[Ffmpeg-devel-irc] ffmpeg-devel.log.20120511

burek burek021 at gmail.com
Sat May 12 02:05:03 CEST 2012


[01:17] <tg2> libavcodec/libavcodec.so: undefined reference to `x264_encoder_open_124'
[01:17] <tg2> keep getting this
[01:17] <tg2> ./configure --enable-libx264 --enable-nonfree --enable-gpl --prefix=/usr/local/
[01:19] <cbsrobot> tg2 you hit the chicken and egg problem
[01:20] <cbsrobot> tg2: read http://ffmpeg.org/trac/ffmpeg/wiki/MacOSXCompilationGuide - section x264
[01:21] <cbsrobot> but I guess the switch is now called --enable-lavf
[01:22] <cbsrobot> or --disable-lavf
[01:23] <tg2> why is it complaining about x264 lib
[01:23] <tg2> when it's installed adn at the right version
[01:23] <cbsrobot> didn't you test the threaded lookahead patch on x264 ?
[01:23] <tg2> libavcodec.so is referencing another version of x264
[01:24] <tg2> ?
[01:24] <tg2> yeah I did
[01:24] <tg2> now ffmpeg is throwing this error when I try and compile
[01:24] <tg2> which it wasn't before
[01:24] <cbsrobot> yeah
[01:24] <tg2> but htis is on a new box
[01:24] <tg2> that never had the v.125 version
[01:24] <cbsrobot> read the link !
[01:25] <cbsrobot> 1. install x264 with --disable-lavf
[01:25] <tg2> k
[01:25] <cbsrobot> the install ffmpeg with --enable-libx264
[01:25] <tg2> was missing the --disable-lavf step
[01:25] <cbsrobot> then reinstall x264 with --enable-lavf
[01:26] <cbsrobot> np
[01:26] <tg2> odd bug
[01:26] <tg2> what causes it?
[01:26] <tg2> dependancy loop?
[01:26] <cbsrobot> x264 depends on ffmpeg epends on x264
[01:26] <cbsrobot> *depends
[01:27] <tg2> ok
[01:39] <tg2> still getting that error
[01:39] <tg2> even after recompiling x264 with the --disable-lavf
[01:39] <tg2> and --enable-shared
[01:39] <tg2> i think it only starts happening afte ryou install libav
[02:26] <tg2> damn
[02:26] <tg2> yeah had ot manually remove all instances of libx264 (libav put some .120 versions in some odd places) then recompile/install x264 and recompile ffmpeg
[03:12] <CIA-63> ffmpeg: 03Michael Niedermayer 07master * r8616c44606 10ffmpeg/ffmpeg.c: 
[03:12] <CIA-63> ffmpeg: ffmpeg: fix deinterlace
[03:12] <CIA-63> ffmpeg: Fixes ticket1295
[03:12] <CIA-63> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[03:26] <Compn> tg2 : get all of the cores working yet?
[03:26] <Compn> ehe
[03:38] <CIA-63> ffmpeg: 03Michael Niedermayer 07master * r940be96540 10ffmpeg/libavcodec/dcaenc.c: 
[03:38] <CIA-63> ffmpeg: dcaenc: fix out of array read
[03:38] <CIA-63> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[04:08] <tg2> compn :D
[04:08] <tg2> 680 fps
[04:08] <tg2> using 32 cores
[04:08] <tg2> on my split demo
[04:08] <tg2> Array
[04:08] <tg2> (
[04:08] <tg2>     [fps] => 625
[04:08] <tg2>     [frames] => 75752
[04:08] <tg2>     [size] => 353566720
[04:08] <tg2>     [seconds] => 3775.2
[04:08] <tg2>     [q] => 10677
[04:08] <tg2>     [qct] => 480
[04:08] <tg2>     [bitrate] => 731.68044077135
[04:08] <tg2>     [avgq] => 22.24375
[04:08] <tg2>     [pct] => 57.5
[04:08] <tg2>     [eta_seconds] => 89.6192
[04:09] <tg2> thats a 118m rip @ 480p
[04:09] <tg2> in 150 seconds
[04:09] <tg2> going to try it on some bigger stuff and higher resolutions
[04:09] <tg2> but as of now it's working quite well
[04:10] <tg2> wrote a script to cut the video into slices no smaller than 5 minutes
[04:10] <tg2> and convert in parallel
[04:10] <tg2> then stitch it together
[04:11] <tg2> just playing around with bitrates etc
[04:11] <tg2> thats on normal profile, not ultrafast or superfast or whatever
[04:12] <Compn> interestings
[04:13] <tg2> working pretty well
[04:13] <tg2> i'm just trying it with various video types
[04:13] <tg2> to make sure the quality is acceptable
[04:13] Action: Compn guessing it fails on files without keyframes
[04:13] <ohsix> and with wildly varying bitrates
[04:14] <tg2> you can always do 2-pass on it
[04:14] <tg2> anyway, stop hatin ;)
[04:15] <ohsix> i'd hate on a guy riding a sport bike with no helmet and shorts/flipflops too
[04:15] <tg2> either way
[04:15] <tg2> one can always select ideal candidates before deciding to encode in chunks or in sequence
[04:15] <tg2> traditionally
[04:17] <tg2> if you have a 2hr video, it will cut it into a maximum of 6 parts
[04:18] <tg2> that's 20 minutes per chunk
[04:18] <tg2> which should be sufficient to equalize vbr
[04:18] <tg2> not 2-pass quality
[04:19] <ohsix> until you're all out of bits and your face gets ground off
[05:04] <tg2> -ss terribly unreliable when placed before -i
[05:04] <tg2> lol
[14:03] <CIA-63> ffmpeg: 03Michael Niedermayer 07master * r3700f655c5 10ffmpeg/libavformat/mpegtsenc.c: 
[14:03] <CIA-63> ffmpeg: mpegtsenc: check max_delay in write_packet_internal().
[14:03] <CIA-63> ffmpeg: Checking it during PES writing is too late.
[14:03] <CIA-63> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[14:03] <CIA-63> ffmpeg: 03Michael Niedermayer 07master * r9c409e568f 10ffmpeg/libavformat/mpegtsenc.c: 
[14:03] <CIA-63> ffmpeg: mpegtsenc: fix delay type
[14:03] <CIA-63> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[14:18] <CIA-63> ffmpeg: 03Michael Niedermayer 07master * r6bb35f4351 10ffmpeg/libavformat/mpegtsenc.c: 
[14:18] <CIA-63> ffmpeg: mpegtsenc: fix 10l typo, wrong stream
[14:18] <CIA-63> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[15:08] <CIA-63> ffmpeg: 03Michael Niedermayer 07master * r44391f706b 10ffmpeg/libavfilter/lavfutils.c: 
[15:08] <CIA-63> ffmpeg: lavfutils: ff_load_image: Initialize context
[15:08] <CIA-63> ffmpeg: should fix ticket1264
[15:08] <CIA-63> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[15:47] Action: Compn wonders why rtmp options are being committed one at a time
[15:47] <Compn> theres no reason not to combine them and commit them all at once?
[15:49] <av500> that's optional
[15:52] <Compn> the options ?
[15:52] <Compn> there are two types of rtmp servers, one that require no options like -app or -playpath, and then the other serverse require everything from -pageurl -swfurl plus securetoken stuff and live
[15:53] <Compn> not important
[15:53] <Compn> just strange
[15:54] <av500> I meant combining the commits of the options is optional
[15:55] <kriegerod> hi. i'd like somebody to help me to figure out why RTMP output causes no correct av_*write_frame() fail, in case of RTMP server shutdown. Only SIGPIPE arrives, which is unacceptable for me.
[15:56] <kriegerod> I've tried doing send() with flag MSG_NOSIGNAL on libavformat/tcp.c in tcp_write(), not a solution yet
[17:00] <ubitux> Daemon404: if you don't have libutvideo, make checkheaders doesn't pass
[17:00] <Daemon404> it's entirely pointless
[17:00] <Daemon404> from a practical standpoint
[17:00] <Daemon404> that file is never touched unelss libutvideo is explicitly enabled and has passed teh configure check
[17:01] <ubitux> ah my bad
[17:01] <Daemon404> what im saying is, make checkheaders is broken from a design standpoint.
[17:01] <Daemon404> :)
[17:01] <ubitux> no i'm wrong with this patch
[17:01] <Daemon404> ?
[17:01] <ubitux> it's just like OS2THREADS
[17:01] <ubitux> i have to skip it according to the enabled stuff
[17:01] <Daemon404> ah
[17:01] <ubitux> i'll resend a patch :)
[17:02] <Daemon404> =p
[17:03] <ubitux> basically, sth like SKIPHEADERS-$(HAVE_LIBUTVIDEO) += libutvideo.h
[17:03] <CIA-63> ffmpeg: 03Michael Niedermayer 07master * r30fc710556 10ffmpeg/libavformat/utils.c: 
[17:03] <CIA-63> ffmpeg: lavf: print a matching tag in case of code tag validation failure
[17:03] <CIA-63> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[17:03] <CIA-63> ffmpeg: 03Michael Niedermayer 07master * r75a9479b39 10ffmpeg/libavformat/utils.c: 
[17:03] <CIA-63> ffmpeg: lavf: extend the avi raw hack to mov
[17:03] <CIA-63> ffmpeg: Fixes ticket1194
[17:03] <CIA-63> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[17:13] <RushPL> hello, could anyone tell me why this line is necessary in ffplay.c:2063  "dec_channel_layout = (dec->channel_layout && dec->channels == av_get_channel_layout_nb_channels(dec->channel_layout)) ? dec->channel_layout : av_get_default_channel_layout(dec->channels);"? why using "dec->channel_layout" is not sufficient and requires additional check?
[17:44] <michaelni> RushPL, thats a odd question
[17:44] <michaelni> i mean "why is line X in file Y needed?"
[17:45] <michaelni> is there some problem that you suspect is caused by that line ?
[17:46] <RushPL> no, but why wouldn't using dec->channel_layout be sufficient? this is supposed to somehow normalize it but I don't understand in what way.
[17:46] <RushPL> dec->channel_layout is supposed to say whether it is MONO, STEREO, 5.1 etc. why is not sufficient?
[17:47] <ubitux> i guess it's checking weither the channel layout is set or not
[17:47] <ubitux> and if not set, it will set the channel layout to the default one according to the number of channels
[17:48] <ubitux> maybe it should be internal to the lib
[17:48] <RushPL> so hm, it is optionally set then?
[17:48] <ubitux> (setting a default channel layout if not set)
[17:48] <RushPL> so should I assume its value can be wrong?
[17:48] <RushPL> and channels is always right?
[17:49] <RushPL> (well obviously it must be right since we get so many channels of data :P)
[17:53] <CIA-63> ffmpeg: 03Carl Eugen Hoyos 07master * r6d376346eb 10ffmpeg/libavcodec/libopenjpegdec.c: 
[17:53] <CIA-63> ffmpeg: Allow decoding of j2k images with bpp<8 and 8<bpp<16 with libopenjpeg.
[17:53] <CIA-63> ffmpeg: Fixes one of several problems described in ticket #1279.
[18:09] <RushPL> ubitux: I guess that is good enough for me knowing that channel_layout might not be set. Thank you.
[18:09] <Daemon404> [mjpeg @ 0x9e7fa80] [IMGUTILS @ 0xbfd1b804] Picture size 29528x29528 is invalid
[18:10] <Daemon404> lol awesome
[18:10] <Daemon404> (this is really that big)
[18:10] <nevcairiel> i hope its just one image and not a movie :d
[18:10] <Daemon404> yeah
[18:10] <Daemon404> one image
[18:11] <Daemon404> imagemagick can luckily handle it
[18:11] <Daemon404> (along with 6 gb of ram)
[18:12] <ohsix> tell it where swap is and it will use the tile cache
[18:12] <Compn> why have 6gb if you never use it, thats all i'm sayin!
[18:47] <CIA-63> ffmpeg: 03Michael Niedermayer 07master * re20f46481b 10ffmpeg/libavcodec/mpeg12.c: 
[18:47] <CIA-63> ffmpeg: mpeg12: print an error when there was a frame in extradata.
[18:47] <CIA-63> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[18:47] <CIA-63> ffmpeg: 03Michael Niedermayer 07master * rde1824e970 10ffmpeg/libavcodec/ (mpeg12.c mpeg12.h): 
[18:47] <CIA-63> ffmpeg: mpeg12: fix logic that prevents extradata from being parsed twice.
[18:47] <CIA-63> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[21:04] <CIA-63> ffmpeg: 03Mans Rullgard 07master * r4f1500689d 10ffmpeg/ (avconv.c tests/ref/fate/vc1-ism): (log message trimmed)
[21:04] <CIA-63> ffmpeg: avconv: use lrint() for rounding double timestamps
[21:04] <CIA-63> ffmpeg: Converting the double to float for lrintf() loses precision when
[21:04] <CIA-63> ffmpeg: the value is not exactly representable as a single-precision float.
[21:04] <CIA-63> ffmpeg: Apart from being inaccurate, this causes discrepancies in some
[21:04] <CIA-63> ffmpeg: configurations due to differences in rounding.
[21:04] <CIA-63> ffmpeg: Note that the changed timestamp in the vc1-ism test is a bogus,
[21:04] <CIA-63> ffmpeg: 03Mans Rullgard 07master * rc776531aef 10ffmpeg/libavcodec/vqavideo.c: 
[21:04] <CIA-63> ffmpeg: vqavideo: change x/y loop counters to the usual pattern
[21:04] <CIA-63> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[21:04] <CIA-63> ffmpeg: 03Mans Rullgard 07master * r2eba6898c9 10ffmpeg/libavcodec/arm/dsputil_neon.S: 
[21:04] <CIA-63> ffmpeg: arm: dsputil: prettify some conditional instructions in put_pixels macros
[21:04] <CIA-63> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[21:04] <CIA-63> ffmpeg: 03Mans Rullgard 07master * re54e6f25cf 10ffmpeg/libavcodec/arm/dsputil_neon.S: (log message trimmed)
[21:04] <CIA-63> ffmpeg: arm/neon: dsputil: use correct size specifiers on vld1/vst1
[21:04] <CIA-63> ffmpeg: 03Michael Niedermayer 07master * r9eb99ab25b 10ffmpeg/: (log message trimmed)
[21:04] <CIA-63> ffmpeg: Merge remote-tracking branch 'qatar/master'
[21:04] <CIA-63> ffmpeg: * qatar/master:
[21:04] <CIA-63> ffmpeg:  fate: employ better names and add a convenient shorthand for vp6 tests
[21:04] <CIA-63> ffmpeg:  arm/neon: dsputil: use correct size specifiers on vld1/vst1
[21:04] <CIA-63> ffmpeg:  arm: dsputil: prettify some conditional instructions in put_pixels macros
[21:04] <CIA-63> ffmpeg:  vqavideo: change x/y loop counters to the usual pattern
[22:33] <CIA-63> ffmpeg: 03Michael Niedermayer 07release/0.5 * rbde4b66063 10ffmpeg/libavformat/nsvdec.c: (log message trimmed)
[22:33] <CIA-63> ffmpeg: nsvdec: Fix use of uninitialized streams.
[22:33] <CIA-63> ffmpeg: Fixes CVE-2011-3940 (Out of bounds read resulting in out of bounds write)
[22:33] <CIA-63> ffmpeg: Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
[22:33] <CIA-63> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:33] <CIA-63> ffmpeg: (cherry picked from commit 5c011706bc752d34bc6ada31d7df2ca0c9af7c6b)
[22:33] <CIA-63> ffmpeg: Signed-off-by: Alex Converse <alex.converse at gmail.com>
[22:34] <CIA-63> ffmpeg: 03Alex Converse 07release/0.5 * r056c909d9d 10ffmpeg/libavformat/nsvdec.c: (log message trimmed)
[22:34] <CIA-63> ffmpeg: nsvdec: Be more careful with av_malloc().
[22:34] <CIA-63> ffmpeg: Check results for av_malloc() and fix an overflow in one call.
[22:34] <CIA-63> ffmpeg: Related to CVE-2011-3940.
[22:34] <CIA-63> ffmpeg: Based in part on work from Michael Niedermayer.
[22:34] <CIA-63> ffmpeg: Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
[22:34] <CIA-63> ffmpeg: (cherry picked from commit 8fd8a48263ff1437f9d02d7e78dc63efb9b5ed3a)
[22:34] <CIA-63> ffmpeg: 03Michael Niedermayer 07release/0.5 * rc4e8c99507 10ffmpeg/libavformat/dv.c: (log message trimmed)
[22:34] <CIA-63> ffmpeg: dv: Fix null pointer dereference due to ach=0
[22:34] Last message repeated 1 time(s).
[22:34] <CIA-63> ffmpeg: Fixes part2 of CVE-2011-3929
[22:34] <CIA-63> ffmpeg: Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
[22:34] <CIA-63> ffmpeg: Reviewed-by: Roman Shaposhnik <roman at shaposhnik.org>
[22:34] <CIA-63> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:34] <CIA-63> ffmpeg: 03Alex Converse 07release/0.5 * r5a92aa378d 10ffmpeg/libavformat/dv.c: (log message trimmed)
[22:34] <CIA-63> ffmpeg: dv: Fix small stack overread related to CVE-2011-3929 and CVE-2011-3936.
[22:34] <CIA-63> ffmpeg: Found with asan.
[22:34] <CIA-63> ffmpeg: Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
[22:34] <CIA-63> ffmpeg: Signed-off-by: Alex Converse <alex.converse at gmail.com>
[22:34] <CIA-63> ffmpeg: (cherry picked from commit 2d1c0dea5f6b91bec7f5fa53ec050913d851e366)
[22:45] <CIA-63> ffmpeg: 03Michael Niedermayer 07release/0.5 * rf73106256d 10ffmpeg/ (RELEASE VERSION): 
[22:45] <CIA-63> ffmpeg: fix release number
[22:45] <CIA-63> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:45] <CIA-63> ffmpeg: 03Michael Niedermayer 07release/0.5 * r123e925956 10ffmpeg/libavformat/mm.c: 
[22:45] <CIA-63> ffmpeg: mmdemux: dont set pkt->size to an invalid value.
[22:45] <CIA-63> ffmpeg: Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
[22:45] <CIA-63> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:45] <CIA-63> ffmpeg: (cherry picked from commit 0c97fd336e17535239ab44d755a0d957dc2688f3)
[22:46] <CIA-63> ffmpeg: 03Michael Niedermayer 07release/0.5 * ra55db1fc49 10ffmpeg/libavcodec/x86/dsputilenc_mmx.c: 
[22:46] <CIA-63> ffmpeg: dsp: fix diff_bytes_mmx() with small width
[22:46] <CIA-63> ffmpeg: Fixes Ticket1068
[22:46] <CIA-63> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:46] <CIA-63> ffmpeg: (cherry picked from commit 73089eccd3e48539555349b36d8aabbf1cea416e)
[22:46] <CIA-63> ffmpeg: 03Michael Niedermayer 07release/0.5 * r96c6b3a11c 10ffmpeg/libavcodec/h261dec.c: 
[22:46] <CIA-63> ffmpeg: h261: check mtype.
[22:46] <CIA-63> ffmpeg: Fixes out of array read
[22:46] <CIA-63> ffmpeg: Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
[22:46] <CIA-63> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:46] <CIA-63> ffmpeg: (cherry picked from commit ec3cd74f2dab8e3e8234ccb994132b23d3098585)
[22:46] <CIA-63> ffmpeg: 03Michael Niedermayer 07release/0.5 * r5e3cd42b6d 10ffmpeg/libavformat/4xm.c: 
[22:46] <CIA-63> ffmpeg: 4xmdemux: Check chunk size
[22:46] <CIA-63> ffmpeg: Fixes over reading the header array
[22:46] <CIA-63> ffmpeg: Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
[22:46] <CIA-63> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:46] <CIA-63> ffmpeg: (cherry picked from commit 474e31c904f766b6989fe614c3fb093e697c847f)
[22:46] <CIA-63> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[00:00] --- Sat May 12 2012


More information about the Ffmpeg-devel-irc mailing list