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

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

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

#rockbox log for 2005-07-29

00:05:17 Join BBub [0] (belzebub16@dsl-084-059-238-036.arcor-ip.net)
00:06:29BBubthere is a error in the cvs ;( metadata.c:849: error: `track' undeclared (first use in this function)
00:07:25 Join preglow [0] (~c39fb6c6@labb.contactor.se)
00:08:08preglowred builds :/
00:09:11BBub[00:06:31] <BBub> there is a error in the cvs ;( metadata.c:849: error: `track' undeclared (first use in this function) <- ;)
00:09:44preglowyes, someone's done poor testing before committing
00:09:56 Quit phaedrus961 (Remote closed the connection)
00:10:32 Quit preglow (Client Quit)
00:10:49 Join prethom [0] (~c39fb6c6@labb.contactor.se)
00:11:12prethomi cant fix it, unfortunately
00:11:59prethomonly need to change 'track' to 'entry', it seems
00:12:12prethomi sincerely doubt he even compiled this before he commited
00:14:21 Join webguest39 [0] (~513f6968@labb.contactor.se)
00:14:32amiconntrack is a struct trackinfo, which has a member id3 that is a struct mp3entry
00:14:57 Join ashridah [0] (ashridah@220-253-122-228.VIC.netspace.net.au)
00:15:13prethomis someone planning to start renaming all the id3 stuff to something more generic?
00:18:14amiconnI wonder (1) why people don't compile their changes before committing (2) why they don't even check the compile status on the site and (3) what the default genre fix (hack?) has to do with the last-tag-in-flac problem
00:18:44Learwell, he did chang a few more things...
00:18:50Learbut I'll make it compile at least...
00:18:58prethomi believe people generally do compile, the trick is convincing yourself even the most trivial change can go wrong
00:19:13prethomand then again making yourself never commit if you're in a hurry
00:19:22prethomi've gotten burned on that last one myself a time or two
00:19:43amiconnOnce should be enough to learn that...
00:22:03prethomyes, 'should' being the troublesome part
00:22:46 Part atmcsld ("Leaving")
00:24:01prethomis playback on h1x0 the way it is on archos now, btw?
00:24:02LearThere, fixed. Seems to work in the simulator too...
00:24:19Learand what exactly do you mean by that? :)
00:24:25prethomi really dislike the seeking behaviour, which is more or less the only place i feel the iriver firmware still does a better job
00:24:37Learthat it pauses you mean?
00:24:37Rickwhat's wrong with the seeking?
00:24:55RickI hate how iriver seek does that live seek or whatever
00:25:01prethom1. no sound while seeking, 2. it doesn't start prebuffering if you're clearly seeking outside the current playback buffer
00:25:02Ricktakes FOREVER to seek through 3 hour tracks
00:25:21prethomstill, it's sensible enough as it is
00:25:26prethombut i'd like a live mode as well
00:25:28Lear2. why start prebuffering when you don't really know where you'll end up anyway?
00:25:29prethomfor those short seeks
00:25:40prethomlear: that's what i mean with it being sensible
00:26:04prethombut i still do a ton of shorter seeks where i know where i want to skip to by the sound
00:26:07amiconnPlayback is not yet fully like on archos
00:26:10prethomwith rockbox that's a pain
00:26:39prethomhow does it differ?
00:26:40amiconnI don't care about sound while seeking
00:27:04amiconnI want a fast & exact seek
00:27:35prethomsure, me too, but i usually dont know the time i want to seek to, i know what music i want to seek to
00:27:49prethomsince i listen to quite long mixes from time to time
00:28:00amiconnI have to evaluate the exact differences, here are 2 that straight come to mind
00:28:27amiconn(1) Stop in file browser doesn't work. It's supposed to stop playback if there's music playing
00:28:51prethomwell, that should be easily fixable
00:28:56amiconn(2) The last few seconds of a track don't update the wps
00:29:15prethomyeah, that's true
00:29:30amiconn(3) Long seeks (*way* longer than the buffer mem) still don't work correctly
00:29:40Leari suspect all those "while playing" loops could have something to do with 2...
00:30:35amiconnAnd of course, the other way 'round, the runtime db is still not hooked up on archos...
00:33:18HClyes
00:33:30HClwe need to get that fixed
00:33:34HClsometime soon.
00:33:40prethomshouldn't be that hard, no?
00:33:49HCli dunno archos, but probably not
00:34:05amiconnI once tried, but failed
00:34:10HCl :/
00:34:14HClwhat failed about it?
00:34:29amiconnHowever, I believe this was due to the runtimedb init problem I introduced....
00:35:01HClmight very well be
00:35:17HClbrb
00:35:49amiconnUgh! I just converted mandelbrot.rock to 64 bit arithmetics for a test. The binary size almost doubled...
00:36:22prethomahaha
00:36:29prethominteger or float?
00:36:36amiconnlong long
00:36:58prethomstrange, both are dependent on support functions, not much is inlined, i believe
00:37:47prethomemac optimise it! :PPpPPppPP
00:38:04amiconnWho's the emac guy? ;)
00:38:13amiconnThe speed is still bearable, btw
00:38:23 Quit webguest39 ("CGI:IRC (EOF)")
00:38:34prethomthe emac guy just left screaming
00:38:37Learemac doesn't help much here... You only get 32 (or possibly 40) bits of result anyway...
00:38:56amiconnNow I can zoom like mad without precision probs
00:39:03 Join courtc [0] (~courtc@adsl-158-44-25.asm.bellsouth.net)
00:39:24prethomlear: yeah, i dont know shit about how mandelbrot calculations work
00:39:27 Join TCK [0] (TCK@81-86-99-206.dsl.pipex.com)
00:41:29amiconnI should probably implement adaptive maximum iteration count
00:41:29 Quit TCK- (Read error: 110 (Connection timed out))
00:41:42prethomthe web client seems to drop sentences from time to time
00:41:46 Join webguest39 [0] (~513f6968@labb.contactor.se)
00:41:59prethombut yeah, i've mentioned it before, but the sim resampler FRACMUL routine is wrong
00:42:12Learplayback works fine for me...
00:42:18prethomwith resampling?
00:42:24Learyep.
00:42:27prethomqueer
00:42:47Learif you now what the emac does it fractional mode, it makes sense...
00:43:03prethomwell, i wrote it
00:43:30prethomthe emac frac multiplier doesn't return the top 32 bits
00:43:49Learyes it does. see the coldfire programmer's manual.
00:44:04prethomit shifts the result once first
00:44:08prethomleft
00:44:21Learonce as in one bit?
00:44:21prethomso you actually get (x*y) << 1
00:44:39Learactually, in frac mode, there's no shift at all (like you can get in integer mode)...
00:45:12prethomare you absolutely 100% certain? i've done countless tests on this, and all pointed to there being a shift
00:45:18prethomand it makes sense for fractional numbers
00:45:43Learwell, I'm playing a 32 kHz ogg in the simulator right now. works fine (except for the occasional gap, that is...).
00:45:45prethomi've also written quite a few lines of asm that assumes there being a shift
00:46:00LearI've seen comments to that effect at least...
00:46:15prethomprobably my comments ;)
00:46:32amiconnDoes the fractional mode work on 0.32 signed fixed point numbers?
00:46:39prethomsigned 0.31
00:47:09prethomdepending on your notation, heh
00:47:12amiconnAh, yes, so range is -0.5 <= x < +0.5 ?
00:47:21prethomone sign bit, 31 frac bits
00:47:38prethomrange is -1 <= x <= 0.999999999
00:47:41Learthe frac thing is, it multiplies two numbers, both essentially being multiplied by a large factor. but the result should only be multiplied once, so you have to divide the result of a mul.
00:48:19amiconnprethom: For that range, the shift doesn't make sense to me...
00:48:22Learso, two numbers, e.g. shifted left by eight, to multiply them you must write (x * y) >> 8 for the right result.
00:48:41prethomsure, i know this
00:49:09prethomamiconn: the leftmost bit is NEVER used, apart from the case of -1 * -1
00:49:32amiconnI do also know how fixed point works... that's what mandelbrot.rock uses
00:49:59prethomyou can get the leftmost bit, it's part of the upper extension byte
00:50:09amiconnI also know that depending on the int.fract split, it might be necessary to shift beforehand
00:51:25Learsee page 1-10 in the coldfire programmer's manual. Shows quite clearly that you get the top 32 (40 via extension byte) bits from a 32 by 32 mul.
00:51:52 Quit webguest39 ("CGI:IRC (EOF)")
00:52:00prethomlear: i suggest you do some direct tests, then, mul two numbers and see what you get out, i'll bet you find the answer to be shifted left once from what you expect
00:52:24prethomif this isn't correct, i might as well just resign from the project on the grounds of being too stupid to have anything work
00:52:51prethom'cuz i just couldn't understand what the emac unit was giving me in the start
00:52:52Learthe volume during playback ought to be noticably wrong then, wouldn't it?
00:52:58prethomlear: indeed
00:53:09LearI don't think it is...
00:53:41prethombut please, hack up a plugin to see what a direct mul gives you, i'd love to have this verified
00:54:11prethomif i'm not correct, it's just quite simply a miracle that any of my asm opts work at all
00:54:24prethomsince a shifting error can account for quite much in some places
00:54:28prethomgod knows i've hunted them
00:55:43prethombut like you say, the docs clearly states the top 32 bits are being returned
00:55:51prethombut the motorola docs contain heaps of errors
00:56:56prethomall of my libmad asm optimisitations correspond directly with other decoders in term of quality and amplitude, so i _should_ be right
00:57:08amiconnFound 2 more things: (1) The long long arithmetics do also work on archos, just rather slow. (2) The critical zoom factor where thelack-of-precision artifacts become visible is where I would expect it
00:57:34prethomare the plugins compiled with -O ?
00:57:37Leari'll see if I can make a quick hack now then...
00:57:52prethomperhaps O2 would give some nice boosts for stuff like mandelbrot and jpeg
00:58:49amiconnI don't really trust gcc in what it does with higher optimisation settings
00:59:17amiconn...at least for non-x86
00:59:25prethomi don't think it screws too badly up with highlevel code
01:00
01:02:27prethomat least it does wonders for the codecs
01:03:57solex_If rockbox crashes on my iriver, and I get an error message with some memory addresses, should I report that here?
01:04:25prethomsure
01:04:49prethomdo we bundle a map file along with the daily builds, anyone? :P
01:05:13solex_prethom: Ok, I will take notes in the future.
01:07:15Leardoubt it...
01:07:23prethomas long as you know which build you use, it _might_ at least give some info
01:08:21prethomugh, i need to rewrite the resampler
01:08:54Learwhy?
01:09:01prethomi designed a couple of quick filters the other day, and it seems like iir filter based resampling might be within reach
01:10:14prethomlear: i need to make downsampling work in-place, and i need to make it resample stereo data directly, right now it calculates all the loop variables twice when it doesn't really have to
01:12:01Cassandralear: Good work on the replaygain stuff, btw. It's nice.
01:12:09Learthanks.
01:12:10CassandraBeen listening all today on random.
01:12:35amiconnThe 64-bit division is a monster on SH1 - ~1.3 KByte of code...
01:12:43prethomahahaha
01:13:05prethomit's a pretty big routine on most non-64-bit platforms
01:13:10Learamiconn: no surprise. I guess 32-bit divide is pretty bad too (which explains why rolo was so darn slow in the beginning)...
01:13:13amiconnI'm pretty sure this is compiled C code, using multiple 32 bit operations
01:13:41amiconnI *think* the SH can do better
01:13:49prethomit really should
01:13:56prethom1.3kb of code for a div is over the top
01:14:05Leardoubt it, it doesn't even have a proper div instruction!
01:14:12prethomi'm sure the gcc people will accept your patch ;)
01:14:18prethomlear: for real?
01:14:28amiconnIt doesn't offer direct division of 16 or 32 bit numbers, but a 1-bit-division-step instruction
01:14:36prethomahh
01:14:44prethomthat's not _very_ unusual, at least
01:14:49Learfor a 16 bit divide, you call something called div0, then 15 calls of div1. or along those lines at least...
01:14:55Learamiconn: exactly.
01:15:13prethom3dnow has some instructions along those lines
01:15:20amiconnI think that can be used for >32 bits as well, only the shifting may get a bit tricky
01:15:23Cassandrabtw Am I the only one getting Rockbox randomly freezing on track change. Pressing stop closes the WPS, but the resume data is gone.
01:15:24prethomfor calculation square roots and divisions
01:15:33amiconnUsing multiple registers it should be possible though
01:17:18amiconnMandelbrot doesn't need division in the calculation loop, fortunately
01:17:20Learprethom: well, yes, for the same result in the simulator, I do need to use >> 31...
01:17:26prethomi rock
01:17:29prethom:-)
01:17:57Learlooked good when I tested (and sound okay now too), so I didn't consider that sign bit...
01:18:18Learand the brief motorola docs didn't help either. :)
01:18:24prethomwell, the extra shift doesn't help things
01:18:33prethomit had me puzzled for a week
01:18:58prethomand another week went until i fully understood that was the internal format, not just movclr.l magic
01:19:03amiconnThere is one thing that I like about the coldfire asm: It has shift-by-n instructions
01:19:24amiconnThat helped a lot for grayscale...
01:19:29Learit isn't really an extra shift; more like the proper result when multiplying 31-bit fractional numbers, so to speak.
01:19:37prethomyes
01:19:45prethomthat is how it's usually done other places as well
01:20:15Learso the simulator macros needs some tweaking, but I wonder if anything else in dsp.c does...
01:20:31prethomnothing in the resampler does
01:20:46prethomi wrote that and FRACMUL, so should work peachy
01:21:21prethomi wonder if the sim FRACMUL mistake is mine, though
01:21:31prethomprobably is
01:22:11 Join bagawk [0] (~lee@bagawk.user)
01:22:34Learnope, I changed that, from something like (x*y)>>1, which wasn't much better when you think about it. :)
01:22:47bagawkamiconn, what happened to lcd_clearpixel?
01:23:04prethomno, my initial take was seriously damaged
01:23:08bagawkamiconn, also warning: "HAVE_LCD_BITMAP" redefined
01:23:50amiconnbagawk: (1) See http://www.rockbox.org/twiki/bin/view/Main/GraphicsAPI lcd_clearpixel() is no longer needed
01:24:00prethomthe routine assumes you work with long long, but there's a long cast there clear as day, heh
01:24:00amiconn(2) Dunno what you're doing...
01:24:33Learno, it was x*y << 1...
01:24:42prethomyup
01:24:54prethomthat's what it would be if you used 64 bit precision all the way
01:25:44LearI said "like" above, didn't bother about exact types...
01:26:31bagawkamiconn, so a replacment is bg_drawpixle(x,y);?
01:26:41prethomdidn't get to test sim audio those days anyway
01:27:03amiconnbagawk: No, you use lcd_drawpixel with a suitable drawmode
01:27:49bagawkahh
01:28:01Learhm... the fracmul_8 thing ought to be correct. In effect, I'm multiplying a s.31 with a s7.24, so that should still be shifted right 24 bits...
01:28:30Learso the comment was misleading, and fracmul in sim was off...
01:29:39prethomyup
01:29:52prethomfracmul_8 is correct, yes
01:30:03prethomat least looks like it, i'm not too awake right now
01:30:16Learhm.. wait a minute, I did look at values from the resampler in the simulator, and it did look right when I shifted right by 32... (I got nice values between samp-1 and samp, so to speak)
01:30:58Learbut there could've been shift errors in my debug prints, I might add... :)
01:31:40Learoh well, nothing I'll bother to commit right now anyway. :)
01:32:09prethomno, sure, but just put it in some day, the less confusion about the frac mul, the better
01:32:28prethomi did consider writing a quick emac howto, but never got around to it
01:34:32Learhe, I could've used that. took some time before I understood how it worked / found all relevant documentation about it...
01:34:55prethomyes, me too
01:35:30 Quit bagawk ("Leaving")
01:38:45prethomi better start thinking about going to bed
01:40:12Learhm.. wonder if starting the comment with "Multiply two S.31 fractional integers" makes it much clearer for "outsiders"...
01:40:21CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
01:40:21*Lear should also go to bed soon...
01:42:04Learbtw, got that butterfly replacement working, but it wasn't much of an improvement... (not too surprised really)
01:42:14prethomno, me neither, really
01:42:46prethomif you're serious about codec improvement, you should try to integrate that special cache dependant function section i was talking about
01:42:53prethomi'd do it myself, but no dev env yet
01:44:20prethomi'm moving to another flat in a couple of weeks, think i'll be set for more coding then
01:44:44prethomi wont cry leaving this 28k8 hell behind
01:46:41LearI'm not sure what that would do really, and where to use it, so...
01:47:36Learbut throught that code i verified an errata in the cpu (at least I think I did...) :)
01:48:08Learwhat, only 28k8? but that's still better than what I had when I changed broadbad isp...
01:49:22Learfor some reason (bad modem/adsl equipment still installed), my 56k modem did not give me any good speed - or reliable connection for that matter...
01:50:27prethomit's for grouping all performance critical functions together so they dont step on each others feet in the cache
01:50:39Learregarding the errata, when getting the result from the emac, try to put an instruction after the last mac/msac and the first move from accumulator.
01:50:43prethomthis is a 56k modem, but the connection's so bad it reverts to 28k8
01:51:06prethomyeah, good old emac fetch stall
01:51:08Learokay... still no idea on how to do it... :)
01:51:21prethomi've got a imdct patch in the patch tracker that does it
01:51:36Learthe one that's supposed to be faster, you mean?
01:51:41prethomyes
01:52:17prethomthe imdct routine itself doesn't do it, but i added a special .cachetest section that's supposed to hold all the critical functions
01:53:38prethombut i gotta go to bed
01:53:42prethomsee y'all later
01:53:48Learbut there is still an errata regarding the stall, doing a mul on acc1 shouldn't stall a fetch of acc0...
01:53:55prethomahh, yes
01:54:13prethomdoesn't matter which acc
01:54:16 Quit tvelocity ("Leaving")
01:54:31Learso you just mean put all "critical" functions in one section?
01:54:36prethomyep
01:54:40 Quit ghode|afk ()
01:54:45prethomso they end up mapped right next to each other in memory
01:54:56Learyep, got it.
01:55:07prethomthis way the chance of them sharing the bottom part of the address is lesser
01:55:34prethomi believe the cache uses the bottom 12 bits as a key
01:55:40prethomand it's 1 way associative :/
01:56:26prethomso if there's a collision, the cache line is flushed, no mercy
01:56:47prethomfor the case of the imdct patch, this effect was _really_ noticable
01:57:05prethomi had no idea why it was so slow until linus suggested that
01:57:14Learsomething to look at for vorbis; might not need icode then...
01:57:42prethomyup, i'm hoping for at least some small gains
01:57:48Lear(which basically has the same effect)
01:58:08prethomwell, more or less, the cache will still be flushed when other non-tremor code is called
01:58:11prethombut it's close enough
01:58:23***Saving seen data "./dancer.seen"
01:58:46prethomiram code will always be faster, but as long as we have important data to stuff into it, that's what it should be used for
01:59:01prethombut you're also right in there not being too much important data left in tremor
01:59:18prethomperhaps there's something left that is dynamically allocated
01:59:22 Quit Moos (" HydraIRC -> http://www.hydrairc.com <- Try something fresh")
01:59:40prethomanywho, i need to go
01:59:41prethomlater
01:59:47 Part prethom
02:00
02:00:15Learbye, I'm off too.
02:00:33 Quit Lear ("Chatzilla 0.9.68.5 [Firefox 1.0.6/20050720]")
02:23:22 Quit hicks (Remote closed the connection)
02:36:13 Join tvelocity [0] (~tony@ppp44-adsl-240.ath.forthnet.gr)
02:43:17 Quit DreamTactix291 (Read error: 110 (Connection timed out))
02:51:36 Join OPP [0] (~OPP@adsl-69-208-168-20.dsl.emhril.ameritech.net)
02:51:37OPPhello
02:52:14OPPdo you guys think the h1xx stuff can be ported to h3xx
02:54:48 Quit OPP (Client Quit)
03:00
03:15:42 Quit Chamois (" HydraIRC -> http://www.hydrairc.com <- Try something fresh")
03:21:19ashridahnevar!
03:22:52t0maskik
03:22:54t0mas*lol
03:28:10 Quit Strath ("Client closed")
03:29:31t0masomg
03:29:46t0massome idiot on a dutch channel is talking about his new idea...
03:30:03t0mas"When I'm 20, I'm going to kill myself... if life doesn't prove it's cool"
03:30:15t0mas"How should life fo that?"
03:30:32t0mas"I should have a great chick, nice car... a lot of money... and I should be a famous hacker."
03:30:46t0mas"Dude... you can kill yourself right now... it's just impossible :P"
03:36:32 Quit Cassandra (Nick collision from services.)
03:37:29 Join Cassandra [0] (~cassandra@82-70-230-150.dsl.in-addr.zen.co.uk)
03:58:26***Saving seen data "./dancer.seen"
04:00
04:01:17 Join DreamTactix291 [0] (~dreamtact@adsl-19-235-161.bna.bellsouth.net)
04:03:47 Join OPP [0] (~OPP@adsl-69-208-168-20.dsl.emhril.ameritech.net)
04:03:53OPPanyone here
04:05:39 Join QT_ [0] (as@area51.users.madwifi)
04:05:46BBubyup
04:06:16OPPnice
04:06:24OPPhows the h3xx port goin
04:07:18BBubnoone is really working on it afaik
04:07:43OPP?
04:07:44OPPreally
04:07:45OPPwow
04:07:51OPPi thought it was
04:09:27BBubmaybe with the IaudioX5 port going forward that will improve
04:09:32BBubas they both have colour-displays
04:10:29OPPwell
04:10:31OPPthe h120
04:10:35OPPis like the same
04:11:25BBubit is in most parts
04:11:37BBubthe cpu goes up to 140 mhz instead of 120
04:11:43OPPahh
04:11:51OPPso there shouldnt be much trouble porting it
04:11:53BBubbut i guess the colour-implementation will take some time
04:11:57OPPya
04:12:49BBuboh wait, seems like the h1x0 cpu can also go up to 140 mhz
04:13:00OPPo
04:13:05OPPdo you have an h120
04:13:07OPPor h140
04:13:15BBubh110 ;)
04:13:25BBubmaybe a h130 soon
04:13:27OPPis that 10 gig?
04:13:30BBubyup
04:13:38OPPo
04:13:53BBubthe first hdd-player from iriver
04:14:09OPPi want one
04:14:39BBubyeah, its a really nice player
04:14:44BBubbest sound i heared so far
04:14:52BBubthe h10 is a little worse ;(
04:14:54OPPi had one
04:14:57OPPbut i sold it
04:14:59OPPa long time ago
04:15:04OPPand now i have an h320
04:15:09BBubwell, i got mine for 2 years now
04:15:13BBubwont give it away ;)
04:15:19OPPya
04:15:25OPPim lookin for them everywhere
04:15:29BBubi never liked the h3x0 having no remote
04:15:36OPPthey made one for it..
04:15:40 Quit QT (Read error: 113 (No route to host))
04:15:41OPPlcd remote
04:15:47OPPbut i dont like remotes
04:16:19BBubi find the player too unhandy to get it out of your pocket while biking for example
04:16:25OPPoh
04:16:29OPPi just use playlists
04:16:42BBubk
04:18:50OPPgrr and im moving
04:18:54OPPhave to wait after that
04:19:18OPPmaybe the h3xx port will get released in the next couple months
04:19:28BBubi would think so
04:19:34OPPreally?
04:19:36OPPthats oon?
04:19:38OPPsoon*
04:19:55OPPhmm ill hang on to it then
04:20:11BBubas i said they are also working on a IaudioX5 port which also requires colour-support
04:20:23BBubso i guess that will push it further
04:20:35OPPoh
04:20:37OPPya
04:20:49OPPisnt all the hardware already figured out for the h3xx
04:20:58BBubyup
04:21:43OPPthats good to hear
04:22:50OPPwhats the process after that making the boot loader
04:23:22BBubyeah
04:23:32OPPand then its done?
04:23:49BBubwell, lets say it will boot ;)
04:24:22OPPo
04:24:22OPPok
04:24:34BBubbut as the dsp is the same it should work right from the start
04:25:09OPPahhh
04:25:20OPPso once the boot loader is made its using the same fw
04:25:45OPPi have some ideas for it :-D
04:25:48OPPsome basic some not
04:25:48 Quit edx (Read error: 110 (Connection timed out))
04:26:35OPPbeing able to get like 20 fps instead of 10 would be great
04:26:39OPPfor vids
04:26:55BBubthat will take some time i guess
04:27:05BBubthere is no video in rockbox yet
04:27:14OPPya
04:27:30OPPeverything in the h1xx should stay
04:27:34OPPits all genius
04:27:44BBuband again they would need integer-based codecs for the videos to decode
04:27:51OPPya
04:27:55BBubso that will be a little harder
04:27:59OPPactually
04:28:04OPPi could live without videos
04:28:07BBubbut i find the jpeg-support for the h1x0 quite nice
04:28:14BBubits really good qualy for the display it has
04:28:18OPPi mean you can always go back to the original firmware
04:28:23BBubsure
04:28:27OPPto watch vids
04:28:29BBubi also do it because of the remote
04:28:38BBubwhich only shows the rockbox logo so far ;(
04:28:53OPP:-(
04:29:47OPPbe happy, this things been out less then a month and its already way better then ipodlinux'
04:29:58BBubhehe
04:30:20BBubipod is for fashionists anyway ;P
04:31:02OPPya
04:31:15OPPiriver is for t3h n3rds and people who like high quality music
04:31:22BBubexactly
04:31:36OPPalthough i saw a girl with an h120 yesterday
04:31:40OPPi knew her pretty well
04:31:51OPPseemed like she was a fashionist
04:32:41BBubheh
04:35:26OPPare you a coder
04:39:26BBubnope
04:39:26 Quit DreamTactix291 (Read error: 104 (Connection reset by peer))
04:39:37BBubi am not experienced enough at c
04:39:43OPPo
04:55:48OPPwell im psyched
04:58:31 Join DreamTactix291 [0] (~dreamtact@adsl-19-235-161.bna.bellsouth.net)
05:00
05:22:28OPPhey dt :-D
05:40:31 Quit tvelocity ("Leaving")
05:41:00 Quit DreamTactix291 (Read error: 110 (Connection timed out))
05:42:46 Quit OPP ()
05:58:30***Saving seen data "./dancer.seen"
06:00
06:11:57 Join webguest16 [0] (~1895b05c@labb.contactor.se)
06:12:44webguest16hey everyone, I was just thinking of a feature...
06:13:01webguest16can we add some menu option in the view playlist screen to "delete all by playing"
06:14:30webguest16oops, sorry, I meant "delete all BUT playing"
06:30:11 Join koniu_ [0] (~koniu@host-ip194-145.crowley.pl)
06:47:37 Quit webguest16 ("CGI:IRC")
06:59:13 Join DreamTactix291 [0] (~dreamtact@adsl-19-235-161.bna.bellsouth.net)
07:00
07:27:56 Join Bru [0] (HP_Adminis@ip68-98-50-58.ph.ph.cox.net)
07:28:25 Quit Bru (Client Quit)
07:42:27 Nick QT_ is now known as QT (as@area51.users.madwifi)
07:47:23 Quit DreamTactix291 (Read error: 110 (Connection timed out))
07:58:33***Saving seen data "./dancer.seen"
08:00
08:00:26 Join Nibbler [0] (~sven@port-212-202-193-100.dynamic.qsc.de)
08:08:34 Quit koniu_ (Read error: 113 (No route to host))
08:37:53 Join Cka3ka [0] (~Cka3ka@ool-4355b700.dyn.optonline.net)
08:37:57Cka3kahey hey
08:38:08Cka3kaanyone around at this late hour
08:38:15Cka3kaspecifically anyone with rockbox on their iriver?
08:38:32amiconnLate? It's 08:38 a.m. here ...
08:38:41Cka3kalmao its 2am here
08:38:42Cka3ka230
08:38:43Cka3kahehe
08:38:46Cka3kai go to bed at 5
08:38:52Cka3kaso i guess its early morning?
08:38:53Cka3kaeither way
08:39:11Cka3kaso man you have an iriver?
08:42:40Cka3kaeh ill pose the question openly: frankly, i just want a clock on my iriver, thing is that iriver doesnt inherently have a "sleep" mode like the ipod that lets it be off but still keep track of time without resetting time. BUUUT with iriver coding hd spindowns and sleepmodes, even on the iriver build, i was wondering if possible to get the whole clock idea going, if so we can build on that by adding a callendar/date and from there you have a
08:42:48Cka3kadamn that was long, someone better read that shit =-)
08:44:07Cka3ka.....no reply isnt cool lol
08:48:18 Join koniu_ [0] (~koniu@host-ip194-145.crowley.pl)
08:50:34amiconnI don't think a clock makes sense on iriver. As soon as you shut down the time will be lost
08:51:12amiconnI mean, that will inevitably happen.
08:51:45amiconnEither you forgot that you don't want the normal shutdown, or the iriver runs out of battery power while in 'deep sleep'...
08:52:06amiconnIt would be very cumbersome to set the clock that often
08:52:14ashridahaah, finally. 1.5 years into this damned degree, and i finally get a challenging programming assignment :)
08:52:41amiconnWhat may be feasible is a wake-up alarm, i.e. "wake me up after x hours" with deep sleep mode
08:54:17Cka3kamm
08:54:21Cka3kasorry im back
08:54:28Cka3kai dont understand why not replace off with deepsleep
08:54:44Cka3kaand doesnt the same crap happen with ipod, its inevitable, battery dead=no time
08:54:49amiconnBecause it eats battery power
08:54:52Cka3kaso big deal, i havnt killed my baterry in forever
08:55:02Cka3kayea, how much? to run 1 tiny program
08:55:12Cka3kalow level program* id expect
08:55:20Cka3kai really think almost none at all =-\
08:55:43Cka3kabesides the iriver rarley completley drains
08:55:51Cka3kait gives you low bat warning and still saves some
08:55:56Cka3kathat you cant really use
08:56:00Cka3kawhich is smart id say lol
08:56:04Cka3kajust in case
08:56:05Cka3ka=-)
08:56:09amiconnHmm?
08:56:20Cka3kathe iriver, it shuts down before bat is at 0
08:56:40amiconnI can run my battery all the way down so the HD can't spinup any more
08:56:44Cka3kai guess so it doesnt screw anything up with a shutdown that just jacks the power
08:56:44amiconn(with rockbox)
08:57:00amiconnI don't use the iriver fw, so I can't tell what iriver does
08:57:06Cka3kai guess, i tried today
08:57:10Cka3kaneither do i anymore
08:57:20Cka3kai clicked the on button and it said warnin low battery
08:57:25Cka3kai plugged into wall
08:57:27Cka3kaand it ran
08:57:30Cka3kaso i dunno
08:57:55Cka3kabut still you have 16 hrs of battery while playin music; thats thousands times more in deepsleep with clock
08:57:59Cka3kadont you think?
08:58:04amiconnI didn't use the iriver fw at all, except for some tests
08:58:11Cka3kaand then you can have a button combo say play and left to turn off completley
08:58:19amiconnI actually got my iriver for the purpose of rockbox hacking
08:58:32Cka3kai had to use it, i got my ihp before rockbox was useable ;-)
08:58:34Cka3kanice
08:58:55Cka3kaare you part of the team? if so props to an amazing job
08:59:19amiconnThere is no fixed 'team' but I think I've done quite a lot for rockbox
08:59:28Cka3kathanks a lot then =-)
08:59:38Cka3kaim using it as we speak
08:59:39amiconn...not only on iriver. I have 3 different archoses as well...
08:59:45Cka3kawow
08:59:45Cka3kalol
08:59:54 Join DreamTactix291 [0] (~dreamtact@adsl-19-235-161.bna.bellsouth.net)
09:00
09:00:06Cka3kawhen i was looking what to buyi thought of archos due to rockbox being fairly done for them
09:00:11Cka3kabut i decided too bulky
09:00:14Cka3kaand whent ihp
09:00:37Cka3kabulky and somehow extremley lacking any hotness (the ihp isnt the pinacle, but has a nice case)
09:01:07Cka3kabut yeah man, i think that a tiny program can be built ONLY to keep track of time
09:01:21Cka3kaso like it watches seconds mins hours days months years
09:01:37Cka3kai think its really easy to prog that much; its the deep sleep that would be hard
09:01:40amiconnThe thing is that even a tiny program consumes power, while full off consumes *no* power
09:02:05Cka3kathas what im saying, replace full off (make it an option) with deep deep sleep
09:02:14Cka3kathat drains tiny bit of power for sake of clock
09:02:18amiconnI can store my iriver in off state for weeks, then power on again and have full battery capacity available
09:02:33amiconnI couldn't do that with deep sleep mode
09:02:42Cka3kahence the options button
09:02:46Cka3kaoption*
09:02:54Cka3kait can be toggleable i guess lol
09:02:57amiconnThe question is how low we can get the power consumption
09:03:14amiconnI have some ideas what to do, but it requires measurements
09:03:16Cka3kawell look; i use my iriver non stop as do a lot of people that have onnly 1 mp3 player
09:03:25Cka3kaand i haev to charge every other day
09:03:34Cka3kabat lasts me 2 days
09:03:40Cka3kaim gonna be charging anyways
09:03:46Cka3kaand im not always at 0 when i charge
09:03:50Cka3kawhich i know is bad but i cant help it
09:04:03Cka3kalike if im goin to school i cant go with like 1 % bat life just so itll go to 0
09:04:13Cka3kathing is most people can spare that battery
09:04:26Cka3kaand in terms of useabillity i guess
09:04:33ashridahuh. correct me if i'm wrong, but lipoly batteries aren't supposed to have any real memory effect anyway, are they?
09:04:52Cka3kayou could even make it so that total off or (power save mode or something) is done by clicking off and another button simultaneosly
09:05:07Cka3kagod so many spelling errors, i cant see what im typing, falling asleep, no more caffine
09:05:59Cka3kasee i dunno, arent supposed to, but my last irivers battery (which i got replaced cus i fell off a bike and the screen cracked...they said they dont care andll replace it) the battery practically died in 1 day
09:06:22Cka3kai mean it went away so much faster as compared to when i bought it
09:06:40amiconnWe can shutdown harddisk power (we already do that), LCD screen, and even the DRAM if running code in IRAM only
09:06:40Cka3kaand that was only after a year of charging it like so
09:06:52amiconnWe can also put the CPU at minimum clock
09:07:13Cka3kasounds like youre almost there
09:07:20Cka3kaid totally do it but i completley lack the skill
09:07:31Cka3ka:-\ working on that though
09:07:35Cka3kasooo need programming classes
09:07:48ashridahi've been using mine for almost a year and a half. i haven't noticed any major changes in battery life. i rarely have to go the full distance tho
09:08:33Cka3kabtw side note, do you like santana? if so recomend an album other than supernatural
09:08:39 Join amiconn_ [0] (~jens@p54BD5DBB.dip.t-dialin.net)
09:08:46Cka3kacus ive gone through that a lot and i have like 17 others and am swamped to listen to it all
09:08:47Cka3kalol
09:08:49Cka3kaa bit at a time
09:08:50Cka3ka=-)
09:08:51 Quit amiconn (Nick collision from services.)
09:08:51 Nick amiconn_ is now known as amiconn (~jens@p54BD5DBB.dip.t-dialin.net)
09:09:14amiconnThe minimum CPU clock should be even lower than the 11 MHz PLL bypass
09:09:23amiconnIt seems the PLL can be uses to clock down, but that needs measurements whether the saved CPU core power is worth the slightly extra power to run the PLL
09:09:48Cka3ka<<−− already a mile behind you
09:09:54Cka3kano clue what pll is
09:10:27ashridahphase locked loop
09:10:44ashridahnice cheap frequency synchronisation circuit
09:10:57ashridahlook it up on wikipedia
09:11:04Cka3kagood call
09:11:05Cka3kagive me a sec
09:11:06ashridahpretty sure there's an entry on them there
09:11:56amiconnThere's also a simplöe by-2 divider iiuc, meaning we should be able to clock at 5.5 MHz without using the PLL
09:12:29Cka3kamm
09:12:36Cka3kamay i add a totally different approach?
09:12:44Cka3kajust an idea of course
09:13:18Cka3kacolarado broadcasts an extremley low frequency high bounce radio sygnal that contains the time (atomic clock sync)
09:13:20 Join Aison [0] (~hans@zux166-181.adsl.green.ch)
09:13:26Cka3kaif the iriver's radio can be used to pick that up
09:13:30Cka3kathen its just as good?
09:13:49Cka3kaoooo and to add to the previous idea (not atomic)
09:14:09Cka3kayou can have reg clock run while iriver is on (like an app) and when in sleep just time how long its been in sleep
09:14:24Cka3kawhen on use simple code to add the timer in seconds to last saved clock
09:14:44Cka3kashould take even less cpu/power than keeping track of current time/date
09:14:48Cka3kawhat do you think
09:15:11 Join webguest47 [0] (~c2cbc979@labb.contactor.se)
09:15:15Cka3kaso itll be like turned off last at hh:ss mm/dd/yy + seconds its been off
09:15:52Cka3kato the new ones entering, im tryin to help come up wit ideas on having a clock on the iriver, despite its total off ...thing lol
09:16:12amiconnOf course it would only count up in deep sleep; the tricky part is how to hand over the updated value when rebooting rockbox at poweron
09:16:50Cka3kawell wouldnt it be using ram to keep the seconds in check
09:17:00Cka3kaso why not just somehow add those seconds to the clock
09:17:08Cka3kaessentiall like moving the min arm forward by hand
09:17:13amiconnThe radio chip can only receive FM radio broadcast range, VLF time service reception is impossible
09:17:33Cka3kaaww thas too bad; heheh worth an try though
09:17:43Cka3kaonly fm? i thuught am too lol
09:17:44Cka3kao well
09:18:35 Join deadArchos [0] (~Roman@adsl-bn150-156-158-212.bluetone.cz)
09:19:02amiconnhttp://www.rockbox.org/twiki/pub/Main/DataSheets/TEA5767HN_2.pdf Philips TEA5767HN: Low-power FM stereo radio for
09:19:02amiconnhandheld applications
09:19:17Cka3kaso yea man i gotta go to sleep and i figure im not much help anyways
09:19:18Cka3kaahh
09:19:27Cka3kai just wanted to sorta throw out that idea
09:19:41Cka3kaits 3 and i shoulda been sleeping by 1 or so =-\ i got a tournament comin up
09:19:44Cka3kanight all
09:20:06Cka3kaif you come up with anything amiconn email me?
09:20:44Cka3kajust out of curiosity so i know if my idea went anywhere
09:24:57 Quit Cka3ka ()
09:40:09*amiconn just got an idea.
09:41:30amiconnI think it *is* possible to measure the PLL relock delay, what would be deemed impossible....
09:42:42amiconnWe know that the CPU runs at 11 MHz until the PLL relocks, so we could set up a timer before setting the PLL on bypass, then set the PLL for the new frequency
09:43:08amiconnSince we are busy-waiting for the relock anyway, we could then read the timer count immediately after relock
09:43:34amiconnThat should allow to calculate the relock delay...
09:44:22amiconnBetter still, read the timer repeatedly within the busy-wait loop, then use the last read value after leaving the loop
09:45:06amiconnShould reduce the error margin, because the counter counts much faster after leaving the loop (by the factor of new_clock / 11 MHz)
09:58:01 Quit DreamTactix291 (Read error: 110 (Connection timed out))
09:58:34***Saving seen data "./dancer.seen"
10:00
10:05:19 Quit webguest47 ("CGI:IRC (EOF)")
10:05:36 Join austriancoder [0] (~5078751e@labb.contactor.se)
10:05:46austriancodermorning
10:07:50amiconnafternoon ;)
10:07:56amiconnNo, morning :)
10:08:10austriancoder*g*... maybe
10:08:42austriancoderamiconn: what are you doing at the moment?
10:17:14austriancoderhave you time to check something?
10:18:10amiconnBusy with boring real-life stuff :(
10:18:29austriancoderah ok
10:34:53 Join Yokalosh [0] (~Yokalosh@cpc2-cbly2-4-0-cust241.glfd.cable.ntl.com)
10:35:43Yokaloshthe solder points are different on the jbr10 to the ones on the player in the guide made in 2000 arent they?
10:35:48austriancoderwhat is De-emphasis in connection with audio codec?
10:41:38Yokaloshmy dc in kinda became detached to the board
10:41:48Yokaloshi should be able to take it appart and solder it back on right?
10:47:22Yokalosh?????
10:48:17 Join preglow [0] (~c39fc175@labb.contactor.se)
10:48:26Yokaloshpreglow :)
10:48:50preglowac: de-emphasis is removing funny high frequency amplification, you'll almost never need it
10:49:25preglowyokalosh: hi
10:50:35Yokaloshis there a updated verion of dassassembling the archos for a jukebox recorder 10?
10:50:40austriancoderpreglow: ok
10:50:49amiconnhttp://en.wikipedia.org/wiki/Deemphasis
10:50:58austriancoderpreglow: have you time to check some stuff for me?
10:51:52preglowdepends, i dont have a development environment and am on a slow modem
10:52:06 Join [-AIR-] [0] (~air@82-35-84-182.cable.ubr03.dals.blueyonder.co.uk)
10:52:47austriancoderpreglow: you dont need a dev enviroment
10:52:48austriancoderhttp://nopaste.php-q.net/150129
10:53:16austriancoderhttp://www-s.ti.com/sc/ds/tlv320aic23b.pdf
10:56:25austriancoderi feel not safe with bitwise operations
10:57:18[-AIR-]hey...here: http://www.rockbox.org/requests.shtml surely "WPS Right Justify" in "configuration" is complete?
11:00
11:00:13 Quit koniu_ ("Lost terminal")
11:01:05preglowac: what's with all the 0 << x ? they'll all be zero
11:02:14preglowand all the addresses are specified in bits, not hex
11:02:28austriancoderpreglow: could be... i fell not good with such bitwise stuff
11:02:54austriancodercould you maybe try to fix/correct it?
11:03:31preglowdont really have time now, perhaps later
11:03:59austriancoderor could you tell me one correct value and i will try it again?
11:04:14austriancoderhow must #define LIM look like?
11:04:30preglowREG_LHV is supposed to be 0x2 for one
11:04:46preglowyou must convert the values from the bit values in the datasheet to hex
11:04:54preglowas for LIM, gimme a sec
11:05:48preglowLIM should be 1 << 6
11:06:46preglowwrong
11:06:46austriancoderso i set the 6th bit to 1... have i understood this correct?
11:06:47preglow1 << 7
11:06:53austriancoder7th bit
11:07:06preglowi didn't check if it was eight bits, heh
11:07:28preglowbasically it's all 1 << x, where x is the number of the bit, IF the first bit is bit 0
11:07:50preglowi think all datasheets call the first bit bit 0, so shouldn't be a problem
11:07:53austriancoderah ok
11:08:25austriancodersounds easy.. will fix my defines and you can check it later?
11:08:49preglowsure, might be a little while until i come back on, though
11:08:58preglowlater
11:09:01 Part preglow
11:09:52 Nick austriancoder is now known as ac_away (~5078751e@labb.contactor.se)
11:10:16 Quit Yokalosh ()
11:19:07 Quit deadArchos ("UkonÄuji")
11:25:39 Join DreamTactix291 [0] (~dreamtact@adsl-19-235-161.bna.bellsouth.net)
11:35:40 Nick ac_away is now known as austriancoder (~5078751e@labb.contactor.se)
11:48:09 Join a1r [0] (~air@82-35-84-182.cable.ubr03.dals.blueyonder.co.uk)
11:58:35***Saving seen data "./dancer.seen"
12:00
12:05:18 Join hicks [0] (~hicks@zeus.mups.co.uk)
12:06:40 Quit [-AIR-] (Read error: 110 (Connection timed out))
12:13:29 Join hicks_ [0] (~hicks@zeus.mups.co.uk)
12:13:50 Join Lear [0] (~chatzilla@h143n1c1o285.bredband.skanova.com)
12:16:19 Join hicks__ [0] (~hicks@zeus.mups.co.uk)
12:17:58 Join DarkkOne [0] (~chatzilla@cpe-66-68-93-2.austin.res.rr.com)
12:23:05 Join Strath [0] (mike@dgvlwinas01pool0-a225.wi.tds.net)
12:23:09 Quit DreamTactix291 (Read error: 110 (Connection timed out))
12:27:50 Quit hicks (Read error: 110 (Connection timed out))
12:30:00a1ramiconn u there?
12:30:06 Nick a1r is now known as [-AIR-] (~air@82-35-84-182.cable.ubr03.dals.blueyonder.co.uk)
12:30:14[-AIR-]amiconn ?
12:30:50 Quit hicks_ (Connection timed out)
12:40:03 Join memmem [0] (~user@p54A20E5D.dip0.t-ipconnect.de)
12:43:33 Quit DarkkOne ("Chatzilla 0.9.68a [Firefox 1.0.6/20050716]")
13:00
13:07:20 Nick hicks__ is now known as hicks (~hicks@zeus.mups.co.uk)
13:12:07 Quit Strath ("Client closed")
13:14:04 Quit memmem ("ERC Version 5.0.3 $Revision: 1.726.2.17 $ (IRC client for Emacs)")
13:48:52 Join DreamTactix291 [0] (~dreamtact@adsl-19-235-161.bna.bellsouth.net)
13:49:58 Join Strath [0] (mike@dgvlwinas01pool0-a225.wi.tds.net)
13:53:12 Join XavierGr [0] (~XavierGr@ppp12-adsl-233.ath.forthnet.gr)
13:53:57XavierGrhi all!
13:55:59austriancoderhi
13:58:36***Saving seen data "./dancer.seen"
13:58:41XavierGrac do you know what the * does after the type of a variable?
13:59:58austriancoderthis is called a "pointer"
14:00
14:00:39austriancoderhttp://en.wikipedia.org/wiki/Pointer
14:01:05austriancoderthe pointer points to the space-address of the variable
14:02:07XavierGryeah but I thought that the * was put onto the variable not the type. or...
14:02:25 Join leftright [0] (~5087db69@labb.contactor.se)
14:03:03austriancoderint* pVar, int *pVar, int * pVar.. all of them are ok and mean the same
14:03:07Learboth ways works fine, just a matter of style.
14:04:59leftrightthe hiss is qite high with no songs playing, stop selected
14:05:23Learbut if you start involving const into a pointer type, then it becomes slightly more complex... :)
14:05:52leftrighthigh ambient hiss, a lot like tape hiss in the old days
14:05:58Learyeah, I've noticed that too; thought it was at least partially due to my earphones, but when I plugged them into the computer, there was no hiss at all...
14:06:27leftrightvery evident on my IEM
14:07:36austriancoderdoes anybody know something about dsp.c?
14:08:12CassandraI wsih.
14:09:14leftrightmp3 playback crashed twice with yesterdays build, randomly, trying todays build now
14:09:37[-AIR-]today's build...there is one?
14:09:53leftrightthe daily labeled 29th
14:10:22[-AIR-]2005-07-28 22:24
14:10:27[-AIR-]that's the top one
14:11:34Learac: yes...
14:11:45leftrightok, crashes occured with 28th's 09:05 build
14:12:01[-AIR-]yer, mine plays music but anything else freezes.
14:12:11[-AIR-]i can't browse while playing
14:13:10Learaustriancoder: what do you want to know?
14:15:08XavierGrstrange my build (compiled with remote support based on the latest) works like a charm.
14:15:29[-AIR-]mine doesn't :'(
14:15:33XavierGrcan I call the main function of a plugin from its main?
14:16:04austriancoderLear: one moment need to check something
14:16:42Learxaviergr: I don't see why not, but I don't see why either... :)
14:17:41austriancoderLear: it is used to resample pcm streams.. where are the limits of frequences?
14:18:47Learthe resampler can at least take input from 11 kHz to 48 kHz, probably higher frequencies as well, but I haven't tested how much higher...
14:19:02Learcertainly not lower though. :)
14:19:29austriancoderwhouldnt it be better to reset the sample format of the audio codec?
14:19:39Learbtw, it's for scaling, clipping and replaygain too..
14:19:44Learhow do you mean, reset?
14:20:50austriancoderif i look at the datasheet of the tlv320 (audio codec of x5) then i see that it supports samplerates from 8 to 96 khz
14:21:29austriancoderso why dont we simply change the sample rate of the audio codec instead to resample it?
14:22:09*austriancoder will be back in about 8 min
14:22:36Learah, you're talking about the hardware codec. :) the codec in the iriver supports different frequencies as well, but for some reason I'm not aware of, that is not done.
14:22:46Learcould be a limitation in the hardware implementation...
14:22:57leftrightyep its crashed again with todays build, mp3 playback, error IO3:addrrErr at 31022DA4
14:23:31LearI'm listening to mp3 right now, and haven't seen anything... any particular file(s)?
14:24:12leftrightmy stuff is all LAME@−−alt-preset standard with RG
14:24:19crwli got a crash today too when i was banging the next button
14:24:32crwlcan't remember if those were mp3 or ogg files, anyway, it didn't happen again...
14:24:38Learsame here
14:25:17Learbut from the look of my current map file (which might not match) that address is in the fat code...
14:26:03crwli got IO3:addrErr at something, it seemed to happen when it tried to buffer new songs and i was still hitting next
14:28:39 Join RotAtoR [0] (~e@dhcp54-47.calvin.edu)
14:29:43XavierGrso when I run a plugin, is there any variable that holds the filenames on the dir?
14:30:22leftrightyes again, it seems to crash when it needs to buffer the next
14:30:42XavierGrI want to mod jpeg.c to make it rotate automatically to other files (on the same folder) when you press up or down and you are in no zoom mode
14:30:54Learlike hitting next one? twice? or not at all?
14:31:30leftrightno just normal playback, then the HDD red light goes on and then crash
14:31:36XavierGrso I found where the filename is loaded on the plugin and changed the code. But I dont know how I will set the "filename" variable to previous or next file. any pointers?
14:31:42Learxaviergr: I guess you'd have to parse the filename (assuming you get the full path). but I'd put the "decode and show file" in a separate function in that case...
14:31:55leftrighterror code and the HDD light remains on
14:32:24Learextract path name (i.e., remove filename), open and read dir, sort it, locate next/previous file. to put it very briefly. :)
14:32:41leftrightit crashes when it accesses the hard drive for more info
14:32:46austriancoderLear: ah ok.. so i can use the dsp only if the mhz is not supported
14:34:02XavierGrthanks Lear I will try it...
14:34:04Learno such problems here, and it looked like it could be fat related. check the drive, perhaps?
14:34:20Learxaviergr: opendir/readdir/closedir are the main functions to use here.
14:34:42XavierGryeah I can confirm that. No playback problems for me too.
14:34:51CassandraI think I may have tracked down my problem with RB freezing at the beginning of random tracks to missing/moved tracks on the playlist.
14:37:14XavierGrlets say I've got a char array. filepath[] = "Test" will give the variable the value Test?
15:00
15:03:41[-AIR-]http://www.misticriver.net/boards/showpost.php?p=283091&postcount=24 :D
15:05:09austriancodermemmem.. where are you.. i need you
15:07:06 Join BBub_ [0] (belzebub16@dsl-084-059-236-248.arcor-ip.net)
15:07:55 Nick [-AIR-] is now known as memmem (~air@82-35-84-182.cable.ubr03.dals.blueyonder.co.uk)
15:07:55 Quit austriancoder ("CGI:IRC (EOF)")
15:07:58 Join austriancoder [0] (~5078751e@labb.contactor.se)
15:07:59memmemyo
15:08:10austriancoderah.. you are online ;)
15:08:10memmemaustriancoder
15:08:14memmemyer
15:08:16memmemlol
15:08:30memmemhow can i help?
15:08:43DBUGEnqueued KICK memmem
15:08:43memmemhttp://www.misticriver.net/boards/showpost.php?p=283091&postcount=24 DOES THIS HELP???
15:08:47austriancoderi am looking at the moment at the routine to write a value int to a tlv320 reg
15:08:52 Nick memmem is now known as [-AIR-] (~air@82-35-84-182.cable.ubr03.dals.blueyonder.co.uk)
15:08:52DBUGEnqueued KICK [-AIR-]
15:09:12austriancoderhaha
15:11:16 Quit austriancoder (Client Quit)
15:13:06 Quit [-AIR-] ("—I-n-v-i-s-i-o-n— 2.0 Build 3515")
15:17:09amiconnWee! Mu asm-optimised 64 bit multiplication for SH1 is working. Speeds up experimental high-precision mandelbrot by a factor of 2.2 (!) :-)
15:19:17leftrightsounds like you are enjoying yourself
15:20:32XavierGrwhat is mandelbrot exactly. I read the wiki to no avail
15:21:41 Quit DreamTactix291 (Read error: 110 (Connection timed out))
15:22:17leftrightplease dont make sound audible during FF seeking
15:22:56leftrightI disliked that intensely with irivers firmware
15:24:17XavierGramiconn: I have an idea about the jpg viewer. When the user is in no zoom mode, he can press up or down to see the next or previous file in the current dir. What do you think?
15:24:42XavierGrI will try to implement it.
15:24:46amiconnI think that would break the viewer concept...
15:24:51XavierGrwhy?
15:25:10amiconnWell, a viewer is there to act on a file. *One* file
15:25:34 Quit BBub (Read error: 110 (Connection timed out))
15:25:34 Nick BBub_ is now known as BBub (belzebub16@dsl-084-059-236-248.arcor-ip.net)
15:25:36amiconnAnyway, I think this can be really tricky to implement.
15:26:15amiconnYou can extract the directory from the parameter, since all file arguments in rockbox use absolute path
15:26:40amiconnHowever, you have to parse the directory yourself, with several unknowns
15:26:42XavierGrnot SO tricky
15:26:57XavierGrIt will filter the files to jpg only
15:27:21amiconn(1) You don't know how large the directory is; possibly hundreds of entries, not all of which are .jpg
15:28:10amiconn(2) You don't know anything about the order the files are listed in the dir, meaning that you don't know at which position you find your current file
15:28:48amiconn(3) If you parse a directory, you retrieve the entries in the order as they are stored on disk, i.e. unsorted
15:29:46amiconnBasically this means to duplicate functionality of the file browser, and in order to be able to apply any sorting, you need a directory buffer, which probably has to be large
15:30:06XavierGrwell I will try to make a different function for 1 dir reading and sorting. Match the filename with currently open file and then scroll to the files by calling main(other_file)
15:31:09XavierGrit would be fun to have a small acdsee viewer
15:32:24XavierGrbut first I will have to make a dive to the directory, sorting, filename functions. I will have to know them better if I am going to do such thing
15:32:36Learbut don't call main. that means recursive, which can be nasty if you're not careful... :)
15:32:52XavierGrrecursive?
15:33:06XavierGroh I think I got it
15:33:19XavierGryou dont like recursive like gotos and such things
15:33:29XavierGrI love gotos
15:35:15 Join z-lined [0] (~rrr@202.43.230.235)
15:36:05XavierGrso by any means, lets say that I manage to do it. You seem that you don't like the idea of commiting such thing, do you?
15:36:27Learrecursive is not a goto. it is quite useful sometimes.
15:36:46XavierGrmisunderstood then
15:36:57Learbut it can also be a bit tricky to write recursive functions, before you get the hang of it...
15:38:52z-linedcould someone help me with a wps?
15:39:28XavierGrprev_file = "/Image/pic012.jpg"; does this makes a valid filename to the variable?
15:40:05z-linedi seem to have the top line (the battery and sound icon etc) displaying over the top of the wps, i'm using AruSahni wps
15:40:43XavierGrmove to options and remove the statusbar
15:41:35z-linedah great thanks
15:52:44 Join Rori [0] (MO-Pantsu@deadman3000.plus.com)
15:53:11RoriHi. Does anyone know if this has been addressed or has been confirmed as a bug? http://www.misticriver.net/boards/showpost.php?p=282919&postcount=1423
15:53:24 Quit ashridah ("sleep")
15:54:42Learsounds like the effect of the workaround of another bug... :)
15:55:30XavierGrw00t it worked. I gave a string for filename called gray_show(false) and then main(filname) and it loaded the picture perfect!
15:56:46 Join BePe_86 [0] (~kaka@ti311110a080-3468.bb.online.no)
15:58:39***Saving seen data "./dancer.seen"
15:58:48leftrighthmm, mp3 playback has crashed again, with latest build,
15:59:17XavierGrwhy so many people say that? I have the latest build and no problems
15:59:25XavierGrreset settings (backup first) and see
15:59:36XavierGrthough I doubt that this is to blame
15:59:43leftrightI'm not inaging it
15:59:57leftrightinaging=imagining
16:00
16:00:26 Join memmem [0] (~user@p54A22582.dip0.t-ipconnect.de)
16:01:47 Join muesli- [0] (muesli_tv@hmln-d9b8e188.pool.mediaWays.net)
16:02:39 Quit Lear ("Chatzilla 0.9.68.5 [Firefox 1.0.6/20050720]")
16:04:16XavierGrhi muesli-
16:04:59muesli-hi XavierGr, hi @ll
16:05:15 Quit einhirn ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
16:05:21 Join einhirn [0] (Miranda@carlsberg.heim2.tu-clausthal.de)
16:06:23leftrightright it ctashes with Replay Gain set to 'Album', its OK with RG set to 'Track'
16:07:29leftrightmp3, playback, RG set to Album, skip tracks forward, it crashes soon
16:14:17leftrightyep its definately the RG 'Album' setting which crashes my mp3 playback
16:15:16 Quit einhirn ("Miranda IM! Smaller, Faster, Easier. http://miranda-im.org")
16:15:20 Join einhirn [0] (Miranda@carlsberg.heim2.tu-clausthal.de)
16:25:54muesli-XavierGr could you please update your remote build :D
16:26:15XavierGrit is
16:26:55muesli-cheers :D
16:30:43XavierGrfrom now one look to the first post and check for the date. Sometimes I update it without posting, only editing
16:30:59muesli-just bokkmarked it
16:31:07muesli--k+o
16:31:10XavierGrI will be away to my village until thursday. So no updates till then.
16:31:42XavierGrThen when I come home, I will try to make the jpeg viewer to scroll pics inside a folder, without having to exit first.
16:31:49muesli-we will have to cope with it, but enjoy your stay there
16:31:58XavierGrthanks :)
16:32:02muesli-:D
16:32:16XavierGrbye all see you next week!
16:32:22muesli-cya
16:32:25 Quit XavierGr ()
16:33:09 Join Shebb [0] (~8243722b@labb.contactor.se)
16:33:17muesli-cya mates
16:36:13 Part Shebb
16:38:06 Join DreamTactix291 [0] (~dreamtact@adsl-19-235-161.bna.bellsouth.net)
16:40:02leftrightI cant duplicate this playback bug at will, and its bugging me
16:49:39leftrightI'm going to defrag my player and see if it makes a difference, *shrugs* who knows..
16:55:30 Quit muesli- (Read error: 113 (No route to host))
17:00
17:01:38 Join Tangleding [0] (~Tangledin@ARennes-351-1-35-68.w82-126.abo.wanadoo.fr)
17:22:52 Quit memmem ("ERC Version 5.0.3 $Revision: 1.726.2.17 $ (IRC client for Emacs)")
17:23:22 Part leftright
17:29:40 Quit Tangleding ("Chatzilla 0.9.68a [Firefox 1.0.4/20050511]")
17:41:08 Quit DreamTactix291 (Read error: 110 (Connection timed out))
17:43:12 Quit z-lined (Read error: 110 (Connection timed out))
17:58:41***Saving seen data "./dancer.seen"
18:00
18:28:12 Join sjutg [0] (~chaos@212.200.112.4)
18:28:53sjutgfuck you hi-hi-hi-hi-hi!!!
18:29:50sjutgcrash_ you are a trash hi-hi-hi-hi-hi!!!
18:30:30sjutgcrashd you are too hi-hi-hi-hi-hi!!!
18:31:34 Part sjutg
18:33:58Rorilol wth?
18:34:35Roridrama
18:35:30 Join Mhula [0] (~hdaalz@cpc1-stok3-4-0-cust240.bagu.cable.ntl.com)
18:36:07Mhulahi all, can anyone tell me what power adapter I should be using?
18:36:14Rori5v
18:36:26Mhulathanks :)
18:36:56Rorinp
18:37:19Roriguessing or iriver h1x0 so yes
18:37:21Slasherijust make sure you have the correct polarity
18:37:22Rorifor
18:37:32Slasherithat is very importCant or you will burn your player
18:37:57Slasherihmm.. are there something wrong with my irc-client?
18:37:57RoriI've swapped polarity befor on one of those interchangeably plugs and never had a burnout on anything
18:38:10Slasheridid you see typos on my last message?
18:40:01Rorithat sjutg just pm'd me asking where I live
18:41:17RoriI hate windows that steal focus
18:41:48Roriyou are typing something then boom! you don't know what you said ok to etc
18:42:42Slasherihmm, could you please copy/paste in private the last messages i said?
18:43:30RoriI was not here so sorry
18:43:37Slasherieh..
18:43:53Slasheriok, never mind
18:52:36 Quit TCK (Read error: 110 (Connection timed out))
19:00
19:02:20 Join DreamTactix291 [0] (~dreamtact@adsl-19-235-161.bna.bellsouth.net)
19:05:34 Join bagawk [0] (~lee@bagawk.user)
19:09:28 Quit courtc (Read error: 104 (Connection reset by peer))
19:10:16 Join leftright [0] (~5087db69@labb.contactor.se)
19:11:49leftrightRori: a several users have burnt out their players by swapping polarity, its bad advice to be nonchalant about polarity
19:14:53leftrightSlasheri: <Slasheri> that is very importCant or you will burn your player
19:18:22Slasherileftright: ok, so there is some bug with my irc client :/
19:18:29Slasherijust tried unloading a few scripts
19:19:26leftrightthey're obviously not keeping you busy in the army :)
19:19:54Slasherihehe, i just have ~two days of holiday ;)
19:19:59Slasheriweekends are free
19:20:10leftrightwhat, you've only been there 4 days
19:20:40Slasheriyes, 4 full days.. friday is a partial day and then free after that
19:20:51Slasheribut that's the "civil service"
19:21:01leftrightheh, whats the army coming too, next they'll give you tickets to vegas
19:21:11Slasherireal mil service is much harder and i would not like that so i selected this
19:21:17Slasheri:D
19:21:54Slasheriwe have 33 free travel-with-return tickets to anywhere in finland during that service ;)
19:25:19leftrightwhere do I sign up, do they have pretty girls in the army
19:26:02 Join courtc [0] (~courtc@adsl-158-45-197.asm.bellsouth.net)
19:34:41Mhulathe girls in the army are 'fit' :)
19:35:06Mhula.. on a non-related note, my sister used to be in the army..
19:41:59 Quit einhirn (Read error: 104 (Connection reset by peer))
19:43:54 Join austriancoder [0] (~5078751e@labb.contactor.se)
19:44:46 Join einhirn [0] (Miranda@carlsberg.heim2.tu-clausthal.de)
19:54:02 Join [-AIR-] [0] (~air@i-195-137-117-187.freedom2surf.net)
19:58:43***Saving seen data "./dancer.seen"
20:00
20:00:27 Join solex [0] (~jrschulz@d009251.adsl.hansenet.de)
20:01:32 Join ryan_j [0] (~ryanj@208.34.18.4)
20:03:22 Quit ryan_j (Client Quit)
20:05:25 Join ryan_j [0] (~ryanj@208.34.18.4)
20:12:19 Join koniu_ [0] (~koniu@host-ip194-145.crowley.pl)
20:15:07 Quit solex_ (Read error: 110 (Connection timed out))
20:17:51[-AIR-]hey people.
20:18:04DreamTactix291yo
20:18:16zewho're you calling people
20:18:28zeany relation to the band air?
20:21:08[-AIR-]no.
20:21:29[-AIR-]ze what's with the attitude<−−−−not meaning to be rude
20:21:37CtcpIgnored 2 channel CTCP requests in 9 seconds at the last flood
20:21:37*[-AIR-] is away eating chubby pies
20:21:54zeweird
20:21:58zethat could almost be a song
20:22:04ze*/ whats with the attitude
20:22:12zenot meaning to be rude
20:22:19zeair is away eating chubby pies */
20:23:45zethats a great line too
20:23:51ze'air is away eating chubby pies'
20:24:07zeit almost sounds like a social commentary on the state of the US
20:24:12zeor something
20:31:17DreamTactix291haha
20:42:34 Join lodesi [0] (~moi@d01v-62-34-91-138.d4.club-internet.fr)
20:49:13bagawkIf anyone is willing to try a custom rockbox version (for recorder v1 only atm): http://leepil.dyndns.org/rockboxv1.zip
21:00
21:06:34 Quit Mhula (" HydraIRC -> http://www.hydrairc.com <- IRC with a difference")
21:12:35bagawkServer with the custom bild is going down, If anyone else wants to try: email me bagawk _AT_ gmail d0t com I can give a build for any of the archos models :)
21:12:41 Quit bagawk ("Leaving")
21:24:28 Quit DreamTactix291 (Read error: 110 (Connection timed out))
21:27:02 Part leftright
21:35:17 Quit RotAtoR ()
21:38:05 Quit jwalk ("CGI:IRC")
21:58:48***Saving seen data "./dancer.seen"
22:00
22:00:22 Quit austriancoder ("CGI:IRC (EOF)")
22:07:28 Join Moos [0] (DrMoos@m20.net81-66-158.noos.fr)
22:07:49MoosHello all !
22:10:04Moosvery quite this night :)
22:11:07ryan_jYes it is, only it's afternoon for me
22:11:34Mooshehe scuse europe here (France)
22:17:35koniu_Moos: where in France are ya?
22:18:37Moosin Paris :)
22:18:41 Join austriancoder [0] (~5078751e@labb.contactor.se)
22:18:45austriancoderhi all
22:18:56MoosHello
22:19:04ryan_jhey
22:19:38koniu_a good place to be :) do you folks also have these terrible heat waves?
22:20:00austriancoder34 celsius here
22:20:14austriancodernow about 28 in night
22:20:30koniu_at least it's not only me who's being scorched ;)
22:20:47koniu_here it's still over 30 though
22:20:58koniu_(poland)
22:21:36Moos~30°C here
22:22:03Moosbut every day it's an other weather
22:22:09Mooshere :)
22:23:06RoriLOL http://www.boingboing.net/2005/07/28/microsoft_genuine_ad.html
22:23:22koniu_oh, i've recently came back from the isles, the weather there was different every 15 minutes :). but at least it wasnt ever over 35 and no a/c
22:24:52Moosin few weaks Morocco planed for me :)
22:26:16austriancoderan somebody download this file? http://www.freescale.com/files/platforms/doc/user_guide/SCF5250_USERS_MANUAL.pdf
22:26:31austriancoderby me it hags at 44%
22:27:09koniu_austriancoder: for me too
22:28:14koniu_but wget -c does the trick
22:28:26koniu_austriancoder: http://sheket.org/SCF5250_USERS_MANUAL.pdf
22:28:52ryan_jGotta love wget :-)
22:30:00 Join Musicmad [0] (~Musicmad@port547.ds1-oebr.adsl.cybercity.dk)
22:30:11austriancoderkoniu_: merci
22:30:27Musicmadhi all - any news on fm for iriver?
22:30:39*austriancoder will look into dma handling of iaudio
22:38:40 Join memmem [0] (~user@p54A211E4.dip0.t-ipconnect.de)
22:45:40 Join OnkelJonas [0] (vircuser@ip230.rev112.brygge.net)
22:49:59 Join DreamTactix291 [0] (~dreamtact@adsl-19-235-161.bna.bellsouth.net)
22:50:02 Join pit [0] (~pit@p54BD2C3F.dip0.t-ipconnect.de)
23:00
23:09:43 Join courtc_ [0] (~courtc@adsl-154-32-231.asm.bellsouth.net)
23:12:18OnkelJonas*yawn* ... i am teh bored :P
23:14:06 Quit courtc (Read error: 110 (Connection timed out))
23:24:50 Quit pit ("Verlassend")
23:27:45 Quit DreamTactix291 (Read error: 110 (Connection timed out))
23:31:09 Quit [-AIR-] ("—I-n-v-i-s-i-o-n— 2.0 Build 3515")
23:45:18 Quit ansivirus (Read error: 110 (Connection timed out))
23:48:21 Part Musicmad
23:52:14 Join leftright [0] (~5087db69@labb.contactor.se)
23:52:31leftrightanyone here to help me verify a bug ?
23:53:13leftrightthe bug is...
23:54:29leftrightfilw tree mode, add one song to the playlist, then whilst that song is playing right click the joystick as if selecting a new song
23:55:30leftrightit freezes the display and song, as if its paused, the only way to unfreeze it is to press stop
23:56:44ryan_jWorks fine here. Which build are you using?
23:56:55leftrighttodays
23:57:23ryan_jOh, is it iriver or archos?
23:57:26leftrightadd only one song to the playlist, not more
23:57:29leftrightiriver
23:57:33leftrighth140
23:58:52***Saving seen data "./dancer.seen"

Previous day | Next day