`Bien (Read error: 110 (Connection timed out))
19:59:45preglowthere you go
19:59:49linuxstbAre you subtracting the size of the Ape/id3v2 tags from the filesize before calculating the bitrate?
20:00:29linuxstbI don't think any of the parsers do that, but maybe they should.
20:02:10preglowlinuxstb: as a matter of fact, i don't, but i know of the issue
20:02:19preglowlinuxstb: i just can't be bothered now, that's all, i plan to leave in a TODO :)
20:07:15 Join ghode|afk [0] (
20:10:41preglowhow do i make rockbox snprintf print unsigned numbers?
20:12:54preglowdoesn't look it's possible, to me
20:16:25_FireFly_mybe i have a snprintf implementation where i can look at
20:16:31_FireFly_on my drive
20:17:51preglowno worries, i did it another way
20:18:04preglowturned out it was me acting the fool, as usual
20:18:15_FireFly_:).. i have a vsnprinft found ;)
20:27:00lamedwhat's the GMT time on the server?
20:30:12lamedarg... how do i dump make error messages into file instead of screen?
20:30:41thegeek_> file
20:31:09lamederror messages are send to screen
20:31:50amiconn2> file is for stderr redirection iirc
20:32:07_FireFly_just tested :)
20:33:12lamedyeah really, that part I already know. how do i redirect errno?
20:36:21 Quit ghode|afk (Read error: 110 (Connection timed out))
20:37:22preglowerrno is a variable
20:39:10lamedhow do i redirect error messages to a > file
20:39:20preglowwith 2>
20:39:25preglowlike amiconn said
20:40:22lamedthat "iirc" part was misleading <amiconn> 2> file is for stderr redirection iirc
20:40:51preglowamiconn usually does recall correctly ;)
20:44:34linuxstbI normally use &> which redirects both stdout and stderr to the file.
20:45:15_FireFly_or > 2&>1
20:45:57linuxstbIs that any different?
20:46:03preglowwell, yeah
20:46:08preglowthat redirects stderr to stdout
20:47:03lamedgreat advice! i think there should be a 'wiki dev tip page'. honestly, working for quite some time with cygwin, i still donno how to delete a folder with it content and stuff :(
20:47:17_FireFly_rm -r :)
20:47:31linuxstblamed: Search google for "bash tutorial"
20:47:32_FireFly_btw man <command> if manpages are installed
20:48:03_FireFly_or search man <command> on the net
20:48:53lamedyes, they do. its just takes a while to understand what's going on in them... but ok i see you are on a roll so i won't stop you :D how (the #$F#) do you clear screen?!?! (dos cls)
20:49:05lamed(they do. = man pages)
20:49:14lamedTHANK YOU!
20:49:16linuxstbor type the command "clear"
20:50:27amiconnMrf. TiMiD: You wanted to make the code smaller, not bigger, right? ;)
20:50:51_FireFly_currently old code resists
20:50:58lamedhow do you search a function or a struct definition placement when you only have a call? (i mean, there are lots of files included)
20:51:00_FireFly_e.g. statusbar in status.c
20:51:49amiconnWhat do you mean? Do you want to know where it is defined?
20:53:33lamedyes, & where it really is.
20:53:49linuxstbI would just use "grep" to search .h files for the definition. I believe it's like the DOS find command
20:53:54***Saving seen data "./dancer.seen"
20:54:29_FireFly_linuxstb: this works only if the function/struct is defined in an header
20:54:37 Join RotAtoR [0] (
20:54:45lamedthat what i do, only in windows. but that kind of suck because i'll get many results and many times the init function isn't in the same header file
20:55:25linuxstb_FireFly_: But it's only available to other .c files if it's defined in a header.
20:55:28amiconnI like gui methods better, I use the windows file search, with a pattern of *.h (or *.c *.h) and the name as search string
20:55:51_FireFly_if you know the return type then you could search like that "void main"
20:56:26lamedfirefly: that's a good idea
20:57:14lamedamiconn: aren't all files in rb source code are .h / .c ...?
20:57:19_FireFly_or an good editor which can search in a directory
20:57:44_FireFly_yepp but mostly all c-functions are in .h /.c files
20:58:17amiconnlamed: No they aren't. There's also .S, .pl etc, and aprat from that I search the tree, which includes the build dirs
20:58:53linuxstbYou can easily do a recursive grep - something line "grep -r myfuncname apps/ firmware/"
20:59:25lamedfirefly: i asked for a zillion times what editors are you using, and if there is a good editor to gdb from a gui... responds usually were "emacs?" "i mostly use notepad"
21:00:12_FireFly_i use currently jedit as editor and insight as gui-frontend for gdb
21:00:34lamedamiconn: true, but what i'm looking for is in .c / .h files, so it usually won't be found elsewhere... anyways i'm just asking if there are better ways :)
21:01:58lamed" Insight is a graphical user interface to GDB, the GNU Debugger" -sounds good. aspeacially now that I know how to work with command gdb ...
21:01:58amiconnYes, but since I know that, I tell windows not to look into the others. That saves time, the build dirs contain many files
21:02:12_FireFly_a other is ddd
21:02:38lamedyes, i was wondering what was that stupid name for a program doing.
21:06:13preglowi wonder why title and artist doesn't show up in the id3 screen for my musepack files
21:08:44lamedi believe i ren out of stupid questions for today
21:12:04 Join dpassen1 [0] (n=dpassen1@resnet-233-61.resnet.UMBC.EDU)
21:20:35 Join hardeep [0] (i=hardeeps@norge.freeshell.ORG)
21:28:14preglowso, what to do with this musepack volume level bug
21:28:40 Join arkascha [0] (
21:29:05_FireFly_take a fly flap and smash it :)
21:29:57preglowi tried, and now i've cracked my lcd
21:33:08preglowit looks like the settings are correct, but no, they can't be
21:33:13preglowi only get half-scale output
21:34:05 Join DrMoos [0] (
21:34:05 Quit Moos (Read error: 104 (Connection reset by peer))
21:34:31_FireFly_maybe a *2 forgotten or a /2 to much ??
21:34:52preglowwell, yes, obviously
21:35:12preglowbut problem is, if i fix it that way, we get some very slight clipping
21:35:56lamed_firefly_: I don't assume you use windows? -insight isn't reported to be very stable on it.
21:36:24_FireFly_yepp i'm mostly under linux :)
21:38:42_FireFly_it seams that you have either wait for the 5.4 release or use a cvs-snapshot and comile it yourself
21:45:04preglowwell, i've got a "hack" going now that gives the correct output volume
21:45:12preglowit's not the way it should work, but it works
21:45:26preglowshould i commit it?
21:48:26lamedi'll check ddd
21:52:33 Quit _FireFly_ ("Leaving")
21:53:09*preglow summons Lear
21:55:45 Quit lamed ("CGI:IRC")
21:57:12 Join uski [0] (n=uski@
22:06:00 Nick DrMoos is now known as Moos (
22:09:48linuxstbpreglow: How bad is your "hack" for musepack?
22:12:13preglowlinuxstb: it's just sets sample depth to one value lower
22:12:19preglowlinuxstb: like the way it was
22:12:35preglowlinuxstb: but now i've fixed it to handle clipping so we wont have any overflow
22:16:36linuxstbI say commit it then.
22:20:01preglowbut i think i'll commit it together with this metadata patch
22:20:17preglowhowever, i don't know the fixed point format of the peak data
22:20:27preglowlooks like it's 24 bit frac part, but if i use that, it doesn't work
22:23:39preglowthink i've got it nailed
22:23:41preglowso i'll commit it
22:24:23preglowi still wonder why musepack ape tags don't show up, though
22:28:53 Quit Kohlrabi (Nick collision from services.)
22:28:57 Join Kohlriba [0] (
22:29:31 Join lImbus [0] (
22:30:39RotAtoRmusepack apev2 tags show up for me fine
22:31:59lImbushi all
22:32:16RotAtoRi can see all the usual information on the show id3 info screen
22:32:17preglowwonder what i'm doing wrong, then
22:32:33preglowin a little while you'll have bitrate as well
22:33:00RotAtoRthis is good news, i have a large musepack collection
22:33:00preglowplus some other snacks
22:33:09preglowreplaygain as well
22:33:13preglowplus i'll fix the volume bug
22:33:17RotAtoRvery nice
22:33:31preglowRotAtoR: do you have any sv6 files?
22:33:47RotAtoRi'd might be hard to check though
22:34:14preglowfoobar tells you what it is
22:34:17preglowbut it's a bit of a bother
22:34:18RotAtoRi don't know of any good way other than checking individual files in foobar
22:34:35RotAtoRand I have hundreds of musepack albums :p
22:34:56lImbusI just wanted to say that I just felt hart to swallow, because my mouth had been gaping all the time whilst reading the discussion with that new guy on the dev-list. i support the already existing core developers :-)
22:35:44preglowRotAtoR: i'd appreciate it if you gave my new commit a spin
22:35:55RotAtoRsure, no problem
22:36:03RotAtoRany specific thing I should look for?
22:36:14*amiconn is still fighting against breakpoint ogic and emulator mode :/
22:37:01preglowRotAtoR: well, if it displays bitrate, length and sample rate properly
22:37:11preglowRotAtoR: if it supports replaygain properly, if you even use that at all
22:37:24RotAtoRyes, all my files are replaygained
22:37:31RotAtoRi'll check that too
22:46:12preglowand does gapless musepack work?
22:46:21preglowi would be very surprised if it did
22:46:41RotAtoRi'm don't think i ever specifically checked
22:47:47preglowme neither
22:47:54preglowi don't even have any musepack encoded albums
22:50:08RotAtoRhmm, i can't find a single non-sv7 encoded file in my collection
22:53:58***Saving seen data "./dancer.seen"
22:57:25 Quit lImbus (" gottago")
22:57:28RotAtoRyou can find sv4 and sv5 encoders here:
22:57:42RotAtoRbut no sv6 encoders that i can find
22:59:46preglowdoesn't matter, the bitstream is the same
22:59:49preglowthanks, think i'll try that
23:08:25 Quit ender` (Read error: 104 (Connection reset by peer))
23:10:58RotAtoRpreglow: ok, trying your changes
23:11:40RotAtoRsample rate, track length, and bitrate for the first track i tried are the exact same as they appear in foobar
23:12:04RotAtoRreplaygain is not quite right...
23:12:36RotAtoRthe show id3 info page shows the track number as the track gain and part of the title as the album gain
23:14:27RotAtoRall other tags appear correctly
23:14:35RotAtoRtrying more files now...
23:18:33RotAtoRmiraculously gapless playback seems to be perfect for the first files a tried
23:19:00preglowtrack number as the track gain?
23:19:39RotAtoRyes, the appear to always be segments of other tags
23:19:47RotAtoRthe replaygain fields, that is
23:21:21RotAtoRthe trackgain is always the part of the track number tag
23:21:32RotAtoRand the album gain is always part of the artist tag
23:21:56RotAtoR(so far, at least)
23:23:02MoosHello guys
23:23:34Moospreglow: I'm just back, let me few seconds for MPC files tests :)
23:23:39preglowMoos: seems there's a bug
23:23:42preglowwill fix tomorrow
23:24:55RotAtoReverything else is looking good though :)
23:25:38preglowRotAtoR: could you send me a faulty file?
23:26:44RotAtoRsure, what way should i send it?
23:27:10RotAtoRall my files so far exhibit the same replaygain bug
23:28:28RotAtoRooh, this file's album gain is part of the track title tag, something different!
23:34:51Moospreglow: :(
23:36:06preglowMoos: got a couple of friends and some beers here, hard to code ;)
23:36:13preglowRotAtoR: if you can dcc that's ok
23:36:25RotAtoRok, i'll give it a try
23:36:30preglowRotAtoR: pm in the forum should work to
23:36:34preglowbut now i'm gone
23:36:48Moossorry, just your code cause freeze
23:37:02Moosfreeze the HD
23:37:41Moosin 4 files tests all sv7
23:37:45Moos2 bugish
23:38:08Moosand 2 have got the good bitrate like foobar ones
23:38:52RotAtoRweird, i've been testing dozens of my sv7 files, no freezes so far...
23:40:01Mooshere playback doesn't start for 2 albums track and hard disc freeze in spinning
23:40:21RotAtoRouch! :o
23:41:39RotAtoRpreglow: did you do any optimization? I'd swear some of my files that used to boost 80-90% are now boosting 40% !!
23:41:56Moosand the id context screen doesn't shows the right info, it did drunk a lot of beer, infos are mixed like you :)
23:42:23Moosgood to hear :)
23:42:56Moosaparently preglow go away ;)
23:43:11 Quit dpassen1 ()
23:47:56RotAtoRpreglow: I haven't check musepack performance in the last week or so, but something has definitely improved, even my hardest q10 files only boost 40-50% where as I know that they used to boost ~90%
23:48:56RotAtoRnow the codec buffer can actually fill in one shot rather than limping along for several minutes trying to fill
23:56:08amiconnTiMiD: The new settings handling code has the same bug as the menu code had: The current value isn't voiced when you enter the setting. Changing the value does voice the new value.

Previous day | Next day