[Ffmpeg-devel-irc] ffmpeg-devel.log.20161017

[burek]

[00:59:05 CEST] <philipl> BtbN: remind me how you force the framerate so that deinterlaced transcoding works
[00:59:18 CEST] <BtbN> -r as input option
[00:59:38 CEST] <BtbN> the deinterlaced framerate, that is.
[00:59:43 CEST] <BtbN> For 25i, you'd pass 50
[01:00:13 CEST] <philipl> cool. that works
[01:01:42 CEST] <philipl> So, I poked at the avi seeking thing more. The best solution I've got so far is (specifically for mpv) to turn on genpts.
[04:40:41 CEST] <cone-222> ffmpeg 03Shivraj Patil 07release/3.1:d89979e86b32: avutil/mips/generic_macros_msa: rename macro variable which causes segfault for mips r6
[04:40:41 CEST] <cone-222> ffmpeg 03Shivraj Patil 07release/3.1:d2566b124af8: Support for MIPS cpu P6600
[04:40:41 CEST] <cone-222> ffmpeg 03Hendrik Leppkes 07release/3.1:263add4462a4: ffmpeg: remove unused and errorneous AVFrame timestamp check
[04:40:41 CEST] <cone-222> ffmpeg 03Moritz Barsnick 07release/3.1:fc36e692c4ac: tools: fix grammar error
[04:40:41 CEST] <cone-222> ffmpeg 03Moritz Barsnick 07release/3.1:30c80e81d2a5: lavc: fix typos
[04:40:41 CEST] <cone-222> ffmpeg 03Moritz Barsnick 07release/3.1:f12c0da09b04: lavfi: fix typos
[04:40:41 CEST] <cone-222> ffmpeg 03Moritz Barsnick 07release/3.1:7fefd776682a: cmdutils: fix typos
[04:40:42 CEST] <cone-222> ffmpeg 03Michael Niedermayer 07release/3.1:08eef74a39d7: avformat/utils: Update codec_id before using it in the parser init
[04:40:43 CEST] <cone-222> ffmpeg 03Moritz Barsnick 07release/3.1:6109c10b81d6: doc: fix various typos and grammar errors
[04:48:18 CEST] <cone-222> ffmpeg 03Michael Niedermayer 07release/3.1:675258764dce: Update for 3.1.5
[05:02:58 CEST] <cone-222> ffmpeg 03Shivraj Patil 07release/3.0:ef5b120e2dd6: avutil/mips/generic_macros_msa: rename macro variable which causes segfault for mips r6
[05:02:59 CEST] <cone-222> ffmpeg 03Shivraj Patil 07release/3.0:c993a11e56ab: Support for MIPS cpu P6600
[05:03:00 CEST] <cone-222> ffmpeg 03Hendrik Leppkes 07release/3.0:27f1fb125143: ffmpeg: remove unused and errorneous AVFrame timestamp check
[05:03:01 CEST] <cone-222> ffmpeg 03Moritz Barsnick 07release/3.0:6457346e6a15: tools: fix grammar error
[05:03:02 CEST] <cone-222> ffmpeg 03Moritz Barsnick 07release/3.0:17b8e7799f64: lavc: fix typos
[05:03:03 CEST] <cone-222> ffmpeg 03Moritz Barsnick 07release/3.0:8baf2d8fadc1: lavfi: fix typos
[05:03:04 CEST] <cone-222> ffmpeg 03Moritz Barsnick 07release/3.0:88f52f2f8f21: cmdutils: fix typos
[05:03:05 CEST] <cone-222> ffmpeg 03Michael Niedermayer 07release/3.0:047e0049d1cc: Changelog: update for recent commits
[07:43:53 CEST] <atana> michaelni: hi
[07:45:15 CEST] <atana> I am not sure how to use fft function of ffmpeg. I tried on a test code https://dpaste.de/es8K. could you point me to an example or something
[12:48:35 CEST] <michaelni> atana, example code for all the ffts should be in libavcodec/tests/fft.c, (ignore all code in it that doesnt apply to avfft)
[12:50:41 CEST] <cone-192> ffmpeg 03Hendrik Leppkes 07master:656feb641de3: mov: move stsd finalization to an appropriate place
[13:23:37 CEST] <atana> michaelni: I have made changes to `libavfilter/avfilter.c` and libavfilter/avfilter.h in the ffmpeg repo cloned from github. after configuration I installed it using make install. Now, I want to test the filter by running a driver program. How do I compile the program? I ran `gcc -g pp.c -o pfilter -Iinclude -Llib` but it says undefine reference to function peakPointsFilter however I have included it in libavfilter/avfilter.h file
[13:23:47 CEST] <atana> am I missing some flag?
[13:32:03 CEST] <michaelni> atana, the easiest way to test a filter is with ffplay or ffmpeg command line programs
[13:32:36 CEST] <michaelni> but either way maybe you are missing a line for the new filter in libavfilter/Makefile
[13:33:24 CEST] <michaelni> ./ffplay somerandomvideo.mp4 -vf myfilter=myfiltersarguments
[13:34:09 CEST] <atana> I haven't done any changes in libavfilter/Makefile. I will check
[13:35:02 CEST] <michaelni> better ffplay example: (now with audio inszead of video filter) ./ffplay random.mp3 -af myfilter=myfiltersarguments
[13:36:55 CEST] <atana> are myfiltersarguments separated by comma?
[13:39:19 CEST] <ubitux> atana: myfilter=arg0:arg1:arg23=val23:arg57=val57
[14:22:58 CEST] <atana> michaelni: I could't figure out what to include in Makefile. This is the git diff of what I have done https://dpaste.de/fReB . function `peakPointsFilter` is included in avfilter.h and is defined in avfilter.c but still on compilation `gcc -g pp.c -o pfilter -Iinclude -Llib` it says undefine reference to function peakPointsFilter
[15:00:38 CEST] <cone-192> ffmpeg 03Aman Gupta 07master:9ea91e4114e8: lavc/videotoolboxenc: implement a53cc
[15:00:39 CEST] <cone-192> ffmpeg 03Rick Kern 07master:aa413b810a7f: lavc/videotoolboxenc: flush/free frames on close
[15:00:40 CEST] <cone-192> ffmpeg 03Rick Kern 07master:9875695e2cee: lavc/videotoolboxenc: Update a53cc handling
[15:00:41 CEST] <cone-192> ffmpeg 03Rick Kern 07master:d3874b74f35b: lavc/videotoolboxenc: Error log formatting.
[15:03:36 CEST] <barb_> Hello, I'd like to apply for outreachy but the deadline is today and I won't have the time to finish my qualification task. Does anybody know if maybe I could start contributing today(and finish something consistent by the end of the week), could I still apply and mention that in my application? I'm asking because since the decisions are only final in November 08, so there would be still time to evaluate me.
[15:06:02 CEST] <Compn> barb_ : sure, but we dont have qual tasks set up
[15:06:10 CEST] <Compn> i dont think
[15:06:47 CEST] <Compn> but yeah it cant hurt to submit an application
[15:07:06 CEST] <Compn> we dont mind if you do the qual task after the application
[15:07:13 CEST] <Compn> to answer your question :)
[15:09:52 CEST] <barb_> Oh, great, so.. I talked to Paul Mahol about the XPM decoder project and he mentioned something about extending bitdepth support to some filter
[15:10:19 CEST] <barb_> but then mentioned that I could pick another project myself, like creating a QR generator
[15:10:45 CEST] <barb_> do you have any other ideas for tasks?
[15:11:10 CEST] <barb_> maybe more appropriate to the QR generator project
[15:13:29 CEST] <barb_> Or maybe you guys have any particular filter(s) in min for this addition
[15:13:55 CEST] <barb_> *in mind
[15:55:02 CEST] <cone-192> ffmpeg 03Andreas Cadhalpun 07master:eb751f06db9f: matroskadec: fix NULL pointer dereference
[16:31:42 CEST] <Compn> barb_ : ah, no idea :)
[16:31:55 CEST] <Compn> qr generator you say?
[16:31:58 CEST] <Compn> hmmm
[16:32:06 CEST] <Compn> never used a qr code myself :)
[16:32:26 CEST] <Compn> i am upset they dont just print the website under the qr code, its quicker for me to type it in :P
[16:32:46 CEST] <Compn> if we are talking about the same QR thing... :D
[16:35:13 CEST] <Compn> barb_ : there was someone who wanted to pay a developer to work on 4k visual perception filter
[16:35:44 CEST] <Compn> e.g. to cut encoding bits by discarding changes that the human eye cannot really tell the difference of
[16:35:59 CEST] <wm4> lolwut
[16:36:23 CEST] <Compn> like the audio encoding does, audio ... whatever its called
[16:37:03 CEST] <Compn> wm4 : comcast guy is hiring
[16:38:24 CEST] <Compn> https://en.wikipedia.org/wiki/Perceptual_audio_coder
[16:38:31 CEST] <Compn> that , except for video
[16:41:04 CEST] <wm4> that should be within an encoder, not in a separate libavfilter filter (which is what your statement sounded like to me)
[16:42:03 CEST] <barb_> hahaha yes, we're talking about the same QR code hahaha
[16:42:19 CEST] <barb_> I agree, there are link shorteners for a reason
[16:42:45 CEST] <barb_> but they are quite popular nowadays so... hahaha
[16:43:02 CEST] <barb_> and it seems to have a pretty cool algorithm behind it
[16:45:08 CEST] <barb_> hmmmm, interesting (the discarding thing)
[16:46:53 CEST] <Compn> wm4 : you dont think its possible to want a filter to introduce as a step in between raw source and a standard video codec that does not have perceptual feature ?
[16:47:15 CEST] <Compn> or whatever hardware encoder that does not have that*
[16:47:27 CEST] <Compn> who knows, i dont
[16:47:34 CEST] <Compn> but if you want a job i have his contact info :D
[16:48:13 CEST] <kierank> What has a qr generator got to do with FFmpeg
[16:48:17 CEST] <kierank> Or multimedia
[16:49:26 CEST] <barb_> Compn : I'm not sure I'm fit for the job, I've seem some stuff about it in college, but very superficially, but who knows
[16:50:16 CEST] <barb_> Compn : So, yes, I'd like to get his contact, maybe we can arrange something
[16:51:22 CEST] <barb_> kierank : Maybe because it's a data encoding process that has a media as an output?
[16:52:30 CEST] <Compn> its text input though ?
[16:52:42 CEST] <Compn> >forgot how qr works :D
[16:52:42 CEST] <barb_> I believe so
[16:52:54 CEST] <Compn> we dont usually deal with text inputs
[16:53:12 CEST] <Compn> some subtitles stuff
[16:53:33 CEST] <barb_> mmm
[16:54:38 CEST] <barb_> so, do you think it would be better to drop this idea?
[16:56:20 CEST] <Compn> kierank doesnt like any of the project ideas in gsoc or opw.... :P
[16:56:32 CEST] <barb_> hahaha really? why?
[16:56:50 CEST] <Compn> we are an ornery bunch here.
[16:57:06 CEST] <funman> can swresample do S32 -> S16 without going through FLTP as intermediate format?
[16:57:32 CEST] <Compn> funman : rgb ? yuv ?
[16:57:42 CEST] <Compn> err audio
[16:57:51 CEST] <Compn> my brain is frozen today.
[16:58:30 CEST] <funman> Compn: palettized audio
[16:58:57 CEST] <funman> problem is it does 48000 fps right now (floats per second)
[17:00:48 CEST] <funman> it's not clear whether swr can run without an intermediate sample format
[17:01:04 CEST] <barb_> Compn : haha did you guys try to suggest other ideas? Or it was not up to you?
[17:01:37 CEST] <barb_> Compn : Like, there are not many of them in the official page
[17:02:26 CEST] <Compn> anyone can suggest ideas, its difficult to think up ideas 
[17:03:02 CEST] <Compn> that can be done in that timeframe
[17:03:41 CEST] <barb_> mmmm, got it
[17:05:25 CEST] <Compn> by someone who is new to the code, as there are some advanced concepts in multimedia
[17:06:56 CEST] <barb_> sure
[17:07:13 CEST] <barb_> Paul also suggested a new audio denoising filter using wavelets
[17:07:40 CEST] <barb_> would it be more useful to the purpose of the organization? 
[17:08:46 CEST] <barb_> (I keep mentioning him and not asking him directly because I only have his email, so it's a bit complicated to get this kind of insight in such short notice)
[17:09:16 CEST] <Compn> better ask the admins, i'm just here to take up space
[17:09:35 CEST] <Compn> michaelni or durandal_170 (paul's nickname, not here atm)
[17:09:44 CEST] <barb_> haha ok, but thank you anyway :)
[17:09:45 CEST] <Compn> or whoever else is opw admin / mentor
[17:09:51 CEST] <michaelni> atana, you should add a avfilter as in a new file with AVFilter struct and implemenattion not adding code directly to avfilter.c
[17:09:52 CEST] <Compn> no problem
[17:10:04 CEST] <michaelni> atana, see git log of libavfilter for examples on how to do that
[17:10:22 CEST] <wm4> "palettized audio" 
[17:10:23 CEST] <wm4> what and why
[17:12:32 CEST] <funman> michaelni: should swr be able to run without needing an intermediate sample format?
[17:14:39 CEST] <wm4> funman: what for
[17:14:47 CEST] <wm4> this is just going to make everything more complex and slower
[17:15:01 CEST] <funman> wm4: for s32 -> s16
[17:15:58 CEST] <michaelni> for simple sample convert (no resamping) it should not need an intermediate sample format
[17:16:15 CEST] <michaelni> as in int32 -> int16
[17:16:50 CEST] <funman> michaelni: ok thanks
[17:17:09 CEST] <kierank> funman: try with avresample =p
[17:17:29 CEST] <funman> kierank: obviously i thought of it
[17:18:05 CEST] <wm4> I think I've seen code that makes swr use floats only if resampling or rematrixing is involved
[17:22:15 CEST] <funman> yeah seems i'm not setting channel layouts correctly, that's why it goes through float
[17:30:21 CEST] <barb_> michaelni : any input on the QR Code discussion? Or any suggestion on a new denoising filter?
[17:36:01 CEST] <michaelni> barb_, QR generator (and reader maybe too?) could be interresting projects if pauls (or someone else) is ok with mentoring these
[17:37:26 CEST] <michaelni> about audio denoising i am not upto date with that area so i cannot comment
[17:39:37 CEST] <funman> michaelni: works perfectly now, no more intermediate format
[17:41:42 CEST] <barb_> michaelni: Oh, I see, so I think I'll commit to the QR generator/reader. Thank you! :)
[17:44:03 CEST] <wm4> why does ffmpeg need to be able to read/write Q codes
[17:44:05 CEST] <wm4> *QR
[17:47:38 CEST] <kierank> ^ this
[17:47:52 CEST] <kierank> hey look michaelni, this is how we ended up with a web server last time
[17:48:09 CEST] <kierank> but no we will shoot ourselves in the foot again
[17:49:07 CEST] <kierank> barb_: I'm sure you mean well but I can't see how a QR code reader and writer fits in ffmpeg
[17:49:12 CEST] <wm4> do we still have a web server?
[17:50:00 CEST] <barb_> =S
[17:50:20 CEST] <barb_> I understand, but do you have anything else in mind?
[17:50:31 CEST] <barb_> Like a new feature? 
[17:51:02 CEST] <cone-192> ffmpeg 03Carl Eugen Hoyos 07master:f04c27fe7ce9: lavc/videotoolboxenc: Enable a53cc by default.
[17:51:37 CEST] <barb_> Since I couldn't finish a qualification task on time, I had kind of given up, then today I decided to check with you guys.
[17:51:45 CEST] <barb_> And try it anyway
[17:52:31 CEST] <barb_> So I needed something new, cause I suppose the suggested projects already have enough submissions
[17:52:59 CEST] <iive> are qr codes in question these qube barcodes?
[17:53:19 CEST] <michaelni> I see noting wrong with a QR reader or generator, it fits in the many diverse filters we have in libavfilter
[17:53:33 CEST] <wm4> I don't know about the project status, but everyone is welcome to contribute any time
[17:53:54 CEST] <nevcairiel> its not much of a filter, is it? cant exactly take video and make a QR code out of it, or audio for that matter
[17:54:17 CEST] <wm4> libavfilter is a really weird framework for such arbitrary things, and you'll end up with an awkward user interface for low quality implementations
[17:54:18 CEST] <michaelni> you can scan a video for QR codes and extract metadata 
[17:54:54 CEST] <michaelni> also QR codes can get a overlaid human readable text by such filter for example
[17:55:27 CEST] <kierank> wm4: demand a vote
[17:55:52 CEST] <wm4> who cares
[17:55:53 CEST] <michaelni> id consider it fun if i could on the flick of a button see QR codes in a video in human readable form
[17:56:06 CEST] <wm4> "flick of a button"
[17:56:13 CEST] <wm4> more like writing a 10KB command line
[17:56:14 CEST] <michaelni> switch
[17:56:32 CEST] <wm4> lol
[18:16:23 CEST] <cone-192> ffmpeg 03Andreas Cadhalpun 07release/3.1:1a43626fdf30: configure: fix detection of libopenjpeg
[18:16:24 CEST] <cone-192> ffmpeg 03Andreas Cadhalpun 07release/3.1:a22155dacd93: libopenjpegenc: stop reusing image data buffer for openjpeg 2
[18:16:25 CEST] <cone-192> ffmpeg 03Andreas Cadhalpun 07release/3.1:d391719be19b: libopenjpegenc: fix out-of-bounds reads when filling the edges
[18:16:26 CEST] <cone-192> ffmpeg 03Andreas Cadhalpun 07release/3.1:a2d3e7392d2d: Changelog: update for recent commits
[18:17:22 CEST] <cone-192> ffmpeg 03Andreas Cadhalpun 07release/3.0:ada229e66fea: configure: fix detection of libopenjpeg
[18:17:23 CEST] <cone-192> ffmpeg 03Andreas Cadhalpun 07release/3.0:de42af2beef7: libopenjpegenc: stop reusing image data buffer for openjpeg 2
[18:17:24 CEST] <cone-192> ffmpeg 03Andreas Cadhalpun 07release/3.0:0887f1835d69: libopenjpegenc: fix out-of-bounds reads when filling the edges
[18:17:25 CEST] <cone-192> ffmpeg 03Andreas Cadhalpun 07release/3.0:f309b698e6b7: Changelog: update for recent commits
[18:26:58 CEST] <cone-192> ffmpeg 03Andreas Cadhalpun 07master:c143a9c96ff9: aiffdec: fix division by zero
[18:41:21 CEST] <kierank> wm4: are we getting a qr code filter then
[18:42:01 CEST] <JEEB> it didn't go anywhere
[18:42:07 CEST] <wm4> yeah, there was no further discussion
[18:42:10 CEST] <JEEB> as in, the discussion stopped after the /part
[18:42:33 CEST] <JEEB> or more like, had finished before that but anyways
[18:43:28 CEST] <JEEB> also is the way to go with XML parsing now libxml2 or libexpat?
[18:43:42 CEST] <JEEB> if someone would start working on a DASH demuxer
[18:44:52 CEST] <wm4> if I do the latter, I'll use libexpat
[18:45:01 CEST] <JEEB> ok
[19:10:05 CEST] <RiCON> probably better to wait/decide with hds demuxer before starting to work on that?
[19:11:08 CEST] <RiCON> if the hds author goes with libxml2 and dash uses expat one would end with two separate xml libs in ffmpeg?
[19:12:11 CEST] <cone-192> ffmpeg 03Carl Eugen Hoyos 07master:df9b151a1bdc: doc: Update x264 option a53cc, forgotten in 93e04102.
[19:15:35 CEST] <Chloe> We obviously just need to write our own xml lib within libavutil
[19:17:14 CEST] Action: JEEB gently stabs stabs Chloe 
[19:17:17 CEST] <BtbN> If there is already an xml parser ready to be used via configure, used by other parts of ffmpeg, using it seems a reasonable requirement.
[19:19:22 CEST] <RiCON> since people can choose  between 4 tls libs why not choose between 2/3 xml libs too, one might say
[19:19:52 CEST] <Chloe> barb_: I see you were interested in the XPM encoder/decoder, why don't you just do that?
[19:20:07 CEST] <JEEB> BtbN: yes - that's why I was wondering which way the whole thing was heading
[19:20:19 CEST] <JEEB> because I sure as hell wouldn't want to rewrite my XML parsing code
[19:21:52 CEST] <barb_> Chloe : I'll be applying for that project as well, I was just going for a plan B, just in case somebody else wants to do it too.
[19:24:52 CEST] <nevcairiel> afaik we currently have funding for only one outreachy applicant either way, so your best chances would likely be to put all the energy into the most promising project
[19:25:21 CEST] <nevcairiel> (unless some more funds appeared when I wasnt paying attention)
[19:25:36 CEST] <wm4> <RiCON> if the hds author goes with libxml2 and dash uses expat one would end with two separate xml libs in ffmpeg? <- not like I care
[19:25:53 CEST] <Chloe> I care, pls no
[19:26:23 CEST] <wm4> I encouraged that contributor to use libexpat, and whether others agree or not or whether they agree the patch in whatever form is not up to me
[19:27:55 CEST] <Chloe> XML isn't like TLS, there is no benefit for using multiple libraries, also 3/4 of the TLS libraries are API compatible (afaik). There is little reason to use yet another XML library
[19:29:32 CEST] <nevcairiel> the only real benefit of supporting both openssl and gnutls is licensing and system availability - which is the same argument one might make for any kind of library
[19:29:44 CEST] <nevcairiel> not that it makes it any better
[19:29:57 CEST] <BtbN> So libavxml it is
[19:31:30 CEST] <barb_> nevcairiel : mmmm, I see...  
[19:33:14 CEST] <Plorkyeran> the only unusual thing about TLS libraries is that none of them have a reasonable license
[19:33:28 CEST] <Plorkyeran> so it's easier to argue for giving the user the choice of which unreasonable license they want
[19:34:48 CEST] <nevcairiel> gnutls is  lgpl, is that so bad? its the mostl iberal license you can build ffmpeg under as well
[19:35:36 CEST] <Plorkyeran> oh right it is
[19:36:02 CEST] <Plorkyeran> I have gnutls mentally lumped under "awkward license" because it was an issue for using it in ios apps
[19:36:03 CEST] <wm4> more importantly all tls libs suck
[19:36:07 CEST] <Plorkyeran> but that's an unusual case
[19:36:47 CEST] <JEEB> btw, was the internal representation of subtitles in lavc ASS?
[19:36:50 CEST] <Compn> i thought openssl sucks and was full of buggy holes, and thats why people were moving to polarssl and gnutls :P
[19:37:06 CEST] <JEEB> so if you want to support various styled text related things you have to parse ASS tags?
[19:37:13 CEST] <JEEB> (if you encode subtitles)
[19:37:38 CEST] <Compn> i think ffmpeg originally converted all srt and subtitle tags to ass internally, yes
[19:37:46 CEST] <Compn> but i think that has since been changed.
[19:38:14 CEST] <nevcairiel> it should still do that if you use a subtitle "decoder"
[19:38:22 CEST] <JEEB> ok, I'm just thinking of implementing a subtitle format and the first thing of course that comes to mind is the internal representation
[19:38:33 CEST] <Compn> ah
[19:38:34 CEST] <JEEB> from which I would then create the output format
[19:38:48 CEST] <nevcairiel> ubitux knows all the details of course, but from what I understand internally all text subtitles are "decoded" to ass syntax
[19:38:55 CEST] <JEEB> ok
[19:39:36 CEST] Action: Compn members the old days when people complained about the "ass" name
[19:44:00 CEST] <ubitux> JEEB: yes, internal decoded representation of text subtitles is ASS
[19:44:13 CEST] <ubitux> text subtitles decoders output ASS
[19:44:25 CEST] <wm4> "decoders"
[19:44:32 CEST] <ubitux> (and demuxers just split the original format)
[19:44:55 CEST] <ubitux> encoders have access to an ASS parser which allows them to define a bunch of callbacks
[19:45:08 CEST] <ubitux> for handling different styles
[19:48:05 CEST] <JEEB> ok
[19:48:19 CEST] <JEEB> like italic and crap like that I guess
[19:48:27 CEST] <Compn> yep
[19:48:43 CEST] <Compn> italic, underline, bold...
[19:48:53 CEST] <Compn> colors
[19:49:04 CEST] <JEEB> alright
[19:55:18 CEST] <cone-192> ffmpeg 03Muhammad Faiz 07master:92de89d14acb: avfilter/firequalizer: optimize gain/gain_entry command
[19:55:19 CEST] <cone-192> ffmpeg 03Muhammad Faiz 07master:923488bdaaab: avfilter/showcqt: add font option
[20:46:45 CEST] <cone-192> ffmpeg 03Andreas Cadhalpun 07master:9959a52b14bc: astdec: fix division by zero
[20:46:46 CEST] <cone-192> ffmpeg 03Andreas Cadhalpun 07master:bc7e128a6e8e: westwood_aud: prevent division by zero
[23:31:18 CEST] <atomnuker> michaelni: could you test the v3 of the AAC SIMD for crashes, should be up on the ML soon?
[23:36:52 CEST] <rcombs> ubitux: might be a good time to make the internal format """pseudo-ass""", wherein backslash-escapes are Actually A Thing
[23:37:34 CEST] <ubitux> rcombs: we'll probably do that when assv5 is out ;)
[23:43:51 CEST] <michaelni> atomnuker, will do
[00:00:00 CEST] --- Tue Oct 18 2016