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 2012-10-08

00:18:41 Quit lebellium (Quit: ChatZilla 0.9.89 [Firefox 16.0/20121002073616])
00:24:59 Join Provel [0] (
00:43:49 Quit Keripo (Quit: Leaving.)
00:47:06 Quit bertrik (Remote host closed the connection)
00:48:41ParkerR_rasher, If you are who I think you are, I want to say thanks for the Android port.
00:56:15 Quit scorche (Disconnected by services)
00:56:20 Join scorche` [0] (~scorche@rockbox/administrator/scorche)
00:57:22 Join perrikwp [0] (
00:59:32rasherParkerR_: Please don't - I did not do any work related to the Android port. I mere distribute builds of the port.
00:59:56 Quit ender` (Quit: The early bird may get the worm, but the second mouse gets the cheese.)
01:00:32ParkerR_rasher, Well thanks for the builds
01:00:51ParkerR_I wouldn't be using it without them
01:01:04rasher:) No problem, but the actual work was done by other people
01:01:19ParkerR_I see
01:04:18 Quit prof_wolfff (Ping timeout: 260 seconds)
01:07:15ParkerR_So this adds an old Hero CDMA to my list of Rockbox devices. Other two are a Nano 1gand Nexus 7
01:07:25ParkerR_*1g and
01:11:42 Quit perrikwp (Ping timeout: 255 seconds)
01:15:48 Quit FOAD (Quit: I'll be back)
01:15:59 Join FOAD [0] (~foad@
01:16:11 Join speckmade1 [0] (
01:16:56 Quit FOAD (Client Quit)
01:17:14 Join FOAD [0] (~foad@
01:20:44 Join perrikwp [0] (
01:22:06 Quit LambdaCalculus37 (Quit: LambdaCalculus37)
01:23:20 Join mikroflops [0] (
01:27:04 Quit mikroflops_ (Ping timeout: 264 seconds)
01:35:13 Quit mgottschlag (Ping timeout: 260 seconds)
01:42:04 Quit pamaury_ (Ping timeout: 264 seconds)
01:56:18***Saving seen data "./dancer.seen"
02:16:59 Quit zahari_bgr (Read error: No route to host)
02:22:43 Join zahari_bgr [0] (~zahari.yu@
02:44:05[Saint] the manual just plain wrong, or, do we really ignore "folder.jpeg" as valid album art?
02:44:55[Saint](this {according to the manual} is the only naming scheme case that ignores the .jpg and .bmp cases, and it doesn't explain why)
02:54:46 Join mikroflops_ [0] (
02:58:16 Quit mikroflops_ (Read error: Operation timed out)
02:58:27 Quit mikroflops (Ping timeout: 248 seconds)
02:59:12 Quit factor (Quit: Leaving)
02:59:59 Join factor [0] (~factor@
03:00:10ParkerR_[Saint], So it only accepts jpeg?
03:00:58[Saint]Apparently, ...but, I'm having a ahrd time believing that this single case only accepts jpeg and not bmp or jpg.
03:03:13 Join mikroflops [0] (
03:03:35[Saint]IIRC, various media players use "folder.jpeg" as their album art location if not embedded, but that doesn't seem like a reason to ignore other extensions.
03:03:50*[Saint] shrugs.
03:04:16 Join mikroflops_ [0] (
03:04:39 Part mikroflops_
03:07:24ParkerR_[Saint], Well jpeg used to be the way most images were until they kinda went with jpg. Maybe it just hasn't caught up
03:08:55[Saint]I'm kinda leaning towards it being added solely because various media players use this format, as WMP set the style.
03:56:19***Saving seen data "./dancer.seen"
04:06:07 Quit pixelma (Disconnected by services)
04:06:07 Join pixelma_ [0] (pixelma@rockbox/staff/pixelma)
04:06:07 Join amiconn_ [0] (amiconn@rockbox/developer/amiconn)
04:06:08 Quit amiconn (Disconnected by services)
04:06:09 Nick pixelma_ is now known as pixelma (pixelma@rockbox/staff/pixelma)
04:06:12 Nick amiconn_ is now known as amiconn (amiconn@rockbox/developer/amiconn)
04:14:46 Quit pixelma (Read error: Connection timed out)
04:14:56 Join pixelma [0] (pixelma@rockbox/staff/pixelma)
04:30:55 Join TheSphinX_ [0] (
04:34:18 Quit TheSphinX^ (Ping timeout: 260 seconds)
04:39:55 Join baegle [0] (~baegle@unaffiliated/baegle)
04:40:23baegleHello, not seeing this anywhere else. Any idea why I'm only getting the option to install the development release of Rockbox when I use the utility?
04:40:32 Quit the-kyle (Quit: Leaving.)
04:42:15[Saint]baegle: which target?
04:44:42[Saint]Only stable targets are displayed by default.
04:44:58 Join the-kyle [0] (
04:44:58[Saint]The Fuze+ isn't stable, you need to enable the display of non-stable targets.
04:46:02[Saint]"Configuration - Show disabled targets"
04:46:39[Saint]Ooooohhh...heh, I totally misread your post.
04:47:22[Saint]I also made a wrong. Bad me.
04:49:06baeglehmmmm, that didn't add any releases to the available list
04:49:11baeglestill only development snapshot
04:50:18[Saint]Yeah, I don't even know what I was thinking with that...I need to wake up, heh. I don't think there *is* a release for this target, totally makes sense, but isn't exactly obvious.
04:52:21[Saint]iirc, only stable targets get RC and release builds.
05:01:21 Join Rower85 [0] (husvagn@
05:07:46 Nick scorche` is now known as scorche (~scorche@rockbox/administrator/scorche)
05:10:38 Quit Epicanis (Quit: sleepytime, y'all)
05:19:29 Quit shamus (Read error: Connection reset by peer)
05:22:33 Join shamus [0] (
05:31:37baeglegot it
05:35:03 Quit TheSeven (Disconnected by services)
05:35:12 Join [7] [0] (~quassel@rockbox/developer/TheSeven)
05:35:51*[Saint] apologises for his earlier display of brain-fartery.
05:36:10[Saint]I said so many wrong things, it killed kittens.
05:37:00[Saint]A quick backpedal and some actual correct info later, all is well.
05:37:37baegleI was ready to toss this POS out the window before Rockbox
05:37:40baeglenow I'm happy
05:37:50baegleAnd you can use that on the website as a testimonial
05:44:03ParkerR_baegle, Sweet
05:55:48baegleSo I've got an sd card for my player, and I want to have the simplest possible setup for an idiot proof playlist for an event
05:56:05baegleI figure, there's no need to organize by artist or album, just a single playlist
05:56:20***Saving seen data "./dancer.seen"
05:56:37baegleBut it looks like to use it, I have to create a playlist on the card, and then use File -> MMC01 then open the playlist
05:56:49baegleif I put everything in root, if I put it all in music, I have to go another step further
05:57:13baegleIs there an easier way? I notice that the Playlist Catalogue on the main menu doesn't notice that I have a playlist on the SD card.
06:00:40ParkerR_baegle, I think that playlist catalog is for playlist you make on the device not pls files
06:01:58 Part amayer
06:02:21baegleIs there a way to put a playlist on the main menu?
06:02:54ParkerR_Not sure
06:03:05[Saint]Not without recompiling from source, no.
06:03:33[Saint]Playlist Catalogue should be one the main menu, so, it's only one extra step.
06:04:14baegleHow do I add a link to a playlist on an SD card to the catalogue?
06:05:01[Saint]I take it this is a playlist you created manually, and then exported?
06:05:14baegleI created it on my computer by hand
06:05:15[Saint](which will probably fail, due to relative paths anyway)
06:05:26[Saint]Yeah, best to create on device.
06:05:32baeglenfw I'm gonna do that
06:05:37[Saint]Then, it'll "Just Work (TM)"
06:05:37baegleWay too much music
06:05:56[Saint]But, you said it's all in one dir, no?
06:06:11baegleOrder matters
06:06:13[Saint]Bring up the context menu on the containing directory, add to playlist, done.
06:06:24[Saint]Oooooh, ordering. Yeah, sucks.
06:06:55baeglelooks like I can maybe just do \<MMC1>\ ?
06:07:38[Saint]I think slashes mess it up, create a playlist on device from a file contained in the sdcard to check the formatting.
06:08:22[Saint]You'll also need to escape dirs that have spaces in the names: "/dir/my\ dir\ with\ spaces/file.file" etc.
06:09:27[Saint]I'm not sure how people go about exporting playlists and making sure the paths still make sense, I just use a dynamic playlist, yeah.
06:20:24baegleOK, how do I stop the current list and load another one?
06:24:06baegle[Saint]: Creating the playlist through the player doesn't work either.
06:24:39[Saint]can you define "doesn't work" please?
06:24:44[Saint]In what way does it fail?
06:24:52baegleNothing to resume
06:25:04baegleit adds the song to a new playlist as requested
06:25:08baeglebut the playlist won't play
06:27:06baeglewhy does "select" ask to delete a playlist anyway, That's the most counter intuitive process
06:27:50[Saint]Sounds like you're just blowing away the dynamic playlist.
06:28:00[Saint]I assume you get a warning to this effect?
06:28:21baegleThat's exactly what I'm seeing. Didn't understand the message initially
06:29:06[Saint]You can create a static playlist from the current dynamic playlist from the context menu, iirc.
06:29:19baegleI think I might have gotten it
06:30:16[Saint]The hard thing to grasp, is that *all* Rockbox playback is playlist based. It is a little non-obvious.
06:30:45[Saint]Just launching an arbitrary file from the file explorer adds it to the dynamic playlist, which you can then save, or discard, etc.
06:31:18[Saint]Reading our Fine Manual may shed some light on the entire process for you.
06:31:50[Saint]I only use the database and dynamic playlist, so..creation of playlists and their management isn't a strongpoint of mine.
06:32:31baegleI'm just looking for a straight shot solution to a simple problem
06:32:33baegleI think I've managed it
06:33:07baegleAfter being on the brink of returning the Fuze+ because it wouldn't play my flac, I think I've got enough control with Rockbox to do my thing
06:33:50*[Saint] is aware of very few media players that play flac out of the box.
06:34:21[Saint]alac in the iPods is the closes I can think of off the top of my head.
06:34:52 Join speckmade [0] (
06:35:28 Quit speckmade1 (Quit: Leaving.)
06:37:02baegleFuze+ has it listed on the box
06:37:36[Saint]Ah, I wasn't aware. It specifically lists flac, but, doesn't play it?
06:37:47[Saint]...nice work SanDisk.
06:38:39baegleAnd it wasn't a 16 vs 24 bit issue either
06:38:58baeglesince these were 16 and the forums specifically identify 16 bit as the only type that works
06:39:07baegleanyway, the box is now rocked
06:39:12baegleand I can make a playlist for my wedding
06:39:23 Join speckmade1 [0] (
06:39:44 Quit speckmade (Client Quit)
06:39:45[Saint]Mazel tov!
06:44:24 Join perrikwp_ [0] (
06:46:46 Quit perrikwp (Ping timeout: 256 seconds)
07:02:42 Part baegle
07:06:32 Join Keripo [0] (
07:07:39 Quit XavierGr ()
07:24:44 Quit factor (Read error: Connection reset by peer)
07:25:45 Quit sciopath (Read error: Connection reset by peer)
07:26:14 Join sciopath [0] (
07:37:27 Join factor [0] (~factor@
07:56:24***Saving seen data "./dancer.seen"
08:01:52 Quit amiconn (Remote host closed the connection)
08:01:53 Quit pixelma (Remote host closed the connection)
08:03:01 Join pixelma [0] (pixelma@rockbox/staff/pixelma)
08:03:01 Join amiconn [0] (amiconn@rockbox/developer/amiconn)
08:19:20 Join mortalis [0] (
08:34:14 Quit Belzebub (Ping timeout: 260 seconds)
08:36:52 Join amithkk [0] (u4289@2buntu/writers/amithkk)
08:43:09 Join Belzebub [0] (
08:46:39 Join ender` [0] (
08:46:47 Join Zagor [0] (
08:46:47 Quit Zagor (Changing host)
08:46:47 Join Zagor [242] (~bjst@rockbox/developer/Zagor)
08:52:31 Quit Belzebub (Ping timeout: 255 seconds)
08:52:46 Join Belzebub [0] (
08:56:40 Quit Belzebub (Read error: Connection reset by peer)
08:57:44 Join Belzebub [0] (
09:06:06 Join mgottschlag [0] (~quassel@reactos/tester/phoenix64)
09:17:46 Quit zahari_bgr (Read error: Connection reset by peer)
09:21:01 Join pamaury [0] (~quassel@
09:21:01 Quit pamaury (Changing host)
09:21:01 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
09:21:08 Quit Belzebub (Changing host)
09:21:08 Join Belzebub [0] (~torrentow@unaffiliated/blown-engine)
09:23:43 Join zahari_bgr [0] (~zahari.yu@
09:38:31 Join ender1 [0] (
09:38:31 Quit ender` (Read error: Connection reset by peer)
09:42:59 Quit Belzebub (Ping timeout: 256 seconds)
09:50:59 Join petur [0] (~petur@rockbox/developer/petur)
09:51:31 Join LinusN [0] (
09:51:31 Join Belzebub [0] (
09:56:25***Saving seen data "./dancer.seen"
10:06:12 Quit mgottschlag (Ping timeout: 246 seconds)
10:13:59 Join wodz [0] (
10:14:29 Join speckmade [0] (
10:17:13 Quit pamaury (Ping timeout: 240 seconds)
10:17:14 Quit speckmade1 (Ping timeout: 246 seconds)
10:17:46 Quit Belzebub (Ping timeout: 240 seconds)
10:28:10 Join Horscht [0] (
10:28:10 Quit Horscht (Changing host)
10:28:10 Join Horscht [0] (~Horscht@xbmc/user/horscht)
10:32:27 Join Belzebub [0] (
10:36:52 Quit Belzebub (Read error: Connection reset by peer)
10:37:28 Join Belzebub [0] (
10:38:28 Join mgottschlag [0] (~quassel@reactos/tester/phoenix64)
10:47:18 Quit Belzebub (Changing host)
10:47:18 Join Belzebub [0] (~torrentow@unaffiliated/blown-engine)
10:50:11 Join Kiwicam [0] (~quassel@
10:50:32 Nick Kiwicam is now known as Guest80204 (~quassel@
10:50:43 Quit Guest80204 (Client Quit)
10:51:34 Join Kiwicam_ [0] (~quassel@
10:57:21wodzThere is something I don't get about elfs - for cube on CF symtab have 52 entries, while only 16 are referenced when relocating. I thought −−strip-unneeded should eliminate unused in relocation symbols
10:57:37 Nick Jack87 is now known as Jack87|Away (Jack87@nasadmin/admin/jack87)
11:06:38 Join pamaury [0] (
11:06:38 Quit pamaury (Changing host)
11:06:38 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
11:21:11 Quit speckmade (Quit: Leaving.)
11:21:15 Join speckmade1 [0] (
11:27:19 Quit speckmade1 (Ping timeout: 255 seconds)
11:35:52 Join speckmade [0] (
11:38:10 Quit speckmade (Read error: Connection reset by peer)
11:38:10 Join speckmade1 [0] (
11:40:22[Saint]Kiwicam_: Holy sh*tballs!
11:40:38[Saint]You have 127 sub-images in "G"?!?
11:42:38 Quit speckmade1 (Client Quit)
11:43:29Kiwicam_127? One fo revery pixel across the Clip +
11:43:38Kiwicam_0 to 127
11:43:53Kiwicam_Ground Movement.
11:45:25[Saint]Surely there's a better way to do that?
11:45:40 Join speckmade [0] (
11:45:54[Saint]It could be done by displaying a few image "tiles" in random locations.
11:46:03Kiwicam_Probably not.
11:46:56[Saint]You can display alignment tags now?
11:47:11*[Saint] encourages Kiwicam_ out of *-community
11:47:29Kiwicam_[Saint]: It's been ages since I looked at this stuff. I'm rusty.
11:48:30 Join Thra11 [0] (~thrall@
11:48:36Kiwicam_Alignment Tags?
11:49:01 Quit bluebrother (Disconnected by services)
11:49:07 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
11:49:34[Saint]Kiwicam_: re:
11:49:37[Saint]"<Kiwicam_> Finally got a Clip ZIp. So I tried out my theme I made for it ages ago and then realised that it's showing two %ar's instead of a %ac & %ar."
11:49:59[Saint]I wondered how it would be "showing" alignment tags.
11:50:17[Saint](which is what those tags are)
11:50:21Kiwicam_Oh S! Sorry two %pr's instead of %pc & %pr
11:52:14 Quit fs-bluebot (Ping timeout: 246 seconds)
11:53:32 Join fs-bluebot [0] (
11:56:29***Saving seen data "./dancer.seen"
11:56:38wodzcan someone explain me what this snippet from pictureflow.c is actually doing?
11:57:49wodzplugin_get_buffer() and plugin_get_audio_buffer() return the size of the buffer granted
11:59:10 Join speckmade1 [0] (
11:59:51 Quit speckmade (Quit: Leaving.)
12:00:12 Quit Keripo (Quit: Leaving.)
12:01:45 Quit speckmade1 (Client Quit)
12:01:51 Join speckmade [0] (
12:04:23[Saint]wodz: I assume it's not going to become any more apparent to me if I actually take context from above and below?
12:05:10wodzyou can try apps/plugins/pictureflow/pictureflow.c:2946
12:05:28amiconnwodz: You mean the #ifndef SIMULATOR part?
12:07:52amiconnThat's easy to explain. It's necessary for those targets where pictureflow is loaded as an overlay
12:08:58amiconnOtherwise the plugin might overwrite itself with data. plugin_get_audio_buffer() doesn't know whether an overlay already occupies a part of the audio buffer.
12:09:49gevaertsZagor: can you move the uploaded 3.12 files to the proper location one of these days and do the necessary magic?
12:09:53amiconnplugin_get_audio_buffer() just returns the same buffer over and over, no matter how often it's called
12:12:24wodzamiconn: ok, Could you explain me a bit how overlay is working actually?
12:14:55amiconnThe overlay loader (pictureflow.rock) gets the audio buffer, then loads the overlay (pictureflow.ovl) there
12:15:26amiconnThis is a mechanism to support large plugins on targets with a small plugin buffer, at the price of having to stop playback
12:16:02amiconnThe .ovl header is the same as that of an ordniary plugin, apart from the different load address
12:17:43wodzok, so if we hypothetically switch to relocatable elfs we may get rid of this and load the plugin at the 'right' address in the first place
12:19:48 Quit [Saint] (Remote host closed the connection)
12:21:20 Join [Saint] [0] (~quassel@rockbox/user/saint)
12:31:07 Quit Thra11 (Quit: kthxbai)
12:32:04 Quit Horscht (Quit: Verlassend)
12:41:01 Quit speckmade (Quit: Leaving.)
12:42:53 Join speckmade [0] (
12:44:19 Quit speckmade (Client Quit)
12:44:28 Join speckmade [0] (
12:48:48Zagorgevaerts: will do!
12:48:53 Join Thra11 [0] (~thrall@
12:51:48wodzTorne: Could you look at g#326 and comment?
12:51:50fs-bluebotGerrit review #326 at : ELF loader for plugins by Marcin Bukat (changes/26/326/4)
12:52:07Tornei will try and look later
12:52:42wodzTorne: I guess elf with .symtab but without .strtab is nonstandard, right?
12:53:25Torneinasmuch as the one refers to the other, probably
12:57:47 Join Thra11_ [0] (~thrall@
13:00:20 Quit Thra11 (Ping timeout: 246 seconds)
13:08:06 Quit Thra11_ (Ping timeout: 246 seconds)
13:11:48 Quit petur (Quit: *plop*)
13:20:13 Join Thra11_ [0] (~thrall@
13:26:17 Join lebellium [0] (
13:26:40 Quit speckmade (Ping timeout: 245 seconds)
13:31:37 Join mgottschlag2 [0] (~quassel@2a00:1398:200:200:868:7a8b:ea6f:402e)
13:35:24 Quit mgottschlag (Ping timeout: 246 seconds)
13:37:21 Join speckmade [0] (
13:48:02 Join Horscht [0] (
13:48:02 Quit Horscht (Changing host)
13:48:02 Join Horscht [0] (~Horscht@xbmc/user/horscht)
13:56:33***Saving seen data "./dancer.seen"
14:01:22 Quit RAThomas (Quit: ChatZilla 0.9.89 [Firefox 15.0.1/20120905151427])
14:04:30 Join Epicanis [0] (
14:16:19wodzamiconn: reading ovelay.c now I understand - overlay is loaded at the end of audiobuff as opposed to regular plugins which are loaded at the begining of pluginbuff
14:19:46 Quit speckmade (Read error: Connection reset by peer)
14:23:07 Join speckmade [0] (
14:25:09 Quit zahari_bgr (Read error: Connection reset by peer)
14:29:16 Join amayer_ [0] (
14:30:56 Quit wodz (Quit: Leaving)
14:31:23 Join zahari_bgr [0] (~zahari.yu@
14:33:19 Join einhirn [0] (
14:33:40 Join markun [0] (~markun@rockbox/developer/markun)
14:39:14 Quit einhirn (Quit: Miranda IM! Smaller, Faster, Easier.
14:48:32 Join sciopat [0] (
14:51:03 Quit sciopath (Ping timeout: 246 seconds)
14:53:39markunis opus decoding in real-time for all rockbox targets?
14:54:50EpicanisIt's been a few days since I tried it last, but last I tried on my Sansa C240 it was close, but not quite (hugely improved from the initial release, though).
14:55:38Epicanis(Had gone from 2-seconds-playing/1-second-pause to 8-seconds-playing/half-second-pause from initial release to when I last tried it)
14:56:29the-kylemarkun: I think it's realtime on most if not all targets, but there are still some ram issues on the clipv1. I'm not entirely sure about everything, as I've only tried it on the clipv1 and clip+.
14:56:59gevaertsmarkun: n1s summarised at
14:57:35gevaertsSo I'd say for 64kbps files realtime on everything except PP, which is nearly there
14:57:48gevaertsNo ideas about other bitrates
14:59:27the-kyle128 is supposed to be near realtime on most targets. 256 still has a long way to go if I understand correctly, but I don't usually use 256 even on my clip+, so haven't tested it on anything yet.
15:00:08the-kyleI think 128k is still going to have some problems on slower targets.
15:01:57markunso it sounds like it will drain batteries pretty fast
15:02:48the-kyleI do notice a little decrease in battery life on my clip+, but it doesn't drain extremely fast.
15:06:01markunthe-kyle: good to know since I also primarily use the clip+
15:06:10gevaertsthe-kyle: does 256kbps make sense for opus?
15:07:06the-kylegevaerts: By all accounts, probably not. I think 180 is pretty much tops, but 256 may have some benefit in some applications. Probably not for portable though.
15:07:48the-kylemarkun: Opus works extremely well on the clip+.
15:08:17the-kyleIt actually worked well on the clip+ at the initial release.
15:08:33gevaertsI'd have thought opus would make most sense at the lower end, where (at least according to it makes a lot of sense, but above 128kbps codec choice doesn't seem to make much quality difference
15:08:44*the-kyle patched it in a day before the release.
15:10:07the-kyleApparently there is a difference at 128k between Opus and Vorbis, but on portable devices, above 128k doesn't really make much of a difference.
15:10:52the-kyleOpus is the best I've heard at or below 128k.
15:11:20the-kyleThis is even taking into account the proprietary codecs.
15:11:37gevaertsYes, that's what that graph seems to show, except below 7k or so
15:11:56gevaertsBut still 7k to 128k is a large range to be good in
15:12:15the-kylebelow 7k, codec2 shows a lot of promise, but it's still not frozen.
15:12:47gevaertsThat's for rather different usecases again though I expect
15:13:24the-kyleCodec2 is supposed to be for ham radio and other very tight bandwidth speech applications.
15:13:34markunbelow some bitrate I guess it's just the SILK codec?
15:13:48the-kyleNot really practical for most portable audio devices.
15:14:05gevaertsCodec2 could be good for speech support on lowmem devices like the clip :)
15:15:17the-kyleTrue, and who wouldn't jump at the chance to read an audiobook at 2.4k? Well, it does sound a bit computerized, but it was intelegible at that bitrate.
15:15:55*the-kyle likes 12k and 16k Opus for audiobooks.
15:16:00markunTTS is the best codec for audiobooks of course ;)
15:16:38the-kylemarkun: This is true, of course, but there is not yet a working tts on Rockbox.
15:17:01the-kyleThey generally take a bit more ram also.
15:17:34gevaertsAt 2.4kbit, a 32MB device would have to spin up lightly less than once per day
15:17:38gevaertsGood enough for me :)
15:17:43markuneSpeak being GPLv3 is not a problem anymore, right?
15:17:54EpicanisI was wondering if codec2 might actually make some speech MORE intelligible...
15:17:59the-kyleI think someone was working on eSpeak, but I'm not sure about its status.
15:18:41the-kyleIt would solve a lot of problems if it can be made to work, including the whole .talk clip thing.
15:19:41markunI use espeak as TTS on my android phone, but I'm usually the only one in the car who understands the directions of google navigation
15:19:43the-kyleAlthough I think eSpeak is about the only tts that has a low enough memory footprint, so .talk clips and voice files may still need to be used on some targets, or where eSpeak is not the user's preferred tts.
15:20:59the-kyleI use eSpeak to make .talk clips and voice files for Rockbox. My wife uses Pico though, so I patched my tools/configure and tools/ to make it work with Pico.
15:21:50the-kyleThat still doesn't make the Rockbox utility use Pico, unfortunately.
15:22:26 Quit GeekShadow (Ping timeout: 240 seconds)
15:24:52 Join WalkGood [0] (~4@unaffiliated/walkgood)
15:26:51the-kyleAnybody know the status of eSpeak on Rockbox? I saw something in the git log about the possibility of supporting a native tts a month or so ago, but I didn't know if that was eSpeak related, and haven't seen anything since.
15:28:25*the-kyle is looking up the commit that seems to have mentioned it. It had something to do with the way voice data is stored.
15:29:00gevaertsThat's not related to TTS on the device
15:29:43gevaertsIt was about storing the necessary information on the device to allow rockbox utility to generate voice files without needing the rockbox servers, and without being restricted to unmodified builds
15:32:27*gevaerts nods
15:32:36the-kyleThe line of interest was
15:32:38the-kyleFurthermore, putting the voice strings on the target is also a prerequisite for
15:32:47the-kyle eventually supporting a TTS on the player.
15:32:50gevaertsRight. It's also needed for TTS on target, but it doesn't really do much for that
15:32:58*freqmod have made a codec2 interface
15:33:08freqmodif somebody just had ported it to fixed point
15:33:25freqmodbut the codec is still at the changing and experimenting stage
15:34:08the-kyleYes. It's still a bit of a chore to get codec2 to play on the computer.
15:34:30*the-kyle had to encode, decode and then play the resulting wav file.
15:34:36*freqmod too
15:35:11 Join wodz [0] (
15:36:09wodzOk, I finally cleaned up and merged my yesterday work into g#326 - now most plugins run on CF as well.
15:36:10fs-bluebotGerrit review #326 at : ELF loader for plugins by Marcin Bukat (changes/26/326/5)
15:37:22wodzhas any of the active devs have native MIPS target to test on?
15:37:30 Quit Thra11_ (Ping timeout: 260 seconds)
15:41:29 Quit speckmade (Read error: Connection reset by peer)
15:42:10 Join speckmade [0] (
15:51:00mortaliswodz: There are some unnecessary changes in rk27xx.h
15:51:06mortalisin g326
15:51:07fs-bluebotGerrit review #326 at : ELF loader for plugins by Marcin Bukat (changes/26/326/5)
15:55:24 Quit speckmade (Ping timeout: 246 seconds)
15:56:36***Saving seen data "./dancer.seen"
16:01:41 Join speckmade [0] (
16:01:47wodzmortalis: could be - it was taken from bFLT branch which predates lcd dma and cpu scalling stuff
16:04:26 Join Thra11 [0] (~thrall@
16:07:45 Quit speckmade (Ping timeout: 256 seconds)
16:10:11 Quit Thra11 (Read error: Operation timed out)
16:11:47 Join Thra11 [0] (~thrall@
16:12:10amayer_how do you delete a post in the forums?
16:12:13amayer_i tried posting something and it didnt post. then i posted it again and 3 of them poped up
16:12:49amayer_it now only shows up under:
16:12:56amayer_but not the regular topic
16:13:18 Quit the-kyle (Ping timeout: 246 seconds)
16:14:12 Quit mortalis (Quit: Leaving)
16:14:18 Join the-kyle [0] (
16:14:40gevaertsamayer_: you're another victim of our spam filter false positives
16:14:54amayer_well i did post it 3 times by acident
16:15:06gevaertsFor some reason it doesn't like URLs
16:15:17amayer_can someone delete the last 2 and allow past spam?
16:15:23gevaertsI'll recover them all, and then delete them the regular way
16:15:42amayer_i didnt see an option to delete them
16:15:51amayer_gevaerts: thanks for your help
16:17:35gevaertsShould be ok now (I hope)
16:18:42 Quit wodz (Quit: Leaving)
16:19:54amayer_looks good to me
16:20:00amayer_thanks a million
16:21:51 Quit shamus (Read error: Connection reset by peer)
16:22:20 Join shamus [0] (
16:31:22 Join speckmade [0] (
16:34:12 Join y4n [0] (~y4n@unaffiliated/y4ndexx)
16:46:26 Join pamaury_ [0] (
16:47:15 Quit pamaury (Ping timeout: 248 seconds)
16:49:43 Quit speckmade (Quit: Leaving.)
16:50:00 Join speckmade [0] (
16:50:21 Join Topy44 [0] (
16:54:56 Join saratoga [0] (123e0cca@gateway/web/freenode/ip.
16:55:33saratogamarkun: GPLv3 should be fine provided the code is protected by ifdef's so that people have the option of using GPLv2 if they want
16:55:47saratogafor a feature like voice I think that should be a nonissue
16:58:16 Quit pamaury_ (Ping timeout: 264 seconds)
16:58:19 Join speckmade1 [0] (
16:58:29 Quit speckmade (Read error: Connection reset by peer)
16:58:45the-kyleI seem to recall explicit permission being given by Jonathan Duddington, the lead developer of eSpeak, to use eSpeak's GPLv3 code in Rockbox, even though Rockbox is GPLV2. If I am remembering correctly, all bases should be covered, in spite of the difference in GPL versions.
16:59:32 Join pamaury [0] (
16:59:32 Quit pamaury (Changing host)
16:59:32 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
17:00:30gevaertsthe-kyle: that permission was easy. GPLv3 is compatible with GPLv2+, without explicit permission. The potential problem is that the resulting binary is explicitely GPLv3(+), and whether or not that's a problem is *our* side
17:01:08gevaertsAlso, as long as we had v2-only code, *that* code was not compatible with v3, whatever the authors of the v3 people say
17:01:14the-kyleAh OK. I wasn't sure where exactly the problem was there.
17:02:05 Join speckmade [0] (
17:02:16gevaertsThe main thing is that (for various reasons) we don't want to move the entire codebase to v3. It's *compatible* with v3 though, so having part of the source be v3 is not a big issue if we document stuff properly
17:02:32 Quit speckmade1 (Ping timeout: 240 seconds)
17:04:56saratogayeah the espeak guy wouldn't actually give us the licensing terms we wanted
17:05:28saratogadoesn't matter now that we're GPLv2 or v3 though
17:05:39the-kyleSo it's safe from a licensing standpoint to use v3 only code in a v2+ project? That was what I thought the problem was.
17:06:05saratogayeah, if you do that you become a v3 project
17:06:11gevaertsWell, it depends
17:06:16gevaertsYour *binaries* become v3
17:06:33 Quit Zagor (Quit: Clint excited)
17:06:53the-kyleBut doesn't that raise some patent issues in some countries with some of the Rockbox codecs?
17:07:19gevaertsSo someone who gets a binary and then exercises his right to download the source might technically get v3 source, while someone who downloads a binary and then unrelatedly decides to also download the source gets mixed-license source
17:08:15the-kyleUn less, of course, I misunderstand some of GPLv3.
17:08:51saratogagevaerts: i don't remember clearly, but i think the reason we didn't put clipping prevention on by default is that for the other 99.99% of people it'll make volume normalization not work as well
17:09:20gevaertssaratoga: 99.99% is a serious claim :)
17:09:22saratogasince it'll scale other people's tracks down if they have some minor clipping
17:10:22saratogaalthough i guess most people won't notice
17:10:41gevaertsI don't have a real opinion on the entire thing, but I'm not too surprised some people are a bit upset by the "But it works fine if you listen to rubbish!" (from their (and my ;) p.o.v.) argument
17:11:23gevaertsTo be honest, to my ears the music where clipping isn't a problem, clipping wouldn't make it sound worse ;)
17:11:46saratogaclipping is really common, but its usually just one or two samples
17:12:00saratogapretty much everything clips in mp3 format for instance, but you never hear it
17:12:06*gevaerts nods
17:12:49saratogai don't really care, it seems like the subset of people who scan their files with replaygain but don't know what it is or hwo to change its settings is probably too small to worry about
17:12:49gevaertsAs I said, I don't have a real opinion. We might want to mention the genre-specific peculiarities in the manual or on the wiki though
17:13:44saratogathe whole idea of defaulting on was mostly that i kept erasing my settings while debugging and hated having to turn it back on :)
17:14:27saratogai do wonder if maybe it would make sense to give our DSP engine an extra dB of headroom though
17:14:41saratogai tried one of those files posted in the forums in foobar2000 and it does not clip regardless of settings
17:14:54saratogai'm assuming foobar just keeps a little more extra headroom then we do
17:15:01evilnickIF this were to have any effect, then it'd be for people who have downloaded music files from some source that does use replygain when the downloader themselves don't.
17:15:31evilnickTherefore the files to clip would be likely to be ones that are copyright infringing anyway
17:15:43evilnickTherefore it doesn't matter :D
17:16:20gevaertsI have no idea if legal file purchase sites tend to use replaygain
17:16:56the-kyleNot that it matters all that much, but I have a 256k Opus file playing in realtime on the clip+. I get a hard freeze on the clipv1 before playback instead of on stop, but I'd never use 256k on a 2GB device anyway.
17:19:02the-kyleI purchase music from Amazon mp3, and I don't think it's using replaygain. It usually shows up as mp3 tags, but I don't see appropriate tags on those files.
17:19:38gevaertsThe mp3 music I've purchased also doesn't seem to have it
17:19:53saratogais volume proportional to intensity or amplitude? i always forget which decibel to use
17:24:39derfsaratoga: BTW, I found out there _was_ a NULL check on the global stack allocation, but it was accidentally moved inside a CUSTOM_MODES ifdef.
17:25:12saratogaso that may result in memory corruption on low memory targets?
17:25:42 Quit zahari_bgr (Read error: Connection reset by peer)
17:25:49 Join zahari_bgr [0] (~zahari.yu@
17:29:07derfsaratoga: Yes. See opus_custom_mode_create() in celt/modes.c.
17:29:30derfThe second #endif at the start there should be moved before ALLOC_STACK;
17:30:41derf(I _believe_ that's sufficient, but we're still investigating all the other ALLOC_STACK usage upstream...)
17:32:55 Quit ser (Remote host closed the connection)
17:38:35 Join ser [0] (
17:39:52EpicanisNot sure if it matters to anyone, but I just tried the latest dev build on my Sansa C240 ("c200v1") with a 96kbit stereo opus file. So far it seems to be working perfectly, no more pauses during playback. Nice!
17:40:22ParkerR_And I;m just sitting here wondering what the heck an opus file is.
17:40:57EpicanisVery nice, high-quality codec (with low-latency so it's also suitable for VoIP)
17:41:05the-kyleOoooh! 96 Kbit! Nice indeed.
17:41:44EpicanisI figure 96kbit should be approximately equivalent to 128-160kbit mp3, so great for most music.
17:42:02the-kyleI would say it's the first truly general-purpose codec. Maybe not, but hey, it does sound good.
17:42:40EpicanisI think that's probably accurate - I can't think of ANY other (free OR proprietary) that doesn't have some flaw that makes it unsuitable for one major use-case or another.
17:43:00the-kyleIf comparing to mp3, 96Kbit probably compares with somewhere between 160 and 180.
17:44:20EpicanisI tend to assume Vorbis is worth one to two steps up in equivalent quality over mp3 (so 96kbit vorbis would be approximately 112-128kbit mp3 equivalent), and Opus is one to two steps beyond that, so 160-180 is probably about right.
17:44:47ParkerR_I just stick with simple mp3s XD
17:45:10the-kylemp3 bleh!
17:45:18Epicanismp3 is too mainstream, I'm getting a player-piano. :-)
17:45:34ParkerR_Carry around the paper rolls
17:46:42EpicanisAs I've been typing, I've been listening to a just-over-9-minute bit of music encoded as 96kbit Opus (transcoded from Ogg Vorbis obtained from Jamendo, I think the original was -q5 or so, so transcoding doesn't seem to have reduced the quality noticeably).
17:47:09EpicanisThe song's just finishing up. There MAY have been a momentary pause (very short, not even sure it was "real") around the 5-minute mark, but otherwise it's played flawlessly...
17:49:23EpicanisSeems to cause some interface glitchiness (playback time updates infrequently, seems to respond slowly to some controls, and I can't figure out how to make playback actually STOP other than waiting for song to end) but it DOES actually work.
17:51:13the-kyleI'm listening to a 48 Kbit Opus file encoded from Flac ripped off an old CD, and it actually sounds much like q1 Vorbis, which encodes at ~80 Kbit usually.
17:51:49EpicanisSo in short, Opus is awesome. :-) (brb, got to run an errand...)
17:54:59 Quit speckmade (Ping timeout: 248 seconds)
17:55:32the-kyleI'm hearing some of the stereo shifting associated with low-bitrate Vorbis, but I had to go all the way down to 48k for it to be more than just a little noticeable, and its
17:55:56the-kyleand it's still mostly a minor annoyance.
17:56:38***Saving seen data "./dancer.seen"
18:00:16the-kyleDisclaimer: I use a framesize of 60ms when encoding at bitrates this low. It seems to improve quality, probably because of the way Opus handles vbr.
18:08:34 Quit scorche|1h (Changing host)
18:08:34 Join scorche|1h [0] (~scorche@rockbox/administrator/scorche)
18:12:24EpicanisI figure for regular music I'll be using 96kbit, for speech (e.g. oggcasts/"pod"casts) probably 48-64kbit depending on how much I'm worried about subtle nuances of sound...
18:12:48EpicanisCan probably get decent audiobook usage all the way down to 24kbit or so, I'd guess (haven't tested yet).
18:13:53 Quit nosa-j (Ping timeout: 260 seconds)
18:14:01 Quit zahari_bgr (Quit: Freedom Chat - Your Home Away From Home | | tIRC script by the Freedom Chat Leets)
18:14:13saratogawhats with people posting bug reports for codecs without a link to the file
18:14:41saratogaits like "a problem exists and I choose not to exchange information about it"
18:14:50saratoga"hope that helps"
18:15:03Epicanis"Like, DUDE! Something went wrong!"
18:15:06bootinfdsdsIs this video possible ?? , Can I have a chat to someone about this Please ?
18:15:24 Quit mgottschlag2 (Ping timeout: 246 seconds)
18:15:44EpicanisThey ought to be teaching how to do bug reports in high school...
18:16:17bootinfdsdsAnyone there ?
18:16:45saratogawhat is your question?
18:17:08bootinfdsdsU2 Ipod doing games like quake etc ... Is this possible ?
18:17:56saratogadoom is possible, sure
18:18:47bootinfdsds...therefore TF-Dooom ? (/Total conversions)
18:18:53bootinfdsdssoz for spaces.
18:19:14saratogai have no idea what that is, but if its doom, then yes
18:19:20bootinfdsds** TC-DooM
18:20:23bootinfdsdsDoes the plug in have a 'loaded' or is it just vanilla doom -style plug-in ?
18:21:27bootinfdsdsIs this DooM or DooM2 ???
18:22:11 Join nosa-j [0] (~m00k@
18:23:28bootinfdsdsHow do I access the command line in the DooM plugin ?
18:23:59bootinfdsdsi.e. in the game ( on the ipod)...
18:25:22amayer_speaking of doom and bug reports has anyone looked at this recently:
18:26:12saratogadoesn't look like it
18:29:37bootinfdsdsamayer_, Yeah, that's pretty standard bug, it sounds to me. I'd like to get this/these bugs fixed so we can move onto other DooM clones etc.
18:30:08 Join saratoga_ [0] (123e0cca@gateway/web/freenode/ip.
18:30:18bootinfdsdsIs there a feature request section to the website ?
18:32:00bootinfdsdsI guess feature are just future bugs, right ?
18:32:20 Join speckmade [0] (
18:32:44 Quit saratoga (Ping timeout: 240 seconds)
18:33:03 Join mortalis [0] (~mortalis@
18:33:05gevaertsIf you want a feature, you can work on it and submit a patch. People submitting feature requests without anyone working on them is just useless noise
18:33:47bootinfdsdsgevaerts, I thought it was called innovation :(
18:34:01saratoga_innovation is when you do something useful
18:34:33Epicanis(I don't think something counts as "innovation" until it's implemented...)
18:34:37bootinfdsdsAye, why not pay someone to do, that, useful
18:38:28 Join n1s [0] (
18:38:28 Quit n1s (Changing host)
18:38:28 Join n1s [0] (~n1s@rockbox/developer/n1s)
18:38:59 Quit speckmade (Quit: Leaving.)
18:39:02 Join speckmade1 [0] (
18:39:13 Join Thra11_ [0] (
18:40:25 Quit Thra11 (Ping timeout: 245 seconds)
18:43:33 Quit speckmade1 (Read error: Connection reset by peer)
18:46:06 Join speckmade [0] (
18:46:09 Quit mortalis (Quit: KVIrc 4.1.3 Equilibrium
18:47:56 Quit saratoga_ (Ping timeout: 263 seconds)
18:49:56 Join speckmade1 [0] (
18:49:57 Quit speckmade (Read error: Connection reset by peer)
18:55:18 Join speckmade [0] (
18:55:30 Quit speckmade1 (Read error: Connection reset by peer)
18:56:19 Join bertrik [0] (~quassel@rockbox/developer/bertrik)
19:05:30 Quit nosa-j (Quit: k)
19:05:43 Join nosa [0] (~m00k@
19:06:01 Nick nosa is now known as nosa-j (~m00k@
19:06:33 Join prof_wolfff [0] (
19:07:11 Join Thra11 [0] (~thrall@
19:08:12 Quit Thra11_ (Ping timeout: 240 seconds)
19:10:05 Join mgottschlag [0] (~quassel@reactos/tester/phoenix64)
19:10:45 Join perrikwp [0] (
19:11:46 Quit Thra11 (Ping timeout: 248 seconds)
19:12:46 Quit perrikwp_ (Ping timeout: 240 seconds)
19:13:00 Join Thra11 [0] (
19:14:45 Quit pamaury (Ping timeout: 256 seconds)
19:15:00 Join pretty_function [0] (~sigBART@
19:17:35 Quit speckmade (Ping timeout: 256 seconds)
19:18:16 Quit Thra11 (Ping timeout: 276 seconds)
19:19:14 Quit WalkGood ()
19:22:32 Quit Horscht (Quit: Verlassend)
19:23:37 Join speckmade [0] (
19:24:00 Quit zoktar (Quit: ZNC -
19:26:27 Join zoktar [0] (
19:27:55 Quit speckmade (Ping timeout: 245 seconds)
19:34:07 Join Thra11 [0] (~thrall@
19:43:09 Quit Thra11 (Read error: Operation timed out)
19:43:37 Join Thra11 [0] (
19:43:41 Join XavierGr [0] (XavierGr@rockbox/staff/XavierGr)
19:44:35 Join Buschel [0] (
19:48:51Buscheln1s: i was not really successful with asm'ing the C_MULC/C_MUL macros. it did not become any faster... but i could find another 0.7 MHz in mdct.c via some re-arranging ->
19:55:13 Join Guest55015 [0] (
19:55:54 Quit Guest55015 (Changing host)
19:55:54 Join Guest55015 [0] (~whiskers7@unaffiliated/whiskers75)
19:56:42***Saving seen data "./dancer.seen"
19:56:45 Nick Guest55015 is now known as whiskers75 (~whiskers7@unaffiliated/whiskers75)
19:56:47 Join pamaury [0] (
19:56:48 Quit pamaury (Changing host)
19:56:48 Join pamaury [0] (~quassel@rockbox/developer/pamaury)
19:57:21 Quit whiskers75 (Client Quit)
20:00:31n1sBuschel: c_mulc gave about 2MHz on cf, even if gcc does a better job for arm i thought it would give some improvement if only by saving a bit of shifting and or'ing
20:02:37 Join whiskers75 [0] (~whiskers7@unaffiliated/whiskers75)
20:02:49whiskers75can someone help me with this:
20:05:06Buscheln1s: maybe i am just out of training in writing assembler :/
20:07:09n1sBuschel: not many ways to get that function suboptimal i think, but i could take a look if you post code
20:07:55n1swhiskers75: sounds like hardware problems to me
20:08:14 Quit nosa-j (Ping timeout: 246 seconds)
20:09:20Buscheln1s: i've thrown it away. i'll give it a try again now, maybe i've got a better day today
20:10:02whiskers75n1s: i think so too :(
20:10:13n1ssometimes it also helps looking at what gcc does with the inline asm to spot problems
20:10:28n1slike messed up constraints
20:10:43 Join nosa-j [0] (~m00k@
20:10:48whiskers75n1s: but surely dosfsck would have yielded an error?
20:11:30n1swhiskers75: not necessarily, if your battery theory is correct for example as it will get some power from usb when connected
20:11:56n1sand dosfsck is just checking the filesystem, not the data
20:12:30 Join akaWolf [0] (~akaWolf@unaffiliated/akawolf)
20:13:11 Join speckmade [0] (
20:15:17whiskers75n1s: maybe low power to the HD, or a faulty HD?
20:15:43whiskers75n1s: but it did get the error when plugged in...
20:15:55n1smaybe something like that, it's always hard to tell
20:16:22n1scheck the data then to look for corruption
20:17:22 Quit eckoit (Quit: eckoit)
20:17:56whiskers75n1s: how?
20:18:39whiskers75n1s: it cant be corrupted, as it loads sometimes
20:18:54whiskers75n1s: the error often occurs after a reset or power loss
20:19:57whiskers75n1s: would the debug menu help?
20:20:02 Quit nosa-j (Ping timeout: 248 seconds)
20:23:12 Join nosa-j [0] (~m00k@
20:24:54 Quit whiskers75 (Quit: Lost terminal)
20:30:04 Join eckoit [0] (
20:35:54 Quit mgottschlag (Ping timeout: 246 seconds)
20:44:16Buscheln1s: -> ~0.8 MHz slower
20:46:16n1sBuschel: btw, doing 32 bit noninterleaved is quite complicated after all since the codec does a bunch of processing on the uotput
20:46:51Buschelnot good :(
20:51:56Buscheln1s: any idea why the asm code is slower?
20:53:43 Join eyfour [0] (
20:57:25 Quit nosa-j (Ping timeout: 240 seconds)
20:58:15 Quit nsx (Changing host)
20:58:15 Join nsx [0] (jm@unaffiliated/nsx)
21:01:14 Join nosa-j [0] (~m00k@
21:01:37 Join saratoga [0] (123e0cca@gateway/web/freenode/ip.
21:02:39 Quit eckoit (Quit: eckoit)
21:04:07n1sBuschel: hmm, it looks fine to me...
21:09:31Kiwicam_I figured out what the problem was with KDE Kate somehow changing the format of plain text files (eg .wps). It's creating a line break for lines longer than 1024 characters. This is it's default.
21:09:33Kiwicam_The solution is Settings/Configure Kate/Open-Save & change the Line Length Limit from 1024 to 0.
21:10:54 Nick Kiwicam_ is now known as Kiwicam (~quassel@
21:11:24 Nick Kiwicam is now known as Guest15791 (~quassel@
21:11:33 Quit eyfour (Quit: WeeChat 0.3.7)
21:12:00 Quit Guest15791 (Quit: - Chat comfortably. Anywhere.)
21:12:47saratogai still wonder about that prerotation step
21:13:48 Join Kiwicam_ [0] (~quassel@
21:14:13 Quit Kiwicam_ (Remote host closed the connection)
21:14:40 Join Kiwicam_ [0] (~quassel@
21:21:43 Quit Kiwicam_ (Remote host closed the connection)
21:22:04 Join KiwiCAM_ [0] (~quassel@
21:23:01n1sBuschel: it's spilling a lot more with that asm
21:24:58 Part KiwiCAM_
21:25:17n1sgcc also seems to have screwed up quite a lot
21:26:19 Quit akaWolf (Quit: my exit)
21:28:59 Join Horscht [0] (
21:29:00 Quit Horscht (Changing host)
21:29:00 Join Horscht [0] (~Horscht@xbmc/user/horscht)
21:29:17 Quit pretty_function (Ping timeout: 260 seconds)
21:29:37 Quit sciopat (Read error: Connection reset by peer)
21:29:52 Join sciopath [0] (
21:30:18saratogawith the power of 2 fft, gcc had an enormous amount of trouble figuring out the register allocation for the butterflies
21:30:28saratogaprobably similar here
21:30:52saratogaload store stuff on ARMv4/5 at least seems bad
21:30:59 Quit y4n (Quit: PANTS OFF!)
21:31:05n1syeah, and adding this sam that clobbers a bunch of regs pushes it completely over the edge, it doesn't even seem to be trying anymore
21:31:42Buschelso, there is not much sense in asm'ing single macros, but the whole thing
21:32:23n1slooks like it unfortunately
21:32:25saratogaprobably not much
21:32:47saratogaon the upside, you can probably make the fft a whole lot faster
21:33:25Buschelthis will be nasty...
21:33:48saratogaisn't it just pre, post, and the 4 and 5 butterflies which need it?
21:34:19Buschelyes, all of them need pretty much the same amount of cpu time: 13%
21:34:53Buscheleach of them, of course
21:34:57saratogaand its not possible to use the pre/post logic from the other MDCT?
21:35:24 Quit ender1 (Quit: The human eye is a wonderful device. With a little effort, it can fail to see even the most glaring injustice.)
21:36:07 Join ender` [0] (
21:42:51Buscheli don't see how, it looks quite different
21:44:15saratogayeah but why is it different?
21:44:20saratogathe spec says its a normal MDCT
21:44:43saratoga(other then needing non-power-of-two)
21:45:08saratoganaively I would think we could use our old mdct.c with the new fft for non-power of two sizes
21:45:09Buschelthat was my understand (we only having 2^n, opus not using 2^n)
21:45:46saratogamaybe derf can clarify?
21:47:05derfThere's lots of ways to factor an MDCT.
21:47:22saratogai'm a little confused by the pre/post rotation code
21:47:31saratogai've never seen it implemented like that
21:48:39derfIf you look at celt_mdct_backward() in, you'll see yet another way.
21:49:30derfThe fun part of that one being that it uses an FFT kernel instead of an inverse FFT.
21:50:47saratogaevery codec i've seen has used a pair of complex multiplies
21:51:18saratogai dont' really understand why this needs a pair of them followed by multiplication by 'sin'?
21:51:24saratogasorry 'sine'
21:51:43derfNo, no, it's one complex multiply.
21:51:48derfAnd then the thing with sine.
21:52:04saratogaah i see
21:52:10saratogagetting crosseyed
21:52:30derf"sine" there is sin(PI/(8*N2))
21:52:46derfcos(PI/(8*N2)) is basically 1.
21:52:51saratogathe goal being to save the extra multiplies?
21:53:04derfThere's even a comment explaining this!
21:53:33saratogai saw that
21:53:42saratogahow accurate is it?
21:54:04 Join mgottschlag [0] (~quassel@reactos/tester/phoenix64)
21:54:34derfcos(PI/(8*120)) = 0.999995
21:54:41derfAnd that's the worst case.
21:54:42n1sBuschel: splitting the store so it requires one less reg helps a little but still slower than before so yeah i see no other way thean full blown asm of the butterflies
21:56:45***Saving seen data "./dancer.seen"
21:58:22 Join petur [0] (~petur@rockbox/developer/petur)
22:01:46 Quit factor (Read error: Connection reset by peer)
22:13:05 Quit fyre^OS (Quit: quit)
22:14:11 Join factor [0] (~factor@
22:26:56 Quit Rower85 (Read error: Connection reset by peer)
22:29:07 Join Rower85 [0] (husvagn@
22:32:07Buscheln1s: i tried to write inline asm for the three consecutive C_MULC calls in the 4x-butterfly. still slower... gotte get some sleep now. tomorrow will start early for me
22:32:11 Join Strife89 [0] (~Strife89@
22:32:24 Quit Buschel (Quit: ChatZilla 0.9.89 [Firefox 15.0.1/20120905151427])
22:46:22bootinfdsdsQu. I have $10k to invest in Rockbox HOW and Where do I do this .. ?
22:46:31 Join Gallomimia [0] (
22:46:32 Quit Gallomimia (Excess Flood)
22:47:10 Join Gallomimia [0] (
22:47:20Bagderget it all in unmarked small bills and meet us midnight in the park!
22:48:27 Nick scorche|1h is now known as scorche|sh (~scorche@rockbox/administrator/scorche)
22:49:59bootinfdsdsRegents park is 10 mins away ... Is that OK ?
22:50:59scorche|shwe can neither confirm nor deny the existence of an agent in that area
22:51:00bootinfdsdsOnly got $ŁK in notes .. I'll have to cheque the rest, sorry.
22:52:35*[Saint] isn't sure that Rockbox classes as an "investment"
22:52:47[Saint]That implies you'll see a return. You won't.
22:53:01Bagderwell, we could convert it into buying into brewery products!
22:53:07[Saint]It'd be a donation, pure and simple. Nothing more.
22:53:19Bagderyeah, the return is not that good
22:53:47bootinfdsdsHow would it be spread across the group ?
22:54:03[Saint]Drunk, bloated, contented, meat-filled devs....sounds good to me :)
22:54:03bootinfdsdsor is it just the website costs?
22:54:39bootinfdsdsDoesn't anyone want to develop a new game or App. ?
22:54:39[Saint]bootinfdsds: various ways, I assume a lot of it would go to "the beer fund", some may go to new developers requesting devices, etc.
22:55:00*[Saint] points out that "the beer fund" pays for more than just beer.
22:55:04Bagderthe costs for the site is covered by a sponsor anyway
22:55:08[Saint] buys meat too! :P
22:55:13gevaertsbootinfdsds: if you want to pay for a specific feature, you can always talk to an individual developer
22:55:30 Join speckmade1 [0] (
22:55:49 Quit speckmade (Ping timeout: 245 seconds)
22:56:04scorche|shBagder: well, your server, at least ;)
22:56:04bootinfdsdsPersonally , I'd like to do a kickstarter icontrol-pad device for the ipod classic ( 4 games / etc).
22:56:26Bagderscorche|sh: ah true, well the other servers are also sponsored at least =)
22:56:28[Saint]scorche|sh: I see what you did there ;)
22:56:44gevaertsscorche|sh: I believe some kind soul also hosts our forums for free
22:56:50bootinfdsdsBadger, whos the sponsor ?
22:56:55scorche|shgevaerts: it isnt free for him! ;)
22:57:01Bagderbootinfdsds: Haxx and scorche
22:57:02gevaertsscorche|sh: there is that :)
22:57:18 Join speckmade [0] (
22:57:30Bagderand a few more probably that I forget just now
22:57:40*gevaerts doesn't think trying to turn rockbox into a gaming device is a good idea
22:57:43bootinfdsdsAnyone =! my kickstarter idea , or have a new one ?
22:57:48scorche|shi have taken over most of rasher's stuff
22:57:57bootinfdsds**Anyone +1 my kickstarter idea , or have a new one ?
22:58:02*[Saint] agrees with gevaerts
22:58:22[Saint]It's a great idea, I'm just not sure that more than a dozen or so other people will think so.
22:58:41[Saint]It's pretty clearly not a gaming device.
22:59:35bootinfdsdsWell, I may be wrong, but I thought it could be directly connected to a projector or HDTV .. and therefore connected to a gaming pad also ?
23:00:10bootinfdsdsMy mum uses it on her tv
23:00:32 Quit speckmade1 (Ping timeout: 260 seconds)
23:01:01*bootinfdsds says : no jokes about bootinfdsds mum, please
23:01:48gevaertsIf you want a homebuilt gaming device that's not very good, base it on a raspberry pi. *much* cheaper, direct USB ports, and more powerful altogether
23:02:15Bagderand way faster than most rockbox targets too =)
23:03:16bootinfdsdsipod more ubiquitious, though.
23:03:39gevaertsSo are cardboard boxes
23:03:47bootinfdsdsand accepted in society.
23:03:49BagderI mean, ipods that people are prepared to run custom software on
23:03:51gevaertsThat doesn't make them into usable gaming devices
23:05:09 Quit amayer_ (Quit: going ~/)
23:06:05bootinfdsdsI think if you showed DooM or another fps/rpg video on kickstarter to show what Rockbox Can Do, would make people buy into the brand more.
23:07:15gevaertsYou can play doom on a toaster these days
23:07:30bootinfdsdsPeople just need to buy the icontrol thingy ( that is already a 1 million dollar hit on kickstarter ) and you have proper used the open-source philosophy of the rockbox project, haven't you ?
23:09:11gevaertsTricking people into investing large anmounts of money in a useless idea is open-source philosophy now?
23:09:33bootinfdsdsWhat is the price of the best 2nd hand ipod that works with rockbox these days, anyway ?
23:12:05bootinfdsdsgevaerts, Attracting a larger crowd of the cornerstone of open-source perpetuity. And you do this though crowd-sourcing/funding these days.
23:12:12 Join linuxstb [0] (
23:12:23gevaertsYou're talking nonsense
23:12:32bootinfdsds**crowd is
23:12:51bootinfdsdsgevaerts, Your in a culvert.
23:14:07 Part LinusN
23:14:43bootinfdsdsI mean , what happens if your sponsors pull out ?
23:15:29gevaertsThen we buy them a beer and they'll reconsider
23:15:36 Join speckmade1 [0] (
23:15:53 Quit speckmade (Quit: Leaving.)
23:16:02 Quit bertrik (Ping timeout: 260 seconds)
23:19:04bootinfdsdsin 50 years ?
23:20:44markunbootinfdsds: I think you can give up now trying to convice gevaerts :)
23:26:07bootinfdsdsmarkun, Yeah, source-code is all we need anyhow.
23:26:26*bootinfdsds akes his borg posistion, again.
23:27:12gevaertsAs long as you remember the GPL...
23:27:26saratogai don't see why you'd use an ipod for that when everyone has smartphones that can easily run doom
23:27:59saratogaapple doesn't put games on the old ipods because theres no demand for it
23:31:36 Quit hype (Quit: ["Textual IRC Client:"])
23:34:34bootinfdsdsIt's not for the 8ipod8 .. Its for the HDTV !
23:35:14 Quit sciopath (Read error: Connection reset by peer)
23:35:22 Join sciopath [0] (
23:35:41saratogathen using an ipod is even dumber
23:35:59saratogapick something cheaper with a GPU and HDMI
23:37:50bootinfdsdsbut I have an ipod ... I'm recycling !
23:38:01 Quit Strife89 (Quit: Heading home.)
23:39:28bootinfdsdsSo do 300 million others.
23:40:34 Join Provel_ [0] (
23:41:25 Join amayer [0] (
23:41:42*bootinfdsds thinks, sometimes, you cant help change a hacker community for love nor money !
23:42:34bootinfdsds**double negitive [love or money]
23:43:01gevaertsIf you want to convince people, having a convincing idea is always a good start
23:43:40saratoga300 million ipods, but very few of those are the classic
23:44:28the-kyle300 million iPods? Sounds like more than there should be. Smiles.
23:45:04 Quit Provel (Ping timeout: 264 seconds)
23:45:20the-kyle300 million iPods == more iPods than people who want them.
23:45:44bootinfdsdsI have them !
23:45:54bootinfdsds** I'll have them.
23:46:02the-kyleYou have 300 million iPods? Where do you store them all?
23:47:12saratogarockbox-community is probably a better place for off topic stuff
23:47:39the-kyleTrue, as per the IRC guidelines.
23:47:57 Join Thra11_ [0] (~thrall@
23:51:04 Quit Thra11 (Ping timeout: 264 seconds)
23:54:26 Quit Thra11_ (Ping timeout: 248 seconds)
23:56:47***Saving seen data "./dancer.seen"

Previous day | Next day