[Ffmpeg-devel-irc] ffmpeg-devel.log.20130205

burek burek021 at gmail.com
Wed Feb 6 02:05:02 CET 2013


[00:02] <cone-831> ffmpeg.git 03Stefano Sabatini 07master:2d22e97f6d73: doc/filters: fix @item syntax in showspectrum docs
[00:13] <cone-831> ffmpeg.git 03Stefano Sabatini 07master:0e2b0033f161: doc/filters: apply various rendering changes to the showspectrum docs
[00:37] <cone-831> ffmpeg.git 03Michael Niedermayer 07master:b77d94dc6cc6: ffv1enc: check for malloc failure
[00:37] <cone-831> ffmpeg.git 03Michael Niedermayer 07master:5a8311513091: ffv1: check for malloc failure
[00:41] <llogan> saste: any comments (since you were the only one with input on ML)? http://pastebin.com/raw.php?i=6B1ege28
[00:44] <saste> llogan, i find annoying when I'm subscribed to a new ML without my explicit approval
[00:45] <llogan> one reason i'm not going to auto-subscribe
[00:45] <saste> so I'd suggest to advertise the current ffserver-user subscribers of the move
[00:45] <saste> you could keep the ML active, but tell all the posters to move to ffmpeg-user
[00:46] <saste> yes an option would be to forward ffserver-user mails to ffmpeg-user
[00:46] <saste> i don't have a strong opinion on that
[00:46] <Compn> i wasnt suggesting subscribing them w/ email subscriptions
[00:46] <llogan> i already made two announcements on ffserver-user.
[00:46] <Compn> only putting their emails in so they could post, automatically
[00:46] <saste> you could leave ffserver-user for a month or so
[00:47] <saste> i don't think there is much traffic anyway
[00:47] <saste> and send a mail every week advertising that the list will be closed in X days
[00:47] <saste> and providing instructions/link to subscribe to ffmpeg-user
[00:48] <llogan> i already said today is it's last day (on jan 27 and again today).
[00:51] <saste> llogan, i think it's safe to remove the list anyway
[00:52] <saste> it would be a bit annoying for the few subscriber to move to an high traffic list
[00:52] <saste> but should be allow a better support, since there will be more feedback
[00:52] <llogan> and they can always disable delivery on their choice.
[00:54] <saste> i'm fine with it
[01:06] <llogan> michaelni: please rmlist ffserver-user at your convenience
[01:45] <michaelni> llogan, ill wait 24h to give people a last chance to raise objections and then delete it
[02:03] <llogan> ok
[03:09] <cone-831> ffmpeg.git 03multiple authors 07master:708ed15d8ccd: libmpcodecs/vf_stereo3d: update to latest version from mplayer
[03:32] <Compn> ehehe
[03:43] <cone-831> ffmpeg.git 03multiple authors 07master:103a2c2b1757: libmpcodecs: update img_format.* & mp_image.* to latest from mplayer
[10:36] <j-b> morning
[11:54] <michaelni> morning
[11:55] <TimNich> moening
[11:58] <durandal_1707> its 12
[12:02] <cone-290> ffmpeg.git 03Diego Biurrun 07master:e66240f22e24: avfilter: x86: consistent filenames for filter optimizations
[12:02] <cone-290> ffmpeg.git 03Michael Niedermayer 07master:0b6f34cc9f28: Merge remote-tracking branch 'qatar/master'
[12:16] <TimNich> not in my timezone
[12:20] <kierank> it is always morning on irc
[12:23] <michaelni> that proofs the earth cant be flat
[12:36] <TimNich> sometimes feels like a moebius loop!
[13:07] <cone-290> ffmpeg.git 03multiple authors 07master:a54737f7ead0: libmpcodecs/cpudetect: update to latest from mplayer
[13:07] <cone-290> ffmpeg.git 03multiple authors 07master:745cb39d0558: libmpcodecs/help_mp: update to latest from mplayer
[13:07] <cone-290> ffmpeg.git 03multiple authors 07master:99a0813a7e0b: libhmpcodec/mp_msg: update to latest from mplayer
[13:09] <michaelni> TimNich, a moebius loop is unlikely because if it was that way that moebius loop could be seen by simply looking up from some places
[13:49] <cone-290> ffmpeg.git 03multiple authors 07master:5d0115118df4: libmpcodecs/vf_divtc: update to latest version in mplayer
[13:49] <cone-290> ffmpeg.git 03multiple authors 07master:e02c713fb149: libmpcodecs/vf_filmdint: update to latest version from mplayer
[13:49] <cone-290> ffmpeg.git 03multiple authors 07master:2b9eb6167b32: libmpcodecs/av_helpers: update to latest version from mplayer
[13:49] <cone-290> ffmpeg.git 03multiple authors 07master:0b36a05c1752: libmpcodecs/vf_fspp: update to latest version from mplayer
[14:12] <cone-290> ffmpeg.git 03multiple authors 07master:011702f5a934: libmpcodecs/vf_ivtc: update to latest version from mplayer
[14:12] <cone-290> ffmpeg.git 03multiple authors 07master:25ed6e99d789: libmpcodecs/vf_ilpack: update to latest version from mplayer
[14:12] <cone-290> ffmpeg.git 03multiple authors 07master:b7e6622971eb: libmpcodecs/vf_mcdeint: update to latest version from mplayer
[14:23] <cone-290> ffmpeg.git 03multiple authors 07master:fdd574ed3f60: libmpcodecs/vf_qp: update to latest version from mplayer
[14:23] <cone-290> ffmpeg.git 03multiple authors 07master:3ca46d0c10aa: libmpcodecs/vf_pp7:update to latest version from mplayer
[14:23] <cone-290> ffmpeg.git 03multiple authors 07master:8f7e3e678b58: libmpcodecs/vf_noise: update to latest version from mplayer
[14:41] <juanmabc> mmm, mplayer pulseaudio got excellent, not even 1%
[14:41] <juanmabc> cpu
[14:59] <cone-290> ffmpeg.git 03multiple authors 07master:fe8bc6ddfcd2: libmpcodecs/pullup: update to latest version from mplayer
[14:59] <cone-290> ffmpeg.git 03multiple authors 07master:14132599a372: libmpcodecs/vf_sab: update to latest version from mplayer
[14:59] <cone-290> ffmpeg.git 03multiple authors 07master:7a4885600ff0: libmpcodecs/vf_spp: update to latest version from mplayer
[15:48] <cone-290> ffmpeg.git 03multiple authors 07master:3ce3d4cc3c0a: libmpcodecs/vf_unsharp: update to latest version from mplayer
[15:48] <cone-290> ffmpeg.git 03multiple authors 07master:13afcdd00eec: libmpcodecs/vf_uspp: update to latest version from mplayer
[15:48] <cone-290> ffmpeg.git 03multiple authors 07master:81e357db6259: libmpcodecs/vf_pullup: update to latest version from mplayer
[16:11] <cone-290> ffmpeg.git 03Michael Niedermayer 07master:2b20397e1fbe: libmpcodecs/vd_ffmpeg.h: remove, its unused and removed upstream
[16:25] <wm4> lol old broken mplayer sources lingering in ffmpeg
[16:27] <cone-290> ffmpeg.git 03Xi Wang 07release/1.1:8d0631c8fa89: mpegvideo: fix loop condition in draw_line()
[16:27] <cone-290> ffmpeg.git 03Luca Barbato 07release/1.1:01050448cfbc: get_bits: change the failure condition in init_get_bits
[16:27] <cone-290> ffmpeg.git 03Janne Grunau 07release/1.1:1a28948eb3e5: videodsp_armv5te: remove #if HAVE_ARMV5TE_EXTERNAL
[16:27] <cone-290> ffmpeg.git 03Martin Storsjö 07release/1.1:62de693a17f9: rtp: Make sure priv_data is set before reading it
[16:27] <cone-290> ffmpeg.git 03Michael Niedermayer 07release/1.1:85e94a30eeb9: Merge commit '62de693a17f9b107be7867d822d5accacd4be544' into release/1.1
[16:33] <cone-290> ffmpeg.git 03Martin Storsjö 07release/1.1:348cd84fc82e: rtp: Make sure the output format pointer is set
[16:33] <cone-290> ffmpeg.git 03Martin Storsjö 07release/1.1:a856623e87d3: rtpenc_chain: Use the original AVFormatContext for getting payload type
[16:33] <cone-290> ffmpeg.git 03Justin Ruggles 07release/1.1:d7e7e12abc7e: lavr: call mix_function_init() in ff_audio_mix_set_matrix()
[16:33] <cone-290> ffmpeg.git 03Reinhard Tartler 07release/1.1:8d55c2441c67: Prepare for 9.2 Release
[16:33] <cone-290> ffmpeg.git 03Martin Storsjö 07release/1.1:3d67f52f9dcc: arm: vp8: Fix the plain-armv6 version of vp8_luma_dc_wht
[16:33] <cone-290> ffmpeg.git 03Michael Niedermayer 07release/1.1:8a4464514fac: Merge remote-tracking branch 'qatar/release/9' into release/1.1
[17:01] <cone-290> ffmpeg.git 03Piotr Bandurski 07release/1.1:22561bc0e963: aasc: fix 16bpp on big-endian
[17:01] <cone-290> ffmpeg.git 03Michael Niedermayer 07release/1.1:6baa54924980: huffyuvdec: Check init_vlc() return codes.
[17:01] <cone-290> ffmpeg.git 03Michael Niedermayer 07release/1.1:75e88db33013: huffyuvdec: Skip len==0 cases
[17:01] <cone-290> ffmpeg.git 03Diego Biurrun 07release/1.1:0e68b6ddce29: Use proper "" quotes for local header #includes (cherry picked from commit 6c1a7d07eb41b4cf36e48ebb560fecd9504fd968)
[17:01] <cone-290> ffmpeg.git 03Michael Niedermayer 07release/1.1:62c9beda0c18: sanm: Check decoded_size.
[17:01] <cone-290> ffmpeg.git 03Michael Niedermayer 07release/1.1:9547034f9120: gifdec: gif_copy_img_rect: Fix end pointer
[17:01] <cone-290> ffmpeg.git 03Michael Niedermayer 07release/1.1:f3d16706060a: vqavideo: check chunk sizes before reading chunks
[17:01] <cone-290> ffmpeg.git 03Michael Niedermayer 07release/1.1:47e462eecc0a: aacdec: check channel count
[17:01] <cone-290> ffmpeg.git 03Michael Niedermayer 07release/1.1:3ef1538121fa: h264: check the pixel format directly and force a reinit on mismatches.
[17:01] <cone-290> ffmpeg.git 03Michael Niedermayer 07release/1.1:d3bec2473972: h264: Only apply error concealment if theres a frame
[17:01] <cone-290> ffmpeg.git 03Michael Niedermayer 07release/1.1:eaa9d2cd6b8c: h264: skip error concealment when SPS and slices are mismatching
[18:26] <wm4> how do I use avcodec_copy_context() when I want to use a codec different from the source context? maybe call avcodec_get_context_defaults3() with the wanted AVCodec after that?
[18:36] <cone-290> ffmpeg.git 03Michael Niedermayer 07master:a8343bfb6a3f: riff: fix infinite loop
[19:01] <cone-290> ffmpeg.git 03Michael Niedermayer 07master:c2dd5a18b27b: ljpegenc: allocate needed scratch-buffer
[21:05] <Zeranoe> Could anyone tell me the current status of GPU encoding/decoding? I see NVIDIA provides a H.264 encoder/decoder that works off of CUDA. I'm interested in starting a LGPL GPU encoder
[21:08] <nevcairiel> in general? only fixed presets, quality below software encoders without a significant speed advantage in many cases
[21:08] <nevcairiel> for encoding, that is
[21:09] <nevcairiel> decoders are usually working quite well, provide bit-exact output, speed is usually below any modern CPU, only real use-case is on embedded/mobile
[21:10] <nevcairiel> and limited to mainstream profiles of course, for ex. h264 only 8-bit 4:2:0 baseline/main/high, both for enc/dec
[21:12] <Zeranoe> nevcairiel: Is the speed below CPU encoding because of limitations of the GPU? Is there a possibility to make an encoder/decoder that would out preform CPU encoders/decoders?
[21:12] <nevcairiel> it always depends what quality you're looking for
[21:13] <nevcairiel> and what kind of gpu you have =p
[21:13] <nevcairiel> latest nvidia models for example have a fixed function encoder, earlier models implemented it in cuda
[21:13] <nevcairiel> the tests i remembered, running x264 on settings that resulted in similar quality and filesize, the speed was pretty similar
[21:13] <wm4> what's the use case for GPU encoding? cameras?
[21:14] <nevcairiel> video conferencing, i guess
[21:14] <Zeranoe> nevcairiel: 
[21:14] <nevcairiel> or if you go into the phone market, camera recordings as well, i suppose
[21:15] <Zeranoe> nevcairiel: Does FFmpeg support  any GPU features?
[21:15] <nevcairiel> ffmpeg supports gpu decoding
[21:15] <nevcairiel> technically, avcodec does, you need some user code to make it work
[21:15] <Zeranoe> nevcairiel: Is it automatically used, or is there special way to test it?
[21:15] <nevcairiel> which the ffmpeg application does not implement
[21:16] <nevcairiel> you could run vlc, it uses it
[21:16] <Zeranoe> nevcairiel: Is there a way to test it on the command line?
[21:16] <nevcairiel> is mplayer considered command line? =P
[21:17] <Zeranoe> nevcairiel: is it only available in other applications?
[21:17] <nevcairiel> gpu decoding is not useful for transcoding because the gpu decoders are so slow in comparison
[21:17] <nevcairiel> the ffmpeg applications dont support it
[21:18] <Zeranoe> nevcairiel: Interesting, so it only really helps with playing a steam? 
[21:18] <cone-290> ffmpeg.git 03rogerdpack 07master:a50049a197a5: dshow: use standardized raw pixel format lookup
[21:18] <cone-290> ffmpeg.git 03rogerdpack 07master:47e88486b4b3: hdyc col
[21:18] <cone-290> ffmpeg.git 03rogerdpack 07master:c540312ac3b5: Merge remote-tracking branch 'origin/master' into combined
[21:18] <cone-290> ffmpeg.git 03Michael Niedermayer 07master:f7b7f021e528: Merge remote-tracking branch 'rdp/combined'
[21:19] <llogan> "hdyc col". not very descriptive.
[21:20] <Zeranoe> nevcairiel: How many encoders support CPU threading in FFmpeg?
[21:20] <nevcairiel> no idea, i dont encode much
[21:22] <J_Darnley> Wasn't a flag added to codecs to state whether they supported frame and/or sliced threads?  Does ffmpeg -codecs tell you?
[21:28] <J_Darnley> Answer: My (slightly) old ffmpeg doesn't but a search shows 59 matches in 39 files.
[21:35] <llogan> ubitux: how is the adventure?
[22:02] <cone-290> ffmpeg.git 03Michael Niedermayer 07master:b0635e2fcf80: movtextenc: fix pointer messup and out of array accesses
[23:04] <cone-290> ffmpeg.git 03Michael Niedermayer 07master:f0eacbc760c8: dvenc: dont fail hard if the timecode is invalid
[00:00] --- Wed Feb  6 2013


More information about the Ffmpeg-devel-irc mailing list