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 2005-09-12

00:04:46 Join stripwax [0] (
00:07:42***Saving seen data "./dancer.seen"
00:23:51 Quit muesli- (Read error: 113 (No route to host))
00:27:21 Quit Moos (" HydraIRC -> <- Leading Edge IRC")
01:04:56 Quit webguest43 ("CGI:IRC (Ping timeout)")
01:09:20preglowcan't really say i've had that much trouble with it crashing either
01:09:27preglowtwice i can remember
01:09:35preglowout in ordinary use, that is
01:09:45preglowonce was really early in the development
01:09:49preglowthe other was on the train, with the stokov
01:09:54preglowand i've never seen that again
01:10:04preglowoh, and another time, but that was rockboy's fault
01:10:59linuxstb've had a quick play with loading/saving config files, but couldn't make it crash. Maybe it's dependent on the codec being played.
01:14:35preglowdont deal with config files too often
01:14:43preglowmore or less only when there's been a change to the config block
01:25:07 Join ender1 [0] (
01:32:22preglowoh, how out of touch i am with most of the rockbox code
01:33:07 Quit ender` (Nick collision from services.)
01:33:09 Nick ender1 is now known as ender` (
01:47:07 Join TCK [0] (
02:07:46***Saving seen data "./dancer.seen"
02:08:24 Quit ender` (Read error: 113 (No route to host))
02:20:12 Quit gromit` (Remote closed the connection)
02:21:17 Quit linuxstb ("CGI:IRC (Ping timeout)")
02:33:23 Quit cYmen ("zZz")
03:06:38 Quit preglow ("leaving")
03:15:06 Join paugh [0] (n=pete@2001:5c0:8fff:ffff:8000:0:3e03:6822)
03:28:02 Quit Maxime (Read error: 110 (Connection timed out))
03:58:59 Join Sucka [0] (
04:06:13 Join QT_ [0] (i=as@madwifi/users/area51)
04:07:50***Saving seen data "./dancer.seen"
04:16:22 Quit QT (Read error: 113 (No route to host))
04:52:24 Quit t0mas (Connection reset by peer)
04:52:25 Quit Rick (Read error: 104 (Connection reset by peer))
04:55:37 Quit Slasheri (
04:55:37 Quit stripwax (
04:55:37 Quit Hadaka (
04:55:37 Quit TCK (
04:55:37 Quit pike| (
04:55:37 Quit Seed (
04:55:37 Quit cv_ (
04:55:37 Quit odd (
04:55:37 Quit HCl (
04:55:37 Quit paugh (
04:55:37 Quit solex_ (
04:55:37 Quit godzirra (
04:55:37 Quit Asku (
04:55:37 Quit Ismo (
04:55:37 Quit Wy (
04:55:37 Quit Febs (
04:55:37 Quit thegeek (
04:55:37 Quit pabs (
04:55:37 Quit einhirn (
04:55:37 Quit atubbs (
04:55:37 Quit stamppot (
04:55:37 Quit NibbIer (
04:55:37 Quit Rob- (
04:55:37 Quit flurble (
04:55:37 Quit merbanan (
04:55:37 Quit crwl (
04:55:37 Quit QT_ (
04:55:37 Quit Sucka (
04:55:37 Quit amiconn (
04:55:37 Quit Strath (
04:55:37 Quit silencer_ (
04:55:37 Quit bluebrother^ (
04:55:37 Quit dwihno (
04:55:37 Quit ansivirus (
04:55:37 Quit Bagder (
04:55:37 Quit igor47 (
04:55:37 Quit ze (
04:55:37 Quit coob (
04:55:37 Quit CoCoLUS (Read error: 104 (Connection reset by peer))
04:56:03 Join CoCoLUS [0] (
04:56:03 Join t0mas [0] (n=Tomas@unaffiliated/t0mas)
04:56:03 Join Rick [0] (i=rick@unaffiliated/Rick)
04:56:03NJoinQT_ [0] (i=as@madwifi/users/area51)
04:56:03NJoinSucka [0] (
04:56:03NJoinpaugh [0] (n=pete@2001:5c0:8fff:ffff:8000:0:3e03:6822)
04:56:03NJoinTCK [0] (
04:56:03NJoincv_ [0] (n=verm@
04:56:03NJoinFebs [0] (
04:56:03NJoinsolex_ [0] (
04:56:03NJoinamiconn [0] (
04:56:03NJoinStrath [0] (
04:56:03NJointhegeek [0] (
04:56:03NJoinpabs [0] (
04:56:03NJoinpike| [0] (
04:56:03NJoinatubbs [0] (
04:56:03NJoineinhirn [0] (
04:56:03NJoinstamppot [0] (
04:56:03NJoinbluebrother^ [0] (
04:56:03NJoinNibbIer [0] (
04:56:03NJoinRob- [0] (
04:56:03NJoinflurble [0] (
04:56:03NJoindwihno [0] (n=dw@
04:56:03NJoinansivirus [0] (
04:56:03NJoinSlasheri [0] (
04:56:03NJoinsilencer_ [0] (
04:56:03NJoinSeed [0] (
04:56:03NJoinmerbanan [0] (
04:56:03NJoincrwl [0] (
04:56:03NJoinHCl [0] (i=hcl@2001:610:1908:8000:290:27ff:feca:8029)
04:56:03NJoinBagder [0] (n=daniel@
04:56:03NJoinigor47 [0] (n=igor47@
04:56:03NJoingodzirra [0] (i=shawn@
04:56:03NJoinze [0] (i=ze@
04:56:03NJoinodd [0] (
04:56:03NJoinHadaka [0] (
04:56:03NJoincoob [0] (
04:56:03NJoinAsku [0] (
04:56:03NJoinIsmo [0] (
04:56:03NJoinWy [0] (
04:59:35 Nick Sucka is now known as Sucka`zZzZz (
05:14:58 Nick paugh is now known as AliasCoffee (n=pete@2001:5c0:8fff:ffff:8000:0:3e03:6822)
05:40:45 Quit AliasCoffee (Read error: 104 (Connection reset by peer))
05:45:13 Join paugh [0] (n=pete@2001:5c0:8fff:ffff:8000:0:3e03:6822)
06:00:31 Join DangerousDan [0] (
06:07:55***Saving seen data "./dancer.seen"
06:16:42 Quit DangerousDan (Read error: 104 (Connection reset by peer))
06:20:17 Join webguest32 [0] (
06:20:29webguest32ok i need help with compiling
06:20:44webguest32bleeding edge doesnt compile on dev tool
06:24:09 Quit webguest32 (Client Quit)
07:03:46 Join LinusN [0] (
08:03:11 Join B4gder [0] (
08:05:01amiconnGod morgon :)
08:05:07B4gderhey ho
08:05:31*B4gder had to wait outside the front door for a good while this morning :-(
08:05:46B4gderbeing a consultant here, they don't trust me with a key
08:06:43 Join ender` [0] (
08:06:49LinusNnobody should trust you with a key :-)
08:07:23B4gderhey, no such jokes before the first coffee, ok? ;-)
08:07:57***Saving seen data "./dancer.seen"
08:11:16amiconnImho good news: My frame count estimation for long recordings seems to work.
08:11:42amiconnOnly problem in my current test version is that I forgot to multiply the time by 10 (mpeg thread counts in ticks but mp3data.c expects milliseconds... my estimated frame count is 1/10 of the reality...
08:12:17amiconnWill fix and start a new test
08:12:42amiconnBut the bad thing is that the recording engine has race problems...
08:13:53LinusNnot that surprising, though
08:14:51amiconnIf you trigger file splits in quick succession, or stop quickly after file split, framecount, filelength (and recording time) may be incorrect... case the first save has not yet finished when the next event is received
08:15:18 Join hubbel [0] (
08:15:54 Quit t0mas (Read error: 104 (Connection reset by peer))
08:16:11amiconnThe code has to store the values at the point of split/stop, then starts saving
08:16:20 Join t0mas [0] (n=Tomas@unaffiliated/t0mas)
08:16:46amiconnSaving does multiple chunks (at least 2 when wrapping, more on 8MB boxes or Ondio)
08:17:22amiconnIf another split/ stop event is received in between, it overwrites the old values before they get used.
08:17:46amiconnI don't know how much effort we should put into this right now
08:18:08amiconnSounds like a bigger rewrite is due, especially if we also want to unify playback...
08:18:11LinusNhow likely is it to happen?
08:19:24amiconnWithout timesplit it's rather unlikely I'd say. I doubt someone will deliberately split in such short intervals, or press "split-stop"
08:19:59amiconnWith timesplit it might happen that the user triggers a split or stop just when the time split triggers automatically...
08:20:10LinusNah yes
08:21:16 Join zhilik [0] (
08:21:34amiconnNothing really bad will happen, only the xing header will be totally off...
08:22:22 Part hubbel
08:22:37amiconnThe thing is that we can't simply lock split/stop events while another split is going on. I'd like to be able to interrupt a split with stop.
08:22:48amiconnI think I have an idea how to work around it...
08:33:54amiconnLinusN: What do you think about my playtime calculation stuff, btw?
08:34:25LinusNi think it's great
08:34:43amiconnI noticed that the archos playback engine also mixes up file and playback offsets, leading to the time display being several seconds off when resuming a vbr track
08:35:29LinusNyes, it only keeps track of the file offset
08:35:38amiconnI observed up to 9 seconds with a 10-minute track (with toc). This isn't introduced by my new code, it also happens with the old
08:36:07LinusNafaik, the resume info only saves the file offset
08:36:16amiconnThe problem is that the code mixes up the real file offset (from the start of the file) with playback offset (from the start of the first frame)
08:37:32amiconnBecause of this the toc search can be off by a full percent...
08:37:42LinusNif the tag is big
08:38:29amiconnDoesn't need to be big, just the chance of being off gets bigger
08:38:51LinusNthe larger the file is, the more inaccurate is the toc
08:38:58amiconnE.g. if the tag is half the size of one percent of the file bytes, the chance of being off is 50%
08:39:29 Join gromit` [0] (n=gromit`
08:39:57LinusNamiconn: isn't this more about toc accuracy than tag size?
08:40:39LinusNi agree that it should take the tag size into account, but isn't the time calculation problem more related with the toc inaccuracy?
08:41:18amiconnI'd expect way higher accuracy when a toc is present, at least with "ordinary" music
08:41:30LinusNas long as the file is small
08:41:40amiconn9 seconds of 600 is more than one percent
08:42:02amiconn...and the toc should allow for at least one percent precision
08:43:49LinusNbut how is the time calculation done?
08:44:43LinusNok i see
08:46:31LinusNok, so we should subtract id3->first_frame_offset
08:49:37amiconnI don't think so
08:50:03amiconnDuring ordinary playback, id3->offset tracks the playback offset
08:50:18amiconnImho we should restore this properly when resuming
08:51:34 Join Zagor [0] (
08:52:06amiconnBtw, id3->filesize also contains the data size without headers, not the true file size. Misleading...
08:53:29 Join linuxstb [0] (
09:05:19linuxstbMorning all. Does enabling FM presets for the iriver constitute a feature or a bug-fix?
09:19:49B4gderI vote for a release asap anyway
09:19:57B4gderlike within days
09:22:01linuxstbAre there any bugs on the ReleaseTodo page that are considered show-stoppers?
09:23:13B4gdernot according to me
09:23:51linuxstbSo if everyone agrees with that, then we just need to finish the documentation?
09:24:25B4gderwell, I think we should release it anyway with everything as it is right now
09:24:51B4gderthe docs could be better
09:25:01B4gderbut I don't think they should delay the release
09:26:14B4gderwhat's your position on this, LinusN and amiconn?
09:29:20LinusNi don't think the remaining issues are that critical
09:29:48LinusNbut i would like to let amiconn finish his mpeg fixes before release
09:47:02*B4gder fights general cluelessness from taking over the world
09:47:20B4gderor something like that
09:47:24ZagorB4gder: a brave fight
09:47:40B4gderme vs debian-devel
09:48:10Zagorheh, what's it about?
09:48:42B4gdershortly put, they want to throw out libcurl built with openssl support and only provide libcurl built with gnutls support
09:48:48*Zagor opens his 37MB unread debian-devel mail file...
09:50:13B4gderwell, I moved my fight to specific bug report entries instead
09:50:31B4gdersince the debian-devel is just too much for my taste
10:05:18 Join ashridah [0] (
10:08:00***Saving seen data "./dancer.seen"
10:11:07amiconnB4gder: I'd like to see some of the critical things on the todo list fixed before a release
10:11:19amiconnImho some of the are critical
10:14:06B4gderI really don't think we should keep a feature freeze much longer
10:14:24B4gderso if these last issues are gonna take multiple days I vote for branching
10:14:32 Quit linuxstb (Read error: 110 (Connection timed out))
10:15:58amiconnI think at least the software poweroff issue and the ahrddrive power in panic() needs to be fixed, in addition to the recording issues
10:16:08amiconnThe fix for splitedit should also be applied imho
10:16:54 Join Lynx_ [0] (
10:21:15LinusNamiconn: the soft poweroff issues are not as critical on the archos as on the iriver
10:21:50amiconnB4gder: I don't think a branch would be a good idea. I detracts developers from bugfixing
10:22:06LinusNi think the splitedit fix might make things worse
10:22:25LinusNsince it might refuse to save even if there is room on the drive
10:22:39B4gderthe bugfixing developers are amiconn and LinusN
10:22:44LinusNjust because the fsinfo lies about the true remaining space
10:22:47B4gdera branch doesn't detract you
10:23:18amiconnLinusN: It uses fsinfo?? That's... strange...
10:23:30LinusNamiconn: you didn't look at the bugfix?
10:23:31amiconnB4gder: It does detract all devs more or less
10:24:04B4gderand the freeze detract them from developing
10:24:20amiconnI'm really disappointed that so few devs are actually bugfixing
10:24:36LinusNamiconn: that's life, unfortunately
10:25:10LinusNi'm disappointed too...
10:25:36amiconnImho new features aren't worth much if the existing code has problems
10:25:58amiconnThe freeze was a really good chance to fix bugs for iriver as well...
10:26:29LinusNand the iriver code is quite bug-ridden
10:26:34B4gderI would say that the explanation is that very few devs consider themselves "core" enough
10:26:49LinusNwhy should only core developers fix bugs?
10:26:50B4gdermost just do "border"-developing
10:27:24B4gderthis is just my theory
10:29:43B4gderso, during this feature freeze I think most people have just backed off and waited
10:29:55LinusNreally bad
10:32:46zhilikHi 2 All! What do I need to do to start developing Rockbox for the HD200? I did the scans, found most of the documentation for the ICs. Now I have to find out what gets connected where, but to do that it seems i'll have kill my player - and I don't want to do that.
10:33:09zhilikMayde there's some place where I can buy a dead one cheap?
10:33:40zhilikI guess the shipping would cost more that the player - I'm in Moscow...
10:34:00B4gderor try find an online forum for your player and post a question/call for dead units and volounteer rockbox hackers
10:34:12Zagorzhilik: are you talking about the mpio hd200?
10:36:17zhilikZagor: If you didn't see them, here are the insides:
10:37:43 Join amiconn_ [0] (
10:39:28zhilikLinusN: Are my private messages getting to you?
10:39:32LinusNwhat kind of hard disk is in there?
10:39:32Zagorwell since it has a BGA cpu it will definitely be difficult to trace the board without removing it.
10:39:39LinusNzhilik: no
10:39:54ZagorLinusN: looks like the hitachi square type
10:40:00LinusNzhilik: probably because of the new freenode policy
10:40:02B4gderfreenode blocks private messages
10:40:10B4gderunless you've "identified"
10:40:28zhilikAnd how do I get "identified"?
10:40:48LinusNhd300 looks cool too
10:41:17B4gderzhilik: you register with 'nickserv'
10:41:30Zagorzhilik: however since it's a coldfire design, it's probably not too different from the irivers.
10:41:31B4gder /msg nickerv help register
10:42:34 Quit amiconn (Nick collision from services.)
10:42:34 Nick amiconn_ is now known as amiconn (
10:43:24Zagorzhilik: can you navigate the player at all without the remote?
10:44:00B4gderthere are buttons on the side
10:44:07zhilikI don't use the remote at all - it's rather primitive and has a bad connector.
10:44:39Zagorahh, the side "volume control" is for navigation
10:45:54zhilikYes. One of the reasons why I want to play around with RockBos is that I want to connect the iRiver remote to it.
10:46:05zhilikTo the HD200 that is.
10:46:06*Zagor spots "Tomato" written on the HD200 LCD :-)
10:46:20zhilikYeah. I even found the documentation for it.
10:46:42B4gder"The only problem is that if you?re trying to sync MP3 files Media Player 10 converts them to WMA"
10:46:59Zagorzhilik: excellent. add or link it to the wiki page
10:47:47zhilikI also found the DOCs for the sound chip
10:47:48B4gderzhilik: how do you update the firmware on it?
10:48:01B4gderflash-update procedure similar to iriver?
10:48:27zhilikNo, I copy the new firmware to the SYSTEM directory
10:48:45B4gderit loads it from disk and uses it?
10:49:00zhilikIt updates itself and deletes the files leavin only unicode.sys of the three.
10:49:30 Quit zhilik ("")
10:51:04 Join zhilik [0] (
10:51:58B4gderzhilik: add that info about hw and URLs to the wiki page as well
10:52:08 Join linuxstb [0] (n=linuxstb@
10:52:13zhilikYeah, just a minute.
10:52:17B4gdersure ;-)
10:52:36B4gderand I'd advice a little post to the mailing list about your work and intentions
10:52:50B4gdergetting some more people involved helps
10:57:36 Quit HCl ("Lost terminal")
10:58:06 Join hcl [0] (
10:58:12 Nick hcl is now known as HCl (
10:59:08LinusNzhilik: the unpopulated ZIF connector is most likely a BDM connector, but it has a different pinout than the iriver (not surprising)
10:59:35LinusNit will be hard to figure out without removing the cpu
11:00:26LinusNi found no hd200's on ebay :-(
11:00:38zhilikWell if we COULD find a dead player, we might also be able toread the program from the flash and dissassemble it - that may help.
11:00:43zhilikNeither did I.
11:04:50amiconnThere's one from Taiwan:
11:05:02zhilikI've included the links to the docs
11:05:29amiconnI wonder whether the HD300 is similar. I found a number of HD300 offers
11:05:31LinusNwow, the hd200 firmware doesn't seem to be encrypted
11:05:47B4gderthat'd be a first
11:06:15LinusNonly a tiny header
11:06:44LinusNwow, not even a header
11:06:52B4gderno checksum?
11:07:04LinusNmust be a checksum somewhere, perhaps in the end of the file
11:08:30amiconnHmm, HD300 seems to be different from HD200...
11:08:37zhilikSome more links (to documentation)
11:09:15zhilikWell, if the firmware is unencrypter, maybe it can be dissambled?
11:10:03LinusNi'm disassembling it as we speak
11:11:24Rickworking on a new device?
11:12:06zhilikis probably for the flash
11:12:11LinusNyes it is
11:12:15B4gderRick: zhilik's mpio hd200
11:12:29zhilikICON.. is probably the icon's :-)
11:12:50zhilikAnd unicode.sys - the main program, loaded from the hard disk.
11:13:24Rickis the player any good?
11:13:29Ricki've never heard of it
11:13:41B4gderRick: coldfire 5249 based one with a 5GB disk
11:14:04zhilikRick: it's OK
11:14:06B4gderseems similar to h1x0 spec wise
11:14:07LinusNi can imagine that unicode.sys isn't code at all, but fonts and string data
11:14:21zhilikThe sound's goos, loud
11:14:34zhilikBut the firmware stinks...
11:14:50zhilikI meant good ofcourse
11:16:14zhilikIf you have a close look, you'll see that one of the capacitor's broken - the one next to the CF connector and crystal resonator C616
11:16:16LinusNwow, hd200_upg.sys seems to be a pure image of the flash
11:16:44zhilikNow taht's GOOD news.
11:21:45LinusNwe can find out most things by disassembling
11:21:53LinusNbut it's tedious
11:22:04zhilikI mailed MPIO with a request for the curcuit (it'd be cool if they gave it, but I guess that's highly improbable) - they didn't even answer.
11:22:20Zagorzhilik: are you a programmer?
11:22:23LinusNhardly surprising
11:22:47zhilikZagor: I guess I am: know C, C++, PICasm, AVRasm.
11:23:12zhilikCurrently working on a project with MSP430,s on IAR C.
11:23:14 Join cYmen [0] (
11:25:45Zagorsounds good. then you're the right man to disassemble this :-)
11:25:46 Join markun [0] (
11:25:51LinusNiar c...shrug
11:26:56zhilikLinusN: what's bad about it? I used avr-gcc, but... gbd...
11:27:25LinusNi have really bad experience with iar and 8051, that's all
11:27:35zhilikBut I guess I'll have to get used to GDB.
11:27:49zhilikWell it's good stuff for the MSP430's.
11:28:10zhilikOnly a bit expencieve.. That's if you pay for it ;-)
11:33:48 Join Moos [0] (
11:37:35zhilikLinuxN: Maybe you could upload the disassembled file to the wiki?
11:38:24 Quit Moos (Read error: 104 (Connection reset by peer))
11:38:35 Join Moos [0] (
11:38:51zhilikOr where can I find a good 68k disassembler? Currently downloading IDA - is it any good?
11:40:04amiconnbinutils have a builtin disassembler...
11:40:29amiconnobjdump to be precise
11:40:35LinusNi use m68k-elf-objdump, works ok
11:40:57LinusNit has some issues with the control register names, but you get used to it
11:41:31amiconnLinusN: Yes, and it has problems with some addressing modes of the emac registers
11:41:53amiconnErm, with addressing modes in instructions using the emac registers
11:42:53amiconnIt disassembles move.l %<reg>,%macsr ok, but e.g. stumbles upon move.l #imm,%macsr
11:43:30 Join [1]Moos [0] (
11:45:10 Quit Moos (Read error: 104 (Connection reset by peer))
11:45:11 Nick [1]Moos is now known as Moos (
11:50:48 Join hshah [0] (
12:04:10 Quit paugh (Excess Flood)
12:04:25 Join paugh [0] (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
12:04:31 Quit paugh (Read error: 54 (Connection reset by peer))
12:08:04***Saving seen data "./dancer.seen"
12:18:35 Join preglow [0] (
12:24:30 Quit hshah ("CGI:IRC")
12:32:15 Join [1]Moos [0] (
12:33:54 Quit Moos (Read error: 104 (Connection reset by peer))
12:33:55 Nick [1]Moos is now known as Moos (
13:15:44 Join muesli- [0] (i=muesli_t@
13:20:43 Quit edx (Read error: 110 (Connection timed out))
13:21:54 Join hshah [0] (
13:24:13 Quit muesli- (Read error: 104 (Connection reset by peer))
13:27:50B4gderI fiddled with the credits list
13:27:56B4gderto make it scroll in a sinus
13:28:02B4gderso that it stops occationally
13:28:12B4gdersince the current version is completely unreadable on iriver
13:28:29LinusNwhat a good bugfix! :-)
13:28:30B4gderturned out pretty good
13:29:11B4gderwell, showing a list of names is rather useless when you can't read the names ;-O
13:29:23amiconnB4gder: I thought about scrolling in from left & right alternating per line, delaying the lower lines a bit
13:29:27amiconnThe whole thing paged
13:29:47LinusNor maybe having them fly in from the horizon :-)
13:30:12amiconnWould look a bit like a zipper
13:30:16LinusNor write them on each side of the cube
13:30:33amiconnLinusN: Are *you* going to implement texture mapping? ;)
13:30:57B4gderthat'll be in his h300 bootloader for sure!
13:31:03LinusNin any case, being abla to pause the scroll would be good
13:31:16B4gderyes, pausing makes sense
13:31:38LinusN:-) chinese magic
13:33:07hshahthis whole rockbox thing is getting better and better every day :)
13:33:10amiconnIt seems the lcds are differing in response time a bit per unit
13:33:24amiconnI *can* read the credits list on iriver (but hardly)
13:33:47hshahi can JUST about read them on mine
13:34:02hshahthats if i strain my eyes though...
13:34:34amiconnLCD response times are also temperature dependent, meaning there's a higher chance for the credits to be readable in summer...
13:35:26LinusNif we only had an RTC... :-)
13:35:29preglowit's really hard to read here
13:35:40hshahjust curious - is there an optimum operating temp for the iriver, so for example u get more out if it in terms of battery in different temps?
13:35:41preglownow please, rotozooming credits!
13:36:31hshahgreat... i decided to gulp down a bar of dairy milk, and now i feel sick :(
13:38:23*preglow also eagerly awaits second_reality.rock
13:39:49 Join webguest22 [0] (
13:40:01webguest22rockbox doesn't compile on the dev tool
13:40:52preglowmore details, please
13:40:56webguest22latest builds
13:41:08webguest22i get error messages
13:41:14preglowcvs update
13:41:22webguest22make[3]: ***No targets specified and no makefile found. Stop.
13:41:25preglowlatest build builds fine here
13:41:27webguest22make[2]: ***[libmad] Error 2
13:41:31webguest22make[1]: ***[build-codecs] Error 2
13:41:35preglowtried a reconfigure?
13:41:36webguest22make: *** [all] Error 2
13:42:01preglowfrom the build dir
13:43:36amiconnLinusN: If nobody else jumps in, I'll open up my iriver and do some power measurements with deactivated DRAM, HD, LCD and low clock. After the 2.5 release of course
13:44:15LinusNamiconn: have fun ;-)
13:44:23Moosrelease coming soon now :)
13:44:51hshahit is...
13:45:18preglowamiconn: pure iram operation?
13:46:20preglowwould be fun with a quasi rtc
13:54:50 Quit webguest22 ("CGI:IRC")
13:54:55 Join webguest22 [0] (
13:55:01webguest22arg.. got disconnected
13:55:08webguest22you mean "rbconf"?
13:55:38webguest22i mean, it's not like the first time i'm compiling using the dev tool
13:55:47webguest22it just doesnt work on the latest builds
13:55:52LinusNwhat is "rbconf"?
13:55:54B4gderso re-run configure first
13:56:09B4gderif you done it before, you ran configure before
13:56:19B4gderjust do it again, to get an updated makefile
13:57:01webguest22well, i did this on a fresh directory so i dont see why i have to do rbconf again
13:57:06B4gderrbconf sounds like some weird BC devkit "convenience" thing
13:57:08preglowsomething isn't working
13:57:10preglowthat's why
13:57:33webguest22B4gder> yeah that's correct
13:57:39LinusNthose "convenience" things make life hard for us developers
13:57:51webguest22yup it does :p
13:57:51LinusNand in the long run, for the devkit users
13:57:54amiconnpreglow: I don't intend to implement quasi rtc (that'd be too awkward to operate), but a power-on timer
13:57:54B4gderwebguest22: well, it works for us and the cvs table is green
13:58:05preglowamiconn: why too awkward?
13:58:14B4gderwebguest22: so you'd need to get more detailed
13:58:41B4gderyou using source from cvs or tarball?
13:58:42webguest22i can't really get more detailed than what i've already said but, here i go
13:58:52webguest22bleeding edge
13:59:41preglowwhy not just use the precompiled bleeding edge binaries?
13:59:55webguest22because i need to patch this thing :(
13:59:57preglowthat's what i do, unless i've got something uncommited
14:00:18webguest22but this case, im getting errors even without the patches applied
14:00:49B4gderI'd guess that there's a file or two missing in the tarball
14:00:55preglowamiconn: still, i guess a power on timer is really what i want anyway
14:01:01B4gderso again, we'd need more details
14:01:24webguest22ill copy paste the errors again
14:01:30webguest22make[3]: ***No targets specified and no makefile found. Stop.
14:01:36preglowwont help much if its the same errors, we can still read them
14:01:37webguest22make[2]: ***[libmad] Error 2
14:01:39B4gderwhat are the last few lines _above_ the errors=
14:02:10webguest22AR RANLIB /home/guest/rockbox-bleeding/buil/libcodec.a
14:02:19webguest22make in libmad
14:02:24webguest22and then error
14:03:02webguest22btw, it works fine right upto the playback.c fix
14:03:56webguest22fix where if your player is paused and you go to the next track, it's still paused
14:04:05webguest22 iRiver: Correctly keeping playback paused when skipping tracks.
14:04:52LinusNhmmm, apps/libmad contains no source files, and no makefile
14:05:44LinusNnone of the codec libs contain anything
14:05:46webguest22hmm.. weird
14:05:49LinusNin the tar file
14:07:38webguest22so what should i do?
14:07:57webguest22as of now, i'm just using "rockbox-daily-20050911"
14:08:06***Saving seen data "./dancer.seen"
14:09:02preglownot much you can do apart from waiting for a tarball that contains what you need
14:09:19preglowthat, or using cvs, which you shouldn't be unless you're capable of helping yourself
14:10:25webguest22hmm do you think it could be the problem, or incapability of the dev tool?
14:10:50B4gderyou mean dev kit?
14:10:53preglowno, not your fault, it seems
14:11:14webguest22yeah dev kit
14:11:35webguest22for example, i patch the source with this korean patch some guy made
14:11:40webguest22and it patches convbdf.c
14:11:49amiconnUsing cvs is the better way anyway. The tarballs sometimes cause problems which wouldn't happen with cvs
14:11:51preglowlibmad is quite obviously not included in the tarball
14:11:56webguest22it compiles fine on cygwin, but refuses to compile on devkit
14:12:04amiconn(like the leftover file problem Joerch got last week)
14:12:14preglowisn't bc's devkit really cygwin in disguise?
14:12:19webguest22oh i guess on this one, the missing file is the only problem
14:12:23webguest22yeah thats what i thought
14:12:37webguest22but i keep on getting the message that im missing "ctype.h" file
14:12:43webguest22which doesnt make sense
14:12:44amiconnpreglow: Yes, it's a cut-down (and outdated) cygwin
14:13:27webguest22i tried to replace ctype.h file from the latest cygwin to devkit but that didnt do any good
14:13:55webguest22so i jus replace again the convbdf to original before make on tools dir
14:14:12webguest22anyways, i guess ill just use cvs
14:14:57preglowamiconn: how nice...
14:19:03B4gderwebguest22: I fixed that package manually now
14:19:56 Join webguest43 [0] (
14:20:00webguest43hi !
14:20:13webguest43i have a question about recording:
14:20:42webguest43the iriver firmware lets me choose between: internal mic, external mic, line in
14:20:51webguest43optical in
14:21:11webguest43in rockbox i get only mic and line
14:21:59webguest43is that the internal or external mic ?
14:22:16preglowline and mic is the same connector
14:22:26webguest43thanks, that was my next question
14:22:27webguest22B4gder: lovely
14:22:30preglowonly difference is a gain change, which i dont know if iriver does implicitely or not
14:22:37webguest43so its actually a line/mic/spdif plug ?
14:22:40webguest22B4gder: so can i dl it right now then?
14:23:18webguest22ok ty :)
14:23:19preglowwebguest42: yes
14:23:25 Quit webguest22 ("CGI:IRC")
14:23:40webguest43is the pcm recording stable enough to record 2 hrs straight ?
14:24:43preglowwebguest43: might be, you'll have to find out yourself
14:25:01preglowas you've probably seen, recording is in the debug menu, and such it is for a reason
14:26:01webguest43bugger, i am new to recording, and want to record our choir/orchestra performance in two weeks
14:26:18webguest43cant decide if I record with orig. firmware or rockbox
14:26:30preglowif you're worried it wont cope, use the original
14:26:48webguest43its got a time limit though hasnt it ?
14:27:09webguest43i cant really monitor it during record as I have to sing !
14:27:19preglowhmm, it might
14:27:23preglowbut why don't you just do a test?
14:27:38 Quit webguest43 ("CGI:IRC")
14:27:51 Join webguest43 [0] (
14:28:23webguest43looks like thats what is needed :)
14:28:59preglowbtw, i _think_ iriver fw got some enhanced recording features in 1.65
14:29:04preglowmight wanna check that out
14:29:22preglowi hear some people say rockbox recording works ok, but still, no promises, it might crash horribly
14:29:51webguest43the recording glitch on the iriver firmware is another problem
14:30:14preglowyeah, but it's easily fixable with an audio editor
14:32:14 Join caxica [0] (
14:32:30caxicahey there
14:32:53webguest43is there any chance of anyone doing a peak meter while recording ?
14:33:20caxicaI have an Archos Jukebox 6000 and am trying to use a different hard drive in it
14:33:26caxicanew hd is 12 GB
14:33:34webguest43adjusting the gain without it must be difficult
14:34:02caxicathe drive works fine as a mass storage device via the jukebox
14:34:21preglowwebguest43: it'll come
14:34:48caxicabut when I unplug the USB cable and try to power it up it does not go further than the initial boot screen
14:35:04caxicahd does not spin up either to load the firmware on disk
14:35:25caxicaany ideas ?
14:35:51zhilikcaxica: maybe the new hd draw more current?
14:36:42webguest43preglow: that would be very cool, thanks
14:36:47caxicajukebox is running of mains ... if it was a current problem would it work as a mass storage dev ?
14:37:13 Quit NibbIer ("life is like a rental car, you fuck it up, and give it back.")
14:37:23 Join NibbIer [0] (
14:37:28webguest43i'll do a few multiple hour recordings then and report back here
14:38:11webguest43is power consumption while recording known ?
14:38:29LinusNcaxica: how is the drive partitioned?
14:38:39caxicazhilik: hmm new disk says 0.7A original disk says 0.55A is that a big difference
14:38:53caxicaLinusN: 1 big 12G partition
14:39:04caxicausing mkfs.dos under linux
14:39:11LinusNcaxica: does it have a partition table?
14:39:39LinusNmkfs.dos -F 32 ?
14:39:45caxicafdisk -l
14:39:48caxicaDisk /dev/sdc: 12.0 GB, 12072517632 bytes
14:39:48caxica255 heads, 63 sectors/track, 1467 cylinders
14:39:48caxicaUnits = cylinders of 16065 * 512 = 8225280 bytes
14:39:48DBUGEnqueued KICK caxica
14:39:48caxica Device Boot Start End Blocks Id System
14:39:49caxica/dev/sdc1 * 1 1467 11783646 b W95 FAT32
14:40:14caxicawill the -F 32 change much ?
14:40:36LinusNyes, otherwise it will be FAT16 iirc
14:40:58caxicaLinusN: k give me a few minutes to give that a try
14:42:14zhilikcaxica: I don't really know, was just I wild guess. That was the reason why my HDD box didn't work once.
14:44:02caxicazhilik: ok, makes sense, anyone know if a hdd with a rating of 0.7A will run in a Jukebox 6000 ?
14:44:26webguest43i remember seeing power drain measurements once, and would like to do the same while recording, anyone know how to do that ?
14:50:02 Nick Sucka`zZzZz is now known as Sucka (
14:53:47caxicaLinusN, am having trouble writing to the drive with the new format ...
14:54:00caxicaBuffer I/O error on device sdc1, logical block 23211
14:54:00caxicalost page write due to I/O error on sdc1
14:54:00caxicaSCSI error : <9 0 0 0> return code = 0x70000
14:54:00caxicaend_request: I/O error, dev sdc, sector 23275
15:00:08caxicaLinusN, am repartitioning
15:00:43caxicawhat flag do you set on the partition type ? I was using b
15:00:51caxicaie W95 FAT32
15:00:52 Quit hshah ("CGI:IRC")
15:00:56caxicano LBA
15:01:07LinusNshould be "c" iirc
15:01:16 Join edx [0] (
15:01:54caxicak c is W95 FAT32 (LBA)
15:04:44LinusNbtw, make sure that your batteries are in good condition
15:04:57LinusNand also your battery connectors
15:05:02caxicaLinusN, no luck, still coming up with errors, am using an AC adaptor
15:05:25LinusNit still uses the batteries, even if you connect the charger
15:05:29caxicabatteries are not in the device
15:05:42caxicaI guess I should put them in ?
15:06:27LinusNyou connected an ac adaptor to the battery connectors?
15:06:48caxicano to the normal AC in socket
15:07:06LinusNthen you should absolutely insert the batteries
15:08:14caxicawill have to wait for tomorrow, batteries are low, will use a HD adaptor and do the setup of the drive now that I have a little more info
15:08:32caxicawill setup the drive in a computer
15:08:55caxicathanks for your time
15:09:57 Quit caxica ("Leaving")
15:10:53webguest43ah, battery level monitoring ?
15:33:36 Quit Febs (Read error: 110 (Connection timed out))
15:34:30LinusNincredibly suspicious post on mr:
15:36:29preglow_h300_ ?
15:37:09LinusNi guess someone wants his 15 minutes of fame
15:38:21CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
15:38:21*B4gder checks the video
15:38:30LinusNthe video is a laugh
15:38:38B4gderits like one of those UFO videos
15:38:42CoCoLUSwell, he got your attention
15:38:52B4gderits a moving glowing thing
15:38:57coobdid he put vaseline on the lens?
15:39:43coobcan the h300 fw lod images? if so...
15:40:01B4gderI bet it can
15:40:23 Quit ashridah ("sleep (boggle!)")
15:40:37preglowit can, yes
15:40:39coobi cant see anything moving ont he scree
15:40:53B4gderno, but the screen moves a lot! ;-)
15:42:46preglowanyone know of any other free layer1/2 encoders than toolame?
15:43:08LinusNgotta go
15:43:12coobdoesn't ffmpeg have one?
15:43:18 Part LinusN
15:43:20 Quit t0mas ()
15:43:25preglowffmpeg has encoders?
15:44:30cooblibavcodec has an mpeg4 encoder, i don't see why not!
15:45:40coobThe command line:
15:45:40coobffmpeg -i <inputfile.ext> <file.mpg>
15:45:41coobwill create a MPEG file with MPEG-1 video and MP2 audio.
15:45:55coobso yep, i guess, although i dare say it's floating point
15:48:47coobmpeg 2 encoder
15:50:01coobah, mostly fixed point oo
15:51:31preglowooh, niceness
15:51:57*preglow spots a libmad spot needing asm
15:52:51coobpolyphase? or is that already done?
15:53:15coobalso do you guys only use libmad at 16 bits?
15:53:20preglowfirst thing i did
15:53:27preglow16 bits? you mean the output?
15:53:46coobi though libmad rendered at 24
15:53:49preglowit does
15:53:56preglowwe dither it down to 16
15:54:08preglowdac doesn't support dma for bps over 16
15:54:23preglowthat is, we used to dither it down, i disabled it 'cause of being too slow
15:54:24coobdoesn't that seem a little wasteful?
15:54:30preglowwhy wasteful?
15:54:36preglow16 bits are enough for me, certainly
15:54:45coobto render at 24 when the output only supports 16
15:54:57preglowahh, no
15:55:10preglowthe result wont be the same if you only use 16 bits throughout, believe me
15:55:34preglowwe had a limited precision polyphase stage some months backj
15:55:42preglowthe quality really went up when i fixed it to use full precision
15:55:53coobah ok
15:56:01preglowbesides, you'd have to hack libmad quite a bit to make it use another precision internally
15:56:16preglowand you wouldn't save much on most platforms unless you went really low
15:56:36preglowthat is, on arm i believe you save something for every bit you make zero in the muls
15:58:26coobi'm no optimisation freak heh
15:58:33coobi'm lucky if i can get something to work :)
15:58:49coobthere's so much we could do but donn't out of a lack of dedication/time
15:59:03B4gderisn't that always the case? ;-)
15:59:24preglowbut i always use the best precision i can in the lossy codecs
15:59:34cooband a lack people working on the project
16:00:09coobthis is ipods we're talking about, people care more about bootsplash screens than audio playback :)
16:00:31B4gderthat's why you should port rockbox to it
16:01:01 Quit Moos (Read error: 110 (Connection timed out))
16:02:33B4gderas then preglow would finally do his arm assembler ;-)
16:02:55preglowoh, and joyus times it would be
16:03:00preglowjoyous, even
16:04:33*preglow prods Slasheri
16:07:12amiconnpreglow: I had an idea about realtime mpeg encoding. On amiga, there's Pegase, an (now) open-source layer 1 + layer 2 encoder
16:07:31amiconnIt runs at approx. 50% realtime on 68060/50.
16:08:03amiconnI checked the source, it's asm optimised. Unfortunately it uses the fpu...
16:08:09***Saving seen data "./dancer.seen"
16:16:23 Join paugh [0] (n=kickback@2001:5c0:8fff:ffff:8000:0:3e03:6822)
16:16:30preglowamiconn: hrmph
16:16:41preglowamiconn: decent quality?
16:17:10amiconnYes, the quality is quite good
16:18:41amiconnCheck Pegase_src.lha (the newer one)
16:19:26preglowand it's c++...
16:19:48preglowuses the fpu quite extensively, hrmph
16:20:48amiconnIf we could use some of the optimisation ideas... I think we have a higher chance to get layer 2 encoding realtime than layer 3
16:21:03preglowoh, yes
16:21:04preglowby far
16:21:29amiconnJust for comparison: lame runs at 3..10% realtime on 68060/50...
16:21:40preglowlame would also be a bad encoder to choose
16:21:46amiconn(10% with 128kbps cbr, 3% with −−preset standard)
16:22:02 Join Lynx0 [0] (
16:23:28preglowunfortunately, i'm not all that interested in lossy encoding (with the possible exception of speex) on portables myself
16:23:31amiconnLayer 2 isn't all that bad at adequate bitrates
16:23:32preglowwe'll see
16:23:38preglowi think layer2 is quite good
16:23:56preglowi've got some 192kbps layer2 files that are excellent quality
16:24:02amiconnDAB uses layer2 at 192 kbps
16:24:20coob128 also
16:24:20amiconn(and sometimes 128 kbps)
16:24:29coobnearly all dab stations in the uk are 128
16:24:37coobbecause the multiplexes are jammed
16:24:43coobi think there's like 2 192 stations
16:24:56preglowi also think i prefer the layer 2 artifacts to the layer 3 ones
16:24:59preglowno glitching imdct
16:25:44amiconncoob: Here in germany practically all stations use 192 kbps
16:26:06B4gderpeople there actually use dab?
16:26:18amiconnpreglow: The pegase encoder is quite good, imho it's even better than toolame
16:26:36amiconnB4gder: I have a DAB car stereo...
16:26:48*preglow never tried dab
16:26:53coobyes but how many stations do you have on how many multiplexes?
16:31:39amiconnabout 20 stations on 4 multiplexes in Berlin
16:32:08coobi think we have about 30 on 5
16:34:01preglowwhy can't people use _either_ tabs or spaces?
16:35:15coobwelcome to open source :)
16:36:36preglowone indent = two spaces, two indents = one tab, three indents = one tab, two spaces
16:38:14 Quit Lynx_ (Read error: 110 (Connection timed out))
16:38:14 Nick Lynx0 is now known as Lynx_ (
16:41:50Zagorin rockbox we only use spaces
16:42:46amiconnZagor: In theory... yes
16:42:59amiconnCheck our sources, there are tabsssss....
16:43:14Zagorcorrection: in rockbox we should only use spaces
16:43:28coobtabs are where its at :)
16:43:51Zagorwe've already decided this long ago
16:44:02preglowZagor: codecs use tabs a lot
16:44:06Zagordocs/CONTRIBUTING even :-)
16:44:10amiconnZagor: I know, but tabs keep sneaking in for whatever reason
16:44:10preglowZagor: and it's been decided not to change that
16:44:40Zagorpreglow: yes, codecs are a special case since it's not our own code
16:44:46amiconnI throw them out whenever I encounter them...
16:45:14Zagoramiconn: I know, it's not so strange. I'm just pointing out that we already know how we want it, so nobody has to wonder.
16:45:38preglowi don't mind tabs, i do mind using both tabs and spaces
16:45:41preglowthat's just nasty
16:46:06Zagortabs are nasty as soon as you reindent, print or pretty much do anything other than look at the code
16:47:05preglowi even like tabs for being able to change the indent size people use
16:47:24preglowZagor: why nasty when reindenting?
16:47:41coobthey're actually easier to use :/
16:47:48Zagorbecause when you add four spaces in front of a tab, it doesn't do what you want it to
16:48:02coobso don't mix tabs and spaces
16:48:07coobyou're not supposed to.
16:48:10Zagorno. don't use tabs.
16:48:20Zagorredefining the size of tabs breaks EVERYTHING
16:48:22preglowyeah, but then you're suddenly mixing the two
16:48:33preglowas long as you just use either, i don't see much of a difference
16:48:35coobno it doesn't
16:48:36preglowZagor: no?
16:48:43preglowZagor: not if you just use tabs
16:48:44coobi suggest you learn what a tab actually is.
16:48:58preglowbut anywho, gotta get food
16:49:31coobusing tabs is like using <div>'s to markup a page, where as using spaces is like using <img src="spacer.gif">
16:49:37Zagorcoob: fine, grab some code with tab=3 and put it on the web. or print it. or prettyfy it. have fun.
16:50:04Zagorcoob: don't start. we have a rule.
16:50:22Slasheripreglow: hi
16:50:38coobi suggest you calm the fuck down, this is tabs and spaces we're talking about, i wasn't suggesting anything change
16:50:51coobif you're mixing tabs and spaces, then you don't know what a tab is.
16:51:14Zagori'm not mixing. i'm saying "use only spaces in rockbox"
16:51:40coobi wasn't saying otherwise.
16:51:58Zagornice. then we agree!
16:53:23 Join webguest44 [0] (
16:54:55webguest44Slashei: would you be ineterested in implementing replay gain 'Track' mode if shuffle is selected, as a feature, please :)
16:55:28webguest44Slasheri even
17:04:18Slasheriwebguest44: Hmm, maybe but there must be an option also to enable/disable that behaviour
17:04:37linuxstbpreglow, Another layer-2 encoder is mp2enc - I think it's part of the mjpegtools project
17:05:52webguest44Slasheri, yes, true, but in my humble, honest unassuming opinion, it would be a great feature
17:05:57Slasheribtw, does anybody know how i could set replaygain _tags_ for oggs/mp3s under linux?
17:06:15crwlSlasheri, for oggs with vorbisgain
17:06:23crwlif there's similar tool for mp3 files, let me know too :P
17:06:25Slasheriwebguest44: Hmm, i will try remember that after the feature freeze is over
17:06:32Slashericrwl: hehe, thanks ;)
17:06:58B4gderthe replaygain situation seems to be... eh, f*cked up
17:07:06B4gderthere's only one tool
17:07:15B4gderand it screws the tags for just about everyone
17:07:16webguest44foobar also does ogg gaining
17:07:23Slasherihmm :/
17:07:47B4gderso I would say the world needs a tool that works
17:07:57webguest44doesnt vorbis gain write the info to the header, and foobar to the tag ?, not sure
17:08:11B4gdernot that _I_ use it ;-)
17:08:48crwlwebguest44, no, vorbisgain uses them REPLAYGAIN_blahblah tags
17:08:57crwland iirc it's the original tool to do the job
17:09:21webguest44does it store the info in ape tag or V2 tag ?
17:09:28crwlof course not
17:09:35crwlit naturally uses vorbis comments
17:09:56webguest44silly me
17:10:06linuxstbI know no-one's asked, but it's worth saying that metaflac can calculate replaygain for FLAC files - also using vorbis comments (I assume).
17:11:48 Part webguest44
17:22:50 Quit B4gder ("Lämnar")
17:28:05 Join lostnihilist [0] (
17:28:11crwlyes, that's true
17:28:35crwlsomebody should hack the commandline mp3gain to write id3v2 replaygain tags...
17:30:40amiconnI think mp3gain is the better solution. It doesn't require any tags or special support in the player
17:30:59amiconn(Not that I would care. I never used mp3gain or replaygain, whatever)
17:31:18preglowSlasheri: oh, hi, you didn't actually start on a wavwriter, did you?
17:31:30crwlamiconn, but it has that annoying problem with non-replaygain-aware players that mp3 files will then sound much more silent than oggs/flacs/whatever else you have
17:31:50crwland it's also less accurate and doesn't support track and album gain at the same time
17:32:23preglowindeed not
17:32:33Slasheripreglow: i haven't had time yet to do that..
17:32:37preglowit's also got bad grain granularity
17:32:46amiconn1.5 dB
17:32:48preglowSlasheri: no, was just wondering, i might try it myself tonight
17:32:56preglowSlasheri: any pointers on how to quickly achieve it?
17:33:00Slasheripreglow: ah, sounds good :)
17:33:35SlasheriHmm, if you would like to do it easily, you should probably write a pcm driver that pcmbuf uses
17:33:57Slasherijust look the simulator audio playback code to get clue what functions you would need to implement
17:34:22preglowoh, i think i know how to do that part
17:34:28Slasherigood :)
17:34:34preglowwill just need to hook it up to a context menu item or something
17:34:41preglowa lot of code i'm not at all familiar with
17:38:08webguest43hmm, would someone have some time for the recording menu ? Would it be possible to move the "monitor" button from the remote to the main unit ?
17:38:34webguest43'cause you dont need the remote for anything else than this button
17:39:01preglowwell, isn't it there for a reason? are there any main unit buttons left?
17:42:03webguest43well, I am not very sure what the "play-pause" button does in record mode
17:42:21webguest43and pressing the stick-thing does not do anything
17:43:23preglowbut doesn't the monitor function actually use the remote for drawing?
17:43:43webguest43no, it toggles the audio on/off on the headphones
17:44:22webguest43i guess you use headphones for adjusting gain, and then switch it off to preserver power
17:45:18webguest43you mean the wave thing, which is.. well,... nice :)
17:46:12webguest43does not help a lot for setting gain, as you can hardly see it
17:47:09webguest43but if you look into that, could you please make it check if a file already exists ? it overwrites files when you exit recording, and come back to do another recording
17:47:33webguest43i know its debug, but it seems to work fine, just some rough edges :)
17:48:18preglowcan't look into now, making dinner
17:48:48webguest43would be really nice if you could do that !
17:49:45preglownot promising anything
17:50:04webguest43just a wishlist: 1) check if file exists, if yes increment filename-number 2) move monitor switch to main unit 3) display peak meter on main unit
17:51:29webguest43thats allright, i will do some testing for stability
17:51:49webguest43but i think the main thing is to stop it overwriting files
17:56:08 Join ryan_j [0] (n=ryanj@
18:00:33 Quit zhilik (Read error: 110 (Connection timed out))
18:04:40 Join ]RowaN[ [0] (
18:05:00]RowaN[is it me or has the daily build page been reporting "Build expected to complete around 10:49" all day long?
18:05:17preglowjust you
18:05:21preglowor more specifically, your cache
18:05:56]RowaN[ah yes, stupid IE
18:06:28 Quit ]RowaN[ (Client Quit)
18:08:11***Saving seen data "./dancer.seen"
18:17:30 Quit einhirn ("Miranda IM! Smaller, Faster, Easier.")
18:19:10 Join crash___ [0] (
18:19:15preglowa statement like (void)heh, (void)hah; is correct c, yes?
18:19:42 Quit linuxstb ("Leaving")
18:21:07amiconnpreglow: If you just want to deliberately ingnore multiple parameters, there's an even shorter way
18:21:11 Quit crash__ (Read error: 104 (Connection reset by peer))
18:21:16amiconn(void)bla, blubb;
18:23:47amiconnMost places in rockbox use the slightly longer method of (void)ing each argument separately though
18:25:21preglowamiconn: that's what i wanted, yes, thanks
18:25:57preglow/home/thomj/rockbox-devel/apps/codecs/libmad/layer3.c:1558: warning: statement with no effect
18:26:02preglowthat's what i got then
18:26:38preglowworked fine with the double void
18:26:58 Quit mbr (Read error: 110 (Connection timed out))
18:27:05amiconnHmm, it seems the cast is stronger than the comma operator
18:27:17amiconn(void)(blah, blubb); should work
18:27:56preglow/home/thomj/rockbox-devel/apps/codecs/libmad/layer3.c:1558: warning: left-hand operand of comma expression has no effect
18:28:00preglowthat's what i got then :)
18:28:00 Nick crash___ is now known as crash__ (
18:28:18amiconnBleh, sometimes I hate compilers ;)
18:28:47preglowpretty strange, the longer form is equally without effect
18:29:57amiconnIf the compiler doesn't like the short form, you could use the explicit form as well. Same amount of characters...
18:29:59preglowhow i hate objdumps emac abilities
18:30:12preglowit doesn't even differentiate between mac and msac, sometimes
18:30:39amiconnYes, and it doesn't properly disassemble move.l #imm,%macsr
18:31:13amiconnIt shows that as move.l %a4,%macsr followed by a '?' line (iirc)
18:31:34preglowit also omits the acc0 and acc2 registers
18:37:44HClitunes is evil
18:39:00amiconnHCl: Whom do you tell...
18:40:02 Join zhilik [0] (
18:40:46 Join mbr [0] (
18:44:57 Join hubbel [0] (
18:46:24HClamiconn: people who understand me :)
18:56:39 Part hubbel
18:57:04 Join arkascha [0] (
18:57:11 Join hubbel [0] (
18:57:32 Nick Lynx_ is now known as Lynx_awy (
19:07:34 Join dpassen1 [0] (
19:25:14 Join matsl [0] (
19:34:22 Quit arkascha (Remote closed the connection)
19:42:41preglowoooh, new boards of canada album
19:44:43amiconnThe recording time based frame count estimation is working really good. Precision is better than 10^-4
19:45:39preglowwhat happens if the estimation is off?
19:47:28amiconnThe MAS does VBR recording only, so we always write a xing header
19:48:04amiconnFor a proper xing header we need at least the number of bytes (easy) and the number of frames
19:48:22preglowand you can't count frames properly?
19:48:32amiconnThe latter is a bit problematic. The MAS does have a frame counter, but it saturates at 0x7FFFF
19:48:41preglow_saturates_ ?
19:48:50preglowwhy didn't they make it overflow?
19:48:53preglowwould have been lots better
19:48:57amiconnAsk Micronas...
19:49:50amiconnIt means we can't get the frame count from the MAS for long recordings, or when the MAS was in recording mode for a long time (e.g. when doing prerecording)
19:49:56preglowthe day i send letters to all the annoying people in the world, they'll get their letter too
19:50:12preglowbut yeah, how do players usually react to having a wrong frame count in the header?
19:50:36amiconnWe didn't write a wrong count, we didn't populate the field at all
19:50:42preglowahh, yes, but now
19:50:48amiconn(there's a bit telling whether the frame count is valid)
19:50:52preglownow you write the estimate, yes?
19:51:22amiconnMost players didn't care (they couldn't display the average bitrate though), but some don't like that
19:51:56amiconnNow that I improved the framecount<->playtime conversion precision, it is possible to estimate the frame count from the recording time
19:53:18amiconnBtw, 0x7FFFF equals ~ 03:48
19:53:39preglownot too shabby, then
19:53:57amiconnRockbox is stable enough to record for days on archos...
19:54:11preglowmy brain must be asleep not to automatically assume that
19:54:45amiconnThe 03:48 is at 44.1kHz of course
19:55:06webguest43hmm, sorry as i am just testing long recording, you are not talking about PCM recording are you ?
19:55:22amiconnwebguest43: No, I'm talking about recording on archos
19:55:36webguest43ah, fine then
19:55:55webguest43does anyone know how long 4GB of wav is ?
19:57:07flurbledoesn't it depend on the sample rate?
19:57:13amiconnDepends on the sample precision, # of channels and sample frequency
19:57:40webguest43in pcm recording on iriver
19:57:41Slasheriyou can calculate it: 4*10^9 / (44100 * 4)
19:58:04amiconnFor 44.1 kHz 16 bit stereo it's ~06:45
19:58:20 Join adiamas [0] (n=adiamas@
19:58:20amiconnSlasheri: 4*2^30 ...
19:58:28preglow10^9, hahaha
19:58:32Slasheriamiconn: No, he said GB, not GiB..
19:58:53*preglow learns what GiB means
19:59:09webguest43ah, maybe wont wait that long then.
19:59:18webguest43pcm recording for 2 hours now, no probs at all
19:59:35amiconnSlasheri: In conjunction with bytes, G is always interpreted binary, except by harddisk manufacturers
19:59:39preglowi'd love to see the XiB measurements vanish
19:59:59Slasheriamiconn: and except by _all_ telecommunication devices and data transferring
20:00:31Slasheriall network bitrates are in true SI-prefixes
20:01:15amiconnTell that to the DSL providers. They could save some bandwidth at no cost...
20:01:35amiconnThe 10/100/100 MBit/s are, indeed
20:02:12SlasheriHmm, do you say ISPs are providing for example 512 Kibit/s DSL instead of 512 kbit/s? I haven't seen that to happen
20:02:21amiconnThat's about the only ones I know. And ISDN is 64kbit/s, clearly different from Kbit/s
20:03:52Slasherii think data storage is almost the only place where binary prefixes are explicitly being used
20:05:05zeyeah the *iB stuff doesn't make sense to me
20:05:15ze1024 = 2^10, etc
20:05:37preglowi just think it's annoying and unnecessary
20:05:45webguest43well, having said "no problems"....
20:05:53webguest43i cant stop the recording...
20:06:16preglowhooray! :-)
20:06:17Slasheripreglow: yes, until more people start using that.. fortunaly many programs display the units correctly today
20:06:29amiconnSlasheri: For DSL, I'm convinced they do binary calculation. Otherwise it won't be logical to provide 1MBit/s as the next step above 512KBit/s
20:06:33Slasherize: please check this:
20:06:49webguest43ah, me being idiot, ... hold button ...
20:06:54amiconnWith decimal bitrates, 500kBit/s would be logical...
20:07:16Slasheriamiconn: Hmm, maybe some do.. but i really think not all
20:08:02zebut 1000 = 2^9.96578428466209
20:08:11Slasheriand for example wi-fi and bluetooth maximum theoretical bitrates are in SI
20:08:14***Saving seen data "./dancer.seen"
20:08:27Slasherize: no?
20:08:28preglowanywho, i think we all can agree si units are the easiest and most logical ;)
20:08:36 Join DangerousDan [0] (
20:10:08zebefore all this, harddrive manufacturers (and more recently dvd media manufacturers) were the ONLY ones i've ever seen mean units of 1000
20:10:39zeand honestly, keebee, mehbee? you'd sound like a retard saying that stuff :p
20:10:51webguest43ok, so i repeat 2 hrs recording on h140 worked fine and used 19% of battery
20:12:51preglowi sound like a retard whatever i say, so try to keep it simple
20:13:02preglowwebguest43: so, what was the problem with the stopping?
20:13:18webguest43me retard.... hold button ....
20:13:33webguest43all worked fine, recording looks ok
20:13:41preglowgood to hear
20:13:48preglowi hope hubbel will come back and finish what he started soon ;)
20:14:09Slasherize: that's because most of the units are not visible to user. Mainly the storage thing is the only one user really sees and care about. And it's naturally that file system size is expressed in GiB. How ever there is no logical reason at all to use that binary prefix anywhere else (especially in telecomm.)
20:14:13webguest43if someone could fine-tune the recording GUI that would really be nice
20:14:42Slasheriuh, typos++
20:15:04preglowwebguest43: well, it doesn't really look like it should at all
20:15:19preglowwebguest43: it's going to be integrated with the ordinary recording screen, i'd imagine
20:15:33webguest43"ordinary recording screen " ? ?
20:15:34preglowamiconn: or what do you say? is the recording code very lowlevel?
20:15:49preglowwebguest43: well, yes, some of the archoses have been able to record for a long time
20:15:56amiconnI don't know the iriver recording code
20:15:57preglowso rockbox has a recording screen
20:16:21preglowamiconn: i'm wondering more if the iriver recording will be easy to integrate in the existing recording facilities
20:16:24amiconnThe archos recording code itself is part of the mpeg thread, as is the playback
20:16:32preglowor if lowlevel and gui code is very intertwined, etc
20:16:54amiconnThe recording screen is pretty highlevel, it mainly sends messages to the mpeg thread (via functions)
20:17:18amiconnThen there's the peakmeter, which needs to be adapted for recording
20:18:16amiconnTriggered recording integration might need some more thinking
20:18:37preglowhmm it monitors the peak values and acts accordingly.
20:18:47preglowi would think the existing peak meter should be easily adaptable
20:18:48amiconnIt's also app code
20:19:26webguest43well, in my humble opinion, make a simple peakmeter so one can easily adjust gain, some code so it does not overwrite its own files and a few better buttons and all is well
20:19:30amiconnThe messages are something like "start recording", "split and start a new file", "stop"
20:19:47preglowwebguest43: i'd also like to see some kind of volume limiter for those unexpected peaks
20:20:11amiconnTime split and recording size monitoring (for splitting when the file becomes too large) is done at app level, buffer flushing is low level
20:20:24webguest43nice, but once it stops overwriting its last files its fully useable
20:20:50amiconnwebguest43: There is a reason why recording is in the debug menu.
20:21:02amiconnYou can be sure that this is not the final thing...
20:21:06webguest43sure, but not much is missing to make it a useable feature
20:21:35webguest43just preventing it to overwrite files and moving the monitor button on the main unit
20:22:00webguest43all the rest is bells and whistles, (well peakmeter would be nice too)
20:26:44preglowoh, yes
20:26:48preglowand we'll need encoders
20:26:53preglowthe wavpack one is already ready
20:27:02webguest43what for.
20:27:07preglowmore disk space
20:27:15 Join linuxstb [0] (
20:27:15webguest43i did record in mp3 a few days ago.
20:27:16preglowmore recording on a battery charge
20:27:28webguest43now i have to convert to wav to do the editing and then go back to mp3
20:27:50preglowwavpack is lossless
20:28:02webguest43all right, point taken
20:28:16webguest43can it be edited without extracting ?
20:28:35preglowcan't quite remember, depends on the editor, i guess
20:28:45preglowcool edit pro/adobe audition has direct support, i _think_
20:29:41 Join solex [0] (
20:30:52webguest43still, once the gui is improved its fully useable
20:32:48preglowyes, sure
20:33:12preglowbut gotta watch som election tv, brb
20:35:47webguest43you german ?
20:38:03webguest43ah, well i hope TV is very boring :)
20:41:10preglowit always is
20:41:34webguest43good !
20:41:53webguest43quick question: why can I set the battery capacity in the settings ?
20:42:03webguest43or better what does it affect ?
20:44:18preglowthe runtime estimation, i guess
20:44:29preglowwhich is pretty much rubbish for iriver anyway
20:45:44 Quit solex_ (Read error: 110 (Connection timed out))
20:51:20webguest43ok, it not about charging or something
20:52:30webguest43could you not do a calibration thingie ? my laptop has that, i till fully charge, then do a full discharge while simulating normal operation, in the end it should get a good estimate
20:54:39preglowiriver charging is strictly hardware controlled
20:55:59webguest43ah ok, another thing, i just found about 300 wma's on my iph140, anyone know a good/simple wma => mp3 converter to exorcise them ?
20:57:43preglowthe only wma i had, got converted to wavpack
20:57:51preglowon the grounds of it already being lousy quality
20:58:13linuxstbI think foobar will convert anything to anything.
20:58:55paughlead to gold!?
20:59:01preglowahh, yes
20:59:07preglowfoobar is good for transcoding
20:59:16linuxstbpaugh: If you can find the right plugins
20:59:33 Join wintermute [0] (
21:02:25 Quit TCK (Read error: 104 (Connection reset by peer))
21:03:31webguest43being dumb again, cant find any transcoding in foobar2000 ?
21:06:26 Quit ryan_j ("leaving")
21:10:24hubbelhave anyone else had bmp images in the WPS inverted? maybe photoshop saves the images badly
21:11:03preglowgoing to work more on recording soon? :>
21:13:09 Nick Sucka is now known as Sucka`away (
21:13:25hubbelpreglow: sometime, but not in the coming weeks i'm afraid.. shouldn't be that big of a job to integrate into the normal recording screen though..
21:28:17 Quit Nilisco (Client Quit)
21:29:01webguest43hmm, cant find any way to do wma=>ogg or mp3
21:29:08webguest43help anyone ? please ?
21:30:16preglowi'm in linux, so can't help
21:31:37webguest43well, got linux, so shoot !
21:32:07webguest43what a bummer, must have been my ex-girl encoding in WMP
21:32:38wintermuteUse foobar2000.
21:33:19wintermuteWait, are you using Linux or WIndows?
21:38:27linuxstbUnder Linux (or probably Cygwin), you could try this wma2wav converter:
21:38:45linuxstbIt uses the "ffmpeg" open source WMA decoder.
21:39:22 Join Nilisco [0] (
21:39:45linuxstbThis is the decoder we're waiting for someone to convert to fixed-point....
21:41:20webguest43im on windows, i have foobar2000 installed, but cant find any conversion in the gui ????
21:41:46webguest43foobar 0.8.3 that is
21:44:30 Quit webguest43 ("CGI:IRC (EOF)")
21:44:40 Quit wintermute ("CGI:IRC")
21:47:09linuxstbwebguest43: I think you just select some files, right-click and use one of the menu options there. But I've only used foobar a couple of times.
21:59:43 Join webguest43 [0] (
21:59:51webguest43hi, sorry, had to reboot
22:00:22 Quit webguest43 (Client Quit)
22:00:26 Join webguest43 [0] (
22:00:57webguest43but found a nice tool to do wma => ogg:
22:01:16webguest43its free, they just have to charge for the mp3 encoder
22:01:30webguest43which is disabled after 30 days if not paid
22:02:12webguest43fine by me, doing ogg and flac now, long live open source and gnu
22:06:25 Quit webguest43 ("CGI:IRC")
22:08:18***Saving seen data "./dancer.seen"
22:15:10 Join webguest43 [0] (
22:23:23 Join xNibbler [0] (
22:24:10 Quit NibbIer (Read error: 104 (Connection reset by peer))
22:24:23 Quit zhilik ("")
22:39:02 Quit Zagor ("Client exiting")
22:44:59 Quit webguest43 ("CGI:IRC")
22:45:12 Quit matsl (Remote closed the connection)
22:45:31 Quit hubbel (Read error: 110 (Connection timed out))
23:05:29 Join XavierGr [0] (
23:25:43 Join muesli- [0] (
23:26:57muesli-hi king otto ;)
23:27:35muesli-xavierakles ;)
23:27:37XavierGrHi :)
23:28:16XavierGrSo any news about the upcoming 2.5 release? I have lost some episodes in this effort...
23:39:47 Quit paugh ("Leaving")
23:41:55muesli-XavierGrwill there be a new rls from you?
23:44:07 Join Moos [0] (
23:44:51MoosRe guys
23:47:23muesli-bonjour madame :-)
23:47:51Mooswhy madame, monsieur,no ;)
23:48:05muesli-just kidding ;)
23:48:17muesli-how do you write madmoiselle!?
23:48:48muesli-cheers ;)
23:51:37 Join Mojito [0] (
23:52:18Mojitolast thursday i ordered a iriver h340 :)
23:52:28Mojitonow i've found the rockbox project
23:52:35Mojitoand i think its very great! :D
23:52:46Moosthe best ;)
23:53:00XavierGrmuesli: Currently I gave up my remote build, too many changes to cope with.
23:53:14Mojitonow i download cygwin ;)
23:53:17muesli-buhuuuu *cry*
23:53:25XavierGrIf I find some time later in this week I will start making the remote build again from scratch
23:54:01XavierGrThough I doubt to succeed making it correct, maybe I will stick with the quick hack again, time will tell.
23:54:23Moosmaybe one core devloper will can help you :)
23:54:34Mojitoif i want code for the h340, can i write only in assambler?
23:55:20muesli-XavierGr doenst matter how buggy it is..the main thing is that theres support matter how it works
23:56:56muesli-does rbx using a specific language anyway?
23:57:13 Join stripwax [0] (
23:57:33stripwaxany had experience of iriver warranty repairs of an ihp with rockbox installed on it?
23:57:36XavierGrmuesli: Problem is that I don't have a clue on how to programm it using a nice way, even thinking of it makes me wonder...
23:57:52XavierGrdid you broke yours?
23:58:07stripwaxXavierGr - is that to me?
23:58:22 Join ashridah [0] (
23:58:39Mojitocan i code with java for a jukebox like irivers ?

Previous day | Next day