[Ffmpeg-devel-irc] ffmpeg-devel.log.20140530

burek burek021 at gmail.com
Sat May 31 02:05:02 CEST 2014

[00:12] <cone-339> ffmpeg.git 03Diego Biurrun 07master:65d5d5865845: dsputil: Move SVQ1 encoding specific bits into svq1enc
[00:12] <cone-339> ffmpeg.git 03Michael Niedermayer 07master:ea0931fb962f: Merge commit '65d5d5865845f057cc6530a8d0f34db952d9009c'
[00:19] <llogan> ubitux: i stole some of your examples
[00:20] <ubitux> llogan: you should show what ffprobe displays before the eval
[00:20] <ubitux> llogan: also, maybe prefer -v error over -v 0
[00:20] <ubitux> so it displays at least the errors when something goes wrong
[00:38] <ubitux> llogan: also r_frame_rate is NOT the frame rate
[00:38] <ubitux> you should definitely not recommend that value.
[00:38] <cone-339> ffmpeg.git 03Diego Biurrun 07master:3832a6590156: dsputil: Move ff_alternate_*_scan tables to mpegvideo
[00:38] <cone-339> ffmpeg.git 03Michael Niedermayer 07master:c814a6c778b4: avcodec/x86/svq1enc_mmx: Add author attribution
[00:38] <cone-339> ffmpeg.git 03Michael Niedermayer 07master:a9a312d299c7: Merge commit '3832a6590156e6abc575bed51cbffa78617ee5fd'
[00:51] <llogan> ubitux: ok. what is that then?
[00:51] <ubitux> llogan: http://ffmpeg.org/faq.html#AVStream_002er_005fframe_005frate-is-wrong_002c-it-is-much-larger-than-the-frame-rate_002e
[00:54] <llogan> feel free to add other stuff
[00:55] <cone-339> ffmpeg.git 03Diego Biurrun 07master:256da0770e49: dsputil: Move mspel_pixels_tab to the only place it is used
[00:55] <cone-339> ffmpeg.git 03Michael Niedermayer 07master:5c6e94c42bcd: Merge commit '256da0770e495176d1b2699ec6e9c7993c2a6d7b'
[01:21] <cone-339> ffmpeg.git 03Diego Biurrun 07master:054013a0fc6f: dsputil: Move APE-specific bits into apedsp
[01:21] <cone-339> ffmpeg.git 03Michael Niedermayer 07master:40f3a87c10d5: Merge commit '054013a0fc6f2b52c60cee3e051be8cc7f82cef3'
[01:30] <cone-339> ffmpeg.git 03Diego Biurrun 07master:cb52a17cb622: dsputil: Move Bink-specific add_pixels8 to binkdsp
[01:30] <cone-339> ffmpeg.git 03Michael Niedermayer 07master:d35899ccce39: Merge commit 'cb52a17cb62297802b0a77e5af9850d87c3f925b'
[03:02] <cone-339> ffmpeg.git 03Diego Biurrun 07master:368f50359eb3: dsputil: Split off quarterpel bits into their own context
[03:02] <cone-339> ffmpeg.git 03Michael Niedermayer 07master:98a6806fddc2: Merge commit '368f50359eb328b0b9d67451f56fda20b3255f9a'
[04:02] <UtUser> I'm not sure if I'm selecting these frames right for my outut file
[04:02] <UtUser> I want to trim down the bugged video to jut a few frames for testing but I'm way off
[04:03] <compn> mencoder -nosound -ovc copy input -frames 5 -o output.avi
[04:03] <UtUser> awesome
[04:03] <UtUser> mencoder...
[04:08] <cone-339> ffmpeg.git 03Michael Niedermayer 07master:b15a3cc575ed: qpel_template: Restore author attribution
[04:08] <cone-339> ffmpeg.git 03Michael Niedermayer 07master:8c891d90cac9: avcodec/x86/qpeldsp_init: Restore author attribution
[04:08] <cone-339> ffmpeg.git 03Michael Niedermayer 07master:772d46d3b33c: avcodec/qpeldsp: Restore author attribution
[04:13] <UtUser> so how do I copy frames 331-332?
[04:37] <relaxed> UtUser: #ffmpeg is for user questions
[04:38] <UtUser> this is for a bug report
[04:38] <UtUser> nevermind anyways
[04:38] <UtUser> get off my bck
[04:49] <UtUser> well, it's definitely the internal encoder
[04:53] <UtUser> *decoder (I've had along day)
[04:57] <UtUser> http://www74.zippyshare.com/v/73545022/file.html
[04:57] <UtUser> http://uppit.com/vepr33xx9dea/bf1slice.avi
[04:57] <UtUser> JEEB, this is the 1-slice version
[04:58] <UtUser> imported it into VirtualDub and it played fine, so IDK but I think it might be ffmpeg's internal decoder
[11:47] <Varuh> hi guys. I'm stuck with MP4 files with 4k resolution from youtube (ftypdash).
[11:47] <Varuh> all the files which I was able download contain only one video stream, without audio.
[11:47] <Varuh> tell me pls, sre youtube's mp4 dash a/v streams always separated or it is possible get the
[11:47] <Varuh> mp4 (ftypdash) with audio and video streams inside ? 
[12:57] <cone-282> ffmpeg.git 03Carl Eugen Hoyos 07master:1acb5ca02a66: Improve frame size calculation in caf muxer to fix adpcm_ms remuxing.
[12:57] <cone-282> ffmpeg.git 03Carl Eugen Hoyos 07master:ca2ec98cbe1f: Do not overwrite the sample format with AV_SAMPLE_FMT_NONE in libspeex_decode_init().
[13:20] <compn> Varuh : seperated i think
[13:20] <compn> "Youtube-dl can download videos in this format "
[13:20] <compn> is what the internet says
[13:20] Action: compn afk
[13:20] <compn> Varuh  :you can download sd or audio-only version and paste it to your 4k ver :P
[13:48] <Varuh> compn, thanks! that's it.)
[13:54] <cone-282> ffmpeg.git 03Dirk Ausserhaus 07master:d24673a2c804: avcodec/indeo4: Move frametypes into common header.
[13:54] <cone-282> ffmpeg.git 03Dirk Ausserhaus 07master:8bfb4d72dd93: avcodec/indeo4: Decode both parts of IP frames
[14:46] <cone-282> ffmpeg.git 03Michael Niedermayer 07master:b3dfebd64160: avformat/matroskadec: fix off by 1 error in matroska_read_seek()
[15:03] <cone-282> ffmpeg.git 03Christophe Gisquet 07master:884078d2df0f: x86: huffyuvdsp: add SSE2 median prediction
[15:18] <cone-282> ffmpeg.git 03Michael Niedermayer 07master:7be79c76d3f7: avcodec/huffyuvdsp: Change w to intptr in add_hfyu_median_pred() and add_hfyu_left_pred()
[15:25] <cone-282> ffmpeg.git 03Christophe Gisquet 07master:f743fa9c7f87: x86: huffyuvdsp: add_hfyu_left_pred_bgr32
[15:53] <cone-282> ffmpeg.git 03Timothy Gu 07master:bfb488234525: tests: Add MPEG4 nsse test
[16:34] <wm4> congratulations
[16:35] <ubitux> :D
[16:36] <ubitux> i was expecting something more epic
[16:52] <ubitux> huh, the rotation thing is still not done?
[17:06] <cone-282> ffmpeg.git 03Timothy Gu 07master:0b6292b7b85c: x86: dsputilenc: move all the function prototypes together
[17:06] <cone-282> ffmpeg.git 03Timothy Gu 07master:154cee929284: x86: dsputilenc: convert ff_sse{8, 16}_mmx() to yasm
[17:34] <cone-282> ffmpeg.git 03Carl Eugen Hoyos 07master:3ea0d9c8a55b: Fix VDPAU decoders, regression since f2f99f07.
[18:18] <Daemon404> o shit
[18:18] <Daemon404> it hit 300
[18:21] <ubitux> we already had a celebration
[18:21] <Daemon404> aww
[18:21] Action: Daemon404 late
[18:28] <ruggles> unless i'm missing something, it looks like there is something weird going on in avcodec_copy_context() related to a merge from Libav. FFmpeg previously freed dest->priv_data while Libav did not, so merging of efc7df6c1f1 resulted in different behavior.
[18:28] <ruggles> so now it seems to leave a dest->priv_data as a non-NULL freed pointer
[18:56] <cone-282> ffmpeg.git 03Michael Niedermayer 07master:8716605a2605: avformat/nutenc: Suggest genpts if AV_NOPTS_VALUE
[18:56] <cone-282> ffmpeg.git 03Michael Niedermayer 07master:e263af30244c: avcodec/proresenc_anatoliy: Set bits_per_raw_sample
[18:56] <cone-282> ffmpeg.git 03Michael Niedermayer 07master:63b4d6feb883: avcodec/proresenc_anatoliy: Use fdct from DSPContext instead of direct call.
[18:57] <cone-282> ffmpeg.git 03Michael Niedermayer 07master:f6b1cd3936df: avcodec/proresenc_anatoliy: load 4 samples at a time in get()
[18:57] <cone-282> ffmpeg.git 03Michael Niedermayer 07master:82efa24c519a: avcodec/proresenc_anatoliy: simplify get()
[18:58] <ruggles> michaelni: so which behavior do you prefer? preserve it as is done in Libav or free it and set it to NULL, which would then trigger a new allocation in avcodec_open2(). currently it's breaking my application...
[18:59] <ruggles> and i'm guessing it broke avserver too
[18:59] <ruggles> sorry, ffserver
[18:59] <ruggles> ;)
[18:59] <michaelni> theoretically it should be copied i guess
[19:00] <ruggles> that was the previous behavior
[19:00] <ruggles> well, the pointer was copied
[19:00] <michaelni> yes
[19:08] <michaelni> i can look into fixing that btw but not ATM, should have time in 1-2h
[19:09] <ruggles> michaelni: ok, thank you
[19:52] <kurosu> I think qpeldsp broke --disable-inline-asm
[20:02] <kurosu> "rm libavcodec/x86/qpel.o" fixed it, weird it didn't rebuild...
[21:33] <cone-282> ffmpeg.git 03Michael Niedermayer 07master:e0d074556055: avcodec/options: leave priv_data in avcodec_copy_context() intact
[21:34] <michaelni> ruggles, fixed and some patches which would copy the user settable things from priv_data are on the ML
[21:35] <ruggles> michaelni: thanks
[21:35] <michaelni> np
[21:48] <cone-282> ffmpeg.git 03Christophe Gisquet 07master:d136fe6fd7be: x86: huffyuvdsp: fewer functions for x86_64
[22:59] <cone-282> ffmpeg.git 03Martin Storsjö 07master:abb810db0366: Revert "rtpenc_chain: Don't copy the time_base back to the caller"
[22:59] <cone-282> ffmpeg.git 03Michael Niedermayer 07master:692757062a89: Merge commit 'abb810db036628e11a5171134ebe320b187ee6d6'
[22:59] <cone-282> ffmpeg.git 03Michael Niedermayer 07master:1fe40e1b0538: avformat/movenchint: Fix movhinting with ff_rtp_chain_mux_open() from abb810db036628e11a5171134ebe320b187ee6d6
[23:24] <cone-282> ffmpeg.git 03Thierry Fauck 07master:d6d767d93e53: ppc: Fix compilation for ppc64le (ELFv2)
[23:24] <cone-282> ffmpeg.git 03Michael Niedermayer 07master:83b8987df702: Merge commit 'd6d767d93e532246cacf3567e6bcad76a821f838'
[00:00] --- Sat May 31 2014

More information about the Ffmpeg-devel-irc mailing list