[Ffmpeg-devel-irc] ffmpeg-devel.log.20160815

[burek]

[00:25:16 CEST] <Compn> x264 can relicense for $$$
[00:25:23 CEST] <Compn> iirc they have a seperate license ...
[00:25:26 CEST] <Compn> i might be wrong
[02:14:32 CEST] <cone-259> ffmpeg 03Maya Rashish 07master:23f505bee0f8: configure: All sun4 machines are SPARCs, not just sun4u
[02:17:51 CEST] <kierank> Chloe: some of x264asm is more liberal
[02:55:10 CEST] <cone-259> ffmpeg 03Ivan Uskov 07master:b93e2233155e: libavcodec/qsvdec_h2645.c: switch to the new BSF API
[05:54:02 CEST] <Timothy_Gu> Chloe: x86inc is ISC license (more liberal than MIT). x86utils is LGPL
[07:11:33 CEST] <Compn> oh host key changed when box moved.
[09:34:47 CEST] <flux> hmm, so I suppose the part where the github mirror says "Github pull requests should be avoided because they are not part of our review process." still holds true?
[09:36:09 CEST] <flux> ah, of course you mean using the github mechanism for pull requests, pull requests to the mailing list are perfectly fine?
[09:36:21 CEST] <flux> even if said pull request is hosted on gh ;)
[09:37:24 CEST] <nevcairiel> we prefer direct patch mails on the ML just like every other patch
[09:37:34 CEST] <nevcairiel> those dont require any kind of hosting
[09:38:32 CEST] <flux> unfortunately it seems technically difficult to have "git email" work with our mail system..
[10:00:20 CEST] <cone-450> ffmpeg 03Paul B Mahol 07master:ac18114b96fb: avfilter/vf_zoompan: fix pzoom usage
[10:00:20 CEST] <cone-450> ffmpeg 03Paul B Mahol 07master:4899953a479f: doc/filters: add yet another zoompan usage example
[11:35:30 CEST] <crossle> How about this https://github.com/FFmpeg/FFmpeg/pull/208
[11:37:13 CEST] <crossle> FFmpeg videotoolbox encode error when enter background mode on iOS
[13:23:40 CEST] <cone-450> ffmpeg 03Paul B Mahol 07master:ce5ba7707945: avfilter/avf_showspectrum: add some multithreading support
[13:23:41 CEST] <cone-450> ffmpeg 03Paul B Mahol 07master:0ea03dbbf988: avcodec/iff: remove palette swapping for anim
[14:28:59 CEST] <durandal_1707> michaelni: why you haven't used fft when doing find_rect filter?
[15:04:26 CEST] <michaelni> durandal_1707, lazy
[15:07:36 CEST] <michaelni> IIRC the filter was a consulting job and the customer didnt pay, so ive no motivation to make that filter work really well
[15:08:41 CEST] <michaelni> but feel free to improve it
[15:10:22 CEST] <Chloe> Could a int64 wav be generated, or wouldnt that be accepted as a sample?
[15:10:39 CEST] <Chloe> (I mean manually)
[15:10:52 CEST] <nevcairiel> if you j ust make one to get someone to implement it, that kinda defeats the purpose doesnt it =P
[15:11:19 CEST] <Chloe> Well I was looking to implement it anyways
[15:11:26 CEST] <durandal_1707> its audition feature, isn't
[15:12:07 CEST] <durandal_1707> proper way is to add sample format to swresample
[15:13:19 CEST] <Chloe> I cant find anything in the audition docs about int64 samples
[15:13:38 CEST] <durandal_1707> export format
[15:14:10 CEST] <durandal_1707> I doubt its mentiond as int64 exactly
[15:14:36 CEST] <durandal_1707> just integer 64 bit maybe
[15:25:56 CEST] <Chloe> I dont have audition myself, but the person who I've been trying to get a sample from says that they're getting an error saying 32 bit is the max depth. Is a sample really needed to add the format?
[15:27:20 CEST] <durandal_1707> yes, and its already uploaded
[15:32:30 CEST] <durandal_1707> Chloe: what, person have audition but it can't create such file?
[15:32:43 CEST] <Chloe> apparently
[15:33:04 CEST] <Chloe> I've got the sample (on the ticket) from the dodgy site now though
[17:36:09 CEST] <omerjerk> where does the output of ff_dlog go ?
[17:40:09 CEST] <kierank> is there a way of interleaving bytes between two registeres
[17:40:19 CEST] <kierank> that's not punpcklb
[17:47:07 CEST] <cone-283> ffmpeg 03Jan Sebechlebsky 07master:b746ed70efcc: avcodec/bsf: Add list BSF API
[17:49:19 CEST] <ubitux> omerjerk: ff_dlog is compiled conditionally
[19:27:59 CEST] <omerjerk> what does -f lavfi in the command line? I'm encoding the audio.
[19:28:32 CEST] <Compn> there are audio filters iirc
[19:29:38 CEST] <omerjerk> this is the whole command - ffmpeg -f lavfi -i aevalsrc="0.1*sin(2*PI*(360-2.5/2)*t) | 0.1*sin(2*PI*(360+2.5/2)*t)" -c:a als -y out.mp4
[19:30:04 CEST] <omerjerk> when I try to decode the .mp4 created using this command, I get error.
[19:30:37 CEST] <omerjerk> But when I decode a .mp4 compiled with a normal command like ffmpeg -i int.wav -c:a als out.mp4, I don't get any errors
[19:30:46 CEST] <omerjerk> so I'm looking for some advice here. 
[19:30:59 CEST] <ubitux> omerjerk: -f lavfi creates a virtual "format" source taking a filtergraph instead of an url/file as input string
[19:33:16 CEST] <Compn> omerjerk : ask in #ffmpeg
[19:33:28 CEST] <Compn>  unless you plan to develop some patch for ffmpeg ? :)
[19:33:45 CEST] <omerjerk> it's my encoder. So I'm trying to fix the bug. :)
[19:34:08 CEST] <omerjerk> I am just trying to guess where could the problem be.
[19:35:30 CEST] <ubitux> the filtergraph will generate pcm_f64le
[19:35:38 CEST] <ubitux> in a wave, maybe pcm_s16le
[19:36:55 CEST] <omerjerk> so should I modify my command somehow to use pcm_s16le ?
[19:40:59 CEST] <omerjerk> ubitux:, anyways, thanks. :) This was a big help. I can fix the error now.
[19:41:15 CEST] <durandal_1707> omerjerk: what is error?
[19:42:20 CEST] <omerjerk> the encoder is not supposed to work with pcm_f64le 
[19:42:33 CEST] <jamrial> omerjerk: the encoder will receive whatever it needs (one of the .sample_fmts values from the AVCodec struct) from the resampler, regardless of the source
[19:47:28 CEST] <omerjerk> jamrial: I don't think I properly understood what you said.
[19:51:24 CEST] <ubitux> whatever the input audio type, ffmpeg will automatically insert a convert filter to one of the format your encoder is supposed to handle
[19:51:48 CEST] <ubitux> so you have to make sure your AVCodec.sample_fmts contains a list of format you actually support
[20:08:25 CEST] <omerjerk> okay!!
[20:09:17 CEST] <CFS-MP3> michaelni I asked about the GSoC summit the other day (sorry I missed your reply) because I'm going myself, so I'd like to know who to look for :-)
[20:09:30 CEST] <CFS-MP3> I'm going for another org of course
[20:15:00 CEST] <michaelni> CFS-MP3, ohh, ok, understand
[20:15:10 CEST] <michaelni> btw, which org ?
[20:21:33 CEST] <kierank> CFS-MP3: oh you are the scte guy
[20:21:52 CEST] <kierank> it's going to be accepted because ffmpeg has to support everything but personally I don't think scte is appropriate
[20:22:12 CEST] <kierank> or do you just happen ot have the sdame name
[20:27:12 CEST] <CFS-MP3> michaelni CCExtractor
[20:28:31 CEST] <CFS-MP3> kierank Yes, I'm the SCTE guy... glad to hear it will be accepted... can't wait to get it out of the way :-) 
[20:28:40 CEST] <kierank> well i don't think it's a good idea for it to be accepted
[20:28:49 CEST] <CFS-MP3> kierank can you elaborate?
[20:28:49 CEST] <kierank> but #certainpeople will want it however ugly it is
[20:29:04 CEST] <kierank> well it doesn't fit into the ffmpeg codec inside of container model
[20:29:10 CEST] <kierank> scte35 does a ton of other things
[20:29:58 CEST] <CFS-MP3> If there's a way to make it non-ugly I'd be happy to hear and discuss
[20:30:13 CEST] <CFS-MP3> if it's ugly or nothing I guess we'll have to settle for ugly
[20:38:27 CEST] <kierank> or if we were an adult project, say it's out of scope
[20:41:20 CEST] <Compn> CFS-MP3
[20:41:36 CEST] <Compn> CFS-MP3 : [16:06] <michaelni> CFS-MP3b, i think thilo and rostislav ( atomnuker ) or carl will probably go to the summit, why do you ask ?
[20:41:46 CEST] <Compn> theres your reply :)
[20:41:47 CEST] Action: Compn afk
[20:44:27 CEST] <durandal_1707> we are just kids, don't be evil to us :)
[20:52:29 CEST] <CFS-MP3> kierank not sure out-of-scope'ing things is a good idea to be honest... basically it means to tell some people to look for a different solution than ffmpeg
[20:53:00 CEST] <CFS-MP3> You can do that, but then their resources (such as hiring me) will go to that other solution
[20:53:21 CEST] <CFS-MP3> still I'm new to ffmpeg so unless it's a purely technical discussion I'll stay out :-)
[20:54:30 CEST] <kierank> I still don't see how scte35 fits within ffmpeg
[20:54:33 CEST] <kierank> It's not a codec
[20:55:25 CEST] <omerjerk> ubitux: This is the my .sample_fmts - https://github.com/omerjerk/FFmpeg/blob/alsenc_patch/libavcodec/alsenc.c#L3523
[20:55:57 CEST] <omerjerk> I haven't mentioned float here. why does then my encoder receives float sample data?
[20:58:52 CEST] <ubitux> maybe it doesn't, but maybe in some case you receive s16 and on the other the float convert causes s32 to be fed
[20:59:04 CEST] <ubitux> and then your s32 code might be broken
[21:00:38 CEST] <omerjerk> It outputs this - 
[21:00:39 CEST] <omerjerk> Stream mapping:   Stream #0:0 -> #0:0 (pcm_f64le (native) -> mp4als (als))
[21:00:58 CEST] <omerjerk> does this mean my encoder is receiving 64-bit floats?
[21:04:41 CEST] <omerjerk> ubitux:
[21:40:19 CEST] <ubitux> omerjerk: no
[22:04:22 CEST] <cone-097> ffmpeg 03Paul B Mahol 07master:4f226714f569: avfilter/af_astats: add support for other sample formats
[23:10:28 CEST] <cone-097> ffmpeg 03Paul B Mahol 07master:fc3eb173f493: avfilter/vf_atadenoise: add support for more pixel formats
[23:10:29 CEST] <cone-097> ffmpeg 03Paul B Mahol 07master:3282e31baaa7: avfilter/vf_atadenoise: set default size value to 9
[00:00:00 CEST] --- Tue Aug 16 2016