[Ffmpeg-devel-irc] ffmpeg-devel.log.20150316

burek burek021 at gmail.com
Tue Mar 17 02:05:03 CET 2015


[00:00:02 CET] <cehoyos> rcombs: https://trac.ffmpeg.org/ticket/1920
[00:00:19 CET] <cehoyos> I don't think you can "extract" AC-3 from E-AC-3 (but I may be wrong)
[00:07:52 CET] <cehoyos> rcombs: I would have sent a patch to enable xll by default (I think it is useless to have the feature disabled by default) but it crashes badly here;-(
[00:10:18 CET] <kierank> 11:00 PM <"cehoyos> I don't think you can "extract" AC-3 from E-AC-3 (but I may be wrong)
[00:10:20 CET] <kierank> you cannot
[00:10:31 CET] <kierank> there is assisting information in the e-ac-3 for an ac-3 encoder but it's a transcode
[00:10:33 CET] <rcombs> oh, I'm probably confusing Dolby's legal terminology
[00:11:16 CET] <rcombs> because apparently you don't have to pay for a decoder or encoder license to convert e-ac-3 to ac-3 because they don't consider that a "transcode" (but apparently that's purely for legal purposes)
[00:11:17 CET] <cehoyos> FFmpeg does "extract" AC-3 from TrueHD: It shows two separate streams
[00:12:01 CET] <rcombs> cehoyos: if it crashes badly, that's probably why, then :3
[00:12:02 CET] <kierank> ye
[00:12:02 CET] <kierank> s
[00:12:52 CET] <cehoyos> That makes no sense imo
[00:13:07 CET] <cone-415> ffmpeg 03Luca Barbato 07master:66028b7ba6b4: udp: Use AVOptions
[00:13:08 CET] <cone-415> ffmpeg 03Michael Niedermayer 07master:4f3f5ee1ac93: Merge commit '66028b7ba6b411ba12ef553e9c8f1f4a4fe27710'
[00:13:11 CET] <rcombs> cehoyos: the legal nonsense?
[00:13:29 CET] <cehoyos> No, I think it just wasn't tested (this is unfortunately not unususal)
[00:13:37 CET] <rcombs> rip
[00:18:00 CET] <cone-415> ffmpeg 03Himangi Saraogi 07master:9a60b1fad02c: libswscale: fix compiler warnings enumerated type mixed with another type
[00:18:01 CET] <cone-415> ffmpeg 03Michael Niedermayer 07master:3e486162267f: Merge commit '9a60b1fad02cb783b895b2145c3dafc01f7b337c'
[00:22:03 CET] <kierank> no way
[00:22:07 CET] <kierank> untested patches in ffmpeg
[00:22:07 CET] <kierank> wow
[00:22:09 CET] <kierank> shocker
[00:24:43 CET] <cehoyos> The patch wasn't committed to FFmpeg...
[00:25:49 CET] <jamrial> the couple samples i tried decoded fine
[00:27:16 CET] <cehoyos> Yes, I found some that do not crash in the meantime: It was just the very first file I tested that crashed.
[00:27:41 CET] <cehoyos> But I see many warnings
[00:28:41 CET] <cehoyos> And I am not sure why the channellayout is not set for DTS-MA (or is it?)
[00:47:41 CET] <rcombs> cehoyos: I get a warning saying it was guessed
[00:47:55 CET] <rcombs> and it was merged in ffmpeg an hour or so ago
[01:34:54 CET] <cone-415> ffmpeg 03Michael Niedermayer 07master:21d25441c6ea: avcodec/dcadec: Only upsample as much data as is there
[01:34:55 CET] <cone-415> ffmpeg 03Michael Niedermayer 07master:e802abd6e309: avcodec/dcadec: if XLL is enabled use xll channels
[02:00:46 CET] <cone-415> ffmpeg 03Micah Galizia 07master:7859618affe5: avformat/hls: store cookies returned in HLS key response
[03:45:29 CET] <cone-415> ffmpeg 03Mark Reid 07master:e5755893786e: libavformat/mxfdec: fix MXFPackage->comment_refs memory leak
[04:10:18 CET] <cone-415> ffmpeg 03Michael Niedermayer 07master:1e4d0498df66: avformat/mov: Disallow ".." in dref unless use_absolute_path is set
[04:10:19 CET] <cone-415> ffmpeg 03Michael Niedermayer 07release/2.6:f6327e5fa618: avformat/mov: Disallow ".." in dref unless use_absolute_path is set
[05:53:42 CET] <cone-415> ffmpeg 03James Almer 07master:bdd179c8cbe8: x86/proresdsp: remove unused macro
[05:53:43 CET] <cone-415> ffmpeg 03James Almer 07master:e8374d72027c: x86/proresdsp: remove ff_prores_idct_put_10_sse4
[09:34:12 CET] <j-b> hello
[10:56:13 CET] <cone-057> ffmpeg 03Stefano Sabatini 07master:83ee820a1678: lavfi/sendcmd: improve error reporting and robustness in case of missing commands
[11:00:24 CET] <ubitux> any idea why libav decided to drop av_dlog?
[11:01:24 CET] <ubitux> it's not supposed to be changed into an av_log in many places because of performance
[11:03:06 CET] <ubitux> ah, it's split, ok
[11:03:58 CET] <ubitux> but mmh well, i don't see why libavfilter won't need that new ff_dlog
[11:04:14 CET] <ubitux> performance relevant code isn't only in lavc
[11:05:00 CET] <ubitux> ok so in fact av_dlog still exists, it's just renamed and get a more constrained usage (limited to lavc)
[11:05:50 CET] Action: ubitux really fails to see the point then
[11:08:38 CET] <termos> I'm curious about this line http://ffmpeg.org/doxygen/trunk/mux_8c_source.html#l00521 is AVStream::pts not suppose to be deprecated? I've had random crashes happening here, and I wonder if it's some uninitialized AVFrac::val that's causing this frac_add to be executed
[11:08:58 CET] Action: saste really fails to point the sea then
[11:24:54 CET] <ubitux>     { "localport",      "Local port",                                      OFFSET(local_port),
[11:24:56 CET] <ubitux>     { "local_port",     "Local port",                                      OFFSET(local_port),
[11:24:58 CET] <ubitux> just as predicted...
[11:26:15 CET] <rcombs> localport at kiekaku.doori
[11:27:27 CET] <ameyamarathe18> hi
[12:06:08 CET] <wm4> ubitux: so why did nobody ask them to consider the other already existing name?
[12:07:00 CET] <ubitux> wm4: it's more tricky
[12:08:30 CET] <ubitux> wm4: basically ffmpeg has localport and localaddr, and my assumption is that they added local_port option, then looked at ffmpeg if there was other options, and saw localaddr; so copied that to complete, and thus ended up with local_port and localaddr
[12:09:14 CET] <ubitux> (i might be wrong, but given how consistent they try to be all the time, i can't think of another reason why they would end up with local_port and localaddr)
[12:13:33 CET] <ubitux> saste: any comment on the patch 5/5? :P
[12:26:03 CET] <cone-057> ffmpeg 03Luca Barbato 07master:66a0ac0a802d: mkv: Add support for DVB subtitles
[12:26:04 CET] <cone-057> ffmpeg 03Michael Niedermayer 07master:18a964799636: Merge commit '66a0ac0a802dd794acc1ab8559d3db9e98755b57'
[12:26:33 CET] <wm4> wasn't there a function that formats fourcc as strings?
[12:32:03 CET] <wm4> av_get_codec_tag_string
[13:04:10 CET] <cone-057> ffmpeg 03Micah Galizia 07master:fca085187940: avformat/hls: refactor repeated HLS option updates
[13:18:35 CET] <cone-057> ffmpeg 03Michael Niedermayer 07master:04c52389d8d4: avcodec/tdsc: Remove ;;
[15:00:45 CET] <cone-057> ffmpeg 03Michael Niedermayer 07master:12cf61c3b261: avcodec/hevc: Fix undefined shifts
[15:00:46 CET] <cone-057> ffmpeg 03Michael Niedermayer 07master:d79f7bf0d63a: avcodec/x86/cavsdsp: remove incorrect LOCAL_ALIGN tmp
[16:29:01 CET] <cone-057> ffmpeg 03zhaoxiu.zeng 07master:332776f6c54c: avformat/cavsvideodec: use avpriv_find_start_code in cavsvideo_probe()
[18:30:02 CET] <ubitux> https://news.ycombinator.com/item?id=9212860 #attentionwhore #spam 
[18:30:20 CET] <ubitux> probably more interesting that templates in C :p
[18:30:26 CET] <ubitux> than*
[18:30:48 CET] <Daemon404> g 42
[18:43:41 CET] <wm4> ubitux: how did you make the last gif?
[18:43:55 CET] <ubitux> legend app
[18:44:33 CET] <ubitux> or well, internal tools used for legend app to get the raw, then encode with ffmpeg
[18:44:37 CET] <ubitux> to be more accurate
[18:44:38 CET] <himangi> michaelni : hi
[18:45:17 CET] <michaelni> himangi, hi
[18:47:14 CET] <himangi> michaelni: I tested the sample  http://samples.ffmpeg.org/benchmark/testsuite1/matrixbench_mpeg2.mpg, but I do not get a  AV desync.. 
[18:51:29 CET] <thardin> neat, better gif support
[18:51:55 CET] <thardin> done  any comparison with imagemagick?
[18:52:40 CET] <michaelni> himangi, ill retry
[18:52:50 CET] <ubitux> thardin: feel free to :p
[18:52:53 CET] <Daemon404> thardin, imagemagick is crappy
[18:52:59 CET] <Daemon404> comparison to libimagequant is better
[18:53:00 CET] <Daemon404> IMO
[18:53:02 CET] <Daemon404> or gimp's lib
[18:54:51 CET] <thardin> I did manage to figure out how to get imagemagick to produce something decent
[18:55:16 CET] <ubitux> it's often a play with depending on the content (how surprising)
[18:55:26 CET] <thardin> convert input* -resize 50% -fuzz 5% -delay 10 +map -layers Optimize output.gif
[18:55:26 CET] <ubitux> with +the options
[18:55:39 CET] <thardin> is what I used last
[18:55:40 CET] <michaelni> himangi, "ffplay matrixbench_mpeg2.mpg -vf detelecine" is definitly having audio and video out of sync
[18:55:51 CET] <ubitux> thardin: i can give you many input where the fuzz will look like shit
[18:56:21 CET] <thardin> prolly yeah
[18:56:50 CET] <michaelni> himangi, frame->pts = outlink->frame_count ... is not correct, the input timestamps arent guarateed to start at 0
[18:59:41 CET] <himangi> michaelni: which value indicates the desync?
[19:01:08 CET] <michaelni> desync as in the sound you hear and the picture you see at the same time do not belong to the same time in the movie
[19:06:30 CET] <himangi> ohh.. I see that.
[19:07:29 CET] <michaelni> need to go afk, ill be back in 30min or so
[19:59:34 CET] <cone-057> ffmpeg 03Michael Niedermayer 07release/2.6:b3e6d3ee7849: update for 2.6.1
[20:20:32 CET] <cone-057> ffmpeg 03Michael Niedermayer 07release/2.6:b2c9cd36d34c: Changelog: update for 2.6.1
[20:27:03 CET] <Guest33432> hi, is there a way to access the list of AVFormatContext of all the outputs from a muxer. i'm working on adding support for WebM live encoding. my intention is to arrive at a command like this: "ffmpeg -i input -f webm output1 -f webm output2 ... -f webm_dash_manifest manifest.xml"; now to generate the manifest, i need to access the AVFormatContext of output1, output2 and so on. is there a way to do that? any help is appreciated!
[20:40:33 CET] <poste9> guys can u help me out finding some implementation of the AVIOContext::write_packet  ?
[20:41:40 CET] <ubitux> you'll find one in nearly every muxer
[20:43:14 CET] <cone-057> ffmpeg 03zhaoxiu.zeng 07n2.6.1:HEAD: avformat/cavsvideodec: use avpriv_find_start_code in cavsvideo_probe()
[21:12:27 CET] <cone-057> ffmpeg 03Rainer Hochecker 07master:786032cad8ec: hevc: avoid unnecessary calls to get_format
[21:27:37 CET] <cone-057> ffmpeg 03Janne Grunau 07master:a9d60c390f35: doc: fate: remove outdated SSH key fingerprint
[21:27:38 CET] <cone-057> ffmpeg 03Michael Niedermayer 07master:7c9ad3c4ea79: Merge commit 'a9d60c390f35f3954821bd635fd31bbb5036b29d'
[21:34:44 CET] <cone-057> ffmpeg 03Janne Grunau 07master:f01c77157789: fate: add explicit support for the toolchain configure option
[21:34:45 CET] <cone-057> ffmpeg 03Michael Niedermayer 07master:6416dce490f7: Merge commit 'f01c77157789b8e3a59ed2c9646faf8299e41641'
[21:44:37 CET] <cone-057> ffmpeg 03Janne Grunau 07master:3e1b5cbc9ab0: configure: handle Apple's armv7s in probe_arm_arch()
[21:44:37 CET] <cone-057> ffmpeg 03Michael Niedermayer 07master:b37852c7d9bc: Merge commit '3e1b5cbc9ab0a61c9bec08a1df1404b9da6ed7ea'
[21:53:34 CET] <cone-057> ffmpeg 03Janne Grunau 07master:faab8f9fcb4f: configure: move cross_compile checks after the toolchain section
[21:53:35 CET] <cone-057> ffmpeg 03Michael Niedermayer 07master:1904704a48b7: Merge commit 'faab8f9fcb4ffeb967dc6872c0f1e9da719106ce'
[22:00:53 CET] <cone-057> ffmpeg 03Martin Storsjö 07master:50ff0c05b691: movenc: Support writing multiple trun atoms
[22:00:54 CET] <cone-057> ffmpeg 03Michael Niedermayer 07master:312eb0d66f24: Merge commit '50ff0c05b691524d1fd044a2f2eefb706932fc98'
[22:10:55 CET] <haasn> This video segfaults https://youtu.be/9ze0RXDU5HQ (using format bestvideo, downloaded with youtube-dl)
[22:11:11 CET] <haasn> At around 6 seconds into playback, I get: 0x00007ffff729a7f0 in ff_vp9_ipred_v_32x32_avx () from /usr/lib64/libavcodec.so.56
[22:11:19 CET] <haasn> Just updated from git master
[22:21:42 CET] <Daemon404> haasn, file a bug or send a drone to BBB's place
[22:36:40 CET] <cone-057> ffmpeg 03Michael Niedermayer 07master:29db8e45fc41: configure: Silence EMMS warnings in ICC
[22:39:51 CET] <Daemon404> ubitux i #1 on hn
[22:39:58 CET] <Daemon404> i hope you are ashamed of yourself
[22:40:01 CET] <ubitux> :D
[22:40:44 CET] <Daemon404> ugh... comments recommending deb-multimedia
[22:40:52 CET] <Daemon404> that thing still exists?
[22:59:49 CET] <cone-057> ffmpeg 03Michael Niedermayer 07master:1b3b018aa4e4: avformat/mpegts: Fix potential pointer overflows
[22:59:50 CET] <cone-057> ffmpeg 03Michael Niedermayer 07master:b425b81fd204: avformat/rtsp: Fix potential pointer overflow in sdp_probe()
[22:59:51 CET] <cone-057> ffmpeg 03srikanth 07master:730f3c49e082: OpenCL uninit bug fix - clear is_compiled flag
[23:00:54 CET] <wm4> Daemon404: yes
[23:36:51 CET] <ubitux> still no one to write us a news for 2.6?
[23:37:01 CET] <ubitux> ...taking the opportunity that 2.6.1 was just released
[23:39:41 CET] <cone-057> ffmpeg 03Martin Storsjö 07master:2cb9c2fc59c9: movenc: Allow interleaving samples when writing fragmented files
[23:39:42 CET] <cone-057> ffmpeg 03Michael Niedermayer 07master:ae65cc7ed364: Merge commit '2cb9c2fc59c9267ad2631c07c81c188058502259'
[23:40:03 CET] <michaelni> ubitux, please write news!
[23:40:22 CET] <ubitux> sorry, not on my schedule :p
[23:40:33 CET] <ubitux> i wrote too much these days
[23:43:23 CET] <cone-057> ffmpeg 03Clément BSsch 07master:38fb183b1256: avfilter/formats: remove unused COPY_INT_LIST() macro
[23:43:24 CET] <cone-057> ffmpeg 03Clément BSsch 07master:93d9ce7cec48: avfilter/formats: use av_realloc_array in ADD_FORMAT()
[23:43:25 CET] <cone-057> ffmpeg 03Clément BSsch 07master:f861d9b2c647: avfilter/formats: proper error handling in ff_channel_layouts_ref() and ff_formats_ref()
[23:43:26 CET] <cone-057> ffmpeg 03Clément BSsch 07master:545b0dd6aa30: avfilter/formats: proper error handling in ff_set_common_*() functions
[23:43:27 CET] <cone-057> ffmpeg 03Clément BSsch 07master:fd682b189264: avfilter: handle error in query_formats() of a bunch of random video filters
[23:48:11 CET] <michaelni> ubitux, posted news suggestion to ML
[00:00:00 CET] --- Tue Mar 17 2015


More information about the Ffmpeg-devel-irc mailing list