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 2012-10-01

00:09:44 Quit Totalled (Quit: PETTAN PETTAN, TSURUPETTAN!)
00:21:55 Quit factor (Read error: Connection reset by peer)
00:22:58 Join factor [0] (~factor@74.195.160.95)
00:26:04 Join the-kyle [0] (~kyle@cpe-024-211-185-030.nc.res.rr.com)
00:31:31 Join Totalled [0] (~Totalled@c-98-245-9-211.hsd1.co.comcast.net)
00:38:49lebelliumOk I opened a ticket for the RDS issues http://www.rockbox.org/tracker/task/12759
00:42:26 Quit bertrik (Remote host closed the connection)
00:44:46 Quit zoktar (Ping timeout: 260 seconds)
00:52:33 Quit lebellium (Quit: ChatZilla 0.9.89 [Firefox 16.0/20120925201946])
00:53:54 Join zoktar [0] (~zoktar@78-72-49-215-no186.tbcn.telia.com)
00:55:47saratogahttp://forums.rockbox.org/index.php/topic,38537.msg206054/topicseen.html#msg206054
00:56:00saratogaanother report of skins impacting USB stability
00:56:05saratogathis time on PP
00:57:01 Quit ender` (Quit: Because nothing says "I have faith in God" like 3 inches of bulletproof glass on popemobile.)
01:00
01:01:39 Quit mgottschlag (Ping timeout: 248 seconds)
01:18:41 Quit soap (Remote host closed the connection)
01:19:16 Quit shamus (Quit: chaos reigns within reflect repent and reboot order shall return)
01:21:27 Join shamus [0] (~shamus@ip-206-192-195-49.marylandheights.ip.cablemo.net)
01:25:02 Join soap [0] (~soap@cpe-174-102-110-153.woh.res.rr.com)
01:25:03 Quit soap (Changing host)
01:25:03 Join soap [0] (~soap@rockbox/staff/soap)
01:35:15 Quit Prodicus (Ping timeout: 248 seconds)
01:38:01 Join Prodicus [0] (~chatzilla@69.169.144.239.provo.static.broadweavenetworks.net)
01:52:46***Saving seen data "./dancer.seen"
02:00
02:00:15[Saint]saratoga: One thing I've noticed, that makes me want to look in the bufflib/skin-engine direction (actually, this started happening almost immediately after bufflib was introduced and used for the skin engine), is that all these troubles seem to involve the .sbs
02:00:29[Saint]kugel: (for logs) ^
02:00:41saratogayeah now that you mention it
02:02:31[Saint]also, this screams memory corruption issues: "I think it may be memory related, since sometimes upon successful (and unsuccessful) connects various graphics on the screen are progressively replaced by tutti-frutti colors and black lines."
02:03:06[Saint]lebellium(sp?) reported similar issues, but never managed to capture a screengrab (iirc), and no one else (that I know of) reported similar issues with screen corruption.
02:03:09[Saint]...until now.
02:04:48[Saint]There's no easy way to roll back and test this either, since most of the themes I'm aware of that fail in this fashion are quite modern and include skin tags that were introduced or modified after bufflib inclusion.
02:05:20[Saint]otherwise I'd just try a build pre-bufflib, but if I did so...the theme(s) would likely fail to parse.
02:05:48[Saint]cutting bufflib out of the picture to test my theory borders on impossibility.
02:06:06[Saint]well.../my/ possibilities.
02:09:37[Saint]afaik, JdGxrdon (don't wanna highlight him and send a mail) is just as clueless about it, but I'm not sure how personally involved he was with the skin-engine/bufflib integration.
02:10:25[Saint]I would *love* to reproduce this, but I just can't. :-(
02:13:55 Quit Rower85 (Quit: Hmmm...)
02:15:35 Quit prof_wolfff (Ping timeout: 244 seconds)
02:16:12[Saint]gevaerts: you around?
02:16:27*[Saint] thinks it may be a bit late for Euros, possibly, but tries anyway.
02:18:26 Quit saratoga (Quit: Page closed)
02:25:51 Join Guinness` [0] (Slayer@c-68-55-111-159.hsd1.va.comcast.net)
02:27:07 Quit Guinness (Ping timeout: 260 seconds)
02:27:08 Nick Guinness` is now known as Guinness (Slayer@c-68-55-111-159.hsd1.va.comcast.net)
02:38:52 Quit sakax (Remote host closed the connection)
02:51:06 Join perrikwp [0] (~quassel@cpe-075-177-082-185.triad.res.rr.com)
03:00
03:20:04 Quit thegeek (Read error: Connection reset by peer)
03:21:39 Join thegeek [0] (~thegeek@171.17.9.46.customer.cdi.no)
03:30:08 Quit XavierGr ()
03:52:48***Saving seen data "./dancer.seen"
04:00
04:14:09 Quit amiconn (Disconnected by services)
04:14:09 Join amiconn_ [0] (amiconn@rockbox/developer/amiconn)
04:14:09 Join pixelma_ [0] (pixelma@rockbox/staff/pixelma)
04:14:09 Quit pixelma (Disconnected by services)
04:14:14 Nick amiconn_ is now known as amiconn (amiconn@rockbox/developer/amiconn)
04:29:47*[Saint] just offered up a contender for "most longwinded forum post".
04:30:11[Saint]on the subject of: Themes - Doing It Right (TM)
04:38:00 Join TheSphinX_ [0] (~briehl@p579CC360.dip.t-dialin.net)
04:41:10 Quit TheSphinX^ (Ping timeout: 252 seconds)
04:50:00amayerHAHA
04:50:02amayer- (line 23) is pretty funny
04:51:01 Quit Epicanis (Quit: I need to go to bed...)
05:00
05:15:21[Saint]amayer: there's some weird voodoo going on in there.
05:19:55amayerI didnt look at it but from what you posted it doesnt sound good
05:21:30 Quit [7] (Disconnected by services)
05:21:38 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
05:35:43 Quit ps-auxw (Remote host closed the connection)
05:45:26 Join ps-auxw [0] (~arneb@p4FF7F89F.dip.t-dialin.net)
05:50:12 Join mulrich [0] (~where.are@130.225.243.68)
05:51:10mulrichHello. I'm having some issues with Rockbox for Android and playlists
05:52:12mulrichI've placed my playlists in the Playlist directory (so they actually show up in the Playlist Catalogue), but that means they no longer work
05:52:23mulrichSo I've tried adding the full relative path
05:52:40mulrichDidn't work. Tried adding <sdcard> in front of them. Didn't work either
05:52:50***Saving seen data "./dancer.seen"
05:53:19mulrichWhat on earth am I doing wrong?
05:57:43mulrichThe device is a Cowon D3 (hence Rockbox for Android)
06:00
06:24:43 Quit eckoit (Quit: eckoit)
06:28:18[Saint]when you say "the full path", what do you mean exactly?
06:28:49 Part amayer
06:29:05[Saint]"/mnt/sdcard/Music/Artist/Album/Track.ext" should work.
06:30:03mulrichHold on, lemme try that
06:30:39[Saint]Alternatively, if you want to know the format the playlists should take. Make a playlist from a file you know is in the external storage, and examine the path in that playlist.
06:31:00mulrichDoes it have to be /
06:31:05mulrichWon't \ work?
06:33:33[Saint]Only if you're using it to escape a space in a path name. As in "/mnt/sdcard/Music/My\ Directory\ Name\ With \Spaces\ In\ It/Track.ext"
06:33:40mulrichFUCK YES!
06:33:42mulrichIt worked!
06:33:50mulrichWith blackslash
06:34:26mulrichThanks man
06:35:18mulrichJust coming off the iPod teat
06:35:31mulrichOnly Apple product I've ever owned
06:35:48mulrichOnly one I ever thought worked properly (mostly)
06:42:01[Saint]A rockbox'ed iPod, one hopes :)
06:44:22mulrichNope
06:44:43mulrichBack when I started out with iPods in 2004, I didn't know about the existence of Rockbox
06:44:56mulrichBy the time I did, my whole music collection was already set up in iTunes
06:45:09mulrichPlaylists and everything
06:45:20[Saint]Rockbox has no problem playing from the iTunes Database...
06:45:32[Saint]TaDa! :)
06:46:44[Saint]Assuming the iPod isn't a Nano 3/4/5/6/7G, Rockbox supports it. And it will happily list tracks synced by iTunes.
06:47:14[Saint]list *and* play, even. (just listing them isn't really that cool... :))
06:48:33*[Saint] thinks he may have just made someone have a "d'Oh!" moment.
06:48:41[Saint]...sorry :)
06:53:24mulrichMy first iPod was a mini
06:53:35mulrichThen I got a 2nd gen nano
06:53:50[Saint]Ooooooh! Two of the best.
06:53:58mulrichBefore having my "current" nano 4th gen
06:54:25mulrichIMO, the last proper nano was the 5th gen
06:54:26[Saint]The Nano2G is really quite insanely powerful. Shame about the Nano4G, though.
06:54:46mulrichThe 6th and 7th gens are just downgrades
06:55:18mulrichMy 4th gen can no longer hold battery
06:55:28mulrichAnd the lock switch is irratic
06:55:33mulrich*erratic
06:55:40[Saint]The newest one is getting back on track. It is back to the form factor of the original Nano, but kept the touchscreen.
06:55:42mulrichIt'll randomly turn off and on
06:56:16[Saint]The newest Nano is really quite cool, but...it doesn't interest me as a target due to the lack of Rockboxification.
06:57:14mulrichWell, shit, that nano wasn't available when I ordered this Cowon D3
06:57:57mulrichI might just go fuck all, and buy one of those nanos when I have money again
06:58:55[Saint]If we continue this line of conversation, it may be best to jump into #rockbox-community
06:59:26mulrichNaw, I should think about getting some sleep
06:59:28mulrichIt is currently 06:59:26
06:59:36[Saint]o/
07:00
07:22:46 Join eckoit [0] (~ryan@50.65.10.24)
07:37:54 Quit shamus (Read error: Connection reset by peer)
07:49:49 Join mortalis [0] (~mortalis@195.34.194.126.kalibroao.ru)
07:52:54***Saving seen data "./dancer.seen"
07:59:07 Join stoffel [0] (~quassel@pD9E43A43.dip.t-dialin.net)
08:00
08:04:17 Quit ps-auxw (Ping timeout: 245 seconds)
08:06:19 Join ps-auxw [0] (~arneb@2001:470:c807:0:1532:4e5f:2ad3:4123)
08:20:06 Join shamus [0] (~shamus@ip-206-192-195-49.marylandheights.ip.cablemo.net)
08:29:07 Quit funman (Ping timeout: 248 seconds)
08:29:35 Join funman [0] (~fun@rockbox/developer/funman)
08:36:25 Join funman_ [0] (~fun@rockbox/developer/funman)
08:37:18 Join Buschel [0] (~chatzilla@p57905E53.dip.t-dialin.net)
08:39:06 Join homielowe [0] (~1854d8a9@www.haxx.se)
08:39:15 Quit funman (Ping timeout: 248 seconds)
08:39:22homielowehttp://hardware.slashdot.org/story/12/10/01/060205/rockbox−−refurbished-mp3-players−−crowdsourced-audio-capture :)
08:40:23homieloweI'd add it to http://www.rockbox.org/wiki/ArticlesAboutRockbox but I've forgotten my wiki password ;) It's been so long.
08:40:31 Quit homielowe (Client Quit)
08:42:18 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
08:42:40 Join ender` [0] (krneki@foo.eternallybored.org)
08:44:54 Join pamaury [0] (~quassel@vit94-1-82-67-248-70.fbx.proxad.net)
08:44:54 Quit pamaury (Changing host)
08:44:54 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
08:53:51bertrikeh, wait what, rockbox on slashdot again
08:58:29 Quit kevku (Quit: KVIrc 4.2.0 Equilibrium http://www.kvirc.net/)
09:00
09:04:13 Join petur [0] (~petur@rockbox/developer/petur)
09:05:33 Join Zagor [0] (~bjst@sestofw01.enea.se)
09:05:33 Quit Zagor (Changing host)
09:05:33 Join Zagor [242] (~bjst@rockbox/developer/Zagor)
09:08:35 Quit Prodicus (Ping timeout: 248 seconds)
09:11:28 Join mgottschlag [0] (~quassel@reactos/tester/phoenix64)
09:16:29 Join einhirn [0] (~Miranda@p4FC74767.dip0.t-ipconnect.de)
09:17:19 Join einhirn_ [0] (Miranda@bsod.vpn.tu-clausthal.de)
09:21:07 Quit einhirn (Ping timeout: 256 seconds)
09:21:47 Quit bertrik (Ping timeout: 245 seconds)
09:49:14 Quit pamaury (Ping timeout: 244 seconds)
09:52:57***Saving seen data "./dancer.seen"
10:00
10:07:26 Join kenty [0] (~d590c461@www.haxx.se)
10:07:37kentyhello people
10:07:58 Quit mgottschlag (Ping timeout: 260 seconds)
10:08:09kentyI wanted to use one of the theme on the site, but didn't works. Is that a known issue? What shoukd I do?
10:09:22kentyDid you guys changed something on the code that breaks them?
10:10:39 Join Rower85 [0] (husvagn@v-413-alfarv-90.bitnet.nu)
10:14:39 Join wodz [0] (~wodz@iwl138.internetdsl.tpnet.pl)
10:20:42 Join lebellium_gs2 [0] (~lebellium@tmo-103-75.customers.d1-online.com)
10:22:11lebellium_gs2Kenty: maybe telling which theme would help the dev
10:23:35lebellium_gs2kenty *
10:25:52kentyI tried with few san sa fuze plus ones
10:26:01kentybut maybe all themes are broken?
10:27:41Zagorvery cool video, using sansa clips w/ rockbox as collar microphones to record dinner conversation (abouth molten salt reactors)
10:30:31lebellium_gs2 kenty : maybe they require the font pack? (usually written in the theme description)
10:33:46 Join mgottschlag [0] (~quassel@reactos/tester/phoenix64)
10:36:01kentylebellium_gs2: thx to trying help me. Yes I dowloaded a freshh font pack when tried
10:36:56lebellium_gs2And fresh build too?
10:37:08 Quit factor (Read error: Connection reset by peer)
10:39:00 Quit mgottschlag (Ping timeout: 246 seconds)
10:40:02kentylebellium_gs2: yeah of course
10:41:30kentyhere an exemple of theme that fail http://themes.rockbox.org/index.php?themeid=256&target=sansafuzeplus
10:42:00 Join mgottschlag [0] (~quassel@reactos/tester/phoenix64)
10:42:07kentyis Alex Parker around? Maybe he can help me
10:42:38lebellium_gs2 kenty : try my theme in attachment of this bug tracker. It should work on a fuze+ http://www.rockbox.org/tracker/task/12759
10:42:39kentyAlex Parker as AlexP?
10:47:01[Saint]I'm looking at the code now, contemplating how energetic I'm feeling and whether or not to compile a sim and check it out.
10:47:27[Saint]The theme code is old, however, it doesn't look suspect to me (yet, at least...but I'm pretty tired).
10:47:34[Saint]I'll compile a sim and take a look.
10:48:25[Saint]kenty: what target are you attempting to load the theme on?
10:48:25[Saint]ie. which device?
10:48:39 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
10:49:56 Join LinusN [0] (~linus@giant.haxx.se)
10:50:41kenty[Saint]: Hey, Tanks you
10:50:46kentysansa fuze+
10:51:34 Join lebellium [0] (~chatzilla@e179142095.adsl.alicedsl.de)
10:51:54 Quit lebellium_gs2 (Quit: Bye)
10:53:32[Saint]ah...shit. seems I don't have the toolchain compiled here.
10:54:12 Join factor [0] (~factor@74.195.160.95)
10:54:35kentyAnd I don't know how to compile myself duh :(
10:54:57kentythanks anyway for the try [Saint]
10:55:11 Part LinusN
10:55:13 Quit nosa-j (Ping timeout: 260 seconds)
10:56:03[Saint]Hmmm? oh, no. It's fine. Compiling now. It'll only taker a few minutes for the toolchain, and a trivial amount for the sim itself.
10:57:13kentyCool, ty
11:00
11:01:04 Join Holyjack [0] (~595d59bc@www.haxx.se)
11:02:16Holyjackhello all
11:02:46HolyjackMy Sansa Clip+ has been dead for a week now
11:03:00Holyjackhow can I be sure if it's bricked or not ?
11:03:04 Join nosa-j [0] (~m00k@184.76.254.130)
11:03:33[Saint]Smashing it with a hammer, vigorously, is a good way...so I hear.
11:03:38[Saint]:)
11:04:45HolyjackI'm using current stable release of Rockbox
11:04:46[Saint]In all seriousness, though. Can you assure that the battery is flat?
11:04:46Holyjackha ha :)
11:04:55HolyjackI guess that is what I want to know :)
11:05:13HolyjackI mean, it's only 3 mm high so that's pretty flat :p
11:05:55wodzHolyjack: I recall someone suggested to plug earphones. If you hear pop/click it mean the device is still running
11:06:16HolyjackI'll try that
11:06:37Holyjackusing my earphones as a stethoscope
11:07:11[Saint]A good way to see that the battery is indeed depleted, and the device is powered off, and not simply hanging, is to hold the power for ~30 seconds and then insert some earphones (or, audio out to the stereo, etc.) and listen for an audible "click".
11:08:13Holyjackno click, no pop
11:08:28Holyjackone more piece of info: when I plug my device, windows says it needs formatting
11:08:40Holyjackit prompts me to start formatting a 4 mb drive
11:08:44Holyjackbut of course formatting fails
11:09:04[Saint]does the device react at all to USB plug?
11:09:14Holyjackno
11:09:19HolyjackWindows yes, the device no
11:09:44wodz4mb is it the magic 'recovery' drive?
11:09:59Holyjackno idea
11:10:14[Saint]Right, yes. Ok...hmmm, yep. It is "bricked", but it is quoted because it is not an exact science.
11:10:17[Saint]not at all.
11:10:24Holyjackit's the only option in the "format drive" window
11:10:33[Saint]We've no idea what the 4MB partition is for, as far as I'm aware.
11:10:37[Saint]No one has worked it out.
11:11:06Holyjackshoot
11:11:09[Saint]No no...recover is a pretty involved process. It's not getting fixed by simply hitting "format".
11:11:26Holyjackyes I read that in the forums
11:11:28[Saint]There is a chance it can be recovered, though. Let me find the info.
11:11:44Holyjackshall I wait for you here ?
11:11:55[Saint]I'll only be a minute or so.
11:12:05Holyjackok thanks !
11:14:06 Join LinusN [0] (~linus@giant.haxx.se)
11:14:50 Join Thra11 [0] (~thrall@119.133.125.91.dyn.plus.net)
11:16:40[Saint]http://www.rockbox.org/wiki/SansaAMSUnbrick
11:16:44[Saint]uuuugh, sorry.
11:17:01[Saint]Chrome managed to eat a whole CPU core and ~8GB of RAM...
11:17:10[Saint]ground my laptop to a halt, lol.
11:18:00Holyjackok so I'm going to have to open my Sansa again
11:18:09[Saint]Anyway, that's it...therein is the combined knowledge of the 4 or 5 people that ever managed to recover a device and document the process.
11:18:19Holyjack4 or 5 ? Only ?
11:18:23[Saint]Please note the warnings.
11:18:34Holyjackis it because unbricking almost fails or because bricking is rare ??
11:19:13[Saint]It's because we've no idea how recovery works, really. It was discovered by accident.
11:19:30[Saint]No one really actually knows the exact combination of magic required.
11:20:19Holyjackok
11:20:38[Saint]As far as I'm aware, the people I know of, including myself have all recovered these devices either fully or partially (some end up without any ability to boot the original firmware), have gone about it a slightly different way.
11:20:56Holyjackwhat kind of screwdriver will unscrew the PCB ?
11:21:01Holyjackhow different ?
11:22:21[Saint]For example, there seems to be 2, possibly 3, different types of recovery (or, ...some form of advanced usage) partitions that can be exposed.
11:23:31Holyjack...and there's no way to tell in advance which one I'll get?
11:23:33[Saint]there's a 4MB partition that no one knows what it is. There's a 32 MB partition that can be exposed that I believe two people have had success with simply dd'ing a stock firmware file to.
11:24:11[Saint]and then there's a ~900 (iirc) MB partition, that as far as I'm aware is the "recovery partition".
11:24:16kenty[Saint]: No luck on that theme?
11:24:51Holyjackwhat does that mean, "recovery partition" ?
11:25:22[Saint]And, it can differ two ways here, also. Some people have had success with dd'ing a stock firmware image to this partition, and others like myself, needed to dd in a rockbox patched firmware and lost the ability to boot the OF entirely.
11:26:10[Saint]We think this may be because some magic the OF needed got trashed.
11:27:02[Saint]The reason I quoted "recovery partition" is because I'm not entirely certain that this is the function of this partition at all...at least not the way(s) we've found to (ab)use it.
11:27:30Holyjackok
11:28:03Holyjackso I might as well try to unbrick it...it can't kill the player more than it already is...
11:28:17[Saint]pretty much, yes.
11:28:20Holyjackare these Philipps screwdrivers I need to the PCB ?
11:28:35[Saint]at this point you're not really losing anything, and can only gain if you're successful.
11:28:54[Saint]yeah, an 00 phillips iirc.
11:29:23Holyjackok thanks
11:30:05Holyjacknow I'm pretty reluctant to Rockbox my Sansa again
11:30:14Holyjackor my next Sansa if I have to buy a new one
11:30:23Holyjackcan I at least try to understand what went rwrong ?
11:30:24kentythat just bad luck
11:30:34[Saint]fyi: a pair of metal tweezers is awesome for bridging the pins you'll need to find later on.
11:30:55kentyRockbox isn't known to brick devices, quite rare cases
11:31:57HolyjackI was told here last week that the latest stable release has bug that disables hard reset
11:32:17Holyjackthat v 3.11 I think
11:34:22 Join lorenzo92 [0] (~chatzilla@host212-110-dynamic.17-79-r.retail.telecomitalia.it)
11:34:48wodzHolyjack: the bug manifested it that you had to wait for battery drainout
11:35:31[Saint]yeah, it's not what you're seeing...with that bug once the battery depleted the device could be charged and it would boot again.
11:36:38Buscheln1s: i made some more optimizations to use OPUS_COPY/OPUS_CLEAR instead of loops in the mdct. i also allocation implementation of my latest patch for all other iram buffers as well. PP needs ~69.5 MHz now -> http://pastie.org/4890247
11:37:38Buscheln1s: i am not sure if iram will be full on cf, let's see ;)
11:39:06Holyjackthe device stopped reponding after I forced it off when the databse re-initialisation was taking way longer than I thought it should
11:39:13Holyjackabout 30mn for 15gb
11:40:40[Saint]kenty: tested here, and...the theme works fine for me.
11:41:04[Saint]using a freshly compiled build...it loads as intended.
11:43:17kentyoh! let's try again. That one right, http://themes.rockbox.org/index.php?themeid=256&target=sansafuzeplus
11:43:34 Join Prodicus [0] (~chatzilla@69.169.144.239.provo.static.broadweavenetworks.net)
11:43:49[Saint]That isn't to say there isn't a problem, but if there is a problem, it isn't one I can test. Because the problem wouldn't be with the theme itself, I know that loads fine, and I don't have the physical hardware to test the real world failuer on.
11:44:46[Saint]and, yes. that theme, I'm playing audio with it via my desktop now.
11:45:04[Saint]menus and wps are all fine here.
11:48:02Holyjackthanks guys anyway
11:48:07HolyjackI'll go and try to recover my player
11:48:12Holyjacksee you around
11:48:15 Quit Holyjack (Quit: CGI:IRC)
11:48:27 Quit fs-bluebot (Ping timeout: 245 seconds)
11:48:53 Quit bluebrother^ (Ping timeout: 260 seconds)
11:49:41 Join fs-bluebot [0] (~fs-bluebo@g224236247.adsl.alicedsl.de)
11:49:42 Quit nosa-j (Ping timeout: 246 seconds)
11:50:33 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
11:53:01***Saving seen data "./dancer.seen"
11:57:42 Join nosa-j [0] (~m00k@184.76.254.130)
11:58:20 Quit Zagor (Quit: Clint excited)
11:59:07 Join rameshwor [0] (~rameshwor@182.50.64.67)
12:00
12:00:14rameshworis rockbox supported in ipodo-nano-4G ?
12:00:51rameshwori wanted to give it a try , and just to be safe if something goes wrong , can it be removed and ipod be changed to previous state ?
12:03:16Torneno, we don't have a build for that
12:03:20 Quit nosa-j (Ping timeout: 245 seconds)
12:07:17 Quit kenty (Quit: CGI:IRC)
12:11:35 Join nosa-j [0] (~m00k@184.76.254.130)
12:21:33 Quit rameshwor (Ping timeout: 246 seconds)
12:35:02 Join benedikt93 [0] (~benedikt9@unaffiliated/benedikt93)
12:39:45 Quit lorenzo92 (Ping timeout: 260 seconds)
12:43:05 Quit shai (Read error: Connection reset by peer)
12:59:39 Join kevku [0] (x@indeed.tastes.like.everything.mm.am)
13:00
13:03:54 Quit petur (Quit: *plop*)
13:05:53 Quit ChanServ (*.net *.split)
13:07:11 Join ChanServ [0] (ChanServ@services.)
13:07:11Mode"#rockbox +o ChanServ " by leguin.freenode.net
13:14:59 Quit mc2739 (Ping timeout: 252 seconds)
13:23:24 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
13:27:51 Quit mc2739 (Ping timeout: 246 seconds)
13:30:06 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
13:41:39 Nick funman_ is now known as funman (~fun@rockbox/developer/funman)
13:42:55 Join Thra11_ [0] (~thrall@84.93.191.26)
13:43:48 Quit mc2739 (Ping timeout: 246 seconds)
13:45:21 Quit Thra11 (Ping timeout: 246 seconds)
13:45:53 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
13:46:34wodzTorne: reading elf spec it seems it is possible to mark specially program header with PT_NOTE which is optional and vendor specific
13:48:53wodzhmm, maybe not for this what we want
13:53:03***Saving seen data "./dancer.seen"
13:59:33 Quit mc2739 (Ping timeout: 240 seconds)
14:00
14:01:45 Join mc2739 [0] (~mc2739@71.20.87.137)
14:01:48 Quit mc2739 (Changing host)
14:01:48 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
14:04:16 Quit Buschel (Ping timeout: 245 seconds)
14:13:06Tornewodz: PT_NOTE is a different type of header, though. there aren't any special flags that i know of..
14:14:17wodzTorne: yeah, I realize this. The spec is a bit confusing
14:17:53 Join dfkt [0] (dfkt@unaffiliated/dfkt)
14:17:59wodzTorne: Do I read correctly that we don't need strtab actually? only symtab will be needed for relocation, right?
14:18:18Torneyou shouldn't need that either
14:18:37Tornei forget :)
14:18:41Tornebut id on't think you need the symbol table at all
14:23:43 Quit nosa-j (Ping timeout: 256 seconds)
14:27:13 Join nosa-j [0] (~m00k@184.76.254.130)
14:28:33wodzTorne: but you need to know symbol value, no?
14:28:44Tornewhy?
14:29:22*Torne has a look
14:30:29 Quit einhirn_ (Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org)
14:31:48Torneoh, yeah.
14:31:57Tornei am confusing the dynamic stuff
14:33:07Tornenot all reloc types need the symbol value
14:33:18Tornebut we will for some
14:42:06 Quit factor (Quit: Leaving)
14:46:47 Join c4ptotc [0] (~c4pt-otc@c-68-80-190-181.hsd1.pa.comcast.net)
14:47:15c4ptotchey i am having trouble getting my ipod classic 160gb into DFU
14:47:45c4ptotcit doesnt seem to do it when i hold menu + ok it restarts and shows a mac apple but then it continues to boot
14:50:00wodzyou mean you are still holding menu+ok after logo show up?
14:50:32benedikt93you need to hold those keys until the screen goes black a second time
14:51:58c4ptotci got it i think
14:52:09c4ptotci just kept holding it past the apple logo
14:52:56 Join willmore [0] (~willmore@c-98-220-135-28.hsd1.in.comcast.net)
14:53:50willmoreSorry for helping slashdot the channel today. Does anyone know what chipset the Eclipse T180 uses? www.eclipsemp3.com uses?
14:57:19willmoreTaking it apart and the most likely chip is by Anyka with part # AK1011Q06405.
14:58:45 Quit nosa-j (Ping timeout: 252 seconds)
14:59:13wodzwillmore: post a photo, maybe someone will know
15:00
15:00:37 Join nosa-j [0] (~m00k@184.76.254.130)
15:02:21willmorewodz, stolen from their web site: http://www.eclipsemp3.com/wp-content/uploads/2011/06/t180.jpg
15:02:40Tornewillmore: a photo of the PCB
15:02:47Tornenot a marketing shot of the device
15:03:17willmoreTorne, do'ah!
15:03:58Tornein high enough resolution/light to read chip markings/etc. a flatbed scanner works well but a photo can be ok if your camera is decent :p
15:05:09 Part c4ptotc ("Leaving")
15:05:13willmoreAny place you guys prefer photos to be put?
15:06:39Tornepost on the forum is most useful
15:07:18wodzhaha, It will be Anyka AK1011 SoC. http://www.anyka.com/enProShow.asp?id=106&sortName=ApplicationProcessor&sortFlag=113
15:07:29wodzARM926ej-s based
15:07:42willmorewodz, yeah, that's the guy.
15:09:46wodzgood luck finding any datasheet
15:10:03willmoreOkay, I'll screw it back together and move on. ;)
15:11:38wodzwithout documentation it will be tremendous task to port rb to this platform.
15:12:05willmoreI understand. The companies that make these don't tend to be very supportive of 'independent' uses for their products.
15:13:58willmoreAre there any of those '$20 at fry's' types of devices that RB supports?
15:14:11willmoreIt seems that most of the devices are pretty high end.
15:14:20willmoreWell, the Sansa line can be had for cheap.
15:15:31wodzTorne: Does .rel.xxx section is placed always after .xxx?
15:15:44Tornei wouldn't count on it
15:19:11willmoreThanks, wodz, Torne.
15:21:19 Quit wodz (Quit: Leaving)
15:22:18 Part willmore ("Leaving")
15:33:21 Quit n17ikh (Ping timeout: 246 seconds)
15:33:37 Join amayer_ [0] (~alex@mail.weberadvertising.com)
15:34:37 Quit Zambezi (Changing host)
15:34:37 Join Zambezi [0] (Zulu@unaffiliated/zambezi)
15:40:01 Join n17ikh [0] (~n17ikh@128.204.195.239)
15:40:01 Quit n17ikh (Changing host)
15:40:01 Join n17ikh [0] (~n17ikh@unaffiliated/n17ikh)
15:44:13 Quit mc2739 (Ping timeout: 252 seconds)
15:46:03 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
15:53:07***Saving seen data "./dancer.seen"
16:00
16:04:04 Quit pamaury (Ping timeout: 255 seconds)
16:07:33 Quit mortalis (Quit: Leaving)
16:36:52 Join Keripo [0] (~Keripo@c-76-22-63-234.hsd1.wa.comcast.net)
16:36:53amayer_has anyone else had problems with adding file to a folder then when you resume play it is one a different audio file?
16:36:55amayer_i was playing a song. paused it. added another song(different name) to the same folder. ejected my ipod. then it was playing a different song(but oddly it was at the same time in the new song as it was in the old song).
16:42:10 Join WalkGood [0] (~4@unaffiliated/walkgood)
16:46:33 Quit Keripo (Read error: Connection reset by peer)
16:52:21 Join y4n [0] (~y4n@unaffiliated/y4ndexx)
16:52:59Torneamayer_: directory based dynamic playlists just store the directory and the index, so if you add/remove files it will resume on the *new* 9th file or whatever
16:53:02Torneat the same time position
16:53:24Tornethis has been this way for a long time :)
16:53:44 Join Buschel [0] (~chatzilla@87.144.94.83)
16:59:01 Join pamaury [0] (~quassel@vit94-1-82-67-248-70.fbx.proxad.net)
16:59:01 Quit pamaury (Changing host)
16:59:01 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
17:00
17:02:24 Quit eckoit (Quit: eckoit)
17:17:41 Quit mulrich (Quit: [19:19:01] <infernus> I love sucking dick!)
17:23:14 Part LinusN
17:31:34 Join Keripo [0] (~Keripo@c-76-22-63-234.hsd1.wa.comcast.net)
17:35:43 Join eckoit [0] (~ryan@96.53.108.182)
17:42:07 Quit Keripo (Quit: Leaving.)
17:53:08***Saving seen data "./dancer.seen"
17:55:40amayer_Torne: thank you. ive never experienced it before. it struck me as kind of odd.
17:55:43amayer_is this the way it was intended to function?
17:55:48 Quit mgottschlag (Ping timeout: 246 seconds)
17:56:12Torneit's not intended to resume the wrong track. but it's not a bug, it's implemented the way it's implemented.
17:56:24Tornewrite a better way :)
17:58:26 Quit kevku (Ping timeout: 260 seconds)
17:59:06 Join kevku [0] (x@indeed.tastes.like.everything.mm.am)
17:59:30 Quit Buschel (Ping timeout: 252 seconds)
18:00
18:16:25amayer_i just wanted to know if the devs knew about it. im fine with it. i dont add files too often(besides podcasts)
18:18:59 Quit ender` (Ping timeout: 248 seconds)
18:41:25amayer_another question: is there a way to adjust the padding between the scroll bar and text?
18:41:26amayer_I couldnt find anything on google or theme documentation.
18:41:36 Join HJ [0] (~595d59bc@www.haxx.se)
18:41:36 Quit HJ (Client Quit)
18:41:48 Join Holyjack [0] (~595d59bc@www.haxx.se)
18:42:39Holyjackhello all
18:42:59HolyjackI'm trying to recover my bricked Sansa Clip+
18:43:21Holyjackbut all I can get Windows to see is a 4mb partition
18:44:00Holyjackwhether I bridge the appropriate pins or not
18:44:09Holyjackso I need help :)
18:59:48 Join n1s [0] (~n1s@nl118-168-30.student.uu.se)
18:59:48 Quit n1s (Changing host)
18:59:48 Join n1s [0] (~n1s@rockbox/developer/n1s)
19:00
19:02:25 Quit eckoit (Quit: eckoit)
19:06:45 Quit guymann (Ping timeout: 255 seconds)
19:09:08pamauryHolyjack: please wait a bit and ask later, the knowledgable people for that issue might not be online at the moment
19:10:08Holyjackok
19:10:23 Join eckoit [0] (~ryan@96.53.108.182)
19:10:37Holyjack[Saint] helped me earlier today; he looks online but probably busy at the moment
19:10:40HolyjackI'll come back later
19:14:44 Quit Holyjack (Quit: CGI:IRC)
19:16:24 Join pretty_function [0] (~sigBART@123.252.213.253)
19:17:05 Quit amayer_ (Quit: going ~/)
19:17:21 Join amayer_ [0] (~alex@mail.weberadvertising.com)
19:19:57n1sBuschel: cool, why did you move the f buffer from the stack though?
19:22:20 Quit amayer_ (Quit: going ~/)
19:22:36 Join amayer_ [0] (~alex@mail.weberadvertising.com)
19:22:53 Quit n1s (Read error: Connection timed out)
19:23:02 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
19:26:44 Join ender` [0] (krneki@foo.eternallybored.org)
19:35:32 Join n1s [0] (~n1s@nl118-168-30.student.uu.se)
19:35:32 Quit n1s (Changing host)
19:35:32 Join n1s [0] (~n1s@rockbox/developer/n1s)
19:36:01 Join Buschel [0] (~chatzilla@p57905E53.dip.t-dialin.net)
19:37:50 Join Holyjack [0] (~595d59bc@www.haxx.se)
19:43:58 Quit Holyjack (Quit: CGI:IRC)
19:44:07 Join Holyjack [0] (~595d59bc@www.haxx.se)
19:45:43n1s<n1s> Buschel: cool, why did you move the f buffer from the stack though?
19:46:24Buscheln1s: no technical reason. I just liked to have a similar implementation for both arrays.
19:46:41Buscheland on PP it's a bit faster (~0.4 MHz)
19:46:51n1sthat's weird
19:47:11BuschelI measured it 2 days ago
19:48:50n1sif iram gets tight on chips with the small iram (some older pp's have the same amount as 5249 iirc) i'd prefer using the stack for one of these buffers to just not having it in iram
19:49:04 Quit WalkGood (Quit: ♪ ♫ ♪ ♫ ♪ ♫ ♪)
19:49:15Buschelyou're free to adapt the patch, of course
19:50:21Buschelbut i am more interested in the impact of the other changes :)
19:50:37Buscheli will be off for 15 mins now
19:53:09***Saving seen data "./dancer.seen"
19:54:55 Quit benedikt93 (Quit: Bye ;))
19:55:16 Join liar [0] (~liar@clnet-p09-185.ikbnet.co.at)
20:00
20:02:26Holyjackhello all. i have asked for help earlier but no one was available so i'm asking again now: can anyone here help me with the unbricking procedure for the SANSA CLIP PLUS. The procedure on http://www.rockbox.org/wiki/SansaAMSUnbrick does not work for me
20:09:45 Join einhirn [0] (Miranda@vpn176.vpn.tu-clausthal.de)
20:12:35 Quit stoffel (Remote host closed the connection)
20:12:49 Quit Holyjack (Quit: CGI:IRC)
20:16:29Buscheln1s: did you already have time to apply the patch?
20:24:02 Join Holyjack [0] (~595d59bc@www.haxx.se)
20:30:06n1sno, had to do some other things
20:32:30Buschelyep, real life needs quite some time as well
20:32:52 Join sakax [0] (~sakax@d8D862498.access.telenet.be)
20:33:12n1si did look at it though, did you check if the changes in the mdct rotate loops actually made a difference?
20:33:34n1sseems to me like gcc should be smart to cse that shift and load
20:33:35 Join saratoga [0] (123e0cca@gateway/web/freenode/ip.18.62.12.202)
20:33:41saratogais there some known problem with buffering of lossless tracks?
20:34:04saratogai'm curious about jhmikes's reply to FS #12160
20:34:07fs-bluebothttp://www.rockbox.org/tracker/task/12160 3APE codec random loud noise during playback (bugs, unconfirmed)
20:34:25Buscheln1s: the OPUS_COPY stuff?
20:35:42n1sBuschel: no, the ones where you do t[i<<shift] and t[(N4-i)<<shift] to vars before the multiplications
20:36:59Buscheln1s: ok. i've measured those with a gain of 0.1 MHz only. the code reads much better imho, therefor i kept it
20:37:39 Quit scorche (Disconnected by services)
20:37:44 Join scorche` [0] (~scorche@rockbox/administrator/scorche)
20:38:05n1sBuschel: ok, i just expected gcc to figure that out
20:40:06 Quit Syconaut^ (Ping timeout: 245 seconds)
20:41:32 Quit pretty_function (Ping timeout: 240 seconds)
20:41:53 Join pretty_function [0] (~sigBART@123.252.215.20)
20:49:43 Join Strife89 [0] (~Strife89@207.144.201.128)
20:50:06 Quit n17ikh (Ping timeout: 245 seconds)
20:52:25n1sinterestingly the cf gcc produces the exact same code but arm-elf gcc doesn't
20:54:07n1sBuschel: do you feel strongly about it?
20:54:29Buschelnot really, i just didn't want to throw it away
20:55:07 Join lebellium_ [0] (~chatzilla@e179142095.adsl.alicedsl.de)
20:56:01Buschelthe real important stuff (when it comes to performance) is comb_filter(), the OPUS_COPY stuff in several locations and moving f2/freq/X to iram
20:56:05n1snothing wrong with doing that but people wanted to not divere too much from the upstream if possible, othoh getting that patch into upstream shouldn't be too difficult
20:56:16n1sBuschel: yeah
20:56:46 Quit lebellium (Ping timeout: 245 seconds)
20:56:48 Nick lebellium_ is now known as lebellium (~chatzilla@e179142095.adsl.alicedsl.de)
20:57:22 Join mgottschlag [0] (~quassel@reactos/tester/phoenix64)
20:57:41n1si wonder what to look at first :)
20:58:12 Join Epicanis [0] (~Epicanis@static-72-95-113-7.port.east.myfairpoint.net)
20:58:22Buscheltry f2/freq/X first, this might scale a lot on cf
20:59:05n1syeah, already did try freq and X, save ~19MHz
20:59:16Buschelhehe :)
20:59:32Buschelso, then try f2
20:59:50n1sso i guess i'll split up that patch as you suggested and push it first, to bring some order into my ever growing mess of branches
21:00
21:00:24n1syeah, i'll push this first though
21:01:58 Join guymann [0] (~c@unaffiliated/guymann)
21:02:09 Quit shamus (Quit: chaos reigns within reflect repent and reboot order shall return)
21:02:42 Join n17ikh [0] (~n17ikh@128.204.195.239)
21:02:42 Quit n17ikh (Changing host)
21:02:42 Join n17ikh [0] (~n17ikh@unaffiliated/n17ikh)
21:02:44n1sif nothing else to make benchmarkign quicker :)
21:03:55 Join shamus [0] (~shamus@ip-206-192-195-49.marylandheights.ip.cablemo.net)
21:04:06Buscheli have overlcocked my PP device ;)
21:04:43Buschelit runs at 100 Mhz since ages, w/o any issues. saves quite some time when doing benchmarks
21:06:10 Quit n17ikh (Client Quit)
21:06:32 Join n17ikh [0] (~n17ikh@unaffiliated/n17ikh)
21:08:10n1sthinking about dropping that change to use OPUS_CLEAR for X as it seems these two loops were used to avoid zeroing unnecessarily, obviously they should have used memset to do that though
21:09:47Buscheln1s: hmm, i do not agree. i think they just avoided to zero out the mssing part to save cpu time. the array is just allocated right before this zeroing out. if a full memset-to-zero is faster than the selective zeroing we should keep it.
21:10:15 Join prof_wolfff [0] (~prof_wolf@213.37.219.103.dyn.user.ono.com)
21:10:41 Quit Holyjack (Quit: CGI:IRC)
21:13:48n1sBuschel: is it faster?
21:14:47Buscheli did not measure it, but OPUS_CLEAR is faster by sure
21:16:42Buschelbut i do not feel strong about this as well
21:22:12 Quit funman (Ping timeout: 240 seconds)
21:22:19n1schecked for two of the test files and it's clearing only 640 bytes with those loops so i'd rather skip this for now and revisit it later
21:22:24 Join funman [0] (~fun@rockbox/developer/funman)
21:23:24Buschelok
21:25:13 Join pamaury_ [0] (~quassel@vit94-1-82-67-248-70.fbx.proxad.net)
21:25:56 Quit pamaury (Ping timeout: 245 seconds)
21:26:21 Join Syconaut [0] (viper@c-60fd72d5.162-1-64736c10.cust.bredbandsbolaget.se)
21:31:04 Quit y4n (Quit: AMIGAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAHAHAHAHAAAAAAAAAAAAHAHAAA)
21:35:58 Join Horscht [0] (~Horscht@p54947084.dip.t-dialin.net)
21:35:58 Quit Horscht (Changing host)
21:35:58 Join Horscht [0] (~Horscht@xbmc/user/horscht)
21:39:39 Quit eckoit (Quit: eckoit)
21:42:26 Quit Strife89 (Ping timeout: 252 seconds)
21:46:39n1sf2 in iram gives another 5.6MHz on cf
21:46:51n1smore than i expected
21:47:20 Join eckoit [0] (~ryan@96.53.108.182)
21:53:12***Saving seen data "./dancer.seen"
21:54:37Buschelsounds reasonable
21:56:07 Join Strife89 [0] (~Strife89@207.144.201.128)
21:56:10 Quit eckoit (Quit: eckoit)
21:56:40n1sand about 1.2MHz on pp
21:57:15Buschelsame here (measured 1.15 MHz)
21:57:37 Join eckoit [0] (~ryan@96.53.108.182)
21:58:42derfWould getting the entire MDCT to be in-place help more?
22:00
22:01:15n1snot sure, the out buffer doesn't seem to be accessed that much
22:01:42derfI meant the FFT, too.
22:01:47n1sah
22:03:09n1swell, we are now failry close to the limit on the fast ram where we have the least so reducing the memory requirement for the fft would help freeing some of that up for other things
22:03:38derfI have code for this in celtdec, but it would be some work to port it to libopus.
22:04:03derf(especially as it isn't fixed-point)
22:05:06 Quit pretty_function (Ping timeout: 240 seconds)
22:12:46 Quit efyx (Read error: Connection reset by peer)
22:13:29 Join efyx [0] (~efyx@91.179.15.23)
22:14:24n1swith the last push opus is fast enough to make 64kbps opus listenable on cf, i guess there's not many cpu cycles left for dsp though
22:18:05 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier. http://miranda-im.org)
22:21:49Buscheli am curious how much the algorithmic changes to comb_filter() will save on cf. on PP those changes gained ~9 MHz.
22:29:20n1smultiplications are usually slower on pp than cf i think so probably a bit less
22:29:31n1sthat gave me an idea
22:32:36n1sheh, yeah this is stupid
22:33:28n1syo know how the arm7tdmi core has early termination of multiplications based on the value of the second operand
22:34:26n1sand we (well, it's my fault) do 32*16 with the 16 bit operand in the *other* place
22:34:53Buschelgood point!
22:35:08saratogaFWIW, ARMv5 has a single cycle multiply for 32x16, so using that would be a good idea
22:35:25n1sfixing this saves 2.5MHz on pp
22:35:51n1ssaratoga: yeah probably, al our armv5's are much faster than the poor 7tdmi cores though
22:35:59saratogayeah
22:38:44 Quit liar (Ping timeout: 246 seconds)
22:39:26saratogais there profile data for opus after IRAM went in?
22:40:07 Quit bootlkjkgf (Quit: you'll probably leave it Late ::: SO GET ON WITH IT ! http://www.ohloh.net/p/systemd)
22:40:09n1ssaratoga: no, i've been meaning to do that but haven't yet
22:41:41saratogathe C_MULC and other complex mul functions could probably be taken from the other MDCT lib's asm
22:41:59saratogai remember that making quite a difference for the other codecs
22:45:29n1syeah, i whipped up a quick c_mulc for cf, saved about 2MHz
22:45:49 Join liar [0] (~liar@clnet-p09-185.ikbnet.co.at)
22:46:43 Join bootlkjkgf [0] (~Prmhfhfx@92.39.204.151)
22:48:41n1sto see some really good speedup of this fft, i think we would need to write the various butterfly functions in asm to use load/store multiple and proper instruction scheduling
22:49:35n1ssome of the opus guys suggested that it would be a good idea to replace it with a different implementation but i don't know enough about this stuff
22:50:10derfWell, that's under the assumption that you already have another well-optimized one lying around.
22:50:21derfMost of them are restricted to powers-of-two, though.
22:50:39 Quit liar (Ping timeout: 246 seconds)
22:50:48n1syeah, we have a good one for powers of two :)
22:54:46 Join lebellium_ [0] (~chatzilla@e179138246.adsl.alicedsl.de)
22:55:55saratogathe implementation doesn't look too bad
22:56:07saratogafaster ones might exist, but finding one for fixed point non power of 2 is probably not easy
22:56:47 Quit lebellium (Ping timeout: 260 seconds)
22:56:55 Nick lebellium_ is now known as lebellium (~chatzilla@e179138246.adsl.alicedsl.de)
22:59:30saratogaits also difficult to figure out what will be fast, since IRAM targets will have very different performance on an FFT then DRAM targets
23:00
23:00:00saratogasince theres a trade off between number of multiplies and how sequential memory accesses are
23:00:58saratogaso ARMv5/6/7 targets with fast multipliers tend to like sequential accesses even if it means more multiplies, while PP has basically free memory accesses but a very slow multiplier
23:01:55saratogawhat sizes does Opus use again for FFT?
23:02:05saratogai wonder which of those butterflies is actually called most
23:02:14derf120, 240, 480, 960.
23:02:21saratogathe distribution might be odd if the large FFT size has a weird prime factorization
23:02:24derfThe 4-point butterfly should be the most-used.
23:03:33 Quit amayer_ (Quit: going ~/)
23:03:40saratogaremind me, does a 960 point MDCT do a 960 point FFT or is it 480 point?
23:05:02 Join VanniX [0] (~VanniX@host104-60-dynamic.51-82-r.retail.telecomitalia.it)
23:05:07derfTechnically, the FFT size should be N/4 (but N for a 960-sample frame is actually 1920).
23:05:17VanniXhi
23:05:29derfSo, yeah, the complex FFT sizes should be 60, 120, 240, 480.
23:05:45saratogausually in MDCT codecs the large size dominates
23:05:46 Quit sakax (Remote host closed the connection)
23:05:52saratogai assume Opus is similar?
23:05:55derfYes.
23:06:19derfI mean, you can run with 2.5 ms frames, in which case you would _only_ use the 60-point FFT.
23:06:30derfBut I don't expect most people to do that for a pocket music player.
23:06:40saratogaok so 4 x 4 x 2 x 3 x 5
23:07:01derfYeah. Not necessarily in that order.
23:07:02saratogaso yeah, that radix 4 one should probably become assembly
23:08:39saratogais there a plan for eventually merging these optimizations with the official opus decoder?
23:08:59derfYou tell me!
23:09:17saratogawell to large extent that depends on the opus people
23:09:21saratogahow do they want to handle this
23:09:53derfI am "the opus people".
23:10:00saratogayeah i thought so
23:10:03saratogaadding asm to embedded decoders is tricky since people will want to use weird compilers
23:10:22saratogaand there are a lot of arm compilers, most of which aren't free to test with
23:10:36derfAs long as it can be disabled, I don't think this is a big issue.
23:10:47saratogavorbis had this problem, where most people use tremor with the asm disables
23:10:57saratogaso Vorbis almost uniformly has awful battery life on portable devices
23:11:01derfIf people want it to work on crazy arches with crazy compilers, they can send patches.
23:11:19VanniXanyone know or work on samsung yp r0?
23:11:28saratogausually what happens is whoever adds vorbis support to a device looks at it for 10 seconds, gives up, and then ships a shitty decoder :)
23:11:38saratogasee for example sandisk
23:11:54derfsaratoga: Hah, yeah.
23:12:18derfWell, I suspect that is large respect just due to no one considering it that important.
23:12:51saratogai think a big part is that c doesn't give a good way to enable ASM in a cross platform, cross compiler manner
23:12:59saratogabut yeah, that too
23:13:16derfWell, ultimately getting it work on stuff I don't have access to will require someone else to care.
23:13:49derfBut I don't see that as any reason not to get stuff fast on things I do have access to.
23:14:33 Quit kevku (Remote host closed the connection)
23:17:04derfAnyway, the unusual requirement I will have eventually is to be able to detect the CPU and switch between multiple versions of the optimized routines.
23:17:23derfBut don't let that slow you down.
23:18:08derf(the reason for that is mostly stuff like Firefox that does not have separate binaries for each target platform, but ships a single binary in the Android store)
23:22:30 Quit Strife89 (Quit: Connection reset by deer.)
23:22:59 Quit n1s (Quit: Ex-Chat)
23:26:08kugellebellium: you said an april build works fine?
23:26:12kugelon the zip?
23:27:05lebelliumyes, I wrote the exact version here FS #12759
23:27:06fs-bluebothttp://www.rockbox.org/tracker/task/12759 3Various RDS issues (Clip Zip, YP-R0) (bugs, unconfirmed)
23:28:33VanniXkugel, i'm working on automount for sdcard
23:28:48kugellebellium: I assume the theme was a bit different back than?
23:28:58kugelVanniX: hey, nice to meet you!
23:29:00VanniXit works simply adding it in fstab
23:29:10VanniXnice to meet you too ;-)
23:29:29VanniXbut how can we umount sd from rb?
23:29:32VanniX:S
23:29:55kugelthere's an umount() system call
23:30:19VanniXgood
23:30:23kugel(i dont mean system("umount ...") but umount("..."))
23:30:29VanniXokok
23:30:49kugelthere's also a respecitve mount system call
23:31:00kugelboth are very similar to the mount/umount unix tools
23:31:13VanniXlorenzo told you about the usb mode?
23:31:26VanniXok, we'll check for them ;-)
23:31:31kugeldid I see right that your modded firmware uses a different filesystem in rb and the of?
23:31:34lebelliumkugel: Not sure what you mean but I tried the latest build and the April build with the same theme version
23:31:48kugellebellium: do you still have that build?
23:32:02lebelliumfor Clip Zip? Sure, it's the only old build I have on my PC
23:32:04VanniXkugel, yes, vfat for rb and rfs for samsung app
23:32:14kugelwhy is that?
23:32:27lebelliumdo you want me to upload it?
23:32:35VanniXrb with vfat is faster
23:32:40 Quit Buschel (Quit: ChatZilla 0.9.88.2 [Firefox 15.0.1/20120905151427])
23:32:46kugellebellium: no, but don't delete it either
23:33:14kugelVanniX: rockbox (as an app) doesnt care about the file system. so i guess it's generally faster
23:33:36VanniXsystem it's faster in disk i/o
23:33:42VanniX;)
23:33:56VanniXand cpu usage is lower
23:34:05kugeli mean i guess vfat is always faster than rfs then
23:34:28VanniXyep
23:34:34kugelunless the of knows some special file system specific tricks
23:34:36lebelliumkugel: but I did not say it's the latest build with RDS working well, between this old April build and the latest builds there are probably dozens or hundreds builds for which RDS is also working well. I don't have the slightest idea which build in particular brought the RDS issue :S
23:35:08kugellebellium: yes i understand that, but it's a starting point for bisection
23:35:17kugelor digging through commit logs
23:36:21VanniXkugel, OF use rfs for a layer between files and file system.. it also deals with system checking and other samsung shitty things
23:37:05VanniXhowever now we're trying to add a stable usb mode
23:37:09kugelre usb, i know lorenzo92 is working on (and debugging) it
23:37:32VanniXwith the script i wrote last week usb mode works
23:37:35kugelusb support within rockbox would be awesome
23:37:44VanniXbut we need to launch it in console -.-
23:38:18VanniXwe also added the g_serial module, you read?
23:38:36kugelyea, I used it already, it's neat
23:39:19VanniXi'm also thinking about adding it in system>debug>g_serial in rb
23:39:25VanniXi'll talk about it with lorenzo
23:39:53kugelsounds like a good idea
23:43:29VanniXwith the physical uart port we can do lots of things :P
23:43:50 Quit Horscht (Quit: Verlassend)
23:47:22 Join amayer [0] (~alex@h62.26.25.72.ip.windstream.net)
23:48:02 Quit pamaury_ (Ping timeout: 246 seconds)
23:48:15 Quit Rower85 (Quit: Hmmm...)
23:53:16***Saving seen data "./dancer.seen"
23:53:29 Quit dfkt (Quit: -= SysReset 2.55=- Sic gorgiamus allos subjectatos nunc.)

Previous day | Next day