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

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

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

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

#rockbox log for 2007-11-20

00:00:58linuxstbWhen did you install the Rockbox bootloader?
00:00:59amiconnJdGordon: around?
00:01:15 Join dandin1 [0] (
00:01:21amiconnI'm not sure I understand your patch (or the comment)
00:01:23chucklast week
00:01:30JdGordonwhich part
00:01:35amiconnWhat needs fixing in the existing formatters?
00:01:54 Join J3TC- [0] (
00:02:02JdGordoncurrnetly the are still using the value as an index.. that needs to be removed
00:02:10 Join advcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
00:02:22 Quit advcomp2019 (Nick collision from services.)
00:02:30 Nick advcomp2019_ is now known as advcomp2019 (n=advcomp2@unaffiliated/advcomp2019)
00:02:30amiconnAh, hmm, I thought you meant all int value formatters...
00:02:36chucki have actually installed the bootloader several time, after updating my ipod version
00:02:43JdGordonno, just the ones using table_setting
00:02:50 Join Soap [0] (n=Soap@rockbox/staff/soap)
00:03:08pixelmachuck: did you only install the build for 60GB/80GB Ipod Videos?
00:04:59pixelmacould you try the other one (the 30GB version)? It is safe for you, in the worst case it doesn't use all available RAM
00:06:45chuckok, i'll try that one
00:06:45*stripwax wonders if translucent selection and drop shadow fonts are achievable on target
00:06:54amiconnForget it
00:06:56stripwaxreckon I can get a feel for target performance from the sim build?
00:07:19amiconnAlpha transparency would be slower - I'd expect a factor of 10 or more
00:07:23amiconnNo, you can't
00:07:31 Quit Redbreva ("ChatZilla 0.9.79 [Firefox]")
00:07:36amiconnThe sim is tons faster than the targets
00:07:37PaulPositionstripwax - even pre=rendered greyscale antialiasing has been refused, so...
00:08:01 Join mirak [0] (
00:08:16stripwaxamiconn - right but I wonder if I can get any meaningful results from a profiler
00:08:16 Quit jhMikeS (Read error: 104 (Connection reset by peer))
00:08:23amiconnPaulPosition: That's due to size though, not performance
00:08:29 Quit mirak (Connection reset by peer)
00:08:30LloreanPaulPosition: Prerendered grayscale antialiasing would only work with very light backgrounds, wouldn't it?
00:08:37stripwaxyeah that's no good
00:08:47PaulPositionamiconn - Ah, right.
00:08:48 Join mirak [0] (
00:08:57Lloreanstripwax: There was a "Fast antialiasing" algorithm on the tracker, I think
00:09:30preglowstripwax: don't expect too valuable profile information on the sim either
00:09:31stripwaxa flat translucent selection should be easy to achieve. and I have a hacky drop shadow routine
00:09:32PaulPositionllorean - It wasn't a fast antialiasing at all, or anyway all I could find was a rejected page for fake-antialiasing by using greyscale fonts..
00:09:38stripwaxpreglow - okie doke
00:09:43PaulPositionOr maybe I understood it wrong. :o
00:09:57*PaulPosition shuts up before saying any more stupidities.
00:09:58stripwaxLlorean- I'm not looking at doing antialiasing
00:10:01pixelmaI still doubt that aliased fonts in the size we currently use will look that much better (maybe later with larger screens and smaller pixels)
00:10:18pixelmaand a larger font of course
00:10:31linuxstbstripwax: BTW, did you notice that about 5 minutes after you updated some themes last week with the new %s tag, we changed it to %m?
00:10:43LloreanPaulPosition: No, I think you're right there.
00:10:46stripwaxlinuxstb - yep, and I already updated those themes
00:10:51stripwaxbut yeah
00:11:02PaulPositionpixelma - I'm going way off topic.. But I'd have love to port the pen and paper wps to the (very) small h10 device. But handwriten pixel-fonts at 8 or 9 pts doesn't work. :p
00:11:03chuckok, i downloaded build 1709 for the 30 gig, and when i tried to play a mp3, i go the following message "data abort at 000089F0 (0)
00:11:48 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
00:11:55linuxstbchuck: When you said other things in Rockbox worked - what have you tried?
00:11:56pixelmahmm... then that's not what I suspected, sorry.
00:11:59stripwaxbut if I was going to take a look at antialiasing, I'd probably take a font four times bigger than I need and scale it down in the lcd_bitmap_mono_part fn
00:12:56chuckgames, themes, fonts, files, even loaded my database from my itunes partition
00:16:52stripwaxchuck - you don't happen to have two rockbox.ipod on your device do you?
00:16:52pixelmacould something be up with the mp3s themselves, did you try other music formats?
00:16:52chuckthere is only one .rockbox folder on my ipod
00:16:52stripwaxchuck - is there a file named rockbox.ipod in the root
00:16:52 Quit petur ("here today, gone tomorrow")
00:17:20chuckpardon my newbie question, but when i display even the hidden files i don;t see a rockbox.ipod. what folder would it be in
00:18:27linuxstbIt should be in .rockbox
00:18:35stripwaxchuck - if you don't see it it isn't there so that is fine
00:18:42 Nick japc is now known as _japc (
00:19:17 Quit davina (Remote closed the connection)
00:19:19stripwaxchuck - can you play back any music at all (something not in .mp3 perhaps)
00:19:57chuckall i have is mp3, apprx 700, some music, some podcasts
00:21:32chuckall i did was build my database from the things that i had loaded on my ipod theu itunes
00:23:48 Part linuxstb
00:24:39stripwaxchuck - are you using rbutilqt to install/upgrade rockbox?
00:25:36stripwaxok, and when you click on Info, does that report the correct version?
00:25:52chucki just did a rest on my settings after installing the build for the 30 gig and my first mp3 has playeed for a minute w/o locking up
00:26:13 Quit ZEA ("leaving")
00:26:36 Quit PaulPosition (Read error: 104 (Connection reset by peer))
00:26:59chuckmy wife is calling me for dinner, can we talk later?
00:27:07stripwaxchuck - ok. resetting the settings may have been what you needed, so if you're up and running now you should put on the 60/80GB build
00:27:21 Quit MethoS- (Remote closed the connection)
00:27:23chucki'll try that
00:27:37chucki sure appreciate your patience and your help
00:28:49 Quit chuck ("Leaving")
00:31:51stripwaxI wonder if the ability to reset settings should be added to rbutilqt, whenever the config version changes. Then again I thought rockbox handled that automatically
00:32:30preglowi'm starting to think this clicking is caused by the speex output not dying out fast enough, we should probably leave a full frame of silence after the end of each clip
00:32:32 Join PaulPosition [0] (
00:32:49 Quit roolku ()
00:33:12Lloreanstripwax: There is no "config version" issue any more...
00:33:31stripwaxLlorean - but what happens when something is no longer compatible?
00:33:56LloreanWhat do you mean?
00:34:02LloreanIf a setting isn't recognized, it's just ignored.
00:34:51LloreanSettings version mattered with the old binary config method, but with the .cfg file it's not the same concern.
00:36:34 Quit Thundercloud (Remote closed the connection)
00:36:36stripwaxTrue - so, I'm presuming, probably incorrectly, that chuck's configs were working at some point but with a rockbox upgrade they cause rockbox to crash. Under what circumstances would that happen? The alternative, I suppose, is that chuck has changed some setting himself which causes rockbox to crash.
00:37:02LloreanUnder no circumstances can that happen unless there's a significant bug in config parsing.
00:37:47stripwaxWe should have asked chuck to mail us his old settings to figure out why rockbox could not play audio without crashing
00:37:56PaulPositionDoesn't the apps/ file help pinpointing data aborts ?
00:38:01LloreanI doubt it actually had anything to do with the settings.
00:38:05LloreanPaulPosition: It can.
00:38:39Lloreanstripwax: He did the reset after installing the 30gb version. I don't recall him saying he tried playing music with it on the 30gb version without resetting
00:38:51 Quit Nico_P (Remote closed the connection)
00:38:52LloreanAnd using the 60/80 version when you only have 32mb of RAM would create the exact symptom he described.
00:38:59pixelmamy idea was that he might have a 60GB with only 32MB RAM and people reported that too, not sure what the reset did there.
00:39:17stripwaxchuck: ok, i downloaded build 1709 for the 30 gig, and when i tried to play a mp3, i go the following message "data abort at 000089F0 (0)
00:39:33LloreanWhat is build 1709?
00:39:48pixelmaI guess it misses a "5"
00:40:01 Join AceNik [0] (n=AceNik@
00:40:02stripwaxI guess so
00:40:12LloreanBut also, it's possible he tried to resume playback rather than actually play a new file
00:40:21pixelmahmm, but we aren't at 15709 yet
00:40:30AceNikguys pleasee include thi patch in svn #7538
00:40:44PaulPositionacenik - thi ?
00:40:53LloreanAceNik: It's not implemented in a way that is acceptable
00:42:09AceNikllorean: ok, bt an you claer one more doubt for me, why is the h10 able to use 24-bit bmps only
00:42:32LloreanAceNik: Ask the patch author...
00:42:47stripwaxpixelma - hm..
00:43:13AceNikLlorean: its not concerning the patch , its a general statement, why is the h10 restricted to 24-bit bmp's
00:43:53Lloreanstripwax: There really should be no way that settings can cause such a crash. If they do, odds are high there's a bug in whatever setting is enabled, rather than the settings themselves being problematic
00:44:01LloreanAceNik: It's not.
00:44:21 Quit stripwax ("Miranda IM! Smaller, Faster, Easier.")
00:44:53AceNikLlorean; whether it be the wps's files, or the backdrops, all only support 24-bit bmp
00:44:58pixelmawhat do you mean it's restricted? The display only uses bit-depth of 16-bit anyways (and Rockbox can scale down to it from higher depths)
00:45:12LloreanAceNik: You can also use 8-bit and 1-bit.
00:45:28LloreanThough not for the backdrop, probably.
00:45:52AceNikpixelma: thats what i didnt know, you gave the answer, tell me is a higher bit better or lower?
00:46:19pixelmaLlorean: I would assume it could also handle that but haven't tried
00:47:12AceNikwil 8-bit bmp's be better in terms of looks than 24-bit
00:47:38LloreanNo.. 8-bit means less colors
00:47:46LloreanThese are things you can google, and really have nothing to do with Rockbox
00:47:50 Quit moos ("Rockbox rules the DAP world")
00:48:05SoapI hate to be the bearer of bad news - but it appears the "conventional wisdom" that the Sansa acheives better battery runtime than the other PP targets is mistaken.
00:48:15 Quit ForgottenMemorie ()
00:48:26preglowjmspeex: hmm, for how many frames can the tail of the lpc filter go on with no input? i guess that's completely up to the coefs...
00:48:37LloreanSoap: Doesn't testing show it generally has a better *ratio* relative to the OF?
00:48:47jmworxpreglow: one frame
00:48:58SoapMind you, this is only the results from the first of my set of tests, but when comparing the Original Firmware to Rockbox I'm only seeing 70% of OF runtime.
00:49:01preglowjmworx: that matches my findings nicely
00:49:24PaulPositionSoap - Which is better than what the h10s and most iPods are getting, relative to OF, no?
00:49:37SoapLlorean, I haven't seen any controlled comparison before. There are some battery benches in the wiki, but no corresponding numbers from the original firmware.
00:49:54preglowjmworx: some of our voice clips end very suddenly, and it sounds like the filter tail is still not done if we cut immediately at the end of the frame
00:50:12SoapPaulPosition, Nobody has done the Nano recently - but the Video is @ 66% post HDD poweroff - and obviously that beast is spinning a disk.
00:50:15preglowso i just do one frame with bits = NULL to help the tail die
00:50:18 Quit ompaul (Client Quit)
00:50:25 Quit spiorf (Remote closed the connection)
00:50:38SoapSo I do not find it hard to believe at all that the Nano is at ~70%.
00:51:16amiconnPaulPosition, Soap: all PP5022 and PP5024 targets get about the same ratio
00:51:22AceNikh10 with the equalizer enabled is fantastic, but on the 20 GB i manage to get only 4 hrs, with eq off it gives atleast 5.30hrs, unlike OF that gives mre than 12 hrs with crappy quality
00:51:22amiconnPP5020 targets are worse
00:51:26jmworxpreglow: Also, because it needs integer number of frames and because of the lookahead, you might have lost some samples in the last frame
00:51:31preglowjhMikeS: seems i need to add that last green delta back, then :/
00:51:38AceNikamiconn: precisely
00:51:41LloreanSoap: I've always thought it was just better than the disk based targets (at least until recent optimizations in that area)
00:51:59preglowjmworx: so i really should do some padding on the encoder side as well, then
00:52:09Soapamiconn, I understand they should - the talk-around-the-water-cooler was that the Sansa was an exception for some reason - speculation abounded that the OF was inefficient or something else.
00:52:13PaulPositionSoap - Interresting. I guess I should try some OF bench on my h10.. Same files, no eq, and a lot of patience to "give a hear" every 15 minutes.. :p
00:52:35SoapPaulJam, yea - that is the resolution of my tests.
00:53:11amiconnPaulJam: Just start a (low quality) recording of the H10's output with your PC. Then you don't need to check every 15 minutes
00:53:24jmworxpreglow: I'd recommend padding one extra frame on the encoder side.
00:53:35preglowjmworx: btw, what was that i needed to do extra for lookahaed to be correct when encoding nDb for use in a wb decoder?
00:53:39Soapamiconn, I do that when I'm at home - but I've been on the road pretty constantly for months now.
00:53:43AceNikguys how about the field strength patch for radios, that seems to be pretty good an addition ?
00:53:45preglowjmworx: s/nDb/nb/
00:53:52jmworxyou can even "encode" it at 250 bps if you care that much about space
00:54:10amiconnLlorean, AceNik: The rockbox bmp loader supports all common bit depths, i.e. 1, 4, 8, 15/16, 24 and 32 bit
00:54:11preglownah, i'll just append a frame and let speex do what it wants
00:54:15Soapand 15 minute resolution is ~ 1.5% resolution @ 16 hours.
00:54:36amiconnExotic bit packings aren't supported for 15/16, 24 and 32 bit though, only the standard ones
00:54:37*PaulPosition 's no PaulJam. Don't insult that honest dev. guy by mixing him up me :o
00:54:51jmworxpreglow: If decoding nb with a wb decoder, you need to double the lookahead value (because the sampling rate's twice as high)
00:54:59 Quit iamben (Read error: 110 (Connection timed out))
00:55:07jhMikeSpreglow: what broke?
00:55:38AceNikamiconn: so how can i makea 32bit bmp, wont it hav better quality, but then why is there a thing where it does not accept bmp's saved from windows i presume they are 32-bit by default
00:55:49preglowjhMikeS: i found part of the reason why some short clips are cut too abruptly
00:55:59preglowjhMikeS: giving a nasty clicky sound
00:56:07amiconnI.e. 15 bit must be XRGB1555, 16 bit must be RGB565, and 32 bit must be ARGB8888 (with the A ignored) or XRGB8888
00:56:11Shaid32bit wont look any different, the extra 8 bits are just an alpha channel.
00:56:25jhMikeSpreglow: should I read back? or is it simple to just say?
00:56:52preglowjhMikeS: when a clip ends very abruptly, the lpc filter in speex doesn't always have time enough to reach zero level
00:57:03preglowjhMikeS: then it'll get cut and you'll have a click
00:57:07amiconnLlorean: Sansa *was* better than the disk based PP502x targets, but no longer is :)
00:57:13jhMikeSsimilar to flush the filter banks out?
00:57:28pixelmaAceNik: re. accepting the setting - have you tried or do you base this on information in the wiki or the manual (because I know there are place which state it wrong)
00:57:35preglowjhMikeS: i tried to fix it a bit by decoding a zero frame after the last proper frame, using the code i removed during the last green delta commit...
00:57:42LloreanAceNik: The screen is only 16-bit, any bitmap more than 16 bit can't look better anyway
00:57:44SoapAnd why is battery-bench recording log entries every five seconds now? It did not use to do that.
00:57:47 Quit Workaphobia (Read error: 110 (Connection timed out))
00:57:51preglowjhMikeS: and that helps for most of the click, but jmspeex says i should pad one frame in the encoder too, and i'll try that
00:58:10jhMikeSI wouldn't think decoding an extra frame would help but not trimming the source so soon.
00:58:55jhMikeSwhy don't I hear clicks in any voice I've got?
00:59:00AceNikamiconn; yes i have tried loading a normal bmp, when i had started using rockbox, but it did not work, then i read up n saved to 24-bit
00:59:06amiconnLlorean: In fact it can, as both 24/32 bit formats and <= 8 bit formats use dithering, while 16 bit does not. But if the 16 bit bmp is pre-dithered, it will look as good as a 24 bit one, and you save a bit of loading time
00:59:16AceNikLlorean: so basically the h10 lcd itself sucks ?
00:59:22jhMikeSwere they encoded with more silence at the end?
00:59:26PaulPositionAcenik - You mean for that SplashScreen patch?
00:59:31Lloreanamiconn: Okay, so how 'bout "can't look better than a properly created 16-bit image"?
00:59:59pixelmaSoap: it records everytime the measured voltage changes, on sansas this seems to happen very often (maybe the readout is not "stable"?)
01:00:08PaulPositionAcenik - Well, 16 bit is over 65k colour. Sucks is a loose term.
01:00:13LloreanAceNik: 16-bit color is 65,536 colors. Is that not enough for you?
01:00:16preglowjhMikeS: no
01:00:25preglowjhMikeS: the clips in question are cut really abruptly
01:00:36preglowlike numerals with espeak voice
01:00:44preglowespecially "two" and "three"
01:00:46AceNikguys is there a way my recordin in rockbox can be improved i do not follow the settigs, everytime i record, it frstly reduces the volume, n the recorded file is pretty low quality n mp3
01:00:59jhMikeSwait...the letter "a" was one with a problem. I think I mentioned it.
01:01:10jhMikeSJust a short "eh"
01:01:28AceNikpaulposition,Llorean: well yeah, but how come its not as bright & better like the ipods obviusly thier lcds are better
01:01:32preglowjmworx: an extra frame appended worked perfectly
01:01:43amiconnpreglow: We need to trim, otherwise the pauses between clips will become unbearable.
01:01:44PaulPositionAcenik - Line in or internal mic? Anyway, recording's settings are (duh!) in settins->recording.
01:01:53preglowjmworx: i assume speex uses 250 bits for that anyway, in vbr mode
01:01:54LloreanAceNik: LCD brightness and sharpness has nothing to do with the bit depth...
01:02:01preglowamiconn: trimming isn't the problem, i got that wrong
01:02:10PaulPosition(And I personnally find my h10 LCD *too* bright. So there... :p )
01:02:26preglowamiconn: unfortunately, this seems to be another bug that needs to update rbspeexenc :/
01:02:34 Join keanu [0] (n=keanu@unaffiliated/keanu)
01:02:49jhMikeSpreglow: does this mean the decoding need to reset on every clip when strung together or do any trimming?
01:02:52keanuWhat could cause usb_find_busses() to return something less than 0?
01:02:52amiconnAnd btw, the talk code plays a silence clip at the end for a reason - if it wouldn't, the end of the final clip would be swallowed on hwcodec
01:03:06AceNikWhat kind of seetings do i need to have recording on mic on h10 like the OF
01:03:11preglowjhMikeS: decoder isn't reset on every clip now?
01:03:13AceNikPaul Position: ok give up, just anted to get to know, because i find, ipod 5G's wps's way better than h10's
01:03:16jmworxpreglow: I don't think Speex can currently reduce the bit-rate *that* quickly
01:03:17amiconn(the mas stops without playing the last few frames when it runs out of data)
01:03:26preglowjmworx: makes sense
01:03:30amiconnI wonder why this silence clip doesn't help for speex
01:03:40preglowjmworx: how would i go about forcing that last frame at 250 bps, then?
01:03:45jhMikeSpreglow: at the start of a clip yes but only once for a string of clips (such as saying "123456")
01:04:00 Quit scorche|w ("CGI:IRC (Ping timeout)")
01:04:06preglowjhMikeS: probably should reset there as well
01:04:14amiconnjhMikeS: Hmm, then the silence clip should do the trick, shouldn't it?
01:04:16LloreanAceNik: Maybe the WPS designers working on H10 ones just aren't as clever then. The only real difference between *what* can be displayed is the size of the screen.
01:04:33preglowthe silence clip should do the trick
01:04:37preglowi didn't know of any such
01:04:38amiconnpreglow: Why would the decoder need to be reset for every clip??
01:05:00preglowamiconn: so that decoder state for one clip doesn't affect another? sounds quite logical to me
01:05:07jmworxpreglow: disable VBR and set mode to 0.
01:05:09preglownot every codec works like mp3
01:05:14preglowjmworx: of course...
01:05:39preglowand even mp3 should have its state reset between streams that don't belong together
01:05:44Soapthank you pixelma. I was assuming it was new behavior. My recent iPod benches also contained more entries than I was used to.
01:06:16SoapBut looking at the voltage levels - I see they are jumping back and forth quite a bit.
01:06:27jhMikeSpreglow: but mp3 prepends and appends silence instead as part of the format
01:06:31AceNikLlorean: no h10 devs are cool enough man, all devs are cool, but still those wps's rok
01:06:39amiconnpreglow: Check talk.c, mp3_callback(), especially lines 326ff
01:06:54jhMikeSthe decoder gets flushed to 0 by the audio itself
01:07:14PaulJamPaulPosition: I'm not a dev. just a fanboy :)
01:07:32preglowso we're currently getting a zero clip? that makes little sense, then, i shouldn't be having these clicks
01:07:56preglowhow is this zero clip encoded?
01:08:01AceNikbye guys im off
01:08:02 Part AceNik
01:08:21amiconnLike normal, just that wavtrim isn't applied (of course)
01:08:22jhMikeSwhy would stringing an encoded 0 clip onto encoded audio work anyway?
01:08:23PaulPositionAceNik - Recording with the mic, I suggest Format:mp3, bitrate:at least 128, channel:mono... Then in the recording screen try with both 0db or 20db gain and see if there's clipping. You're better off with no clipping at all, at worst you'll just have to amplify your file in some post-processing program. Else, if it clips, you'll need filters and such.
01:08:38PaulPosition(I'm assuming voice, of course.)
01:08:43amiconnjhMikeS: Read back 6 minutes
01:08:50preglowjhMikeS: to allow the decoder enough time to let the lpc filter tail die before clipping audio
01:08:58PaulPositionAcenik - Cause to record music, any voc-recorder sucks big time.
01:09:01preglowjhMikeS: if the filter tail isn't dead, you'll hear it when it's cut
01:09:29PaulPositionPaulJam - Well, I'm a fanboy too, alrightee then ;)
01:09:48jhMikeStrue it has delays, but silence appended while encoding would have the proper gapless die out...well in mp3 anyway.
01:09:59preglowjhMikeS: just tried it, works great
01:10:12***Saving seen data "./dancer.seen"
01:10:18pixelmaSoap: maybe that's something else then, just repeating here what I've been told ;)
01:10:20preglowbut i'm thinking this encoded silence might as well be applied in the voice thread instead of in the encoder
01:10:22preglowto not waste space
01:10:39jhMikeSthough mpa would still have dieout time in that case too...yeah
01:10:39amiconnThe silence clip *is* applied in the talk code
01:11:04*jhMikeS doesn't want to waste space and if it work, well, it works.
01:11:18Soappixelma, looking at the numbers what you say is completely reasonable.
01:11:43 Quit ender` (" Join the army, meet interesting people, kill them.")
01:11:51preglowamiconn: i like the fact that espeak encoded " " isn't perfect silence
01:11:54preglowit's pitched...
01:12:17amiconnThe silence clip is pure digital silence, not generated by the tts engine
01:12:46amiconnThe one shipped is 300ms
01:12:51preglowp_silence = get_clip(VOICE_PAUSE, &silence_len);
01:12:57preglowVOICE_PAUSE is " " in english.lang
01:13:00jhMikeSpreglow: why should the voice thread append it? it's done in talk.c already.
01:13:08amiconnpreglow: That's not used - check
01:13:09preglowjhMikeS: then we need to find out why that doesn't work
01:13:14preglowamiconn: ok
01:13:30 Part pixelma
01:13:50preglowjhMikeS: fact of the matter is, if i append one frame, that is 20ms of audio with silence in the encoder, it works ok. no clicks, and if we add 300ms, something else must be wrong
01:13:50jhMikeSthe silence clip isn't available for some reason? is it trying to use it?
01:14:05amiconnif ($id eq "VOICE_PAUSE") { \n print("Use distributed $wav\n") if $verbose; \n copy(dirname($0)."/VOICE_PAUSE.wav", $wav);
01:14:37 Join jurrie [0] (
01:15:02amiconnIf the clip wouldn't be available, voice file creation would fail
01:15:36jhMikeSbut is it being returned and used is what I'd like to know and how big is it?
01:16:28amiconnpreglow: The silence clip is *only* appended if the queue runs out, not in case of shutup(). But that's acceptable imo. shutup() should be fast
01:17:17jhMikeSthe voice thread won't cut itself off when decoding is finish. it just exhausts the data and lets pcm play till that runs out.
01:18:29preglowi can only see two clips passed to voice for each file number say
01:18:33preglowthat would be "file" + number
01:19:37 Nick midkay_ is now known as midkay (
01:20:22preglowinserting a printf at the silence queue inserts does claim it gets inserted after both "file" and number
01:21:50amiconnThere's something that's now quite fishy for swcodec voice (not related to the current problem though)
01:22:28amiconntalk_force_shutup() searches for the next mp3 frame header to cut the clip - obviously no longer correct for swcodec
01:23:06 Join Isolinear [0] (
01:23:07amiconn(all the curr_hd related stuff)
01:23:37preglowjhMikeS: decoder resets inbetween "file" and number
01:23:47amiconnSince swcodec can stop mid-frame and restart without delay, that's most probably not a problem (but wasted code)
01:23:52preglowjhMikeS: didn't you say it shouldn't when clips are concated together
01:24:15 Quit mirak (Remote closed the connection)
01:24:36amiconnThe deal with this stuff is that if the mas has to re-sync, it would swallow the beginning of the next clip - and that's not just one or 2 frames
01:24:56jhMikeSpreglow: if it queues multiple clips, no it shouldn't
01:25:23preglowjhMikeS: well, i don't know how that exact mechanism works
01:25:42preglowin this case it does reset between the clip for "file" and the clip for a number
01:26:49 Join iamben [0] (
01:27:04jhMikeSeach queue slot = one clip. when that slot exausts the data, the next slot should be read from and decoding should continue without interruption.
01:27:11preglowand i can't for the life of me see that the voice thread receives a queued silence clip
01:28:26 Join Thundercloud [0] (
01:28:32preglowevery clip passes through the Q_VOICE_PLAY case?
01:29:21jhMikeSa series of queued clips should just be obtained via the callback
01:29:30preglowget_more, yeah
01:29:31preglowi see now
01:29:36 Quit Zagor ("Client exiting")
01:29:42jhMikeSbut of course the thread must be sent these things in the proper manner
01:30:59jhMikeSit's the same operation as the prevous but doesn't need the Q_VOICE_STOP before the Q_VOICE_PLAY (which identical to just aborting and resarting)
01:31:42preglowit seems "file" gets sent with a chained silence, then number gets sent with a chained silence
01:33:23preglowbut why the hell does adding one frame in the encoder help at all if silence gets added?
01:33:30 Join Mouser_X [0] (
01:33:55jhMikeSdoes the decoder return an error?
01:34:18amiconnEh, this code is already in an #if block...
01:34:24 Join advcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
01:34:35preglowmy debug printfs are starting to get increasingly esoteric
01:34:44preglowjhMikeS: no
01:34:54 Quit advcomp2019 (Nick collision from services.)
01:34:57 Nick advcomp2019_ is now known as advcomp2019 (n=advcomp2@unaffiliated/advcomp2019)
01:35:34preglowjhMikeS: i guess it is possible that the encoder is retaining some data it needs to spit out
01:35:41preglowthat would explain why one frame extra helps
01:36:02preglowjhMikeS: it does make sense, after all, i has some lookahaed
01:36:38amiconnEncoding the very same .wav, both with and without padding, and then comparing the produced speex bitstream might tell you...
01:36:47preglowamiconn: and indeed it shall!
01:39:26*jhMikeS is just going to get rid of the strange unboost cpu on block_w_tmo now but retain the per-thread boost control. much more sane.
01:39:49jhMikeS'course this decision has independent of any voice things
01:40:07preglowamiconn: yes, it does indeed seem to be the case, the file ends in the middle of an impulse if i don't encode an extra frame
01:41:25jhMikeSso the silence clip plays but the clip is abruptly cut off?
01:41:48jhMikeSthe non-silence clip that is
01:42:05preglowproblem is that the silence clip should still be helping more than it is
01:42:16preglowright now it doesn't seem to be helping at all
01:42:37preglowif i do a null decode on a frame after the last frame, it helps, and that is essentially what the silence clip should be doing
01:43:21amiconnIt might be that the silence clip isn't needed at all on swcodec
01:43:25jhMikeSin speex, what's a NULL-decode?
01:43:56preglowjhMikeS: calling speex_decode_int() with NULL as bits poitner, that'll make speex assume we lost a packet
01:44:12jmworxjhMikeS: there's two things here you may be confusing
01:44:12amiconn(it might improve things slightly while music is playing, bevause it lengthens the period of reduced music volume)
01:44:21jhMikeSjmwork: quite likely
01:44:34 Quit Jon-Kha (Remote closed the connection)
01:44:35jmworxThere's the packet loss concealment you get when you pass a null pointer for the bits −− you most likely don't want that
01:44:51preglowgood, then i can keep that code commented out :>
01:45:09amiconnThe packet loss thingy might be useful for talk_force_shutup()
01:45:09jmworxAlso, there's a "null mode" which is mode zero, aka 250 bps, aka null excitation that just let's the LPC filter ring.
01:45:32preglowahhh, i thought they were the same
01:45:46jmworxamiconn: why would you use PLC?
01:45:57preglowhe probably means null mode
01:46:14amiconnIn order to avoid a click when a clip is cut
01:46:31jmworxpreglow: no, the PLC will actually make up a "plausible" excitation using the pitch and other stuff. null mode just drives everything to zero in a clean way
01:46:42preglowjmworx: that's good to know
01:47:13amiconntalk_force_shutup is for when the previous queue contents is flushed immediately and replaced by something new
01:47:14jmworxamiconn: you'd need to decode a "crafted packet" that basically contains a null byte
01:47:45jmworx(the null mode is mode zero, so its content is five zeros in a row)
01:47:53jmworx(I'm talking at the bit level)
01:48:14preglowjmworx: but yeah, if we want the output to die after a clip, the right thing to do would be appending an extra empty frame when encoding, right?
01:48:28jhMikeSyou literally send one byte of "0"?
01:48:35preglowjhMikeS: no
01:48:39jmworxpreglow: You could alternatively do that at decoding time.
01:48:58jmworxchar ch = 0;
01:49:23jhMikeSpreglow: that's looks like one byte of "0" to me :>
01:49:27jmworxspeex_bits_set_bit_buffer(&bits, &ch, 1);
01:49:31 Join Workaphobia [0] (
01:49:41WorkaphobiaSo I was on here a few hours ago looking for small features I could potentially implement. I think this one looks good: (search tool for text viewer) - it hasn't already been taken care of, right?
01:49:44preglowjhMikeS: it s...
01:49:47jmworxpreglow: exactly, one byte of zero contains 8 zeros :-)
01:49:50preglowjhMikeS: i forgot that the mode bits are placed first
01:50:02preglowif you see it in a speex file, it'll be 250 bits, no?
01:50:23jmworxpreglow: 5 bits per frame * 50 frames per second = 250 bps
01:50:38jmworx(not 250 bits, that's an awful waste for encoding zeros)
01:51:04preglowso it's nb/wb bit + 4 bit mode per frame?
01:51:08jhMikeSthat's easy to add in there
01:51:23jmworxpreglow: exactly
01:51:37preglowjmworx: but the encoder would retain nothing of value to flush even if the last frame ended really abruptly?
01:51:49jmworxand there's no need to specify the 3 wb mode bits even if it's a wideband stream
01:52:01jhMikeSwhich bit is the nb/wb bit? the leftmost?
01:52:05preglowthe first
01:52:15jmworxpreglow: what you you mean?
01:52:35jhMikeSis anything compelled to byte alignment?
01:53:01jmworxjhMikeS: It's not exactly a nb/wb bit in that even for wb, you'll see a nb bit first. Then after the wideband, you'll see a "high band" 1.
01:53:02preglowjmworx: i'm just thinking that the lookahead might mean we should allow the encoder to go on for one frame more than we actually have
01:53:06preglowand pad that with zeroes
01:53:35jmworxtrue if you care about all your samples
01:53:47preglowand we do, like i say, the samples end really abruptly
01:53:50preglowand we want those last samples
01:54:12amiconnspeex does not have variable last frame? (just curious)
01:54:13pregloweh, those _clips_ end really abruptly
01:54:23preglowamiconn: no variable frames if you mean size in samples
01:54:30amiconnI do mean that
01:54:38preglowframes are always 160, 320 or 640 samples wide
01:54:47preglowamiconn: gapless is handled by ogg
01:54:50amiconnSo gapless speex would be problematic? (not that it really matters for speech)
01:54:57preglowamiconn: just like for vorbis
01:55:06 Join bsaint [0] (
01:55:17amiconnSo the container just specifies a point to cut?
01:55:32jmworxamiconn: gapless works fine. You just encode (indirectly through Ogg) the number of samples to drop at the beginning and at the end.
01:55:42amiconnThat's not really different to the lame tag for mp3 iiuc
01:55:42preglowjmworx: but then it wouldn't make sense to force the last frame at a low bitrate either, no? since the encoder still has valid data to spit out and all
01:55:53jmworxpreglow: indeed
01:56:45bsaintdoes anyone know where i can get more information on the flash translation layer used in the 1st gen ipod nano?
01:57:09preglowjmworx: second thing, is it possible to pre-compensate for the differing lookahead of nb decoded in wb mode at the encoding side? the decoder won't really know if it has a nb file or wb file, and always decodes in wb mode
01:57:14preglowbsaint: there's a datasheet on it
01:57:20preglowbsaint: it's a separate chip
01:57:22amiconnWhat happens if I use non-standard rates for speex? Since the number of samples per frame is fixed, I'd expect the frame duration to change...
01:57:39preglowamiconn: frame duration in samples remains the same
01:57:50amiconnI mean duration as in time
01:57:57preglowthat changes, of course
01:58:07preglowit pretty much has to if frame size is constant
01:58:37preglowspeex does other rates just fine, it just sounds better at 8/16/32 khz
02:00:24 Join tdruskw [0] (n=tyler@
02:00:41tdruskwHow can I use amarok with rockbox. Sorry if this is off topic.
02:00:54scorcheit is
02:00:59jmworxpreglow: you can say that your narrowband stream is in fact a wideband stream
02:01:46DerPapsttdruskw: maybe someone in #rockbox-community can/is willing to help
02:01:51jmworxpreglow: but why do you insist on having narrowband streams decoded in wideband in the first place?
02:01:57tdruskwDerPapst: thanks
02:02:11preglowjmworx: just as an option for people who are really fixated on file size
02:03:10bsaintpreglow: do you know where i can find nano related datasheets? the apple site has trivial technical data, and the best information i've found is at
02:03:44 Part tdruskw ("Kopete 0.12.3 :")
02:05:52preglowbsaint: i really can't remember where i found that info
02:06:43preglowjmworx: plus, i just thought it'd be cool to take full advantage of the possibilities we have, and that includes nb streams being perfectly valid wb streams :)
02:07:41 Quit DerPapst ("So Long And Thanks For All The Fish!")
02:09:46preglowamiconn: is this guy doing something obviously stupid?
02:10:02bsaintok, well at least now i know that info exists, i should have more luck finding it! thanks
02:10:28preglowbsaint: start by finding the ipod nano chip list
02:10:38preglowsome place in there should ba an ata to flash bridge chip
02:10:43preglowbsaint: why do you want to know, btw?
02:10:58 Nick fxb is now known as fxb__ (
02:12:42 Join advcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
02:13:25bsainti was going to use the ipod nano as the basis for developing a new flash filesystem due to its large capacity, but after i found out that it also uses a ftl i was just looking for more information on it
02:13:33 Quit advcomp2019 (Nick collision from services.)
02:13:35 Nick advcomp2019_ is now known as advcomp2019 (n=advcomp2@unaffiliated/advcomp2019)
02:14:05rasherooh, word from Brian Wolven. Let's get the stuff in svn already
02:14:05 Quit Arathis2 ("Bye, bye")
02:14:39preglowyeah, i should ask him about that
02:16:47rasherand get some sort of license slapped on it as well
02:16:51 Quit z35 (Read error: 113 (No route to host))
02:17:25 Join z35 [0] (
02:17:41 Quit _japc (Remote closed the connection)
02:26:08Mouser_XHow difficult would it be to interface with Rockbox's database? Specifically, if I had "gather runtime data" enabled, and I gave a song a rating of "1," how difficult would it be for a program to gather that data, and match it with the same song on my computer (assuming that the program in question had a database of the songs on my computer)?
02:26:47LloreanMouser_X: A program on a computer can access and update Rockbox's database as easily it can its own, I'd imagine..
02:27:04Mouser_X(I'm wondering, because it'd be handy to mark a file "1" if there's errors/glitches/problems/etc. and match it to the original on my PC, so that I can fix the original.)
02:27:43Mouser_XI'd need to know the format of the database though, correct?
02:27:54Mouser_XAnd, where would I find that information?
02:28:05 Quit Thundercloud (Remote closed the connection)
02:28:09LloreanThe code in Rockbox that handles the database would be the best start
02:28:14LloreanLarge chunks of it can probably be reused.
02:28:24preglowrasher: just got brian's permission to release the scripts under gpl
02:28:35*Llorean cheers
02:28:38rasherpreglow: hurray
02:28:50preglowi assume we'll want to put them in svn too?
02:30:23 Quit scorche (Nick collision from services.)
02:30:55 Join scorche` [0] (n=scorche@rockbox/administrator/scorche)
02:31:09preglowjmworx: looks to me like both nb and wb data are delayed by the same amount when decoded with the same wb decoder, so we should be fine just skipping the amount of lookahead samples that the wb decoder tells us to
02:31:51 Quit Workaphobia (Read error: 110 (Connection timed out))
02:34:21 Nick scorche` is now known as scorche (n=scorche@rockbox/administrator/scorche)
02:36:00 Quit amigan_ (Read error: 104 (Connection reset by peer))
02:36:12rasherpreglow: I'd say put it all in tools/
02:37:20preglownow, let's see if i can figure out this wiki file upload stuff
02:42:20 Join weezerle [0] (
02:42:21*Llorean now has oblib's name.
02:42:33LloreanAnyone think of any reason *not* to commit the Nano fix?
02:43:18JdGordondoes it work?
02:43:26preglowi can't delete files in the wiki?
02:43:47JdGordonLlorean: *couph* FS #8169
02:43:50LloreanJdGordon: It's worked for all testers.
02:44:00 Quit weezerle (Client Quit)
02:44:10preglowgo! go!
02:44:50JdGordonumm... the best reason i can tihnk of it for not commiting is less ussers to support :p
02:45:43 Quit atsea- (Remote closed the connection)
02:46:22LloreanAlright, nano issues should be resolved
02:48:36JdGordondoes it need a bootloader update also?
02:50:27 Join psycho_maniac [0] (i=psycho_m@
02:50:32preglowbootloader runs at 24 mhz, afaik
02:50:51LloreanJdGordon: On other targets does the "Menu" button in the recording screen take you to the Main menu, or the Recording menu?
02:51:21LloreanMy mind says that it should take you to the recording menu, but I don't know what other targets do there.
02:52:08JdGordonI dont know, it should be consistant with the rest, but if power goes to main menu, the down should be rec settings menu.... select and long select? dunno
02:52:09Llorean"menu" is recording menu, "Stop" is main menu
02:52:16 Quit sup (Remote closed the connection)
02:52:20 Join sup [0] (
02:52:33*JdGordon has a quit bite to eat and heads of for the us consulate... ttyl
02:52:55LloreanI don't think there's a use for "Select" in the recording screen
02:53:01LloreanI don't see a mapping that seems like it'd belong on it.
03:03:27LloreanJdGordon: Okay, all better
03:09:38*Llorean made a mistake
03:10:17***Saving seen data "./dancer.seen"
03:13:06LloreanJdGordon: Can you have two buttons do identical things within a context?
03:13:53 Quit Rob222241 (Read error: 110 (Connection timed out))
03:14:31 Quit w0rd54 (Client Quit)
03:14:36 Quit bsaint ()
03:15:07 Quit J3TC- (".•«UPP»•.")
03:15:10 Quit PaulJam (".")
03:15:39 Join w0rd54 [0] (
03:19:47 Join J3TC- [0] (
03:34:02 Quit w0rd54 (Client Quit)
03:38:46 Join w0rd54 [0] (
03:40:48LloreanOh Noes guys!
03:40:53LloreanThere are vulnerabilities in FLAC!
03:41:24krazykitcan i run my own code on your player?
03:41:32LloreanI fear you may be able to!
03:41:51krazykityour player will be a spam bot in my network of spam-daps!
03:42:03LloreanActually, it looks like it's not FLAC that's the problem, but the metadata
03:42:07krazykit...after a network stack is written for rockbox.
03:42:08LloreanWhich, afaik, is vorbis comments.
03:44:55 Join atsea- [0] (i=atsea-@gateway/tor/x-09ebe3878f1a8b45)
03:54:59IsolinearSpeaking of stack, how goes the work on USB?
04:06:21 Join jpt9 [0] (
04:06:32jpt9is there a calendar plugin for the sansa e200?
04:08:20GuerinLlorean: my previously-short-lived 1100mAh x5 battery is getting ~6h with the most recent rockbox
04:08:47Guerini've now reinstalled the suspect x5l batteries (looks like 3x110mAh) to see how that goes
04:08:54Guerin1100, even
04:09:19Guerinthough I must say - ~6h is still about half what I'd expect
04:09:30 Quit jhulst ("Konversation terminated!")
04:09:32LloreanFrom a new battery, that's pretty bad.
04:09:36LloreanHow does it perform in the original firmware?
04:10:42GuerinLlorean: I haven't run a thorough test on the OF.
04:10:45 Quit dandin1 ()
04:11:01LloreanThat's really the only valid point of comparison
04:11:06Guerinreverting to the OF is my last resort; I'll only do it if I get utterly shit battery life on rockbox
04:11:07psycho_maniacjpt9: there is a calendar plugin on flyspray
04:11:08LloreanIf Rockbox is getting >= to the OF, then the battery is a problem.
04:11:36LloreanIt's not unheard of for off-brand batteries to have significantly less capacity than is claimed
04:11:41psycho_maniacand it is also on the wiki.
04:12:08GuerinLlorean: this is an identical Cameron Sino part.
04:12:54LloreanCameron Sino traditionally is expected to be good, so long as the battery was stored properly
04:13:06Guerinthat's the bit I can't guarantee
04:13:17Guerinebay, y'know
04:15:30Guerinam I right in presuming that the only electronic difference between a x5 and a x5l is the battery pack?
04:15:58Guerinbecause this player is actually an x5 with an x5l case and battery
04:16:00 Join Rob2222 [0] (
04:16:30LloreanAs far as I'm aware, yes that is true
04:16:39Gueringood, it seems to be so
04:17:00GuerinI was very pleasantly surprised at how straightforward the x5 is inside
04:26:15 Join minaret [0] (
04:27:02minarethi, i installed rockbox recently, and havent used my sansa much, but i noticed today that it is skipping about every 30 seconds while playing audio. is there anything i can do to fix this?
04:27:53LloreanDoes your WPS have a peak meter?
04:28:08minaretsorry, what is that?
04:28:12 Quit w0rd54 (Remote closed the connection)
04:28:30LloreanA bar that goes moves in time with the music beats
04:28:44minaretwhere would it be?
04:28:53LloreanOn the WPS. The While Playing Screen...
04:29:15minareti know, but where? sorry
04:29:24LloreanI don't understand what you mean.
04:29:32LloreanThe screen you see while music is playing
04:29:34minareti dont see a bar
04:29:41LloreanThen the answer is probably "no"
04:29:48minaretdo you mean the seeker?
04:29:54LloreanI don't mean the progress bar.
04:30:01minaretso, i dont see what youre talking about
04:30:04LloreanAre you using the equalizer?
04:30:13LloreanTry turning it off
04:30:18LloreanThe equalizer can be CPU intensive.
04:30:23minaretah, k
04:30:51minareti think that fixed it
04:31:05LloreanHow much CPU it uses depends on how many filters you use.
04:31:14LloreanSo, try using only 1 or 2 or 3 of the EQ bars instead of all five
04:31:20minaretI'll mess around with that then, thank you
04:32:19minaretalright, its working great now. thanks
04:34:20 Join w0rd54 [0] (
04:34:39 Join keanu|afk [0] (n=keanu@unaffiliated/keanu)
04:36:44 Quit keanu (Read error: 110 (Connection timed out))
04:39:49 Quit minaret ("Quit")
04:42:13 Quit miepchen^schlaf (Read error: 110 (Connection timed out))
04:42:32 Join miepchen^schlaf [0] (
04:51:31 Join Calcipher [0] (
04:55:33 Quit JETC- (".•«UPP»•.")
04:56:15 Join JETC- [0] (
05:06:11 Quit atsea- (Remote closed the connection)
05:10:19***Saving seen data "./dancer.seen"
05:17:20 Join Jon-Kha [0] (
05:18:03*lostlogic has sansa
05:18:08 Join atsea- [0] (i=atsea-@gateway/tor/x-f09b621560f3d439)
05:18:39Mouser_XYay for lostlogic!
05:18:45*Mouser_X does not yet! :(
05:19:45Lloreanlostlogic: And what will you be doing for Rockbox now that you have a new gadget?
05:20:00lostlogicprobably nothing new, but I have a new gadget...
05:20:16lostlogicLlorean: depends what I use that's broken on Sansa :-P
05:20:17LloreanAh, well that's good at least
05:21:22*lostlogic hasta figure out rockbox install on a new device
05:22:59LloreanIt's pretty easy for non-R sansas
05:23:00LloreanWell, very easy
05:23:28lostlogicmine says R
05:23:49LloreanWell then, it's a little less simple
05:24:03LloreanAt least it's easier than it used to be for -R types
05:24:11*lostlogic reads
05:26:14lostlogicwtf? the E200R target won't even compile atm?
05:26:20lostlogicor should I jsut be building E200?
05:26:27krazykitthe latter
05:26:57krazykitit's the same hardware, after all. the former target is for the bootloader and such, i think.
05:29:32psycho_maniaci have no reason to compile my own builds now that the acceleration scroll wheel patch has been committed.
05:29:47Lloreanlostlogic: You could fix the build target. :)
05:30:14jhMikeSAny reason there's an e200r build? I've been throwing the same e200 build on both my devices.
05:30:56jhMikeSjust the bootloader I'd guess
05:32:07LloreanAs far as I know, it's just the bootloader, yeah
05:36:38psycho_maniacso we dont need a e200r build?
05:37:39LloreanWe need it for the bootloader
05:38:10psycho_maniacoh alright. i understand.
05:39:47jhMikeShmmm...wonder if it could just run the same bootloader (auto-detect type)
05:42:43jhMikeSwithout yield_codecs, codecs beat buffering to the full point (yield every 1ms in the ata driver which is actually better anyway).
05:47:14LloreanjhMikeS: I think it's actually the encryption of the bootloader, or signing, or something
05:47:52 Quit qwm (Read error: 104 (Connection reset by peer))
05:48:04 Join qwm [0] (
05:50:27jhMikeSany reason the build must do it and not the installer?
05:50:41LloreanIf it's just that, no not really
05:51:28*jhMikeS should just look at the source to find out :p
05:52:31lostlogicI think I broke my e200r.
05:52:46Mouser_XYou're not the first to say that...
05:52:57lostlogic"Load main image failed" "Switch to Recover mode" <−− I'm newb what I'm do?
05:53:05PaulPositionHah. Read the wiki :p
05:53:09krazykityou read the wiki :P
05:53:32PaulPositionSomething like E200UnBrick or such name...
05:53:36lostlogicugh, I thought I was reading the wiki *reads harder*
05:54:16PaulPosition(or maybe not... lot of contradicting info in the last few days.. Not owing a Sansa, I wouldn't know for sure)
05:55:09jhMikeSwhat? my install went like butter.
05:55:27jhMikeSonly stupid thing was having the wrong usb mode on
05:55:46lostlogicjhMikeS: I dunno −− looked like it was working... but after it "upgraded" itself on the final reboot it did this.
05:57:10jhMikeSso you got the "firmware unlocked" message and rebooted then connected in MSC mode?
06:00:00lostlogicI think I missed a reboot, crap.
06:01:16lostlogicoh well, unbrick works, will try again from the top.
06:01:19 Join webguest73 [0] (i=80cde4d3@gateway/web/cgi-irc/
06:02:46webguest73hey, i'm trying to install rockbox on a 5th gen ipod video 30gb (i'm using MacOS 10.4) but the rockbox installer won't detect my iPod. What am i doing wrong?
06:04:18Lloreanwebguest73: Have you converted it to Fat32?
06:04:30webguest73The ipod?
06:04:51webguest73no, do i need to do that before running the automated installer?
06:05:15Mouser_XWell, since Rockbox won't run on anything but FAT32, it'd be a dang good idea to format it to that first.
06:05:21psycho_maniaconly if its a mac ipod.
06:05:28webguest73it is
06:05:34webguest73how do i format it then?
06:06:14psycho_maniacthere is a write up on the wiki i believe.
06:06:36webguest73okay thanks
06:06:46lostlogicjhMikeS: I'm not getting this firmware unlocked message, although the e200rpatcher application says patching application uploaded successfully, what am I doing wrong now?
06:07:59 Quit krazykit (Remote closed the connection)
06:08:21 Join krazykit [0] (
06:08:45jhMikeShmmm...did you ever?
06:09:05lostlogicno, the text is searching, found, uploaded, press enter to exit
06:09:19jhMikeSit's on the lcd
06:09:31lostlogiclcd is blank whole time :-\
06:10:03jhMikeShmmm...that's not right. mine wasn't.
06:11:01Lloreanlostlogic: Which steps are you following?
06:11:12lostlogic <−− Linux Installer
06:11:39LloreanJust making sure. :0
06:12:45lostlogic*sigh* I feel like such a newb.
06:13:02advcomp2019are you sure it was not a plain e200 with the e200r back
06:13:37lostlogicit boots up and says "Sansa Rhapsody" so I'm guessing not?
06:13:44lostlogichas that happened before?
06:14:01jhMikeSR firmware that's actually plain vanilla firmware?
06:14:04LloreanWe've seen several "not-really-e200R" players that looked like Rs.
06:14:53advcomp2019if it boots with "sansa rhapsody" that is a r series then
06:16:20*lostlogic tries booting a non-R factory mi4 just for shits
06:16:26lostlogicnope, definitely R
06:16:49Lloreanlostlogic: Does your box claim Audible book support?
06:17:29lostlogicno box, woot apparently doesn't do boxes.
06:17:43LloreanAny sign of a "v2" on it somewhere?
06:17:53advcomp2019is there v2 by the back of the sansa
06:17:59lostlogicnope, just a big REFURB
06:18:12*jhMikeS just has a nasty ghetto "REFURB" on the back with a lock/unlock red sticker.
06:19:01lostlogicjust to verify, to get to manufacturers mode (to do the unlock) I have hold on, hold the large round center area "select button" and connect USB, after 10 seconds, I release select and hit enter to continue on the patcher app?
06:19:07 Join Noobasaurus [0] (i=47da7a10@gateway/web/cgi-irc/
06:19:16LloreanPretty much
06:19:42scorchelostlogic: jhMikeS: a fingernail will remove the REFURB
06:19:56lostlogicjhMikeS: do you recall what displayed in manuf. mode before unlocked msg?
06:20:00PaulPositionWould it make sense to go through the plugins and "lcd_set_backdrop(NULL)"-ize those that don't already do? 'cause with backdrops not being confined to WPS anymore, many plugins become almost useless..
06:20:11scorchelostlogic: nothing is displayed in manufactuing mode
06:20:15jhMikeSlostlogic: not particularly, but just a few lines
06:20:17PaulPosition(okay, that's not new.. It is to me though.)
06:20:24LloreanPaulPosition: Backdrops have been not confined to WPSes for over a year...
06:20:32Calcipherhey, is there any possibility of a voice accessible quick menu anytime soon?
06:20:47scorcherun e200rpatcher > drop a few files > done :)
06:20:48LloreanBut yes, plugins should either conform to the current theme (not really possible at the moment) or override it
06:20:48jhMikeSscorche: it turns the display on once the driver connects (at least in windows it did)
06:20:49 Quit Noobasaurus (Client Quit)
06:21:03scorchejhMikeS: once our program is loaded
06:21:23jhMikeSyeah, so it should show some info
06:21:31CalcipherI recently was infuriated by my player not wanting to advance through a folder of songs, and it turned out I mistakenly had accessed and changed the repeat option in the quick menu, from my pocked, without realizing it
06:21:39PaulPositionLlorean - So long as that? I thought... oh well. Just didn't happened upon themes that used them before.. Now I do and god do Sokoban, snake, pong, solitaire, battery_bench and how many others become unreadable... :o
06:21:40jhMikeSI guess that's what I meant
06:22:14LloreanPaulPosition: I'm pretty sure the menu was backdroppable *before* the WPS was.
06:22:24LloreanThey shared a backdrop, then the WPS got the %X tag so it could have one of its own
06:22:48LloreanBut yeah, plugins need to deal with it somehow
06:23:08PaulPositionLlorean - wow... I guess I only used pretty simple, solid backdrop wpses then.
06:23:36lostlogicHmph, I seemed to have had a bad bootloader.bin file somehow. it werked this time.
06:24:15advcomp2019good to hear lostlogic
06:24:20PaulPositionLlorean - I guess if we trust themes-maker to choose good background and foreground colours, just removing the backdrop on entry will do it on the plugins that don't explicitely choose their colour.. −− I was wondering if I start doing it, if I should mind posting it on the patch tracker..
06:24:51LloreanHonestly, think all plugins should explicitly set foreground an background colors
06:24:53*jhMikeS wonders why the checking wouldn't detect that before trying to install it
06:25:04lostlogicahh, beautiful rockbox.
06:25:18jhMikeSchecksums should be dumped and CRCs used
06:25:18 Join kkurbjun [0] (
06:25:22LloreanTrusting a theme author to do it is silly, just because you're either saying "Black can't be used as a backdrop" or you're saying something similar to that at least
06:26:23PaulPositionLlorean - Well, logic does say that a background should contrast with the foreground... (???)
06:26:37kkurbjuncan someone give some pointers on how they would expect the build system for a target with an ARM and a DSP to work
06:26:58PaulPositionLlorean - Sudoku uses bitmap tiles, so it needed explicit colours so as not to look shite. But for the rest.. ?
06:27:03LloreanPaulPosition: Yes, but if you use a backdrop image that the text contrasts with, but your previous theme used a background color, that the new text *doesn't* contrast with, but the new theme didn't reset the color because you should never see it...
06:27:05kkurbjunwe have a free (beer) set of tools that TI is allowing for opensource development
06:27:30Lloreankkurbjun: How we'd expect it to work how?
06:27:32jhMikeSwill the binaries be loaded and fed to the DSP or embedded?
06:28:21 Join eigma [0] (i=eigma@
06:28:25kkurbjunI suppose they could be either, I'll get cat on here and he can comment on what he expects to do since he's writing it
06:28:33kkurbjunspeak of the devil:-D
06:29:10kkurbjunright now the build system treats all the C files as if they should be compiled for one arch
06:29:11PaulPositionLlorean - But if it works with the basic rockbox packaged themes, and it is pure logic, surely the onus would be on the theme makers to make it right? Of course it'd be simpler to make them all black-on-white but people would whine, won't they?
06:29:13jhMikeSI'd think loading them in at boot time from disk would be the best way - no mem use - so long as a program loads quickly.
06:29:36kkurbjuneigma the question was: jhMikeS: will the binaries be loaded and fed to the DSP or embedded?
06:29:41eigmajhMikeS: the DSP images? they're like 6 KB
06:30:06jhMikeSStill 6k I wouldn't want hanging around in RAM all the time. :)
06:30:08LloreanPaulPosition: Well, there's two sides to that.
06:30:10eigmaI suppose we could slurp them from disk, but for right now, I have a neat little tool that generates an #include-able dsp_image_*.h
06:30:22LloreanPaulPosition: One: You tell all theme authors to ALWAYS include a backdrop, foreground, and background color line.
06:30:22PaulPositionLlorean - Wait... You speak of the 'browse to new backdrop' sort of function that people may use... ?
06:30:25LloreanThis doesn't seem to work.
06:30:49LloreanTwo: You assume that people won't make perfect themes, and you safeguard against the hazards in plugins by having them explicitly choose color schemes
06:30:51LloreanMuch like Jewels does
06:30:51 Quit webguest73 ("CGI:IRC (EOF)")
06:30:53eigmaso, I was telling kkurbjun, I don't think we should implement this as "two side-by-side toolchains".
06:30:55PaulPositionLlorean - The backdrop isn't important. Rockboxed doesn't use one.
06:31:07eigmathe DSP images are very small, very well-contained and interface minimally with the rest of the code
06:31:18eigmaa few well-placed custom make rules would do the trick nicely imho
06:31:22LloreanPaulPosition: What does this have to do with any single theme?
06:31:57eigmabut 1) I don't know where to put the rules and 2) I don't know how to properly define the dependencies so that they tie in with the rest of the build system
06:32:03PaulPositionLlorean - ... I meant there's no need to make them include a backdrop like you said.
06:32:08kkurbjuneigma, how would we integrate the images into codecs if we didn't load from the disk/build them all in the main image?
06:32:41jhMikeSI wouldn't want any more effort to compling it and throwing it on my device than there is now that's for sure.
06:33:05kkurbjuncodecs are loaded when needed - I guess the codec could include the dsp code it needs to load
06:33:05eigmaby "load from disk", I understood "load from a file in .rockbox, like .rockbox/pcm_codec.out". right now, it's set up for "#include in dsp-dm320.c", and therefore "bundled with main image"
06:33:33eigmaah, I think you may be thinking too far ahead. I don't want to do dynamic loading/unloading yet
06:33:40eigmajust a single system-wide codec that provides PCM passthrough
06:33:42jhMikeSdepends, perhaps anything could be available if needed
06:34:09eigmaonce we have genius DSP programmers who have implemented MP3, OGG, etc., we can think about dynamic loading
06:34:10jhMikeSwell, it can't stay that simple. the codec "thread" must run on DSP eventually.
06:34:10LloreanPaulPosition: No, they need to include a backdrop LINE.
06:34:21LloreanPaulPosition: Basically, they need to explicitly clear the previous one if they're NOT supposed to have one
06:34:25PaulPositionLlorean - Ah right... my bad.
06:35:13*eigma still thinks generalizing the scheduler to the DSP core is overkill
06:35:30*jhMikeS never said anything about generalizing
06:35:55kkurbjunjhMikeS: do you have any pointers on where rules could be added for a different arch, I'm not very familiar with the build system myself
06:36:08eigmayou're saying that there will be a data structure inside the rockbox scheduler that represents the DSP's thread of execution, correct?
06:36:35 Quit joshin (No route to host)
06:36:40*jhMikeS gets the impression people tend to think simple things are complex. The shared memory access would make it just too easy.
06:37:16jhMikeSeimga: maybe, maybe not. I suppose if I did the work, I'd decide on the best course at that time.
06:37:38eigmahmmm... can you clarify "the codec "thread" must run on DSP eventually"?
06:37:42LloreanPaulPosition: That's one of the greatest problems. Most themes, if they don't set something, just leave off the line entirely, instead of explicitly putting the line there to set it to the default
06:37:47PaulPositionLlorean - Actually, now that I think of it, the theme I just tried has its screen cluttered because the maker used some unofficial patches to set margins in the menus.. On official builds, this wouldn't happen and his backdrop wouldn't cut the mustard. :o
06:38:01*lostlogic loads up music on the e200 −− is 40-45 minutes the norm for fully loading one of these here lil' flash doohickers?
06:38:07PaulPositionLlorean - So in a way, I guess I'm hunting windmills..
06:39:09LloreanPaulPosition: Very few backdrops that work in an official menu should cause problems in a plugin, this is true.
06:39:30jhMikeSeigma: the core code wouldn't know the difference and a codec would load on the dsp and interface with core threads using kernel objects. the threading could be really simple since it wouldn't contain context switching code...just sleep/wake code and the sync primitives.
06:39:55PaulPositionLlorean - So out of courtesy for the guy who asked me for it, I'm gonna do a "remove backdrop" patch but WON'T be putting it on the tracker for it to be rejected. :p
06:40:17LloreanI'm not sure it'd be rejected.
06:40:33LloreanBut I feel that just having them trust the themes isn't the most elegant way
06:40:52jhMikeSeigma: btw, does the DSP have an atomic swp instruction or the like?
06:40:54LloreanI'd rather be able to set a "plugin theme" of some sort, just a foreground and background color for plugins to use, separate from the normal one, or something
06:41:54 Quit eigma (Read error: 104 (Connection reset by peer))
06:42:21PaulPositionLlorean - That's interesting. I'll ponder that and maybe try to find the knowledge required to make something out of that idea.. Sort of like the 'radio wps' that's on the tracker... Then again, those two would add to the settings and thus the binary, no?
06:42:41 Join eigma [0] (i=eigma@
06:42:45eigmasorry about that
06:42:50eigmajhMikeS: sounds like a good direction.. I want to clarify some details though.
06:44:25LloreanPaulPosition: Yes, but at the same time, they're not something that can be handled by say, preprocessing a file on the PC or anything else.
06:44:57 Quit eigma (Read error: 104 (Connection reset by peer))
06:45:06LloreanAnd in terms of plugins at least, it's a real usability concern, and I think the most elegant solution is either having each plugin have its own explicit color choices, or have it read the color choices from settings and use user choices for plugin colors
06:45:33PaulPositionLlorean - Thanks anyway :) Every day I spend some time over here, I feel like I go to bed a lil'bit less stupid than I was getting out of it in the morning.
06:45:56PaulPosition(but now, its (the bed) calling my name real loud so...)
06:46:04PaulPositionSee ya gents :)
06:46:05 Join eigma [0] (i=eigma@
06:46:56jhMikeSeigma: such as?
06:47:04JdGordoneigma: hey, bin2c in rbutil/sansapatcher will generate a .c/h from abinary file. but prob too late :p
06:47:08 Quit PaulPosition ()
06:47:12eigmajhMikeS: hardware mailboxes? interrupts?
06:47:43eigmajhMikeS: and the other thing (which I see from the log you anticipated) is I don't see an atomic CAS or TAS instruction the C54x instruction set −− I may have missed it.
06:48:02jhMikeSmailboxes can generate interrupts?
06:48:19eigmajhMikeS: there is a shared memory are between the ARM and DSP, and there is also interrupts in each direction. that's it. can the current scheduler code make use of this?
06:49:02jhMikeSinterrupts are used to wake cores when all threads are asleep already.
06:49:03 Join webguest06 [0] (i=80cde4d3@gateway/web/cgi-irc/
06:49:37jhMikeSon pp, it uses the processor control interface when it's from one core to another
06:49:51webguest06hey, i've just installed rockbox with the automated installer, and now my ipod shows "can't find rockbox.ipod"
06:50:02webguest06what did i do wrong during install?
06:50:43eigmasounds like a pretty good match to the DM320
06:50:46psycho_maniaci think you just installed the bootloader. extract the file
06:51:09eigmaanyway, we kind of strayed from the original topic of the conversation
06:51:29jhMikeSeigma: for pp is absoltely symmetrical thought since everything is shared but the caches
06:51:33webguest06the automated installer doesn't extract the file?
06:51:41eigma..which is how to integrate TI's build tools into the Rockbox build system
06:52:07eigmaif I wanted to compile a *.c file with the TI build tool by manually writing a Makefile rule, where would I put this rule?
06:52:20 Quit Calcipher (Read error: 104 (Connection reset by peer))
06:52:42JdGordoneigma: I think the best way (for now) is do a seperate rule for it like how the bmp's are done
06:52:45eigmaJdGordon: thanks, I'm actually reading in Texas Instruments COFF files, not raw binary files.. buy thanks
06:52:48JdGordonand compile in the binary
06:53:06JdGordon.. or that
06:53:29eigmathat was Re to your previous message about bin2c btw
06:53:38JdGordonyeah, i figured :)
06:54:00JdGordonwoudlnt it be better having the dsp code in the rockbox binary so we dont have more to read from disk on boot?
06:54:57eigmawell, whether it's in the binary or in a separate file, the net disk read is the same
06:55:20eigmaputting it in a separate file saves main RAM because it can be unloaded as soon as it's put into the DSP RAM
06:55:33JdGordonyeah, but its 6k... not a big deal
06:55:39eigmayeah, I agree
06:56:40webguest06it still gives me the "rockbox.ipod not found" message
06:56:51eigmaI understand that RAM is at a premium, but 1) the image is very small and there's only one.. this will become more of an issue with multiple codecs 2) the m:robe, after all, has 64MB of RAM. 6K/64MB < 0.01%
06:56:54webguest06what do i do?
06:58:08krazykitwebguest06, extract it yourself
06:58:13webguest06i'm using the automated installer
06:58:22webguest06i have the .zip
06:58:33webguest06i just don't know how to install it to the iPod
06:58:47psycho_maniacextract it to the root of the drive of your player.
07:00:13jhMikeSeigma: but I'm realizing there's some sticky details re: thread structs...well, that's what makes it interesting to attempt to make the DSP look like part of the CPU anyway. :)
07:01:35eigmayes, that was the next thing I was going to ask.. can the scheduler handle heterogenous sets of registers, etc?
07:02:01webguest06i copied the zip file to the iPod and extracted it
07:02:06webguest06but i get the same message
07:02:16krazykitwebguest06, do you have a .rockbox directory in the root of the drive?
07:03:16krazykityou should have some path like F:\.rockbox\ with a bunch of folders and some files, including rockbox.ipod
07:03:43jhMikeSeigma: it wouldn't need to since those would be defined based on the core it's made for (as well as what's included). it has no basic need for thread structs to be in an array though so they can be anywhere. I've considered caller-allocated thread struct which would also mean no unused ones and irrelevant placement.
07:03:53webguest06i'm using MacOS 10.4
07:04:06webguest06so all i see is iPod
07:04:19webguest06and there's no .rockbox directory
07:04:55psycho_maniaci havnt been around. did you format the hd to fat32?
07:05:05jhMikeSthreads from two heterogenous cores could block in the same thread queue (to mutex data for instance)
07:05:08webguest06i used my friend's laptop
07:05:15webguest06it's connected to my mac
07:05:21webguest06and the iPod is in disc mode
07:05:32psycho_maniacalright and you DID install the bootloader?
07:05:45webguest06yes the bootloader seems to be working fine
07:05:49*krazykit points out that Macs hide .files by default, so you may need to show hidden files
07:05:57webguest06oh right
07:06:02webguest06this is my friend's mac
07:06:12webguest06so i'm not familiar with how to unhide hidden folders
07:06:12psycho_maniacgood point krazykit
07:06:44krazykitbut i must be off to bed now. good luck.
07:07:05krazykithint: check in Finder's options somewhere, or maybe apple-H or control-H or something.
07:09:15 Join perrikwp [0] (i=982146c1@gateway/web/cgi-irc/
07:10:22***Saving seen data "./dancer.seen"
07:12:36webguest06i can't find it
07:12:45jhMikeSeigma: I dunno, if it hits too many snags, I'll agree...might be overkill...but some simple kernel extension for this situation could be devised. Just thinking about all sorts of possible arrangments.
07:12:52webguest06does anyone know how to show hidden folders in macOS 10.4?
07:13:32psycho_maniacgoogle will tell you
07:14:13 Quit perrikwp ("CGI:IRC (Ping timeout)")
07:25:04 Quit webguest06 ("CGI:IRC (EOF)")
07:26:12amiconneigma: (1) It is never a wise idea to waste a resource just because you seem to have plenty of it on a target. No resources are infinite. (2) Reading the pcm codec from a separate file does take longer than if it would be built into the main binary. That's because of the extra file open().
07:27:17amiconnThat said, I think a separate file would be easier, and btw, I am thinking about doing the same for the pcm codec on archos when integrating pcm playback+recording
07:28:05amiconnSituation is a bit different there because this codec isn't code we have written, it's not even open source
07:28:28 Quit eigma (Read error: 113 (No route to host))
07:29:07jhMikeSwill anyone actually perceive a file open() if the disk already has to be spinning for some reason? I pull that trick to hide probing the disk for recording filename without RTC when the codec must also be loaded.
07:33:48 Join woodensoul [0] (n=noneofya@
07:34:29amiconnA single file as such won't be noticeable, I just wanted to point out the fact that it doesn't come for free. test_disk.rock also measures open() performance. Without dircache (and that's usually the case during bootup, unless dircache needed a foreground scan), open() performance is in the order of 20..50 files/second unless it's a fast flash target
07:34:35woodensoulHey all, anyone around that knows a thing or two about Rockbox on the Archos line and replacing the hard drive in those units?
07:36:19woodensoulI find it hard to locate the docs in the Wiki, but I know I've seen a status report on the Archos line.
07:40:26 Part Llorean
07:41:50 Join Buschel [0] (
07:41:56Buschelgood morning
07:42:57amiconnwoodensoul: . 3 clicks from the rockbox main page...
07:45:02Buschelquestion about the yesterdays submit for ata-timing (fix for nano): in ata_device_init there is now no setting left for non-Nano. As far as I can see this function is called for all targets, not only for nano. Is this correct?
07:45:25woodensoulThanks, that will help. I'm looking for a list of features implemented/unimplemented for the Archos line.
07:46:42woodensoulI think I found it. It's on the 'Why Rockbox' page.
07:47:51woodensoulNo RPG or crossfading support for Archos is a bummer
07:48:54lostlogicis it normal for Sansa to be continuously slowly filling the compressed buffer as opposed to quicly filling the whole thing?
07:49:07woodensoulI was thinking of taking the HDD out of my Creative Zen Xtra and putting it into one of the Archos models so I could have 60GB of Rockbox music.
07:50:58JdGordonlostlogic: seems to be filling quickly on mine...
07:51:18JdGordoncould still be a problem with the ata_is_Active() if thats still being used
07:51:49lostlogicshouldn't be.
07:53:15lostlogicappears to be a performance issue, fighting between pcmbuf and compressed buf.
07:53:34 Quit Buschel ()
07:57:08 Join LinusN [0] (i=linus@rockbox/developer/LinusN)
07:59:45jhMikeSlostlogic: I have the fix.
08:00:01lostlogicjhMikeS: tweaks to yield_codec (removing the god forsaken thing? ;))
08:00:17jhMikeScomplete removal of any sleeps at all
08:00:47lostlogicsure, link me
08:00:58jhMikeSsleeps are nasty
08:01:16amiconnsleeps save power
08:01:37lostlogicamiconn: not if they force filling to take a year and a day.
08:02:23jhMikeSblocks save more
08:03:42jhMikeSit keeps the filling of both nice on mine anyway
08:04:09lostlogichave tested on other targets?
08:04:17*amiconn always thinks sleep == cpu sleep
08:05:39*jhMikeS just randomly switches between one type and another :)
08:06:01 Quit jpt9 (Read error: 110 (Connection timed out))
08:06:20jhMikeSH10 with a nasty slow disk seems to still be ok
08:06:42amiconnjhMikeS: Btw, does you explicit boost control commit mean that boost/unboost can happen more often?
08:06:56amiconnThis is something that should be avoided...
08:07:49jhMikeSno, same thing that would happen otherwise
08:08:25amiconnah ok
08:10:42 Join webguest42 [0] (i=a31d4902@gateway/web/cgi-irc/
08:12:21 Quit webguest42 (Client Quit)
08:12:33 Join Rob222241 [0] (
08:13:19lostlogicminimum volume mute no worky on e200 yet.
08:13:55jhMikeSIf something gets messed up, you can identify the thread at fault in the thread screen (shows a +)
08:14:12 Join ender` [0] (
08:15:30lostlogicman there's a lot or electrical noise in the e200.
08:15:41lostlogicI can easily hear it over music at my 'office' listening level
08:15:52jhMikeSboth mine seem fine for playback
08:15:59TheCollectoras does mine
08:16:03amiconnSpeaking about mute - there is a bug on X5/M5 (TLV320)
08:16:07 Join GodEater [0] (n=bryan@rockbox/staff/GodEater)
08:16:11lostlogicdo you listen at -74dB? :-P
08:16:29jhMikeSummm...not normally
08:16:34TheCollectorI am right now :-P
08:16:40amiconnThe line out on those is a true line out, i.e. unaffected by volume level - but the mute level also mutes line out
08:17:01lostlogicI need to put a permanent attenuator in my earphone cable.
08:17:15lostlogicamiconn: I think on most targets setting the minimum volume level calls audio_mute() or some such
08:17:17TheCollectorwhat earphones do you have?
08:17:18jhMikeSmust have efficient phones
08:17:28lostlogicTheCollector: and you don't hear electrical noise in quiet parts?
08:17:46TheCollectornot that I've recognized
08:18:01*amiconn thinks that lostlogic should try an archos recorder
08:18:04*jhMikeS can hear something when initially starting playback.
08:18:35jhMikeSlostlogic: are you listening to silence tracks?
08:18:43amiconnBetter sound quality than most newer targets, and adjustable down to -99dB in rockbox (could be patched to go down to -112dB)
08:18:50lostlogicjhMikeS: yeah, I can definitely hear the flash access... it's not a lot when not filling
08:19:16lostlogicamiconn: hmm, maybe I should have been more careful with the one that I had and not broken it :(
08:19:24lostlogicjhMikeS: no, just normal music
08:19:48amiconnpixelma reported that there is *very* noticeable electrical noise on c200 when inserting a microSD card
08:19:57lostlogicwell the Sansa is already better than iPod video on volume -74-6 instead of -57-6
08:19:59*jhMikeS really thinks there must be a good deal of variation on e200's and got lucky twice.
08:20:06jhMikeSlostlogic: impediance is?
08:20:44TheCollectorI just reinit'd my db while listening at -45 and can't hear anything
08:21:15lostlogic27ohms, 119db/mw
08:21:47nanoklostlogic: what kind of headphones are those? i haven't tested, but i thik i have no headphones which would make any sense at thatlevel
08:22:08lostlogicWestome UM2
08:22:09jhMikeSmine are 32. 16 ohm seems to have a problem. maybe 27 is a factor too.
08:22:10advcomp2019i have that same noise with very low volume too but the higher volumes i do not hear it
08:22:21nanokmy ffice listening volume is around -18, can go all the way up to -10, maybe down to -25
08:22:36lostlogicadvcomp2019: yeah, the noise level stays constant as volume changes so louder music easily blots it out
08:23:37nanoklostlogic: but i am using the stock earbuds in the office, i cannot use the px100's (way too loud on the exterior)
08:23:37lostlogic-70-ish for me, -50-ish for cycling outdoors
08:23:38jhMikeSthere's quite some value in having a separate codec placed near the jacks.
08:24:22nanoklostlogic: damn, get lower sensitivity headphones, i would say ;)
08:24:33jhMikeSthat's too efficient to be practical
08:25:16nanokthose are not headphones, they must be yanked out from a sonar detecion subsystem :-P
08:25:30nanok(not that it makes any sense)
08:26:02lostlogicnot planning on replacing my perfectly functional $300 earphones any time soon.
08:26:34nanoklostlogic: are they insulated, that sort of studio headphones?
08:26:54lostlogicnanok: foam sleeve in-ear model, if that's what you mean
08:26:55jhMikeSfrom the future
08:27:09nanokaaahm i see
08:27:13lostlogichmm, what button to hold on sansa to insert USB and ignore?
08:27:26nanoklostlogic: middle of the wheel
08:27:27jhMikeSyeah, I'd like that feature too :)
08:27:33lostlogicnanok: thanks
08:27:34TheCollectoroh, nice
08:27:39*nanok has to go
08:27:43*jhMikeS didn't even know that :P
08:27:48nanokor i will not find a parking place at work :(
08:27:49 Join OlivierBorowski [0] (
08:28:34nanokjhMikeS: i locked my sansa at first trying various combinations :) had to hard power-it down
08:29:15nanokas i was saying, the manuals need a bit of brushing up. i need to find some time for that (i am not capabale to code anyway, so that's the least i can do, i guess ;) )
08:29:27lostlogicnanok: would be helpful :)
08:29:28jhMikeSif you locked it doing that, file a bug report
08:29:40lostlogicok, sansa is setup, new toy werks.
08:29:55nanokjhMikeS: i locked it holding another button, trial and error
08:30:05nanokreally got to go
08:30:07jhMikeSit really shouldn't
08:30:09nanoksee you guys later
08:30:27nanokjhMikeS: okay, noted. i will try it again and see if it is worthy of a bug report
08:30:49 Quit Rob2222 (Read error: 110 (Connection timed out))
08:31:19lostlogiclots of good themes out there.
08:31:42jhMikeSlostlogic: did you apply that patch and try it out?
08:31:55lostlogicjhMikeS: yeah, works great, fills ... amazingly fast and no skips.
08:34:06lostlogicfiles from microsd and main drive can be in the same playlist, right?
08:34:20psycho_maniacwow found a really bad typo in the gigabeat manual
08:35:34psycho_maniacWhen this
08:35:35psycho_maniacswtich is to the lwft, the battery is disconnected. This can be used for a hard reset of
08:35:35psycho_maniacthe unit, or if the player is being placed in storage.
08:35:48psycho_maniacwhoops sorry bout that, but thats the typo
08:36:02jhMikeSlostlogic: it just merges them
08:36:33*jhMikeS is tired...thinking "database"...bah
08:36:41jhMikeSdoesn't matter anyway
08:37:03lostlogicjhMikeS: :) I haven't used a multi-storage target before.
08:37:22lostlogichmm, I can turn this into an 8g device for only $65. nice.
08:38:03 Quit ender` (Read error: 104 (Connection reset by peer))
08:38:07 Join ender` [0] (
08:38:29LinusNpsycho_maniac: thanks for noticing, i have fixed it now
08:39:39psycho_maniacno problem.
08:41:11 Quit BigBambi (Read error: 113 (No route to host))
08:45:33lostlogicjhMikeS: just catching up on SVN −− should actually reduce boost-flux in some corner cases (re amiconn's question) definitely better than what Slasheri and I had setup before.
08:46:26 Join linuxstb [0] (n=linuxstb@rockbox/developer/linuxstb)
08:47:14jhMikeSCan't be worse in any case since I placed it before the timeouts that would have triggered. I did remove the need to reboost in buffering because of the sleep during buffering.
08:47:38 Quit ender` (Read error: 104 (Connection reset by peer))
08:47:59 Join ender` [0] (
08:51:13 Quit nanok (Read error: 113 (No route to host))
08:54:05 Quit OlivierBorowski (Remote closed the connection)
08:55:21 Join petur [0] (n=petur@rockbox/developer/petur)
08:55:51 Quit scorche (Nick collision from services.)
08:56:20 Join scorche [0] (n=scorche@rockbox/administrator/scorche)
09:00:36GodEateranyone seen this yet : ?
09:01:49woodensoulDoes anyone know if the only way to get gapless on the Archos units is by the -nogap switch?
09:02:41woodensoulGapless doesn't work with LAME 3.90.3 or later without the -nogap switch?
09:02:55psycho_maniacyou could by a different player
09:04:59amiconnwoodensoul: Gapless only works with -nogap on archos. This has nothing to do with lame version
09:06:56amiconnThat said, you won't notice a gap between usual album tracks even without -nogap, as the gap will be 26ms at maximum, 13ms on average. It's only relevant for live albums where there is no silence at track change
09:07:32amiconn(or similar things, like dj mix album)
09:10:24***Saving seen data "./dancer.seen"
09:11:10woodensoulI mentioned the version number because it's my understanding that that version and later version added the gap tag info which is how rockbox achieves gapless on Software Codec targets.
09:11:51woodensoulamiconn: what about ReplayGain? Not currently supported, but possible for the Archos line?
09:13:41amiconnThere is a patch for replaygain on the tracker that needs testing and a bit of cleanup/improvement
09:13:42 Join Zagor [0] (
09:15:09amiconnIt's on my todo list, but not high priority (since I don't use rg at all)
09:17:16woodensoulI'm looking into getting one of the Archos line to put the 60GB hard drive from my Zen Xtra in it to have 60GB of rockbox music. What's your overall impression of Rockbox on them?
09:17:20 Join OlivierBorowski [0] (
09:17:41woodensoulBattery life, etc.
09:18:26amiconnBattery life is quite good. It's better than in the OF
09:18:29linuxstbZagor: Good morning. I assume you've seen FS #8189 ?
09:18:59woodensoulHow many hours for Rockbox? I don't have one yet, so I don't know what the OF gets.
09:19:15amiconnIf you plan on getting one of the models powered by NiMH cells, I'd strongly recommend to get a set of those new generation cells with significantly reduced self discharge
09:19:53woodensoulThanks. Is the old 6000 capable of displaying the default WPS? I read something about the LCD being different on those.
09:19:56Zagorlinuxstb: yes I have. interesting!
09:21:10amiconnThe Player series has a charcell type lcd
09:21:43amiconnOf course it can display its default wps - but it's a different default than on the other devices
09:23:09amiconnWith 2100mAh cells (typical capacity of the new generation) you can expect around 15 hours of runtime if you don't use the backlight often
09:23:50 Join J-23 [0] (
09:24:00woodensoulDon't all the devices display the first one listed here as the default?
09:24:30woodensoulBut obviously the Archos can't display Gigabeat F WPSs.
09:25:31Bagderwoodensoul: not graphical, character based
09:25:38woodensoul"The old player has a limited LCD with no support for double line height and only four user definable characters instead of eight." What does this mean?
09:26:07Bagderwoodensoul: you should aim for a recorder model really if you're looking into archos models
09:26:20Bagderrecorder v1
09:26:48 Join radinp [0] (n=pradin@
09:27:02Bagderdid anyone read the vulnerability report on FLAC?
09:27:11Bagderit is so badly written...
09:27:27Bagderit just speaks of FLAC and then mentions flaws
09:27:32woodensoulWell I'm actually looking to spend as little money as possible because I have way too many Rockboxed players and I just wanted something to throw the 60GB hard drive from my Zen Xtra into.
09:27:34Bagderbut FLAC is a format, not code...
09:27:45radinpJust curious but is anyone working on the text-viewer search feature?
09:27:55peturit speaks about flaws in a certain lib that is used a lot
09:28:22J-23Search in text-viewer?
09:28:27Bagderah, yes it does down in the end
09:28:41woodensoulBadger: what do those (recorders v1) go for on eBay?
09:28:47peturfixed in libFLAC 1.2.1
09:28:47Bagderwoodensoul: no idea
09:29:02Bagderand we don't use libFLAC...
09:29:28GodEateris our flac code all our own then ?
09:29:35peturso all is well....
09:29:37Bagderno, ffmpeg
09:29:53linuxstbThe decoder is ffmpeg, the flac parsing code and vorbis comments parsers are our own.
09:30:01radinpJ-23: There's a feature request to search for a specific string in the text-viewer plugin. I was thinking of adding a regular expression library in support of this feature.
09:30:36linuxstbIt's obvious that our parsers might well have similar vulnerabilities though - and not just FLAC/vorbis comments...
09:32:08woodensoulBadger: why is the recorder v1 superior to the others archos players?
09:32:12linuxstbBagder: But petur said "all is well" ;)
09:32:43*petur shuts up and concentrates on payed work
09:32:51Bagderwoodensoul: it is really up to each and every one of course, but I like a graphical display and AA batteries
09:33:40woodensoulI will probably leave it in my car and have it plugged in to a cigarette to AC / 3 prong adapter.
09:33:49woodensoulThe other models don't use AA batteries?
09:33:57Bagderwoodensoul: the v2 and fm don't have AA
09:34:00radinpLet me rephrase it, search would be for text_editor plugin not text-viewer.
09:34:09Bagderthe recorder also has slightly better sound than the player
09:34:13woodensoulThe ondio don't, but the 6000 does, right?
09:34:28Bagderwoodensoul: yes, the 6000 ("the player") has AA
09:34:35woodensoulI thought the recorder doesn't have line out.
09:35:50woodensoulIf I were to put in the drive from the Xtra to the 6000, can I plug in the 6000 and format the drive through windows? The Zen Xtra has its own file system.
09:36:20Bagderyes you can
09:36:33Bagderand correct, the recorder has no separate line out
09:36:48 Join spiorf [0] (
09:36:59woodensoulAnd I assume that putting the 6GB drive into the Xtra would also work.
09:37:28woodensoulFor connecting to my car stereo line in, or home stereo, which is better, the recorder or the player?
09:39:02J-23radinp: :) reading ebooks on Rockbox will be more comfortable
09:39:46radinpHey, now there's an idea. How about a bookmarking option?
09:39:51JdGordonLinusN: wasnt the braces rule to just be consistant in the file?
09:40:02 Join davina [0] (
09:40:34LinusNJdGordon: actually, yes, but i took the liberty to "fix" that as well when i removed the tabs
09:41:10JdGordonah there was tabs int here also?
09:41:15linuxstbI thought it was required for the braces surrounding a function? Although that may not be CONTRIBUTING...
09:41:17LinusNplenty plenty
09:41:34JdGordonthe web diff thing doesnt show whitespace changes.. so ok
09:41:44Bagderonly if you select unidiff
09:42:09linuxstbJdGordon: From CONTRIBUTING - "Braces for function declarations are put in a new line under the name"
09:42:29J-23radinp: Yes!
09:44:22linuxstbZagor: Is that last comment in FS #8189 (about index/mask being wrong, and the endpoints being switched) directed at your patch?
09:45:44ZagorI think so, but I disagree with him. if it was that wrong, nothing would work.
09:45:59linuxstbThat's what I would have thought...
09:46:15Bagderzagor: still time to get that code into SVN now, right?
09:46:25JdGordon.... OR.... with the change everything could work bettererer!!!>...
09:47:00linuxstbZagor: Do you think his patch will have the same problem with 512-byte transfers - hence the panics?
09:47:47Zagorlinuxstb: I did a quick test and indeed it seems his patch has the same problem. the first big transfer fails.
09:50:29Zagorthe panic I saw was a stkov(0), which seems rather strange.
09:57:33 Join CaptainSquid [0] (
10:00:02 Join pondlife [0] (
10:00:59 Part J-23
10:01:00 Quit woodensoul ()
10:05:49 Join qweru [0] (
10:06:13markunradinp: the viewer remembers the position in the file when you exit, isn't that bookmarking?
10:06:34 Quit TMM (Remote closed the connection)
10:06:52 Quit courtc (Read error: 113 (No route to host))
10:08:27markunwould it make sense to replace the 12k text buffer in the viewer by rb->plugin_get_buffer ?
10:09:04Bagderyes, I would say so
10:09:21Bagderthen it could be much bigger on most targets
10:09:39markunthe 12k is taken from the plugin buffer anyway, right?
10:09:57markunI'll change it tonight, have to go to work now
10:10:51markunI see one possible problem:
10:11:07 Join courtc [0] (
10:11:53markunif you bookmark with a very large top_ptr and the next time you view the file the buffer is smaller you will read past the buffer
10:14:16JdGordonthe buffer size shuold stay constant for the plugin
10:14:25JdGordonunless something major happens
10:14:49Bagderimho, that sounds like a crappy design of the plugin...
10:15:29markunBagder: we should only have to save 1 pointer, right?
10:15:53Bagderand that's an index into the file, so it would work no matter what the new buffer size is
10:16:07markunI'll try to change that as well, but roolku will not be happy if the bookmark file format
10:16:36Bagderwell, perhaps you can use a dummy for the second value to keep the syntax consistent
10:16:58BagderI didn't check the details, I'm speculating here
10:17:09markunyes, it's easy to do
10:17:22*markun is off to work now..
10:18:33*Bagder added a "not the v2 models" on the sandisk line on the front page
10:19:18*JdGordon would actually like viewer to buffer the entire file into the audio buffer (with some help from MoB so it doesnt have to stop playback to do it)
10:19:45Bagderwhat kind of super huge text files are you using? ;-)
10:19:59linuxstbRockbox manual?
10:20:00JdGordonim not... but 12k isnt really all that large
10:20:23BagderI bet the rockbox manual as text fits in the plugin buffer on most targets
10:20:55Bagderand really, it can't use the audio buffer
10:21:08Bagdermob is nice, but you'd have to fragment the buffer badly
10:21:30JdGordonanother thing that would be ncie for viewer (apart from renamening it) is multiple bookmarks per file and a nice ui for it.. then we could stick the manual on with preloaded bookmarks for the chapters
10:22:34linuxstbIsn't that another project? i.e. a viewer for hyperlinked documents with nice formatting?
10:23:00JdGordoncould be
10:26:35JdGordonis used for rombox ?
10:27:39 Join tkooda [0] (i=goaway@
10:30:47 Quit courtc (Read error: 104 (Connection reset by peer))
10:31:02 Join courtc [0] (
10:34:43 Quit psycho_maniac (" HydraIRC -> <-")
10:37:26 Quit qweru ("moo")
10:37:32markunJdGordon: the plugin buffer on most targets is 512k (32k on the archoses) so we should be able get a bit more than 12k
10:38:06JdGordonhmm... yeah, forgot it was so big
10:38:45markunviewer.rock is 14k in my gigabeat build
10:40:44 Quit courtc (Read error: 104 (Connection reset by peer))
10:42:24 Join courtc [0] (
10:43:32linuxstbmarkun: Regarding "top_ptr" becoming out of range, shouldn't the plugin always check for that case anyway, and just ignore the bookmark (or reduce the value) if it's too big?
10:45:52markunlinuxstb: it should, but I'll change to code to only use top_ptr
10:46:43 Join J-23 [0] (
10:47:03markunlinuxstb: right now it's a 'int' shouldn't it be size_t?
10:47:32 Join Aware [0] (n=bill@
10:48:13amiconnJdGordon: I think the text viewer should stick to using the plugin buffer *only*, in order to not stop playback
10:48:36amiconnHowever, it should use plugin_get_buffer() and use that, as already mentioned
10:51:07linuxstbmarkun: I'm not sure...
10:53:28 Quit Aware ()
10:54:13 Join lee-qid [0] (
10:54:21markunlinuxstb: I just hope they have the same size, so we don't break the bookmarks format..
10:54:54 Quit J-23 (Remote closed the connection)
10:56:15 Quit radinp (Read error: 110 (Connection timed out))
11:10:25***Saving seen data "./dancer.seen"
11:13:13 Join J-23 [0] (
11:13:39 Join PaulJam [0] (
11:14:31PaulJammarkun: here?
11:14:36markunPaulJam: yes
11:15:00markunfound any bugs I introduced?
11:15:20PaulJammarkun: it seems as if your commit in r15695 caused a little bug in the viewer.
11:16:21markunPaulJam: can you tell me what happens?
11:16:39PaulJammarkun: it you have the show scrollbar seeting enabled and open a textfile then the scrollbar is not there. it only appears after you habe entered the viewer options submenu.
11:16:59markunthanks, I'll look into it
11:17:34PaulJam(i don't really see the relations to the changes, but i have tested in the sim, and with the previous rev it works normally)
11:19:30markunto me the relation is also not obvious
11:22:25markunPaulJam: could be that buffer_end is not initialized correctly anymore
11:24:34XavierGrLinusN: ping
11:25:11XavierGrLinusN: Just a heads up FS #8178
11:25:17 Join sqgl_LakeMacq_ [0] (
11:25:30LinusNXavierGr: ah yes
11:25:44XavierGrLinusN: when you have the time of course
11:27:03XavierGrLinusN: thanks :)
11:28:02sqgl_LakeMacq_Anyone here replaced their Archos Jukebox HDD? I did and it now won't shutdown gracefully; need to pull batteries out each time :(
11:28:26LinusNsqgl_LakeMacq_: ouch
11:28:29Zagorsqgl_LakeMacq_: I'd say most archos users have replaced their drives
11:28:46Bagderhehe, indeed or they aren't users anymore ;-)
11:28:46sqgl_LakeMacq_Could it be that this drive draws more power than the previous one?
11:29:06 Part J-23
11:29:09Zagorsqgl_LakeMacq_: that would rather result in problems using it, not problems shutting down
11:29:14sqgl_LakeMacq_bagawk, he he "Archos Losers" ;)
11:29:40*Zagor uses his trusty recorder every single day
11:29:43sqgl_LakeMacq_i mean Bagder.... gee so many ppl here, the autocomplete isn't very useful
11:30:34LinusNsqgl_LakeMacq_: so you can play music without issues, but not shut down?
11:30:36sqgl_LakeMacq_Zagor, that's what i thought. i have no problems using it. Just shutting down. Oh well
11:30:47LinusNwhich model is it?
11:30:49sqgl_LakeMacq_Been using it like this for years
11:31:20sqgl_LakeMacq_Not sure, didn't kno there was more than one
11:31:45Zagorsqgl_LakeMacq_: if you have AA batteries, you have v1
11:31:58sqgl_LakeMacq_Oh yes, i see what you mean now.
11:32:27LinusNsqgl_LakeMacq_: can you shut down if you put back the old hard drive?
11:32:57sqgl_LakeMacq_And the ones with AA batteries are also the oens with laptop sized drive?
11:33:33 Join Bagder_ [0] (
11:33:45 Quit Bagder (Nick collision from services.)
11:33:47sqgl_LakeMacq_The old drive died. hence replacement. IT was 20GB when i bought it. Now 40GB
11:33:57 Nick Bagder_ is now known as Bagder (
11:34:09LinusNsqgl_LakeMacq_: the V2 and FM recorder have the same type of hard drive as V1
11:34:14sqgl_LakeMacq_But this HDD is 1.0 Amps, old one was much less
11:34:51LinusNthat might be a problem, but not when shutting down
11:35:04sqgl_LakeMacq_LinusN, i thought they no longer sell jukeboxes with laptop sized drives 9hence cheaply replaceable/upgradable)
11:35:13 Nick parafin|away is now known as parafin (
11:35:14LinusNthey don't, afaik
11:35:23Bagderwell, not archoses at least
11:35:53LinusNsqgl_LakeMacq_: so you can't even force a shutdown by holding OFF?
11:35:56Bagderand the 2.5" hdd models seem to never grow popular so they live their lives in the shadows
11:36:01sqgl_LakeMacq_OK, two people say it can't be a problem. I'm convinced. Bummer though. Coz pullin out batteries has caused a hairline fracture in board re power line.
11:36:25sqgl_LakeMacq_LinusN, no i cannot force shutdown holdin off. It says "shuttign down" but does not actually do it.
11:36:40 Quit homielowe (Read error: 110 (Connection timed out))
11:36:47LinusNnot even if you keep holding OFF?
11:36:53amiconnsqgl_LakeMacq_: But the ON button works normally in rockbox?
11:37:01*sqgl_LakeMacq_ doing it now
11:37:43LinusNamiconn: yes, that was my next question :-)
11:38:20sqgl_LakeMacq_the on button does the right thing (ie toggles between transport display and file dispplay)
11:38:58sqgl_LakeMacq_but since i have to take batteries out to shut it down, the on button becomes irrelevant in the sense that...
11:39:18sqgl_LakeMacq_when i put batteries back in, it autoboots. No need for ON.
11:39:38austriancoderare some developers with an iriver hxxx player here?
11:39:47LinusNthe reason we ask is that we have seen similar problems when the ON button is stuck
11:39:56LinusNaustriancoder: i have one
11:39:57Bagderautoboot sounds like stuck ON...
11:40:17sqgl_LakeMacq_LinusN, I wish it was just stuck. Good guess, but no.
11:40:24 Join J-23 [0] (
11:40:35austriancoderLinusN: great... can you check if the audiohw_reset(); call in bootloader is needed?
11:40:48LinusNsqgl_LakeMacq_: still, i am sure that the problem isn't the hard drive, but instead something else broke when you replaced it
11:40:53sqgl_LakeMacq_Bagder, if it was stuck on, then ON would not do it's toggling trick when playign a file.
11:41:07Bagderbut why does it boot when you insert the batteries?
11:41:31sqgl_LakeMacq_Good question
11:41:35LinusNaustriancoder: it is, because you will get a nasty humming sound during boot if we don't reset it
11:41:45ZagorBagder: iirc that is normal
11:41:50Bagderit is? oh, ok
11:41:56*Bagder forgets
11:41:59austriancoderLinusN: okay..
11:42:05BagderI haven't used my recorder in a long time...
11:42:46sqgl_LakeMacq_Bagder, ON button also performs the neat little Rockbox trick of resuming play from where it left off when i pullled the batteries out.
11:43:35amiconnBagder: Yes it is normal. The hardware defaults to on
11:43:49austriancoderLinusN: but we could also call audiohw_init() or?
11:45:07austriancoderLinusN: as it calls the reset...
11:46:26sqgl_LakeMacq_Bagder, so someone still makes a player with 2.5" drive? Does RockBox work on any such player?
11:46:40LinusNaustriancoder: i believe that could work fine too
11:47:23 Join Phill [0] (
11:47:33LinusNaustriancoder: however, i don't know if the i2c driver is initiated at that point
11:48:16Zagoroh ffs, what is it with viewer.rock and pointless splashes? on a clean install it says "no valid settings file" for two seconds every time I view a text file...
11:48:36sqgl_LakeMacq_LinusN, If something broke during HDD replacement, then it is surprising that i get to the "Shutting down..." screen at all, no?
11:48:43BagderZagor: kill kill kill
11:48:50LinusNsqgl_LakeMacq_: why so?
11:49:03amiconnDon't the full size Archos AV series devices still use 2.5" hdds?
11:49:14BagderI think they do
11:49:25Bagdersince they've had 160GB drives a while
11:49:34LinusNaustriancoder: looks like it isn't
11:49:36sqgl_LakeMacq_LinusN, because it is a logic problem, and surprising that i have no other logic bugs.
11:49:57amiconnBagder: Yes, even the latest (AV705 wifi):
11:50:05LinusNsqgl_LakeMacq_: well, to me it sounds like a problem with the power supply
11:50:08sqgl_LakeMacq_LinusN, but i am not even a RockBox programmer, so i defer to you.
11:50:08austriancoderLinusN: okay.. I will bring the audiohw_reset back
11:50:45sqgl_LakeMacq_LinusN, yes, you are right, makes sense. Thanks.
11:51:12LinusNsqgl_LakeMacq_: it can probably be repaired
11:51:48sqgl_LakeMacq_austriancoder, are you happy with audio quality of iriver? i am looking for small player for a pal, and was surprised how crap audio on my Creative Muvo is.
11:52:06LinusNsqgl_LakeMacq_: i am very happy with my iriver
11:52:17austriancodersqgl_LakeMacq_: I am sorry.. i own no iriver
11:52:18sqgl_LakeMacq_LinusN, repaired eh? I'll open up and have a look right now.
11:52:26LinusNthe h100 series rocks
11:52:44LinusNstill one of the best daps ever made if you ask me
11:52:50BagderI have tin ears, all my players sound good to me ;-)
11:53:10sqgl_LakeMacq_"tin ears"... not heard that oen before :)
11:53:32 Part J-23
11:53:43sqgl_LakeMacq_I use my Creative for spoken podcasts so am not bothered. But was surprised since they were pioneers in audiocards for PC's.
11:54:16sqgl_LakeMacq_*and* the FM tuner doesn't pick up well at all.
11:58:20linuxstbI didn't think Creative had a reputation for quality soundcards - just popular ones...
12:00:07austriancoderLinusN: i have an other idea... i could do something like
12:00:38 Nick fxb__ is now known as fxb (
12:01:05LinusNaustriancoder: i am behind a firewall that restricts port 8001
12:01:42LinusNhmmm, worked with port 80 as well...
12:02:12LinusNaustriancoder: that could work
12:03:02austriancoderLinusN: fine..
12:04:05LinusNamiconn: there?
12:04:41linuxstbIt doesn't seem very clean to have hardware-specific inits in the bootloader, rather than calling a driver function...
12:06:07LinusNlinuxstb: i don't really mind, since the bootloader is hardware specific
12:06:21LinusNaustriancoder: why is it important to remove audiohw_reset()?
12:06:41austriancoderLinusN: to get a clean api...
12:06:58linuxstbBut if some hardware requires audiohw_reset...
12:07:02LinusNis it cleaner without the reset?
12:07:45LinusNwell, the audiohw_reset() could be moved to the target tree
12:08:30austriancoderLinusN: audiohw_reset is only used by one bootloader.. but wait some minutes.. i will show you then a patch
12:09:11 Join J-23 [0] (
12:10:57 Join Rondom [0] (
12:10:59austriancoderLinusN: - what do you think?
12:11:32LinusNi think that is acceptable
12:11:59LinusNthat code certainly doesn't belong in the uda1380 driver anyway, since it is iriver specific
12:12:10 Part J-23
12:12:25LinusNhowever, it *could* be moved to the target tree
12:12:54austriancoderLinusN: and how would i do this?
12:13:43sqgl_LakeMacq_LinusN, i have just pulled apart my Archos v1 Recorder to check power supply. What should i be looking for? (NB i cannot charge batteries via my power supply... maybe this problem is related).
12:15:03LinusNaustriancoder: by moving audiohw_reset() to firmware/target/coldfire/iriver/audio-iriver.c
12:15:11 Join Phill2 [0] (
12:15:30LinusNsqgl_LakeMacq_: could be related, yes
12:16:08LinusNthere is an 8-pin chip located near the Down and Right buttons on the PCB, right?
12:16:15austriancoderLinusN: I see.. the only question which is left. Whats better/shall i do now 1) move it into bootloader 2) move it into target tree
12:16:39 Join Nico_P [0] (n=nicolas@rockbox/developer/NicoP)
12:16:48 Quit Rick (Read error: 104 (Connection reset by peer))
12:16:49 Quit jhMikeS (Nick collision from services.)
12:16:55 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
12:17:00 Join Rick [0] (
12:17:01LinusNaustriancoder: i think the audiohw_reset() function should be moved to the target tree for all platforms
12:17:05sqgl_LakeMacq_LinusN, below the down button there is, yes.
12:17:25austriancoderLinusN: Okay... will do it this way
12:17:25LinusNsqgl_LakeMacq_: can you see the markings on the chip?
12:18:22LinusNsqgl_LakeMacq_: should be "MC34063A" iirc
12:18:48sqgl_LakeMacq_34063A & 236A written on that chip
12:19:17LinusNsqgl_LakeMacq_: that is the chip responsible for the battery charging
12:19:34 Join Thundercloud [0] (
12:20:02LinusNif you replace that chip, you should be able to charge again, and maybe your shutdown problem goes away too
12:20:43sqgl_LakeMacq_Thanks heap LinusN! Mind you i will have to find someone who is confident with fiddly soldering. I am useless.
12:21:10LinusN"do not try this at home" :-)
12:22:01LinusNsqgl_LakeMacq_: while you're at it, this page might come in handy:
12:22:41sqgl_LakeMacq_How does one apply solder gun to each pin simultaeneously? Or does on use somethin to soak up the solder of each pin? I heard such a material/fabric exists. what is it's name?
12:23:11sqgl_LakeMacq_Oh, a Rockbox wiki! I guess i have not gone to that site for a while.
12:23:43 Quit J3TC- (Read error: 110 (Connection timed out))
12:23:45LinusNthere exists a special alloy that lowers the melting temperature for desoldering, expensive as hell
12:24:43LinusNcalled ChipQuik,
12:25:01LinusNbut you won't need it for this chip
12:25:18LinusNjust remove the solder with a braid
12:26:07sqgl_LakeMacq_"braid"? Sorry i only have a major in electronics... nothing actually *practical*
12:28:18sqgl_LakeMacq_Hey are those metallic "stickers" in battery housing actually conducting?
12:28:40LinusNthey are connected to ground
12:28:49LinusNah, the stickers are not
12:29:01 Quit GodEater (Read error: 110 (Connection timed out))
12:29:14sqgl_LakeMacq_Well the stickers sure are a PIA then
12:29:32sqgl_LakeMacq_(especially when one takes batteries in/out often)
12:30:12LinusNwell, hopefully you won't need to after the repair
12:31:34sqgl_LakeMacq_I was gonna open up my laptop today to resolder loose DC power pin. But i think i better buy this braid before i bother.
12:32:49sqgl_LakeMacq_LinusN, and if i botch the chip replacement, then i have not gone backwards? Just the same as before?
12:34:00 Join moos [0] (
12:35:50LinusNi hope so :-)
12:38:09sqgl_LakeMacq_i know your IP LinusN !! ;)
12:38:10 Quit Phill (Read error: 110 (Connection timed out))
12:38:22LinusNhaha :-)
12:39:07peturnope - LinusN has his cloak on ;)
12:39:18markunaustriancoder: red
12:46:36austriancodermarkun: i know.. i am working on a fix
12:46:56markunaustriancoder: thanks for working on unifying the audio hw stuff again
12:47:20markunwhat do you guys think about using dB's for the balance setting?
12:47:57markunthe linear scale doesn't sound right to me
12:52:03austriancoderLinusN: I think its better to include the fixes for the humming sound directly in the bootloader, else we need to compile audio-river.c for the bootloader and has a negative effect on its final bin size
12:52:06austriancoder#ifndef BOOTLOADER
13:00:24 Join GodEater [0] (
13:01:55 Join n1s [0] (
13:03:12austriancodermay drivers gets also compiled for bootloader... e.g. drivers/tuner/lv24020lp.c - i think this is not nessesary in some cases.
13:05:54 Join barrywardell [0] (
13:08:07*preglow hates vbscript :/
13:08:12amiconnaustriancoder: The linker picks what it needs
13:08:29*amiconn is better at vbscript than at perl
13:09:06amiconnLinusN: sort of
13:09:19austriancoderamiconn: ah okay...
13:09:39preglowamiconn: does the voicebox i uploaded to the wiki yesterday work?
13:09:46preglowamiconn: some guy on the ml says it doesn't
13:10:29***Saving seen data "./dancer.seen"
13:12:01preglowamiconn: btw, i talked to brian and he said we can release under gpl
13:12:21*barrywardell found the problem with his build server...missing zip
13:12:52amiconnpreglow: Didn't even notice that you uploaded a new version...
13:13:29barrywardellBagder: you can re-enable my server whenever you get a chance
13:15:14 Join jac0b-work [0] (n=jac0b-wo@
13:15:19 Quit jac0b-work (Client Quit)
13:17:29*preglow should get rbutilqt going
13:17:30linuxstbAnyone object to me rejecting yesterday's USB patch due to the (C) issue?
13:18:00preglowsure, just make it clear we're ok if he reimplements it
13:19:18linuxstbAre we? I thought Zagor's stack was considered the way to go?
13:20:39*preglow is tired of the ml now...
13:20:55preglowlinuxstb: well, then you should just plain and simple tell him to not submit patches on it
13:21:03preglowif that's out position, i think zagor should commit his code
13:21:18preglowit's silly to keep people from helping if they want to
13:21:19barrywardelllinuxstb: it sounds like he agrees that zagor's version is the way to go
13:21:34barrywardellbut maybe the arcotg_dcd.c bit of the patch is ok to use?
13:25:11preglowZagor: oy, it seems there are people willing to help with usb around, maybe it's time to commit what you have?
13:25:28preglowworking on a patch isn't as fun as it could be
13:27:30markunaustriancoder: thanks, are you planning to be around in future when you are committing stuff?
13:27:59austriancodermarkun: yes
13:28:47markunpreglow: or let's switch to git, then they can all pull the stuff from Zagor :)
13:30:28*preglow never tried git :V
13:31:47markunpreglow: I started using it 2 weeks ago and it's quite nice
13:32:01preglowso they say
13:32:15markunbut probably has it's own problems
13:33:17 Join J3TC- [0] (
13:34:14 Quit Phill2 ()
13:34:36 Join Phill [0] (
13:36:29 Join jpt9 [0] (
13:36:32preglowit does seem nice at handling patches and merging, etdc
13:37:00preglowtrying it out is on my list of stuff to do, but that is quite a list these days
13:45:44 Quit Rob222241 (Read error: 104 (Connection reset by peer))
13:47:29 Join Rob2222 [0] (
13:57:09 Quit Toki_ (Read error: 110 (Connection timed out))
14:04:20 Join Lynx_ [0] (
14:07:47 Quit jpt9 (Read error: 110 (Connection timed out))
14:12:31 Quit parafin ("reboot")
14:17:55 Join Siku [0] (
14:19:56 Quit markun (Read error: 104 (Connection reset by peer))
14:21:06 Join weezerle [0] (
14:21:25 Join Arathis [0] (
14:22:05 Join linuxstb_ [0] (n=linuxstb@rockbox/developer/linuxstb)
14:23:09 Join linuxstb__ [0] (
14:23:31Arathisis there a patch or a planed commit to build the db not from the root dir, but from a specified like /Music on H10s?
14:23:59preglowseriously, where the hell does the sim get its battery readings? :>
14:23:59 Quit linuxstb (Nick collision from services.)
14:24:03 Quit linuxstb_ (Nick collision from services.)
14:24:06 Nick linuxstb__ is now known as linuxstb (
14:25:23linuxstbArathis: No, but someone said they were working on a patch to support ".no_index" files - which would tell the database not to index the folder that file was in (and all sub-folders).
14:25:57 Join markun [0] (n=markun@rockbox/developer/markun)
14:26:23Arathislinuxstb: do you remember who announced that?
14:28:00PaulJamArathis: you can already do something similar by only displaying results that are in a specific oilder in the database search results with a tagnavi_custom.config.
14:28:34PaulJamfolder, not oilder
14:28:40linuxstbArathis: No, that's why I said "someone"... ;)
14:29:00ArathisPaulJam: ah, the custom view .. allways wanted to get some information about it, but never found something
14:29:09Arathislinuxstb: right ^^
14:29:21PaulJamArathis: see the DataBase wiki page
14:29:24 Join parafin [0] (
14:29:52 Nick parafin is now known as parafin| (
14:29:53 Nick parafin| is now known as parafin (
14:30:23Arathisthat's something I could have done before xD I just searched for custom view and similar in the wiki and manual
14:30:27Arathisthanks anyway
14:35:56JdGordonArathis: I tinhk i did a patch simialr to that
14:36:02JdGordonthe .no_index one
14:36:46linuxstbCan any Sansa owners confirm that this looks like a normal E280? The forum posted seems to think it may be a "V2" e280 - (warning: big image...)
14:37:13 Quit nicktastic (Read error: 110 (Connection timed out))
14:37:37tuplanollalinuxstb: Looks same as my e280
14:37:37linuxstbJdGordon: Any reason it wasn't committed? I think I right in saying that most people would like some way to control what files the database indexes.
14:37:39ArathisJdGordon: what do you mean by "similar"?
14:38:23TTThomaslooks the same as my e250r as well
14:38:23JdGordonlinuxstb: I cant remember, I didnt think everyone liked it though?
14:38:33JdGordonArathis: by similar i tihnk i used a differnt filename
14:38:44tuplanollablergh. No codec for: /MUSIC/blablabla/blablabla/The Quiet Place.mp3
14:38:55ArathisJdGordon: :D
14:38:59JdGordontuplanolla: in flames?
14:39:13tuplanollaJdGordon: yeh
14:39:19JdGordonnice :D
14:39:25linuxstbtuplanolla: Do you have a file called /.rockbox/codecs/mpa.codec on your device?
14:39:43 Quit TheCollector (Remote closed the connection)
14:40:03tuplanollalet's see.
14:40:20tuplanollabut it worked earlier today
14:40:43tuplanollaso i'm pretty sure i have it
14:40:48 Join Febs [0] (n=chatzill@
14:41:05linuxstbtuplanolla: It could also be a corrupt filesystem, meaning Rockbox can't read that file.
14:41:10tuplanollayep, there is mpa.codec
14:41:12linuxstb(the mpa.codec file)
14:42:01Arathisbefore I forget to ask again: where's the recording gain saved? I can't find it in any config file and after updating rockbox (including deleting the old dirs) it's reset to a default of -34db or something which is not enough for fm recording on my H10.
14:42:35tuplanollaHow I can make sure it isn't corrupted or how to fix?
14:42:43linuxstbJdGordon: Do you still have that patch around?
14:42:54JdGordonwas just about to say i foud the patch :p
14:43:00linuxstbtuplanolla: Run chkdsk or similar on the device. Or maybe just try to reinstall Rockbox.
14:44:18tuplanollaoh, thanks, there really is errors :O
14:45:51crwli had similar things happen, every once in a while rockbox wouldn't read my settings on boot, it took me a while to notice that there was something wrong with the FAT exactly where ~/.rockbox/config.cfg was
14:46:15tuplanollaHm. Still no codec found
14:46:27tuplanollaTrying rockbox reinstall then
14:46:42preglowjmspeex: there are quite few spots in libspeex where you have loops for copying and clearing, any reason you don't use memcpy and memset? they should pretty much always be afster
14:47:11linuxstbJdGordon: Looks OK to me, but I would prefer something less intrusive than "ROCKBOX_DATABASE_IGNORE_FOLDER" - e.g. ".no_index" ;)
14:47:22jmspeexpreglow: Yes, bad practices. That should be changed.
14:47:29 Join jpt9 [0] (
14:47:37jmspeexpreglow: can you send a list (or a patch)?
14:48:09JdGordonlinuxstb: I was going for the more "painfully obvious" approach :)
14:48:19JdGordonim not a fan on . files
14:48:37JdGordonpartly because you cant create them easily in windows
14:48:50linuxstbI'm just reading the IRC logs from the time of that patch - and that was discussed then...
14:49:41JdGordoncan you link me to the timestamp?
14:49:49*JdGordon doesnt remember the discussion
14:49:59linuxstbNo, I'm viewing my local copy - it's 11th March
14:50:14linuxstbsearch for ROCBOX_DATABASE_IGNORE_FOLDER
14:50:33preglowjmspeex: i'll give it a go
14:52:09JdGordonlol @ self
14:52:26preglowjmspeex: btw, looks to me like both nb and wb data are delayed by the same amount of samples when decoded with the same wb decoder, so we should be fine if we just use the lookahead value we get from the decoder when snipping away samples at the start of the decoded clip
14:53:08 Join joshin [0] (n=josh@unaffiliated/joshin)
14:53:27jmspeexpreglow: There's delay both on the encode side and the decode side.
14:53:57ZagorI have nothing against committing my code, other than a general feeling that it's not nice to commit code that isn't working yet. But if the general consensus is that I should commit, I'll do that.
14:54:04jmspeexat the very least, you get 5 ms for each, either in nb or in wb
14:54:25preglowZagor: just thinking it's easier for people to help out if it's in svn
14:54:43JdGordonlinuxstb: what about rockbox.ignore maybe?
14:54:46markunZagor: I also think it's better to commit (if it doesn't break anything else)
14:54:50jmspeexpreglow: then, you have the QMF delay, which you only get if the encoder or decoder is wb. That delay is 31.5 sample for the encoder and the same for the decoder.
14:54:58 Join stewball [0] (n=WTFOMGBB@
14:55:28linuxstbZagor: The fact that your patch is -4316 lines and +1645 lines sells it to me...
14:55:31 Quit jpt9 ("ChatZilla 0.9.79 [Firefox]")
14:55:31jmspeexpreglow: all these values are for the current implementation and they're subject to change if I improve things
14:56:08JdGordonlinuxstb: LOC is not a good metric!
14:56:19linuxstbI'm shallow...
14:56:21preglowwell, it's less code to read...
14:56:30 Join zicho [0] (
14:56:33JdGordonlinuxstb: straight to managment for you!
14:56:36preglowso it's not an irrelevant metric either
14:56:58barrywardellZagor: did you see the comment in FS #8189 about the index (mask)?
14:57:51barrywardellwith Zagor's patch in svn, we could have proper usb charging on Sansa almost straight away.
14:58:04Zagorbarrywardell: yes, but I don't think he's right. if the mask was wrong, nothing would work.
14:58:20jmspeexpreglow: I've got a patch that does s/\n/ / to the source code. Are you interested?
14:59:03 Join nicktastic [0] (n=nick@unaffiliated/nicktastic)
14:59:10preglowjmspeex: a patch that replaces newlines with spaces???
14:59:16austriancoderZagor: he must we right, else we would have no working msc
14:59:25linuxstbJdGordon: Maybe db.ignore or database.ignore? Rockbox doesn't ignore it, just the database.
14:59:34jmspeexpreglow: Yes, that drastically reduces the LOC
14:59:36Zagoraustriancoder: ?
14:59:44barrywardellZagor: yeah, I was thinking the same thing. Although he did get msc working...
14:59:58preglowjmspeex: sure, in it goes, should improve readability by leaps and bounds
15:00:10JdGordonyeah, and I'm thinking it might be good to add .ignore the the known filetypes
15:00:15linuxstbI didn't think that patch to AC's code worked - it panics?
15:00:20Zagorbarrywardell: is it working for you? I couldn't get it working any more than my code.
15:00:48barrywardell"even one time the whole iRiver disk is accessible via Windows"
15:01:12barrywardellnot reliable, but he did get it working once
15:01:55Zagorbarrywardell: that text is rather contradictory. how is it working if "Remain to do: implement read capacity and write sector in usb_storage.c"
15:02:30Zagorwhen I tried, it failed on the first sector read. just like my code.
15:02:39 Quit male (No route to host)
15:03:48barrywardellTrue. I haven't actually tried the patch, so I'm just going by what he says
15:05:54tuplanollaOk, thanks, now it finds the codec.
15:10:33***Saving seen data "./dancer.seen"
15:13:56 Join DM| [0] (
15:14:14Arathishm .. I set up a tagnavi_custon.config and now rockbox won't boot anymore. I get the rockbox boot screen, but it won't go forther on (H10/20gb)
15:16:37Arathisah, got it
15:16:50 Quit lee-qid (Connection timed out)
15:17:18 Join Llorean [0] (
15:20:10JdGordonlinuxstb: im going to bed.. if you want to commit it, i say go for it :)
15:20:26 Quit JdGordon ("Konversation terminated!")
15:20:46pondlifeHmm, was the tlv320 reset not needed?
15:22:08*Arathis would appreciate the commit of JdGordons patch
15:22:52preglowjmspeex: erm, nb_celp.c, line 698, won't that for loop write outside the bounds of bw_lpc2 by one coef?
15:24:43jmspeexpreglow: you mean this: curve_to_lpc(st->psy, curr_curve, bw_lpc1, bw_lpc2, 10); ? (line 699)
15:26:13pondlifeNico_P: Did you see that nice recipe on FS #8194... ?
15:26:41Nico_Ppondlife: yes... haven't tried it though... have you?
15:26:50pondlifeAbout to
15:27:06pondlifeJust building a sim
15:27:49preglowjmspeex: sorry, i was looking at an old rev, i mean line 709
15:32:55jmspeexpreglow: The answer is yes and no. There *would* be an overflow if st->gamma2<0, but in practice, there's no mode in which that can happen (gamma is always >= 0)
15:33:00 Part sqgl_LakeMacq_ ("Leaving")
15:34:08preglowjmspeex: but it is a "bug", right? there shouldn't be <= there?
15:34:49jmspeexIt's a bug. Basically, I used to include the A(0)=1 coef in the filter and then decided to remove it. Forgot to remove it there.
15:36:28pondlifehaha, repeat one results in buffering reading the same file many times...
15:36:36 Quit animeloe_ ("Leaving")
15:38:01 Join animeloe [0] (n=animeloe@unaffiliated/animeloe)
15:38:12 Join Toki [0] (
15:40:14 Join male [0] (
15:40:17pondlifeNico_P: Tried #8194 with MP3 on the sim, but no problem..
15:40:37pondlifeMind you, it's more likely to happen with a long FLAC by the sound of it.
15:40:55pondlifei.e. when track is bigger than buffer
15:41:17linuxstbTry an iFP sim...
15:41:17pondlifeWould that be another symptom of the rebuffer_handle() issue?
15:41:42pondlifeI'd like a command-line −−wastebuf option for the sim...
15:41:58linuxstbLoad a 20MB voice file...
15:42:18linuxstbBut I doubt you can make speex take up that much room...
15:43:34pondlifeChanging from Repeat One to Repeat None does horrible things though
15:43:50pondlifeThe track advances ok (audio), but the WPS remains stuck
15:44:25 Quit DM| ("*bashes head against keyboard*")
15:46:32 Join Rob222241 [0] (
15:46:33 Quit Rob2222 (Read error: 104 (Connection reset by peer))
15:47:27preglowjmspeex: should i add a speex_memcpy too, or? i don't really get the point in those wrappers anyway, shouldn't the mem* guys be pretty fast on all platforms anyway?
15:48:27pondlifeWooh, nice way to lock up the sim...
15:51:39Nico_Ppondlife: how?
15:52:04pondlifeStarting playback with Repeat One and changing it to Repeat None.
15:52:23pondlifeLet it play to the track transition... The WPS sticks on track 1.
15:52:27pondlifeI let it play for 3 tracks
15:52:34pondlifeThen found the UI was locked
15:53:02Nico_Ppondlife: could you try with a rev prior to bertrik's patch I committed?
15:53:24pondlifeLater, but work calls now..:/
15:53:35pondlifeIt's on Flyspray..
15:54:03 Quit J3TC- (Read error: 110 (Connection timed out))
15:54:11pondlifeI'm letting it play on.. see what happens when it rebuffers!
15:54:12*Nico_P has an exam to prepare :(
15:54:18pondlifeGood luck
15:54:29Nico_Pthanks :)
15:58:28 Join J3TC- [0] (
16:01:37 Quit Zagor ("Client exiting")
16:05:09 Join animeloe_ [0] (n=animeloe@unaffiliated/animeloe)
16:05:11 Join scorche|w [0] (n=42c007b2@rockbox/administrator/scorche)
16:05:48 Quit animeloe (Nick collision from services.)
16:05:52 Nick animeloe_ is now known as animeloe (n=animeloe@unaffiliated/animeloe)
16:10:40 Quit billenium (Connection reset by peer)
16:10:50 Join billenium [0] (
16:11:08 Quit billenium (Connection reset by peer)
16:14:02 Join H10_007quick [0] (
16:15:27H10_007quickbarrywardell: I see that there was a change submited that changed the way the rtc was read on the H10, Does this make the ADC read better, do we not have the problem of jumping readings?
16:17:07barrywardellthe way the rtc was read was not changed
16:17:16barrywardellan alarm feature was added
16:17:29 Join corsair [0] (n=user@
16:17:35barrywardellthere was also a change to reading the ADC which makes the jumpy readings much less
16:17:44 Quit Febs (Read error: 104 (Connection reset by peer))
16:17:48H10_007quickI see
16:17:57 Join DM| [0] (
16:18:06H10_007quickOk because the scroll pad seemed much better after that commit
16:18:24DM|can rockbox read ipod playlists created from gtkpod or on the ipod apple firmware
16:18:27 Join roxfan[zzz] [0] (
16:18:47corsairIt doesn't seems can...
16:18:48GodEaterDM|: no
16:18:49linuxstbDM|: No, Rockbox uses standard m3u playlists, not playlists stored in the ipod's database.
16:18:58H10_007quickhave you put any work into getting full scroll functionality? Just don't want to waste my time if you have already started
16:19:18barrywardellI've thought about it, but haven't started anything. there are a couple of attempts in the tracker though
16:19:42H10_007quickyes one of them is mine
16:19:57H10_007quickI might start working on it again thanks for your time
16:20:38linuxstbbarrywardell: Would you be able to build some e200rpatcher binaries? My changes to display more useful messages aren't in the current release...
16:20:56 Join maddlah [0] (
16:21:01 Quit maddler (Read error: 104 (Connection reset by peer))
16:21:05barrywardelllinuxstb: yeah, sure. I have to go now for a bit but can do it later today
16:21:33linuxstbbarrywardell: Thanks. I don't think the Swedes are around now anyway...
16:21:44LinusNthey aren't?
16:21:53linuxstbYou were marked as away ;)
16:22:00LinusNoops :-)
16:22:07linuxstbIn fact you _are_ marked as away...
16:22:13barrywardelllinuxstb: for mac? linux? windows?
16:22:36linuxstbI could do Windows if you don't have the compiler setup.
16:22:59barrywardellI think I do. I'll let you know if I can't do it
16:23:41 Part corsair ("ERC Version 5.2 (IRC client for Emacs)")
16:23:57 Quit H10_007quick ("ChatZilla 0.9.79 [Firefox]")
16:24:17 Quit CaptainSquid ("Miranda IM!")
16:25:06 Quit Phill (Remote closed the connection)
16:25:25 Join Phill [0] (
16:27:55 Quit roxfan (Read error: 110 (Connection timed out))
16:28:41 Join stripwax [0] (
16:28:58 Quit stripwax (Client Quit)
16:30:46 Join billenium [0] (
16:31:02 Join mf0102 [0] (n=michi@
16:31:36 Join daurn [0] (n=daurnima@unaffiliated/daurnimator)
16:35:51 Quit animeloe ("Leaving")
16:41:32 Join animeloe [0] (n=animeloe@unaffiliated/animeloe)
16:42:59 Join japc [0] (n=japc@
16:50:22 Quit Rob222241 (Read error: 104 (Connection reset by peer))
16:51:25 Join Rob2222 [0] (
16:52:26 Quit daurnimator (Read error: 110 (Connection timed out))
16:57:55 Quit animeloe (No route to host)
16:58:16 Join animeloe [0] (n=animeloe@unaffiliated/animeloe)
17:00:12*GodEater prods at Bagder
17:01:17 Join xufumon [0] (n=opera@
17:03:22 Quit DM| ("*bashes head against keyboard*")
17:07:03xufumonQuestion: Does rockbox support recording with Ipod minis? If so, does it use a 'mic/line in' function for the headphone jack? (I couldn't find information regarding this in the rockbox wiki.)
17:07:49 Quit Rob2222 ()
17:08:07 Join Rob2222 [0] (
17:10:34***Saving seen data "./dancer.seen"
17:21:50 Quit jhMikeS (Read error: 110 (Connection timed out))
17:24:14 Part xufumon
17:24:54 Part LinusN
17:32:18 Quit J3TC- (Read error: 104 (Connection reset by peer))
17:32:48 Join J3TC- [0] (
17:34:13 Join roolku [0] (
17:43:57 Quit Rob2222 (Read error: 104 (Connection reset by peer))
17:44:25 Quit petur ("*plop*")
17:44:34 Join Rob2222 [0] (
17:48:28 Join Rob222241 [0] (
17:48:39 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
17:50:22 Quit OlivierBorowski (Remote closed the connection)
17:56:57 Join OlivierBorowski_ [0] (
17:56:58 Quit OlivierBorowski_ (Remote closed the connection)
17:57:35 Join rotzbouw [0] (i=5840bef7@gateway/web/cgi-irc/
17:57:42rotzbouwhi there
17:58:28rotzbouwsince recently, my ipod video (60gb) takes minutes to mount, on any OS. is this a known problem?
17:58:38rotzbouweven when the ipod has just been reset and is empty.
17:59:01lostlogicrotzbouw: no problem for me on my 30g
17:59:14rotzbouwwhat build do you use?
17:59:20lostlogiclatest svn
17:59:26rotzbouwhm :<
17:59:33lostlogicbut rockbox build has _nothing_ to do with mounting to the OS
17:59:38rotzbouwit's weird, because it really only occurs when rockbox is installed.
17:59:39lostlogicthat's handled by the original firmware bootloaders
18:00:05marazrotzbouw: does it also happen on other computers?
18:00:26marazrotzbouw: does it happen if you boot into the apple firmware and then plug into your computer?
18:00:37rotzbouwor no
18:00:46rotzbouwit doesn't happen if there is no rockbox installed
18:01:03rotzbouwlemme see if it happens with the lock enabled
18:01:11linuxstblostlogic: Rockbox does have a little to do with it - a) It inits the USB controller to distinguish between a charger and a computer at the other end of the cable; b) it reboots into the emergency disk mode, rather than Apple's main OS (which is reportedly slower)
18:01:38marazthis is correct.
18:02:08 Join pixelma [0] (i=pixelma@rockbox/staff/pixelma)
18:02:10lostlogiclinuxstb: sure, I guess I assumed that he was talking about after the USB screen is displayed, may be a wrong assumption.
18:02:11marazthe "rockbox way" of using the USB mode is generally a lot slower than using the "real" apple disk mode.
18:03:17rotzbouwapple firmware is faster
18:03:59lostlogicemergency disk mode is limited to USB1.1
18:04:07rotzbouwbut when rockbox isn't installed at all, it's a lot faster
18:04:31 Join nanok [0] (n=nanok@
18:04:34rotzbouwrockbox is only working in usb 1.1 mode?
18:05:15rotzbouwthat'd be an easy explanation but i can't recall it taking so long when i used it first. might have been the general enthusiasm about the program, though ;>
18:05:22krazykitrockbox doesn't touch emergency mode code.
18:05:49lostlogicrotzbouw: read carefully: Rockbox boots to emergency disk mode, which operates in USB1.1. If you used emergency disk mode w/o rockbox installed, it would be 1.1 as well. The only 2.0 mode currently is from within the apple firmware plugging in USB. You can do this with rockbox installed by rebooting and holding menu to get into the OF and then pluggin in USB.
18:06:24parafini think it's hold, not menu
18:06:30rotzbouwit's hold
18:06:40rotzbouwand yes, it loads faster that way, thanks
18:07:01rotzbouwbut not as fast as if rockbox wasn't installed :<
18:07:15 Quit Rob2222 (Read error: 110 (Connection timed out))
18:07:46parafini don't think this is true
18:07:47linuxstbparafin: It's either...
18:07:49lostlogicrotzbouw: do you have some kind of program that automatically scans the disk? Could it be just the .rockbox folder taking time?
18:08:09rotzbouwwell, I've tried it on mac os x, debian and windows xp
18:08:21parafinbtw, it was cery stupid idea to clear settings if hold is on
18:08:51lostlogicparafin: was very useful in early development :-P
18:09:11 Quit idnar (Nick collision from services.)
18:09:13 Join idnar_ [0] (n=mithrand@unaffiliated/idnar)
18:09:16parafinbecause of that i have to wait while my ipod is booting before placing it in my pocket
18:09:39parafinand it boots not very fast i must say
18:10:32parafinand i cleared my settings a couple of times by mistake
18:10:38parafinthat's annoying
18:11:26crashdipod hold is boot into OF now isnt it?
18:12:02parafinhold in bootloader code boots OF, hold in rockbox booting clears the settings
18:12:06crashdparafin: might be worth updating your bootloader then
18:12:23crashdoh, i see what you mean
18:12:23parafinthat's not bootloader problem but rockbox firmware
18:12:29pixelmafrom what I have seen (5.5G 80GB) Rockbox boots noticably faster than the apple OS if it needs a full reboot (not starting from suspend)
18:12:33crashdrockbox takes like, a second to boot though
18:13:08parafinmay be it's directory scan that takes time, dunno
18:13:27parafinor maybe i should update rockbox
18:13:32pixelmaboot time also depends on your WPS - if it has to seek for a lot bitmaps, especially if they are scattered across the disk due to fragmentation
18:13:34crashddo you have all your album directories in the root?
18:14:14rotzbouwrockbox booting is very fast. rockbox mounting is very slow (due to usb1.1 limitation, as i know now!)
18:14:34rotzbouwanyway, thanks, bye
18:14:36 Quit rotzbouw ("CGI:IRC")
18:15:43parafinOF boots not faster, that's for sure, but that's not the point, i don't use it anyway (only for disk mode)
18:16:30preglowamiconn: did you test the new voicebox?
18:21:55preglowcould anyone at all test the updated voicbox in the wiki?
18:24:27 Join ctaylorr [0] (
18:28:06 Quit linuxstb (Remote closed the connection)
18:28:46 Join linuxstb [0] (n=chatzill@rockbox/developer/linuxstb)
18:29:21 Nick idnar_ is now known as idnar (n=mithrand@unaffiliated/idnar)
18:32:18 Quit Jon-Kha ("Lost terminal")
18:32:35 Join Jon-Kha [0] (
18:39:42 Quit thegeek_ ("( :: NoNameScript 4.1 :: )")
18:43:52 Join BigBambi [0] (n=Alex@rockbox/staff/BigBambi)
18:55:35 Join PaulPosition [0] (
19:00:54 Join Gnu47_ [0] (
19:03:46PaulPositionNot asking for any sort of ETA, but is the 'ui slows to a crawl/songs skip' syndrome of using too many eq bands on the PP50xx targets something that *might* get resolved some day or is it no-can-do?
19:04:39LloreanIt might get solved one day
19:04:45 Quit pondlife ("Read error: 110 (Connection slimed out)")
19:04:48LloreanSongs skipping requires optimization of the codecs.
19:05:15LloreanUI slowing to a crawl will improve a little bit with that, but is more likely to see improvement if someone comes up with ways to take advantage of dual core
19:06:10 Join bertrik [0] (
19:06:46PaulPositionLlorean - That's about what I understood.. But is giving the 2nd core some love something that has any importance given that it's only a quarter, maybe a third of the targets that are dual core
19:07:03amiconnlostlogic: iPod EDM is *not* usb1.1, not even limited to usb1.1 speed
19:07:25amiconnI measured transfer rate to be 2..3MByte/s
19:07:57n1sPaulPosition: coldfire targets have way more headroom for dsp and can use 5 bands eq with no problems with most codecs/bitrates
19:08:05 Quit Gnu47 (Read error: 110 (Connection timed out))
19:08:28*PaulPosition wonders if going back to highschool (maths) then college for some embeded programming lessons make sense at 33y/old :p
19:08:34preglowamiconn: any idea why some people on the ml get "unable to locate encoding exe" with voicebox?
19:08:36amiconn(on 5.5th Gen that is)
19:08:37PaulPositionAren't coldfire single-core?
19:08:49n1sPaulPosition: yes it is
19:08:59amiconnpreglow: No. Doesn't happen with my version, and didn't try the newer one
19:08:59n1s(the ones we use at least)
19:09:08preglowamiconn: the newer one has fixes for dragndrop
19:09:35LloreanPaulPosition: Considering several combinations of codec+EQ+player don't cause slowdown or skipping, it's not exactly a high priority for most people, I imagine
19:10:26LloreanAmong portalplayer targets
19:10:30amiconnThat said, iPod EDM *is* slow, but not due to transfer rate (which is quite low, but e.g. updating Rockbox on my Archos Player is significantly faster than on 5.5th Gen, even though the Player actually *is* USB1.1 only)
19:10:37***Saving seen data "./dancer.seen"
19:10:42preglowcould anyone at all on windows please test voicebox?
19:10:44amiconnThe H10 suffers from the same problem, btw
19:11:31PaulPositionLlorean - Thanks, that's about what I thought.
19:11:51linuxstbPaulPosition: What kind of files are you using that give those problems?
19:12:02 Quit PaulPosition ()
19:12:18amiconnlinuxstb: Hmm, that reminds me - did you look further into the range decoder for ape?
19:12:37linuxstbNot yet... :(
19:16:02 Join markun_ [0] (n=markun@rockbox/developer/markun)
19:17:31 Quit markun_ (Read error: 104 (Connection reset by peer))
19:17:49 Join Bagder_ [0] (
19:18:45 Quit markun (Read error: 104 (Connection reset by peer))
19:19:56 Join markun [0] (n=markun@rockbox/developer/markun)
19:20:58 Join markun_ [0] (
19:22:31lostlogicamiconn: strange, I wonder what makes it slower than the 'full' disk mode
19:23:45 Join Shaid` [0] (
19:23:48amiconnIt *is* much slower. Bu it's still faster than USB1.1 transfer rate wise, although it *behaves* slower than that when transferring many files
19:24:54lostlogicamiconn: Yeah, I hear ya −− so it's something else in the EDM USB stack compared to the OF stack
19:25:02 Join OlivierBorowski [0] (
19:25:12lostlogicor maybe it uses a fully sync. disk mode versus some kind of caching in the OF moded
19:25:23 Join Arathis2 [0] (
19:25:32amiconnpreglow: "Unable to locate encoding executable: \n\n ??"
19:25:42amiconnI wonder what executable '??' should be...
19:25:54 Join Fraser [0] (
19:25:59 Quit J3TC- (Read error: 110 (Connection timed out))
19:26:29 Join J3TC- [0] (
19:26:43*amiconn diffs
19:27:27preglowamiconn: i have no idea... i just assumed brian had tested this himself
19:27:39preglowamiconn: not much changed apart from the dragndrop stuff
19:28:34 Nick Gnu47_ is now known as Gnu47 (
19:29:43amiconnYeah, and that change breaks everything
19:30:10amiconnJust moving that block won't fix the actual problem, but instead extend it from hampering drag&drop mode to hampering all modes
19:31:11amiconnThis is because voiceUtils.vbs uses ExecuteGlobal to read the parameters
19:31:28amiconnThis cannot work for non-bool parameters, like the Encoder selection
19:32:11*amiconn sighs
19:32:20preglowi guess we should just revert back, then
19:32:35amiconnIt needs to read the .ini using the same function as the .hta
19:32:45*preglow doesn't get the extensions...
19:33:24amiconnAnd since voiceUtils.vbs is meant to be a resource for both the .wsf and the .hta, it should probably be moved there somehow
19:33:36amiconnhta == html application
19:33:52 Quit billenium (Read error: 110 (Connection timed out))
19:34:47amiconnwindows script file
19:35:13amiconnWell, in fact ExecuteGlobal *could* be made working, if the encoder would be always quoted
19:35:16 Quit Bagder (Read error: 110 (Connection timed out))
19:35:31amiconnBut executing the contents of an .ini file is dirty, imo
19:36:40 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
19:37:09 Join Aurum_ [0] (
19:37:16amiconnwsf is not limited to vbscript, it can contain vbscript, javascript (or JScript as microsoft calls it), and other 'active scripting' languages like activestate perl. All that even combined
19:37:55preglowwell, ok, but should i just revert the file in the wiki for now?
19:37:55nanokanybody knows, by any chance, the number of zagor's patch for usb, or a keyword?
19:38:10n1sUSB maybe...
19:38:51nanokn1s: a bit wide :)
19:39:30amiconnpreglow: Btw, the '??' results from evaluating (in ExecuteGlobal) the statement "Encoder = speex"
19:39:30n1snanok: searching for just usb among patches yields only 17 results, one of them the one you were looking for
19:39:52amiconnSince 'speex' is neither a reserved word nor quotes, it is treated as a variable name
19:40:07nanokaham, got it
19:40:08amiconnAnd since a variable 'speex' doesn't exist, its content is undefined
19:40:25nanoki discovered the advanced search :)
19:40:31nanokn1s: thank you
19:40:46amiconnIf the script would use "Option Explicit" (as the script I write almost always do), it would bail out...
19:41:14preglowi'd always use that, yes
19:41:22preglowat least i always do "use strict" in perl :>
19:42:23 Quit Shaid (Read error: 101 (Network is unreachable))
19:44:26 Join DerPapst [0] (
19:44:31*DerPapst waves
19:46:42 Quit Arathis (Read error: 110 (Connection timed out))
19:48:12 Join DM| [0] (
19:50:01 Quit Shaid` (Read error: 101 (Network is unreachable))
19:50:30 Nick Arathis2 is now known as Arathis (
19:51:35amiconnOkay, moving some functions over does work... might turn out easier than I thought
19:54:31 Quit hcs (Remote closed the connection)
19:54:47 Join hcs [0] (n=agashlin@rockbox/contributor/hcs)
19:55:38ctaylorrhrm...Wonder if there's something wrong with my MP3 player? I have a Sansa e280 and with every crossfade I get some skipping. I'm very used to my retired x5.
19:55:39barrywardelllinuxstb: what do I do about libusb when building e200rpatcher.exe?
19:57:28 Join MethoS- [0] (
20:00:20 Quit Mouser_X (Read error: 110 (Connection timed out))
20:01:01bertrikctaylorr: I'd like to help reproducing that problem, but can't remember how to enable crossfade
20:01:39ctaylorrbertrik: It's in the menu under settings, general settings, playback, crossfade, enable crossfade.
20:01:50ctaylorrbertrik: (I wonder how one would forget ;) )
20:02:17barrywardelllinuxstb: here are my compiled versions for linux and mac:">
20:02:38bertrikI looked in the context menu and the sound settings menu
20:03:29ctaylorrbertrik: (Of course, I'd rather have too many options than no complaints)
20:03:41BigBambictaylorr: do you have any other audio processing things running, such as equaliser etc.?
20:04:37ctaylorrBigBambi: Yes. Is there some optimizing to do, or is it a limitation of the hardware? Either way's long as it's not a hardware problem.
20:04:47BigBambiIt isn't a problem
20:05:00ctaylorrBigBambi: I have the equilizer enabled, but the wps is pretty simple (plain text).
20:05:05ctaylorrBigBambi: No peak meters.
20:05:08BigBambiOther than the portalplayer targets just aren't fast enough
20:05:17BigBambictaylorr: The full five bnd eq?
20:05:21 Quit PaulJam (Read error: 113 (No route to host))
20:05:24ctaylorrBigBambi: Yes.
20:05:29 Join Lear [0] (i=chatzill@rockbox/developer/lear)
20:05:30BigBambictaylorr: That'll be it
20:05:41BigBambiIf you turn a couple of bands off, does it help?
20:06:07ctaylorrBigBambi: Ah. I just tried to replicate the x5's configuration, assuming this (physically smaller :) ) device would be more powerful.
20:06:13BigBambictaylorr: Optimising would certainly help, at the moment the CPU is just too slow to run everything you are asking of it
20:06:14 Quit Siku (Read error: 131 (Connection reset by peer))
20:06:23BigBambictaylorr: Different architecture for a start
20:06:35 Join Siku [0] (
20:06:40ctaylorrBigBambi: np's. Thanks for that.'s more powerful in different ways.
20:06:42BigBambiThe X5 has a 124 MHz coldfire, the sansa has a dual 80 MHz arm
20:06:58BigBambiBut I don't think we use both cores optimally yet
20:07:57ctaylorrBigBambi: That definitely explains it.
20:11:04 Join Buschel [0] (
20:11:07bertrikouch, data abort while trying to reproduce
20:11:40ctaylorrbertrik: a few of those yesterday.
20:12:19ctaylorrbertrik: But I chalk that up to blindly using unstable (read SVN) builds.
20:12:30*n1s wishes people would stop assuming long is 32 bits... sigh
20:12:48bertrikctaylorr: there is no such thing as stable :P
20:12:50 Join linuxstb_ [0] (
20:12:57ctaylorrbertrik: Windows.
20:13:06 Join hc1 [0] (
20:13:23ctaylorrbertrik: Also, `cat' is pretty stable.
20:13:25XavierGrI see that in some lang files some strings are voices while the same strings are unvoiced in the english lang file
20:13:33 Quit japc (Read error: 110 (Connection timed out))
20:13:36bertrikn1s: how long is long on rockbox targets?
20:13:47BigBambiWindows, stable? Vista has blue screened on my three times today
20:13:47XavierGrShould I follow the english lang file strictly?
20:13:51 Quit linuxstb (Nick collision from services.)
20:13:51BigBambiAny way, Off topic
20:13:55 Nick linuxstb_ is now known as linuxstb (
20:14:06 Quit hcs (Nick collision from services.)
20:14:11 Nick hc1 is now known as hcs (
20:14:12 Part markun_
20:14:32ctaylorr(I joke. Windows is as stable as the stable in the Godfather)
20:14:33n1sbertrik: on our current targets it's 32 but in my sim it's 64, and do you think there are still a big slew of bugs because of that? ;)
20:15:20*preglow wonders why the dithering bugs out now
20:16:03hcs[x] Freak Out
20:16:04LearXavierGr: Probably. That indicates there's no code to make use of any voice string.
20:16:15bertrikwith embedded software I usually use explicit types
20:16:52 Join MethoS-- [0] (
20:17:14n1sbertrik: there was a port (which is now dead) which had 16 bit ints and 32 bit longs, that would be even more broken now because a lot of code assume int == long too...
20:17:22LearHm, should the Sansa bootloader really shut down on hold when inserting the USB cable? Anyone that remebers how the OF handled the same situation?
20:17:44amiconnbertrik: Using explicit types also has its drawback, especially in the embedded sector
20:18:16amiconnIf an int16 would be sufficient for calculations but an int32 won't hurt, it's better to use plain 'int' for performance
20:18:24bertrikyeah, everybody uses a different name for their types, or worse the same name but different bit width
20:19:11LloreanLear: The OF bootloader does the same
20:19:13hcsamiconn: for code size and alignment?
20:19:13 Quit Phill (Remote closed the connection)
20:19:15amiconnUsing int16 explicitly would hamper performance on (some) 32 bit archs like arm, and uing int32 explicitly would hamper performance on 16 bit archs
20:19:23 Quit Siku (Read error: 145 (Connection timed out))
20:19:27LloreanLear: It would pop up a message on the screen saying something like "Keys Locked, Shutting Down"
20:19:34 Join Phill [0] (
20:19:38amiconnSame goes for 'long' and 32 bit vs. 64 bit archs
20:19:57amiconnhcs: I was talking about calculation, not alignment
20:20:15 Join Phill2 [0] (
20:20:51*Lear has barely used the Sansa OF...
20:20:56XavierGrLear: What I don't get is that while in the italian lang there is a voice string for the particular lang id there isn't one for the english lang
20:21:13amiconnpreglow: Seems I have both GUI mode and drag&drop working now
20:21:21linuxstbbarrywardell: Ah, I thought you had built e200rpatcher before, but obviously not... I guess I'm confusing it with sansapatcher.
20:21:23XavierGrI would expect that one of them other is wrong
20:21:25bertrikamiconn: yes I use plain int for simple variables
20:21:28amiconnI also added 'Option Explicit'
20:21:31preglowamiconn: well, that sounds sweet, just upload to the wiki when it's ready
20:21:41amiconnJust testing a bit more
20:21:47bertrikand in some cases you can't avoid depending a little on the native width
20:21:47preglowamiconn: i'm probably soon going to have a new rbspeexenc going, but i can handle that myself
20:22:07*amiconn ponders porting the sample rate 'intelligence' from sapi_voice.vbs
20:22:36bertrikis it possible to diagnose something about a data abort?
20:22:43bertrikperhaps with the map file
20:22:44linuxstbbarrywardell: I built a version a few days ago, so I guess we can just use that -
20:23:26Learpreglow: Do you remember how many bits are affected by the HIGH_PRECISION flag in the ARM EQ code, by any chance? :)
20:23:45preglowLear: i believe a comment in eq_arm.S tells you
20:24:28hcsjhMikeS: I tried out your new spc code on my ipod color, works great with everything I've thrown at it!
20:24:46preglowdamn, i need to think about spc and volume again, i hate them so quiet
20:25:03Learpreglow: Not that I can see. Only talk about "lower bits" or "lower part".
20:25:31hcspreglow: what can one do about that? afaik the full 16 bits is used
20:25:46Buschelquestion: why is FS #7510 (IDE0_CFG for CPU>65MHz) only active for iPod nanos? Does it harm the units, if this is also set for other devices? Sounds like setting the bit dependent upon the clock is the right way?
20:26:00preglowLear: * without this, "shift" of the lower bits will be lost here
20:26:04preglowLear: so you gain "shift" bits
20:26:06linuxstbLLorean: JdGordon said that the OF starts when you insert USB with hold on - see 09:47 here -
20:26:30*linuxstb is curious if anyone has tested that bootloader change
20:26:42 Join thegeek [0] (
20:27:48preglowsweet lord, this mp3 is loud
20:27:51Lloreanlinuxstb: Didn't used to for me.
20:27:56preglowcurrently the max peak i've seen is _five_ times the clip level
20:28:02Lloreanlinuxstb: Maybe depends on FW version?
20:28:56amiconnpreglow: Uploaded and description updated.
20:29:12bertrikctaylorr: crunchy loop now after some experimenting with crossfade and skipping forward
20:29:31amiconnBuschel: It freezes some other PP502x targets
20:29:31ctaylorrbertrik: huh?
20:29:41ctaylorrbertrik: Oh...
20:29:48Nico_Plinuxstb: have you seen the ZVM firmwre updater mcuelenaere posted on the forums?
20:29:51Lloreanlinuxstb: Mine says "Key LOCKED" and on the next line "System Shutdown"
20:29:53Learpreglow: Ah, so shift is 4 or 6 bits. Quite a bit, but that's on the internal format, so there's still 20+ significant bits. Me forgets about enabling the high precision mode again...
20:29:54Lloreanlinuxstb: I've just tested
20:30:31amiconnBuschel: The PP ata controller isn't fully researched yet. The ">65MHz" bit is from ipl, i.e. one should take it with a grain of salt...
20:30:42bertrikand now a data abort at 4000132c, that's in the fiq_handler
20:31:22preglowamiconn: excellent
20:31:29 Quit Phill2 ()
20:31:30preglowamiconn: i'll do an ml post
20:31:45linuxstbLlorean: So does it actually shut down?
20:31:51amiconnpreglow: And my Recorder has shiny new clips now. Generated 3 times to make absolutely sure ;)
20:32:09linuxstbNico_P: No - why do you ask?
20:32:34Nico_Plinuxstb: It uses libmtp to send the file to the ZVM... I was thinking it could be adapted to do the same for the S
20:32:52 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
20:33:15 Join miepchen^schlaf [0] (
20:33:15 Join merbanan [0] (n=banan@
20:33:55Buschelamiconn: ok
20:35:02 Quit nicktastic ("Leaving")
20:35:19barrywardelllinuxstb: thanks. I'd only managed to build sansapatcher.exe before, not e200rpatcher.exe
20:35:31 Quit scorche (Read error: 104 (Connection reset by peer))
20:36:18 Join scorche [0] (n=scorche@rockbox/administrator/scorche)
20:36:21linuxstbbarrywardell: If you're interested, I've documented the tcctool build process for windows in utils/tcctool/README - that's almost identical to e200rpatcher.
20:36:55barrywardellI'll have a look
20:36:58linuxstbNow we need the Swedes...
20:37:37 Join webguest93 [0] (i=53d609bd@gateway/web/cgi-irc/
20:38:02 Part webguest93
20:39:06 Quit Phill (Read error: 110 (Connection timed out))
20:39:15 Quit animeloe ("Leaving")
20:39:43 Join Arathis2 [0] (
20:39:47 Quit MethoS- (Read error: 110 (Connection timed out))
20:41:19linuxstbNico_P: Looks interesting, but I need to ignore the S for now and do other things...
20:42:21Nico_Plinuxstb: ok, no rush... I'll send mcuelenaere a PM though
20:42:25 Join animeloe [0] (n=animeloe@unaffiliated/animeloe)
20:43:14n1sNico_P: is that working?
20:44:04Nico_Pn1s: "This file is a modified libmtp example file, which should upload a firmware correctly."
20:44:24Nico_PI haven't tried very hard to compile it, as I don't have a ZVM to ty it on
20:44:38n1sI get a feeling you haven't tested thouroughly :)
20:44:46 Join Domonoky [0] (n=Domonoky@
20:45:10Nico_PI haven't tested at all :)
20:45:10n1sah, I thought it was for the S
20:45:44Nico_Pn1s: why the feeling? can you see something blatantly wrong?
20:46:24n1s"should" <<<
20:47:12 Quit barrywardell (Remote closed the connection)
20:49:03Nico_Phaha :) these are mcuelenaere's words
20:49:25 Join mirak [0] (
20:50:00linuxstbNico_P: If I was investigating, I would try and find a USB dump from a ZVM update, and compare it with the S update log, to see if they are in any way similar.
20:50:57Nico_Pgood idea
20:53:32n1shmmm, this voice id callback system can't handle decimal numbers...
20:54:18n1samiconn: the various *getlang functions in settings_list.c
20:55:00n1sfor example crossfeed gains are stored internally as centibels and are spoken as 10*displayed value decibels
20:55:23n1sbut we support .5 steps, which is the problem...
20:57:42 Quit Echelon (Read error: 104 (Connection reset by peer))
20:57:54 Quit Buschel ()
20:58:45amiconnYeah, decimal fractions cannot be handled with a simple id
20:58:46 Quit Arathis (Read error: 110 (Connection timed out))
20:58:53n1seither the format of the returned id needs to be changed or make the callback itself do the speaking, I wonder which is best...
20:58:57amiconnYou need a function doing the output
20:59:38n1samiconn: would it be ok if the callback function itself does the speaking?
21:02:28 Join Siku [0] (
21:05:19 Join KoverSrac [0] (
21:08:33KoverSracDoes anyone have experience with iPod+Rockbox+Toyota Yaris iPod connection?
21:09:53linuxstbHave you read this page?
21:10:41***Saving seen data "./dancer.seen"
21:11:11KoverSracNot yet. Thax for the URL, and sorry for asking dumb questions. :)
21:14:02 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
21:14:15 Join SoapSud [0] (
21:14:25 Join miepchen^schlaf [0] (
21:14:42 Join hannesd_ [0] (
21:16:33 Quit SoapSud (Client Quit)
21:18:06 Quit hannesd (Read error: 145 (Connection timed out))
21:18:07 Nick hannesd_ is now known as hannesd (
21:24:02 Quit Siku (Read error: 145 (Connection timed out))
21:25:10n1sok, here is my speaker function, anyone against taking this route?
21:30:12 Join linuxstb_ [0] (
21:30:30 Quit linuxstb (Nick collision from services.)
21:30:35 Nick linuxstb_ is now known as linuxstb (
21:32:12 Quit KoverSrac ("KVIrc 3.2.0 'Realia'")
21:33:27 Join Shaid [0] (
21:36:05 Quit spiorf (Remote closed the connection)
21:38:06 Quit Xerion (Read error: 104 (Connection reset by peer))
21:38:07 Quit amiconn (Nick collision from services.)
21:38:15 Join amiconn [0] (n=jens@rockbox/developer/amiconn)
21:38:59 Quit Aurum_ ("Leaving")
21:39:44 Join przemhb [0] (
21:40:07Learn1s: Not using the list_speak_item callback?
21:40:18LearAnd that list isn't the only one with dB...
21:41:48LearAnd the code is a bit ... backwards. :)
21:42:13 Join FOAD_ [0] (
21:43:09 Join japc [0] (
21:43:17 Quit Fraser ("Leaving")
21:43:18n1sLear: is there a way to use that list callback thing in the settings lists as the callback available in the macro hell that is settings can only return a talk id which can't handle decimals (and this is called as that callback, just it does the work itself)
21:44:33 Join advcomp2019_ [0] (n=advcomp2@unaffiliated/advcomp2019)
21:44:43 Quit advcomp2019 (Nick collision from services.)
21:44:47 Nick advcomp2019_ is now known as advcomp2019 (n=advcomp2@unaffiliated/advcomp2019)
21:44:54 Quit animeloe ("Leaving")
21:45:13LearAh, didn't think of the context. IIRC the speak callback is a fairly recent addition, so it might not be available.
21:45:55 Join animeloe [0] (n=animeloe@unaffiliated/animeloe)
21:48:34 Join Xerion [0] (
21:48:35LearCould UNIT_DB be used to trigger the right behavior perhaps?
21:49:34n1sLear: we would need a new unit then because that is used in voulme ans other lists were we display the actual valu we use...
21:49:57n1sUNIT_CB maybe :)
21:50:27 Join mcuelenaere [0] (
21:52:34Nico_Pmcuelenaere: hi
21:53:09mcuelenaerehow far are you in getting the firmware on the gigabeat?
21:53:14Nico_Pmcuelenaere: I am indeed running your modded sendfile.c, but it couldn't send the file
21:53:22Nico_PError 2: PTP Layer error 02ff: LIBMTP_Send_File_From_File_Descriptor():Could not send object property list.
21:53:25Nico_Pthen segfault
21:53:34 Quit MethoS-- (Read error: 110 (Connection timed out))
21:53:59mcuelenaerejust a sec, I'm searching the file
21:54:16Nico_Pmcuelenaere: thanks for joining btw :)
21:54:22mcuelenaereno problem :)
21:56:42 Quit animeloe ("Leaving")
21:56:45mcuelenaereNico_P: if you run the file multiple times, does it give the same error?
21:57:58Nico_Pmcuelenaere: I can't seem to connect more than once to the device
21:58:30Nico_Pafter doing it once I need to disconnect it, the reconnect it
21:59:08 Quit FOAD (Read error: 110 (Connection timed out))
21:59:09 Nick FOAD_ is now known as FOAD (
21:59:10mcuelenaerethe problem is: I couldn't test it with the ZVM, so I don't know if it works with me either
21:59:38mcuelenaerebut as I can see from the log, both the ZVM and the Gigabeat seem to use identically the same upgrade process
22:00:28Nico_P02ff is PTP_ERROR_IO in case that helps (from ptp.h)
22:00:46Nico_Pthat's good news... a linux firmware updater would really be welcome
22:03:25 Quit merbanan (Remote closed the connection)
22:03:53 Quit scorche|w ("CGI:IRC (Session timeout)")
22:04:37mcuelenaereI'm trying to change some of the parameters of LIBMTP_Send_File_From_File()
22:04:55mcuelenaerebut maybe I'll use LIBMTP_Send_File_From_File_Descriptor()
22:05:47preglowamiconn: another vb bug report on the ml....
22:05:52*preglow continues to hate vbscript
22:06:40 Quit OlivierBorowski (Remote closed the connection)
22:07:06mcuelenaereNico_P: is there some kind of debug function in libmtp?
22:07:10mcuelenaereor debug attribute?
22:07:21Nico_Pmcuelenaere: I'm running it in gdb right now
22:07:40mcuelenaerecan you see the data send to the gigabeat?
22:08:23mcuelenaereI'm guessing something in the ObjectInfo is wrong during the SendObjectInfo
22:08:49Nico_Pmcuelenaere: how would I do that?
22:08:53mcuelenaerebut with the standard LIBMTP_Send_File_From_File() commands I do not have access to the raw object
22:09:10mcuelenaereI don't know :)
22:09:14mcuelenaereI don't really use linux
22:09:27 Quit keanu|afk (Read error: 110 (Connection timed out))
22:09:30mcuelenaerenor gdb
22:10:31Nico_Pmcuelenaere: if you tell me what values you want me to give you I can
22:11:00mcuelenaerevalues of?
22:11:14mcuelenaerethe object property list?
22:11:36 Join animeloe [0] (n=animeloe@unaffiliated/animeloe)
22:12:03Nico_Pin the code... what I can tell you is that it fails at line 4116 of libmtp.c
22:12:13mcuelenaereyes I know
22:12:29mcuelenaereoh you want to custom compile libmtp.c?
22:12:33mcuelenaerewith hardcoded values?
22:13:01mcuelenaereBTW isn't it at l.4175?
22:13:17Nico_Pno, I meant I can give you the values of the vars if it can help (with gdb)
22:13:36Nico_Pthe message is "Could not send object property list"
22:13:45Nico_Pso I guess you're right :)
22:14:09mcuelenaereah yes, could you give those values?
22:14:38Nico_Pwhich ones? the params to the ptp_mtp_sendobjectproplist call?
22:14:38 Quit mcuelenaere (Read error: 104 (Connection reset by peer))
22:15:54 Nick parafin is now known as parafin|away (
22:16:38n1sLear: yeah hacking in a special case in talk.c with a new UNIT_CB worked fine as well, still a bit hacky though...
22:18:17 Quit J3TC- (Read error: 110 (Connection timed out))
22:21:00 Join mcuelenaere [0] (
22:21:08mcuelenaereI'm back
22:21:47 Quit markun ("leaving")
22:22:05 Join markun [0] (n=markun@rockbox/developer/markun)
22:22:14mcuelenaereNico_P: still there?
22:23:03mcuelenaerecould you give me the values of props (4168)?
22:24:59 Nick Bagder_ is now known as Bagder (
22:25:04BagderGodEater: you called sir?
22:25:41ender` <- what i'm doing these days...
22:25:42 Join qweru [0] (
22:26:07 Join lee-qid [0] (
22:26:27 Quit RaRe ("Quit msgs should be longer.")
22:30:44 Join matsl_ [0] (
22:30:49 Quit bertrik ("bye")
22:34:24 Quit Domonoky (Read error: 104 (Connection reset by peer))
22:37:43 Join billenium [0] (
22:37:52 Quit DM| ("*bashes head against keyboard*")
22:39:50 Join MethoS-- [0] (
22:42:10 Join DM| [0] (
22:43:04 Join jhulst [0] (n=jhulst@unaffiliated/jhulst)
22:46:56 Quit mcuelenaere (Read error: 104 (Connection reset by peer))
22:47:08 Join mcuelenaere [0] (
22:47:10 Quit animeloe (Read error: 113 (No route to host))
22:47:37 Quit Tetris-Block ()
22:49:21 Join petur [0] (n=petur@rockbox/developer/petur)
22:51:53linuxstbBagder: Can you do some download server updates?
22:52:34linuxstblinux and mac:"> and win32 -
22:52:49 Join animeloe [0] (n=animeloe@unaffiliated/animeloe)
22:53:44 Quit Lear ("ChatZilla 0.9.79 [Firefox 3.0b1/2007110904]")
22:54:34Bagderthe linux one is now 50K instead of the previous 500K?
22:55:46linuxstb500K seems large, I wonder why...
22:55:46Bagderbtw, does the current one (before this update) have a version number?
22:55:47 Quit miepchen^schlaf (Read error: 104 (Connection reset by peer))
22:56:02BagderI'm thinking of putting the former one in a subdir
22:56:10 Join miepchen^schlaf [0] (
22:56:13 Quit mcuelenaere ()
22:56:28linuxstbBagder: I'm not sure - you could try just running the Linux version...
22:58:43linuxstbIt seems the previous e200rpatcher.linux was statically linked - with everything, including libc...
22:59:08Bagderfiles upload now
22:59:37 Quit matsl_ (Read error: 110 (Connection timed out))
23:02:46 Quit MethoS-- (Read error: 110 (Connection timed out))
23:03:40 Nick maddlah is now known as maddler (
23:05:33 Join digitallo [0] (
23:06:57 Quit digitallo (Client Quit)
23:07:50 Join TMM [0] (
23:09:17 Join Casainho [0] (
23:10:42***Saving seen data "./dancer.seen"
23:12:10 Quit petur ("here today, gone tomorrow")
23:13:26przemhbwhere can I find the function that stores global settings to config file?
23:14:23n1sin apps/settings.c y'now grep is your friend
23:14:55 Join Felixmoulston [0] (i=3d58838f@gateway/web/cgi-irc/
23:14:55 Join scorche|w [0] (n=42c007b2@rockbox/administrator/scorche)
23:15:00FelixmoulstonHi i live in wyoming and go to a high school nearby.
23:15:19FelixmoulstonIm USA.
23:15:52FelixmoulstonAny 5up4 1337 h4ck3rz h3r3?
23:16:00Bagderstop that
23:16:07XavierGrAm I right that all sysfont entries on the lang files shouldn't have non-latin characters in the translation?
23:16:27BagderXavierGr: yes, I think that's the general plan
23:16:28FelixmoulstonYes that's correct.
23:16:38XavierGrBagder: thanks
23:17:18 Quit atsea- (Remote closed the connection)
23:17:42przemhbn1s: thanks
23:18:31FelixmoulstonHey nerds, decode this! 010101010010111101001100010010001000100101111000100010101010101000101010101010011.
23:18:44Mode"#rockbox +o Bagder " by ChanServ (ChanServ@services.)
23:18:53Kick(#rockbox Felixmoulston :Bagder) by Bagder!n=daniel@rockbox/developer/bagder
23:19:02XavierGrgood ridance
23:19:16 Quit jhMikeS (Nick collision from services.)
23:19:22 Join jhMikeS [0] (n=jethead7@rockbox/developer/jhMikeS)
23:19:29 Join webguest801 [0] (i=3d58838f@gateway/web/cgi-irc/
23:19:42Bagderwebguest801: now cut that crap
23:19:43webguest801Hello, is anyone working on the Gigabeat S port?
23:19:57 Join animeloe_ [0] (n=animeloe@unaffiliated/animeloe)
23:20:02XavierGrjeesh translating since langv2 is a lot of work! Already working on this 2.30 hours and I still have around 2000 lines to finish :(
23:20:13Bagderwebguest801: yes, some guys are working on that
23:20:20 Quit webguest801 (Client Quit)
23:20:44moosXavierGr: it take me lot of time too for francais.lang
23:21:09 Quit mf0102 ("Verlassend")
23:21:10scorche|wlooks like i got back to my desk in time...
23:21:20Mode"#rockbox -o Bagder " by Bagder (n=daniel@rockbox/developer/bagder)
23:21:38moosXavierGr: took even, I made it this summer :)
23:22:08 Join atsea- [0] (i=atsea-@gateway/tor/x-b3d0c437d014db79)
23:23:01 Quit animeloe (Read error: 110 (Connection timed out))
23:23:43scorche|wBagder: can you get rid of the bright yellow as an option for the perl IRC log script?
23:24:10BagderI'll rather poke on Zagor for that ;-)
23:24:31scorche|wthat works too :)
23:29:33 Quit animeloe_ ("Leaving")
23:35:20Nico_PBagder: would it be possible to have autolinking of rev numbers in commit messages? and http links?
23:36:16Bagderyeah, I could use the same regex zagor is using for the irc log I guess
23:36:33linuxstbSpeaking of which, could the "since 1 aug 2006" page be split?
23:36:35Bagderfor rev numbers I mean
23:36:36Nico_Pyeah, it's working fine
23:36:47Bagderlinuxstb: split?
23:36:53linuxstbMake smaller...
23:37:02linuxstbSplit into smaller time periods.
23:38:35Bagderbut how would those be presented?
23:38:46 Quit Xerion (
23:38:46 Quit Arathis2 (
23:38:46 Quit kkurbjun (
23:38:46 Quit Soap (
23:38:46 Quit crashd (
23:38:46 Quit crwl (
23:38:46 Quit maraz (
23:38:46 Quit annulus_ (
23:38:46 Quit Weiss (
23:38:46 Quit lodesi (
23:38:56Bagderlike monthly summaries?
23:39:12jhMikeSNico_P, lostlogic: looking to commit =>
23:39:30linuxstbMaybe monthly would be too much, but at least yearly - e.g. "during 2006", "during 2007" ?
23:39:46XavierGrhow would someone describe the word "dithering", I am trying to translate it but I am not sure that I understand the concept
23:39:51Nico_PjhMikeS: does it work well?
23:40:01jhMikeSon every target I have, yes
23:40:13BagderI guess we could have one "the last 6 months"
23:40:31BagderI mean, instead of the fixed since aug 1 2006
23:41:01Nico_PjhMikeS: looks fine to me
23:41:59NJoinXerion [0] (
23:41:59NJoinArathis2 [0] (
23:41:59NJoinkkurbjun [0] (
23:41:59NJoinSoap [0] (n=Soap@rockbox/staff/soap)
23:41:59NJoinmaraz [0] (
23:41:59NJoincrwl [0] (
23:41:59NJoinWeiss [0] (
23:41:59NJoinannulus_ [0] (
23:41:59NJoinlodesi [0] (
23:41:59NJoincrashd [0] (
23:43:28 Join crwll [0] (
23:44:24 Join crashd_ [0] (
23:44:27 Join Weiss_ [0] (
23:44:54 Quit crashd (Remote closed the connection)
23:46:14 Quit Weiss (Broken pipe)
23:46:47 Quit annulus_ (
23:46:47 Quit crwl (
23:46:47 Quit maraz (
23:46:47 Quit Xerion (
23:46:47 Quit Soap (
23:46:47 Quit kkurbjun (
23:46:47 Quit lodesi (
23:46:47 Quit Arathis2 (
23:46:49 Join lodesi_ [0] (
23:46:50NJoinmaraz [0] (
23:46:58lostlogicjhMikeS: that's similar to what I had in there, just with smaller chunks and more frequent ata 'breaks'
23:47:09NJoinSoap [0] (n=Soap@rockbox/staff/soap)
23:47:12jhMikeSyeah, that seemed to be the key
23:47:21NJoinXerion [0] (
23:47:55lostlogicjhMikeS: I'm definitely a fan.
23:47:56NJoinArathis2 [0] (
23:48:09Nico_Plinuxstb: just so you know, mcuelenaere's updater didn't work for me and he hasn't tested it on a ZVM... we've tried to debug it but have had no success
23:48:28rasherJust a reminder to anyone involved: when debugging voice building problems, "V=1 make voice" enables debug output
23:48:28lostlogicjhMikeS: trynig to think if there is any potential issue with filechunk being smaller than guardbuf size
23:48:40 Join annulus_ [0] (
23:48:41Nico_Pbut he said that from the log, he could tell the S has the same update system as the ZVM
23:48:41***Server message 505: 'logbot :Private messages from unregistered users are currently blocked due to spam problems, but you can always message a staffer. Please register! ( )'
23:48:42lostlogicwith regard to the buffering request size guarantees
23:49:15jhMikeSisn't that filled when the data is read?
23:49:36linuxstbNico_P: That sounds promising then - and also for a lot of players that use MTP...
23:49:38jhMikeSnot off disk but during the request
23:50:15jhMikeSyeah, bufgetdata takes care of it
23:50:21Nico_Plinuxstb: indeed. I hope we can get it to work... when it works we can probably make it standalone too
23:50:41DerPapstXavierGr: If you are limited to say 2 colors and you want to make a gradient you use "dithering". colors that aren't availabe are simulated by using different pixel patterns of the 2 colors.
23:51:27lostlogicjhMikeS: yeah, so the guardbuf is OK −− what about the buffer low case when a request asks for more than the available data? this could cause a queue bounce before such a request could be satisfied, whcih was avoided before by filechunk being equal to max request guarantee
23:52:00XavierGrDerPapast would it be similar to describe it as "noisification" (to insert noise on quantized values)?
23:54:01DerPapstyeah.. i guess so.
23:55:13jhMikeSlostlogic: no idea really. why should it in any way depend on chunksizes to be a particular value?
23:58:02jhMikeSBottom line is the threading has to balance properly anyway. Suppose I'm gonna do flat indexes so I'll have to scrutinize all those things anyway.
23:58:40 Quit snake (Read error: 104 (Connection reset by peer))

Previous day | Next day