[Ffmpeg-devel-irc] ffmpeg-devel.log.20120719

burek burek021 at gmail.com
Fri Jul 20 02:05:03 CEST 2012


[02:21] <CIA-41> ffmpeg: 03Paul B Mahol 07master * r6b7849e6da 10ffmpeg/libswscale/ (colorspace-test.c rgb2rgb.c rgb2rgb.h swscale_unscaled.c): 
[02:21] <CIA-41> ffmpeg: swscale: unscaled rgba64->rgb48
[02:21] <CIA-41> ffmpeg: Signed-off-by: Paul B Mahol <onemda at gmail.com>
[02:21] <CIA-41> ffmpeg: 03Paul B Mahol 07master * rb4befca23e 10ffmpeg/libswscale/ (colorspace-test.c rgb2rgb.c rgb2rgb.h swscale_unscaled.c): 
[02:21] <CIA-41> ffmpeg: swscale: unscaled rgba64->bgr48
[02:21] <CIA-41> ffmpeg: Signed-off-by: Paul B Mahol <onemda at gmail.com>
[02:21] <CIA-41> ffmpeg: 03Paul B Mahol 07master * r9d8eac6899 10ffmpeg/libswscale/ (colorspace-test.c rgb2rgb.c rgb2rgb.h swscale_unscaled.c): 
[02:21] <CIA-41> ffmpeg: swscale: simplify unscaled rgb48<->bgr48
[02:21] <CIA-41> ffmpeg: Using av_bswap16 should be enough.
[12:57] <roxlu> hi, asked in ##ffmpeg, but this might be a better place ...
[12:58] <roxlu> Is it possible to i.e. write an application which outputs RGB frames and feed that into ffmpeg and stream that over rtmp? 
[15:35] <Hawk-TAO> morning
[15:35] <Hawk-TAO> someone online?
[15:35] <Hawk-TAO> i have a question
[15:36] <Hawk-TAO> ret = avformat_open_input(&s, input_filename, c->stream->ifmt, &c->stream->in_opts)) opens video
[15:37] <Hawk-TAO> and how dose ffmpeg read video?
[15:39] <Hawk-TAO> anyone helps me?
[15:39] <ubitux> look at the examples in doc/examples
[15:41] <Hawk-TAO> are there examples on it?
[15:41] <ubitux> why don't you just look by yourself?
[15:41] <ubitux> http://git.videolan.org/?p=ffmpeg.git;a=tree;f=doc/examples;hb=HEAD
[15:44] <Hawk-TAO> sorry,i am a fresh
[15:45] <Hawk-TAO> in filtering_video.c?
[15:46] <durandal_1707> decoding_encoding
[15:47] <Hawk-TAO> hmm,i'll read it at once
[15:48] <ubitux> anyone has information on the "AQTitle" subtitles?
[15:49] <ubitux> i don't find any related software nor documentation
[15:49] <ubitux> also, it's quite the contrary for "vplayer" subtitles
[15:50] <ubitux> i'm not sure which app is at the origin of such format
[15:50] <durandal_1707> it is just too old :)
[15:51] <ubitux> :(
[15:52] <ubitux> what about PJS?
[15:53] <ubitux> i can't find much information about these files
[15:53] <ubitux> s/files/subtitle formats/
[15:54] <Hawk-TAO> hi,sir,it use fread() to read data
[15:55] <Hawk-TAO> but ffmpeg does not use fread()
[15:56] <ubitux> use av_read_frame()
[15:56] <ubitux> filtering_video example is simpler
[15:58] <Hawk-TAO> ret = av_read_frame(c->fmt_in, &pkt);this function read the c->fmt_in to pkt
[15:59] <Hawk-TAO> and the video data exist in the c->fmt_in
[15:59] <Hawk-TAO> how does the c->fmt_in get the video data?
[16:00] <ubitux> it's the format context, it has an internal i/o context
[16:00] <ubitux> depending on the protocol
[16:01] <ubitux> (it makes http requests if http://, fread/fwrite if a file, etc)
[16:01] <Hawk-TAO> protocol?but i use local video
[16:01] <ubitux> it's done automatically, why do you care?
[16:01] <nevcairiel> there is also a file protocol =p
[16:02] <Hawk-TAO> automatically?
[16:02] <ubitux> av_read_frame() will call the fread/fwrite for you if it's a file
[16:03] <Hawk-TAO> it can read video data from harddisk automatically?
[16:03] <ubitux> sure
[16:03] <ubitux> did you try the examples?
[16:03] <Hawk-TAO> no
[16:03] <ubitux> "make"
[16:04] <Hawk-TAO> av_read_frame() calls the fread/fwrite?
[16:04] <Hawk-TAO> where is the code?
[16:07] <Hawk-TAO> ubitux,help me please,thx
[16:07] <ubitux> it might actually call read()/write()
[16:07] <Hawk-TAO> but i didnt find the code
[16:08] <ubitux> maybe libavformat/aviobuf.c
[16:09] <ubitux> actually, libavformat/file.c looks more appropriate
[16:09] <Hawk-TAO> which function?
[16:09] <Hawk-TAO> thx so much,i am a fresh from China
[16:09] <ubitux> libavformat/file.c, first function
[16:09] <ubitux> this file defines the "file" protocol
[16:10] <ubitux> use when you're demuxing a file with ffmpeg api
[16:10] <ubitux> anyway,
[16:10] Action: ubitux &
[16:10] <Hawk-TAO> ok,i will track the code,thx
[16:12] <Hawk-TAO> file_read(),this function?
[16:13] <Hawk-TAO> it is static
[16:16] <Hawk-TAO> it cannt be called by other functions not from this file.
[16:19] <Hawk-TAO> ubitux,are you online?
[16:20] <durandal_1707> why you need that function ?
[16:22] <Hawk-TAO> i just wanna know ffmpeg how to read video. 
[16:23] <Hawk-TAO> from harddisk.
[16:23] <durandal_1707> file with video is read with libavformat api
[16:23] <Hawk-TAO> which api?
[16:23] <durandal_1707> video frames are usually read via libavcodec
[16:23] <durandal_1707> Hawk-TAO: told you
[16:25] <Hawk-TAO> any more?
[16:26] <Hawk-TAO> read frames via libavcodec?
[16:27] <durandal_1707> your english skills are extremly low?
[16:28] <Hawk-TAO> low
[16:28] <Hawk-TAO> i cannt understand what you said
[16:29] <Hawk-TAO> sorry
[16:33] <Hawk-TAO> durandal_1707,can you tell me which api or fucntion read video from harddisk?
[16:36] <Hawk-TAO> anyone help me?
[16:37] <av500> libavformat/file.c
[16:37] <durandal_1707> Hawk-TAO: i told you URLProtocols API from libavformat 
[16:40] <Hawk-TAO> in libavformat/file.c?
[16:40] <CIA-41> ffmpeg: 03Nicolas George 07master * r52ab418588 10ffmpeg/ffmpeg.c: 
[16:40] <CIA-41> ffmpeg: ffmpeg: invert using_stdin and rename accordingly.
[16:40] <CIA-41> ffmpeg: There are other reasons not to use stdin than having it as an input.
[16:40] <CIA-41> ffmpeg: 03Nicolas George 07master * r0fe8acf2d6 10ffmpeg/ (Changelog doc/ffmpeg.texi ffmpeg.c): 
[16:40] <CIA-41> ffmpeg: ffmpeg: add -(no)stdin option.
[16:40] <CIA-41> ffmpeg: Allows to disable interaction from standard input.
[16:40] <CIA-41> ffmpeg: Useful, for example, if ffmpeg is in the background process group.
[16:40] <CIA-41> ffmpeg: Roughly the same result can be achieved with "ffmpeg ... < /dev/null"
[16:40] <CIA-41> ffmpeg: but it requires a shell.
[16:42] <durandal_1707> Hawk-TAO: what is your experience with C ?
[16:42] <Hawk-TAO> about one year
[16:44] <durandal_1707> i did not asked how long you know C
[16:45] <Hawk-TAO> just one project.
[16:50] <Hawk-TAO> i use ffserver to deliver video to a client
[16:52] <Hawk-TAO> but i cannt find the ffserver how to read the video from local harddisk
[16:53] <ubitux> doc/examples/filtering_video.c does exactly that by calling av_read_frame()
[16:53] <ubitux> just use that
[16:53] <durandal_1707> i think he needs pointer to file
[16:54] <ubitux> i gave it already
[16:54] <ubitux> ah sorry, ok
[16:54] <ubitux> maybe with a custom AVIOContext then?
[16:55] <Hawk-TAO> av_read_frame() read the c->fmt_in to pkt
[16:56] <Hawk-TAO> how did the c->fmt_in get data? 
[16:57] <durandal_1707> URLProtocol
[16:59] <Hawk-TAO> thx
[16:59] <Hawk-TAO> sir
[17:00] <Hawk-TAO> my EN is so poor
[17:04] <Hawk-TAO> URLProtocol can read video data from local harddisk?
[17:05] <durandal_1707> that file protocol from libavformat/file.c do
[17:08] <Hawk-TAO> but the all functions in libavformat/file.c are static
[17:08] <durandal_1707> url_read is not static
[17:09] <durandal_1707> and it is same as file_read which is static
[17:13] <Hawk-TAO> it is a function pointer
[17:15] <Hawk-TAO> thx so much
[17:17] <Hawk-TAO> please forgive my EN AND C skills
[17:18] <ubitux> durandal_1707: i wonder if an opengl ouput device wouldn't be more useful :)
[17:20] <ubitux> but that's nice to provide that one :D
[17:20] <durandal_1707> ubitux: :))))))
[17:20] <ubitux> don't forget matrixview
[17:21] <durandal_1707> that is screensaver
[17:21] <saste> ubitux: why? libcaca is useful if you want to watch files on server (or if x is borken)
[17:21] <ubitux> durandal_1707: try mplayer -vo matrixview [...]
[17:21] <saste> durandal_1707: I have partially ported the code of jwz for the matrix effectv filter
[17:22] <durandal_1707> A/V is out of sync, but this is ffmpeg/libavdevice bug
[17:24] <ubitux> saste: afaict SDL has an ascii output
[17:24] <saste> ubitux: uh? how to enable that?
[17:25] <ubitux> dunno it was outputted as ascii by default last time i tried to run a personnal sdl app through ssh
[17:25] <ubitux> it was let's say...
[17:25] <ubitux> 1 yr ago maybe
[17:26] <ubitux> try SDL_VIDEODRIVER=aalib maybe
[17:26] <ubitux> but it detected that by default last time
[17:27] <ubitux> i never looked deeper
[17:28] <durandal_1707> and libavdevice api should really change
[17:30] <durandal_1707> i would like to write Xv output too if it is possible without copying
[17:40] <ubitux> we are... overlordz @_@
[18:10] <CIA-41> ffmpeg: 03Paul B Mahol 07master * r5c858a0ac4 10ffmpeg/libswscale/input.c: 
[18:10] <CIA-41> ffmpeg: swscale/input: silence warnings about incompatible pointer types
[18:10] <CIA-41> ffmpeg: Signed-off-by: Paul B Mahol <onemda at gmail.com>
[18:26] <durandal_1707> what is difference between overlord and geek?
[18:40] <iive> durandal_1707: overlords fly in the sky, while geeks bury in basements?
[18:43] <durandal_1707> iive: compare http://ffmpeg.org/pipermail/ffmpeg-devel/2012-July/127939.html with http://lists.libav.org/pipermail/libav-devel/2012-July/031336.html
[18:44] <iive> btw, I was making Zerg joke. (starcraft)
[18:44] <Daemon404> it's probably a reference to how vlc uses ffmpeg
[18:44] <Daemon404> i wouldnt think too much of it
[18:49] <iive> well, j-b is here, you can ask him. eventually you'll get answer :)
[18:51] <j-b> durandal_1707: there is nothing to understand, just I changed a bit the headers
[18:54] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rdaa290580d 10ffmpeg/libavformat/mxfdec.c: 
[18:54] <CIA-41> ffmpeg: mxfdec: simplify code by using av_calloc()
[18:54] <CIA-41> ffmpeg: Reviewed a long time ago by: Tomas Härdin <tomas.hardin at codemill.se>
[18:54] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[18:54] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * re03cd1049e 10ffmpeg/ (3 files in 3 dirs): 
[18:54] <CIA-41> ffmpeg: Fix misspellings of FFmpeg
[18:54] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[18:54] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r83d6ad3616 10ffmpeg/libavformat/avidec.c: 
[18:54] <CIA-41> ffmpeg: avidec: Mark first frame as keyframe in case there are no keyframes.
[18:54] <CIA-41> ffmpeg: This fixes seeking in filecopy.avi of Ticket504
[18:54] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:09] Action: ubitux doesn't like the "geek" term anyway and is happy to be an overlord
[19:15] <CIA-41> ffmpeg: 03Clément BSsch 07master * r652887d6fd 10ffmpeg/tests/fate/subtitles.mak: fate/subtitles: sort tests alphabetically.
[19:15] <CIA-41> ffmpeg: 03Clément BSsch 07master * r55ed91c856 10ffmpeg/libavcodec/pthread.c: threads: fix a potential race spotted by helgrind.
[19:25] <Daemon404> durandal_1707, nice swscale patch btw.
[19:25] Action: Daemon404 goes to resend his other patch (sick last night / this morn)
[19:36] <durandal_1707> michaelni: wow, picked patch from Feb
[22:06] <CIA-41> ffmpeg: 03Diego Biurrun 07master * r9f97af2688 10ffmpeg/libavcodec/x86/dsputil_mmx.c: x86: dsputil: drop some unused CPU flag debug code
[22:06] <CIA-41> ffmpeg: 03Mans Rullgard 07master * rbf1cf4d5a5 10ffmpeg/libavcodec/ (flacdec.c flacdsp.c): 
[22:06] <CIA-41> ffmpeg: flacdec: reverse lpc coeff order, simplify filter
[22:06] <CIA-41> ffmpeg: Reversing the lpc coefficient order simplifies indexing in
[22:06] <CIA-41> ffmpeg: the filter.
[22:06] <CIA-41> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[22:06] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r85044358f6 10ffmpeg/: (log message trimmed)
[22:06] <CIA-41> ffmpeg: Merge remote-tracking branch 'qatar/master'
[22:06] <CIA-41> ffmpeg: * qatar/master:
[22:06] <CIA-41> ffmpeg:  Print full compiler identification, not only version number
[22:06] <CIA-41> ffmpeg:  flacdec: reverse lpc coeff order, simplify filter
[22:06] <CIA-41> ffmpeg:  x86: dsputil: drop some unused CPU flag debug code
[22:06] <CIA-41> ffmpeg: Conflicts:
[22:06] <CIA-41> ffmpeg: 03Mans Rullgard 07master * r45870f8b6e 10ffmpeg/ (cmdutils.c configure): 
[22:06] <CIA-41> ffmpeg: Print full compiler identification, not only version number
[22:06] <CIA-41> ffmpeg: This provides a more precise identification of the compiler used.
[22:55] <CIA-41> ffmpeg: 03Stefano Sabatini 07master * r469b9ce83f 10ffmpeg/doc/protocols.texi: 
[22:55] <CIA-41> ffmpeg: doc/protocols: document "fifo_size" and "overrun_nonfatal" UDP options
[22:55] <CIA-41> ffmpeg: Address trac ticket #1511.
[22:55] <CIA-41> ffmpeg: 03Stefano Sabatini 07master * re1422940f2 10ffmpeg/doc/protocols.texi: doc/protocols: add a note regarding the circular RX UDP buffer
[22:55] <CIA-41> ffmpeg: 03Stefano Sabatini 07master * ra8d8e868c6 10ffmpeg/libavformat/udp.c: 
[22:55] <CIA-41> ffmpeg: lavf/udp: warn in case the circular_buffer_size or overrun_nonfatal options are not supported
[22:55] <CIA-41> ffmpeg: They are only supported if pthread is available.
[22:55] <CIA-41> ffmpeg: 03Stefano Sabatini 07master * r6b80f75848 10ffmpeg/doc/protocols.texi: doc/protocols: apply misc fixes to the UDP protocol section
[22:55] <CIA-41> ffmpeg: 03Stefano Sabatini 07master * rdd13da67b2 10ffmpeg/doc/protocols.texi: doc/protocols: extend/clarify documentation for the buffer_size UDP option
[23:35] <CIA-41> ffmpeg: 03Justin Ruggles 07master * rb0b77b9ca5 10ffmpeg/libavformat/cafdec.c: caf: support either old or new style ALAC magic kuki chunk
[23:36] <CIA-41> ffmpeg: 03Justin Ruggles 07master * rf5a9c35f88 10ffmpeg/libavformat/cafdec.c: 
[23:36] <CIA-41> ffmpeg: caf: fix 'pakt' chunk parsing
[23:36] <CIA-41> ffmpeg: according to the CAF specification:
[23:36] <CIA-41> ffmpeg: "... the value for mChunkSize can be greater than the actual valid content
[23:36] <CIA-41> ffmpeg: of the packet table chunk"
[23:36] <CIA-41> ffmpeg: 03Justin Ruggles 07master * r3bab7cd128 10ffmpeg/libavformat/ (cafdec.c mov.c mov_chan.c mov_chan.h): avformat: move 'chan' tag parsing to mov_chan.c to share with the CAF demuxer
[23:36] <CIA-41> ffmpeg: 03Justin Ruggles 07master * r2ac1737583 10ffmpeg/libavcodec/alac.c: alac: clean up and update comments leftover from reverse-engineering
[23:36] <CIA-41> ffmpeg: 03Justin Ruggles 07master * r6e91f62256 10ffmpeg/libavcodec/alac.c: 
[23:36] <CIA-41> ffmpeg: alac: reduce the number of parameters to bastardized_rice_decompress()
[23:36] <CIA-41> ffmpeg: Use the ALACContext fields directly instead.
[23:36] <CIA-41> ffmpeg: 03Justin Ruggles 07master * ra06fdadd97 10ffmpeg/libavcodec/alac.c: alac: cosmetics: reindent after last commit
[23:36] <CIA-41> ffmpeg: 03Justin Ruggles 07master * rd9837434a9 10ffmpeg/libavcodec/alac.c: 
[23:36] <CIA-41> ffmpeg: alac: limit the rice param before passing to decode_scalar()
[23:36] <CIA-41> ffmpeg: reduces the number of parameters to decode_scalar() and slightly simplifies
[23:36] <CIA-41> ffmpeg: the code
[23:36] <CIA-41> ffmpeg: 03Justin Ruggles 07master * r7e6593e977 10ffmpeg/libavcodec/alac.c: 
[23:36] <CIA-41> ffmpeg: alac: eliminate 2 unneeded local variables in bastardized_rice_decompress()
[23:36] <CIA-41> ffmpeg: x_modified is just unnecessary, and final_val can be removed by simplifying
[23:36] <CIA-41> ffmpeg: the unsigned-to-signed conversion.
[23:36] <CIA-41> ffmpeg: 03Justin Ruggles 07master * r6fd8a28b59 10ffmpeg/libavcodec/alac.c: 
[23:36] <CIA-41> ffmpeg: alac: adjust conditions for updating entropy decoder history
[23:36] <CIA-41> ffmpeg: avoids some unnecessary arithmetic in certain situations
[23:36] <CIA-41> ffmpeg: 03Justin Ruggles 07master * rf2515cd629 10ffmpeg/libavcodec/alac.c: alac: simplify lpc coefficient adaptation
[23:36] <CIA-41> ffmpeg: 03Justin Ruggles 07master * r2fc24b3273 10ffmpeg/libavcodec/alac.c: alac: remove a duplicate local variable
[23:36] <CIA-41> ffmpeg: 03Justin Ruggles 07master * r5177413d20 10ffmpeg/libavcodec/alac.c: 
[23:36] <CIA-41> ffmpeg: alac: conditionally set sign_modifier to 1
[23:36] <CIA-41> ffmpeg: It is already unconditionally set to 0 prior to this, so we can modify it
[23:36] <CIA-41> ffmpeg: only when needed.
[23:36] <CIA-41> ffmpeg: 03Justin Ruggles 07master * r9a6c528e08 10ffmpeg/libavcodec/alac.c: alac: factor out output_size check in predictor_decompress_fir_adapt()
[23:36] <CIA-41> ffmpeg: 03Justin Ruggles 07master * r5138ff143f 10ffmpeg/libavcodec/alac.c: alac: use AVPacket fields directly in alac_decode_frame()
[23:36] <CIA-41> ffmpeg: 03Justin Ruggles 07master * r1193d3fedd 10ffmpeg/libavcodec/alac.c: alac: use get_sbits() to read LPC coefficients instead of casting
[23:36] <CIA-41> ffmpeg: 03Justin Ruggles 07master * r91620a04f1 10ffmpeg/libavcodec/alac.c: 
[23:36] <CIA-41> ffmpeg: 03Justin Ruggles 07master * r2aebac6918 10ffmpeg/libavcodec/alac.c: alac: cosmetics: rename some variables and function names
[23:37] <CIA-41> ffmpeg: 03Justin Ruggles 07master * rf3e5a7844b 10ffmpeg/libavcodec/alac.c: alac: calculate buffer size outside the loop in allocate_buffers()
[23:56] <CIA-41> ffmpeg: 03Nicolas George 07master * r11ce1cf9a7 10ffmpeg/libavcodec/utils.c: lavc: add debug info about skipped samples.
[23:56] <CIA-41> ffmpeg: 03Nicolas George 07master * r0e18ac5611 10ffmpeg/libavcodec/utils.c: 
[23:56] <CIA-41> ffmpeg: lavc: warn when impossible to adjust timestamps for skipped samples.
[23:56] <CIA-41> ffmpeg: It is likely to happen if pkt_timebase was not set.
[23:56] <CIA-41> ffmpeg: 03Nicolas George 07master * r0e57d79b0c 10ffmpeg/libavformat/utils.c: lavf: set pkt_timebase at the same time as stream->time_base.
[23:56] <CIA-41> ffmpeg: 03Nicolas George 07master * rd853ed46b1 10ffmpeg/ffmpeg.c: 
[23:56] <CIA-41> ffmpeg: Revert "ffmpeg: set pkt_timebase"
[23:56] <CIA-41> ffmpeg: This reverts commit 744bd8d08c18d68d92ead091c7e335b4b07459aa.
[23:56] <CIA-41> ffmpeg: pkt_timebase is now set by av_codec_set_pkt_timebase.
[23:56] <CIA-41> ffmpeg: If some demuxer set AVStream.time_base directly,
[23:56] <CIA-41> ffmpeg: it need to be fixed. There is a warning to detect it.
[23:56] <CIA-41> ffmpeg: 03Nicolas George 07master * r885fc05865 10ffmpeg/libavformat/utils.c: 
[23:56] <CIA-41> ffmpeg: lavf: count skipped samples for initial timestamps.
[23:56] <CIA-41> ffmpeg: If skip_samples is set and timestamps are synthesized using durations,
[23:56] <CIA-41> ffmpeg: make them start at -skip_samples (rescaled) instead of 0,
[23:56] <CIA-41> ffmpeg: so that the timestamp of the first undiscarded sample is 0.
[23:56] <CIA-41> ffmpeg: 03Nicolas George 07master * rcd08900393 10ffmpeg/libavcodec/utils.c: 
[23:56] <CIA-41> ffmpeg: lavc: update pkt_duration for skipped samples.
[23:56] <CIA-41> ffmpeg: Also: factor the the computation of the timestamp difference.
[23:58] <CIA-41> ffmpeg: 03Stefano Sabatini 07master * r7e17389c1e 10ffmpeg/libavdevice/sdl.c: 
[23:58] <CIA-41> ffmpeg: lavd/sdl: use verb in options help message
[23:58] <CIA-41> ffmpeg: An help message tells what an option does, not what its argument is.
[00:00] --- Fri Jul 20 2012


More information about the Ffmpeg-devel-irc mailing list