[Ffmpeg-devel-irc] ffmpeg-devel.log.20130927

[burek]

[00:26] <BBB> is it a known issue that ffplay can'y play video only files?
[00:26] <BBB> or rather, it sucks ass at it unless you say -sync video
[00:31] <beastd> BBB: report a bug. i think marton would have a look
[02:57] <cone-388> ffmpeg.git 03Michael Niedermayer 07release/0.10:20854f9bffd2: avcodec/parser: reset indexes on realloc failure
[02:57] <cone-388> ffmpeg.git 03Michael Niedermayer 07release/0.10:e28812439484: avcodec/flashsv: check diff_start/height
[02:57] <cone-388> ffmpeg.git 03Michael Niedermayer 07release/0.10:ef8145270f4a: avcodec/rpza: Perform pointer advance and checks before using the pointers
[02:57] <cone-388> ffmpeg.git 03Michael Niedermayer 07release/0.10:1a311ad99a57: jpeg2000: check log2_cblk dimensions
[02:57] <cone-388> ffmpeg.git 03Michael Niedermayer 07release/0.10:5230f1529ae6: matroska_read_seek: Fix used streams for subtitle index compensation
[02:57] <cone-388> ffmpeg.git 03Michael Niedermayer 07release/0.10:9300b1f64e5b: avcodec/pngdsp: fix (un)signed type in end comparission
[02:57] <cone-388> ffmpeg.git 03Michael Niedermayer 07release/0.10:a99aff4e4bbe: avcodec/dsputil: fix signedness in sizeof() comparissions
[02:57] <cone-388> ffmpeg.git 03Michael Niedermayer 07release/0.10:4a5bb426e26e: avcodec/mjpegdec: Add some sanity checks to ljpeg_decode_rgb_scan()
[02:57] <cone-388> ffmpeg.git 03Michael Niedermayer 07release/0.10:f3dc3bef4b40: avcodec/truemotion2: Fix av_freep arguments
[02:57] <cone-388> ffmpeg.git 03Michael Niedermayer 07release/0.10:f6057c5a62ea: avcodec/ffv1enc: update buffer check for 16bps
[02:57] <cone-388> ffmpeg.git 03Michael Niedermayer 07release/0.10:f66ecdb1b4de: avcodec: add emuedge_linesize_type
[02:57] <cone-388> ffmpeg.git 03Michael Niedermayer 07release/0.10:0d1ae06fe95e: update for 0.10.9
[03:10] <cone-388> ffmpeg.git 03Mashiat Sarker Shakkhar 07release/0.10:89d56f3be110: vc1dec: Do not use random pred_flag if motion vector data is skipped
[03:26] <cone-388> ffmpeg.git 03Michael Niedermayer 07fatal: ambiguous argument 'refs/tags/n0.10.9': unknown revision or path not in the working tree.
[03:26] <cone-388> Use '--' to separate paths from revisions
[03:26] <cone-388> refs/tags/n0.10.9:HEAD: avcodec/parser: reset indexes on realloc failure
[07:24] <Jerry123> May I ask a question?
[07:26] <Jerry123> my ffmpeg code is 2.0.1 version. I use mingw compiler under windows.
[07:26] <Jerry123> My configure options are : $ ./configure --prefix=/mingw --disable-yasm --cpu=i686 --enable-static --disab
[07:26] <Jerry123> le-shared
[07:27] <Jerry123> but that still link the share lib: libiconv-2.dll
[07:27] <Jerry123> I want to link the static lib. How can I do it?
[07:29] <JEEB> Static is enabled by default, and shared disabled. But that is related to ffmpeg, not its dependencies. Also I feel the easiest way to have whatever link against a static lib instead of a shared one is to not have the dll.a there, just the .a
[07:29] <JEEB> dll.a is the library that links against the dll (shared), .a is generally static
[07:29] <JEEB> also I have no idea how your toolchain works
[07:30] <JEEB> or where this dependency is coming from
[07:30] <JEEB> if it's ffmpeg itself then you can most probably also just disable iconv
[07:30] <JEEB> if it's not ffmpeg itself then you will have to look somewhere else in your dependencies
[07:30] <JEEB> also you really, really do not want to disable yasm
[07:31] <Jerry123> Thanks a lot, I will try.
[07:31] <JEEB> those are runtime-checked optimizations that range everywhere from MMX to AVX(2), and they /do/ make ffmpeg faster on pretty much every machine
[07:32] <Jerry123> I see.
[07:36] <TheRyuu> use this toolchain instead of whatever mingw one you're  using: http://xhmikosr.1f0.de/tools/MSYS_MinGW-w64_GCC_481_x86-x64_Full.7z
[07:36] <TheRyuu> that *only* has static libs
[07:36] <TheRyuu> (to link against)
[07:37] <Jerry123> OK thanks
[07:51] <Jerry123> I delete all *.dll.a files, and then relink, that's wokes well. :)
[08:12] <nevcairiel> TheRyuu: yes, it mostly just uses the default things to open files
[08:12] <nevcairiel> although addtionally there is also some custom avio logic to open a stream, but thats only used if required
[08:14] <TheRyuu> yea, I mostly figured out the workings of it after poking it a bit
[09:01] <durandal_1707> wm4: so what you think about replaygain filter?
[09:02] <wm4> durandal_1707: is it somewhere?
[09:03] <durandal_1707> i posted link here yesterday
[09:43] <Jerry123> hi all
[09:44] <Jerry123> I use mingw+ffmpeg+sdl. use static sdl lib, sdl lib needed -liconv -lm -luser32 -lgdi32 -lwinmm options.
[09:44] <Jerry123> but the config.mak can not included these options.
[09:44] <Jerry123> I must add it by hand.
[09:51] <durandal_1707> michaelni: if you not gonna relicense noise to lgpl, i will write addgrain filter
[09:59] <cone-106> ffmpeg.git 03Martin Storsjö 07master:8b09d917e7dc: tls: Add options for verifying the peer certificate
[09:59] <cone-106> ffmpeg.git 03Michael Niedermayer 07master:e6e71fd7b563: Merge commit '8b09d917e7dc7d7f2ace31419f802d4ff518236c'
[10:42] <durandal_1707> wm4: tried it?
[10:42] <wm4> no
[10:42] <durandal_1707> why?
[10:42] <wm4> I don't know where it is anyway
[10:47] <cone-106> ffmpeg.git 03Martin Storsjö 07master:5c53bf7aaf03: http: Pass options through to the nested protocol
[10:47] <cone-106> ffmpeg.git 03Michael Niedermayer 07master:689a1bd917d5: Merge commit '5c53bf7aaf03748464cbf978bffe7ffdb71112b1'
[11:38] <durandal_1707> should contrast/gamma be added to hue, or new filter is preffered?
[11:39] <cone-106> ffmpeg.git 03Martin Storsjö 07master:705b748e8d86: tls: Add support for listen mode
[11:39] <cone-106> ffmpeg.git 03Michael Niedermayer 07master:f31c36e55266: Merge commit '705b748e8d8612385c96428ae36ed0d42a170d93'
[11:43] <saste> durandal_1707, depends
[11:44] <saste> probably as an extensions, if there is enough code to share with hue
[11:45] <cone-106> ffmpeg.git 03Martin Storsjö 07master:1af78b865b66: rtpdec_qt: Add an accidentally removed allocation return value check
[11:45] <cone-106> ffmpeg.git 03Michael Niedermayer 07master:47c5c942f823: Merge commit '1af78b865b66acb8020c54262f12aa41ad47e05c'
[11:45] <durandal_1707> i pressume it means new filter
[11:54] <cone-106> ffmpeg.git 03Martin Storsjö 07master:06ebc0bf9a64: lavf: Allocate arrays with av_realloc if they will be realloced later
[11:54] <cone-106> ffmpeg.git 03Michael Niedermayer 07master:328a5b93d3a4: Merge commit '06ebc0bf9a6401733a4ce1310325de19f631819a'
[12:19] <cone-106> ffmpeg.git 03Martin Storsjö 07master:d872fb0f7ff2: lavf: Reset the entry count and allocation size variables on av_reallocp failures
[12:20] <cone-106> ffmpeg.git 03Michael Niedermayer 07master:72eddc10fa1f: Merge commit 'd872fb0f7ff2ff0ba87f5ccf6a1a55ca2be472c9'
[12:48] <cone-106> ffmpeg.git 03Reimar Döffinger 07master:779951cd2c1b: lxf: Support version 1 files
[12:48] <cone-106> ffmpeg.git 03Carl Eugen Hoyos 07master:839df90c718d: lxf: Support 16-channel files
[12:48] <cone-106> ffmpeg.git 03Michael Niedermayer 07master:e82071e7279a: Merge commit '839df90c718dcab9b9e91ca3c7e73479b3e8103c'
[12:56] <Compn> durandal_1707 : i'd like it to have a better name 'hue' but thats just me bikeshedding
[12:56] <Compn> i dont know that 'eq' is any better
[13:09] <cone-106> ffmpeg.git 03Paul B Mahol 07master:7e52080cb183: pcm: support 24-bit/32-bit little-endian planar
[13:09] <cone-106> ffmpeg.git 03Michael Niedermayer 07master:b654aa6beb41: Merge commit '7e52080cb18332cb1da9fe133498750b6b4b0fb3'
[13:17] <cone-106> ffmpeg.git 03Paul B Mahol 07master:a529fa2f4dd3: lxf: remove deplanarization hack
[13:17] <cone-106> ffmpeg.git 03Michael Niedermayer 07master:fc424f37a37e: Merge commit 'a529fa2f4dd3292ba7d9cd3e6ba76425b1ba7aef'
[13:32] <cone-106> ffmpeg.git 03Luca Barbato 07master:5c349ad08a9f: lxf: check the nb_streams instead of relying on padding
[13:32] <cone-106> ffmpeg.git 03Michael Niedermayer 07master:d8d235ffe9c1: Merge remote-tracking branch 'qatar/master'
[15:02] <Jerry123> I use mingw+ffmpeg+yasm Yasm generate .obj file, but AR needed .o file, how correct it? 
[15:04] <durandal_1707> saste: how to use sendcmd with X filter, if X filter is used several times in row?
[15:05] <cone-106> ffmpeg.git 03Paul B Mahol 07master:4d86a945dfe7: avfilter/af_ladspa: refactor code setting input control ports values
[15:05] <cone-106> ffmpeg.git 03Paul B Mahol 07master:110a2f446d54: avfilter/af_ladspa: allow changing controls dynamically
[15:05] <cone-106> ffmpeg.git 03Paul B Mahol 07master:191ee4aefbde: avfilter/af_ladspa: set output channel layout for passthrough case
[15:15] <michaelni> durandal_1707, i dont mind if you use my code in vf_noise.c 15985 bytes / d0b86b7749de940eb00d3d47cdff71e54ee9e07d under LGPL
[15:15] <Jerry123> YASMFLAGS must add a compiler option: --objext=o 
[15:16] <Jerry123> If -fwin32 , yasm's object's  file suffix is .obj
[15:18] <durandal_1707> michaelni: is that hash from mplayer?
[15:23] <Daemon404> Jerry123, we do not need that
[15:23] <Daemon404> we explicitly use -o
[15:23] <Daemon404> yasm -f elf -m amd64 -g dwarf2 -I./ -I.// -Pconfig.asm -I libavcodec/x86/ -o libavcodec/x86/hpeldsp.o libavcodec/x86/hpeldsp.asm
[15:23] <Daemon404> e.g. like that
[15:23] <Daemon404> (this holds true on windows too)
[15:24] <Jerry123> I think that's a configure script's bug.
[15:24] <Daemon404> i cant reproduce it on windows here.
[15:25] <Jerry123> ffmpeg version 2.0.1, configure file, line 3924
[15:25] <Daemon404> im using git HEAD
[15:25] <Daemon404> lets see...
[15:27] <Jerry123> YASMFLAGS="-f  ought to change to YASMFILAGS="--objext=o -f
[15:27] <Daemon404> library.mak
[15:27] <Daemon404> line 25
[15:27] <Daemon404> where it is actually used
[15:27] <Daemon404> $(SUBDIR)x86/%.o: $(SUBDIR)x86/%.asm $(DEPYASM) $(YASMFLAGS) -I $(<D)/ -M -o $@ $< > $(@:.o=.d) $(YASM) $(YASMFLAGS) -I $(<D)/ -o $@ $<
[15:27] <Daemon404> -i is *always* passed.
[15:28] <Daemon404> er -o
[15:28] <Jerry123> I see.
[15:28] <Jerry123> but my yasm compiler was generated .obj file
[15:28] <Daemon404> what ar you confgre flags
[15:29] <Jerry123> that's surprise.
[15:29] <Daemon404> and yasm version
[15:29] <Jerry123> my yasm is the latest
[15:29] <Daemon404> from where
[15:29] <Daemon404> which binary
[15:29] <Jerry123> yasm 1.2.0
[15:29] <Daemon404> thats not what i asked
[15:29] <Jerry123> vsyasm.exe
[15:30] <Daemon404> i believe vsyasm is patched.
[15:30] <Daemon404> it is not vanilla yasm
[15:30] <Jerry123> sorry my English is poor
[15:30] <Daemon404> and ou should not be using it
[15:30] <Jerry123> I see.
[15:30] <Daemon404> http://www.tortall.net/projects/yasm/releases/yasm-1.2.0-win32.exe
[15:30] <Jerry123> thanks I will try it.
[15:38] <Jerry123> <Daemon404> I thank I find out the reason, my 1.2.0 version's yasm, have no -M option.
[15:55] <saste> durandal_1707, you need to use the "unique" filter instance name, something like "parsed_filter_X"
[15:55] <saste> that's why we should allow the user to specify a unique name, which will not depend on the position in the parsed graph
[15:59] <cone-106> ffmpeg.git 03Michael Niedermayer 07master:df21537dfbb8: avformat/tls: fix {} error for the GNUTLS case
[16:41] <michaelni> durandal_1707, its a sha1 from a file
[16:49] <cone-106> ffmpeg.git 03Georg Martius 07master:e064d0aaf622: avfilter/vidstabtransform: update optzoom range values for recent vid.stab update.
[17:01] <mark4o> durandal_1707: lutyuv can already do gamma, contrast, and brightness adjustments
[17:02] <mark4o> although it requires entering formulas currently
[17:02] <durandal_1707> command?
[17:02] <mark4o> adding options like lutyuv=gamma=0.9:brightness-1.1:contrast=1.2 shouldn't require any change to filter_frame
[17:02] <mark4o> -vf lutyuv=y=gammaval(0.5) is listed in the docs
[17:03] <mark4o> for gamma
[17:08] <durandal_1707> -vf drawgraph ?
[17:25] <mark4o> -vf lutyuv=y=clip(val * x + y)  for contrast/brightness adjustment
[17:27] <mark4o> -vf lutyuv=y=clip(val * 1.5 - 20)
[18:04] <durandal_1707> saste: i wonder about doing something like frei0r nervous filter, it cache several frames and than pick random one from the list
[18:50] <saste> durandal_1707, why not
[18:51] <durandal_1707> saste: another is audio spectrum filter that is like showspectrum but it interprets data differently
[18:59] <durandal_1707> saste: about nervous, i'm just not happy with that name
[19:09] <cone-106> ffmpeg.git 03Michael Niedermayer 07master:9790a03e67d3: avformat/mxfdec: check sample size more completely
[20:25] <cone-106> ffmpeg.git 03Michael Niedermayer 07master:f7bea731d955: avcodec/mlpdec: Fail if the input is too small
[20:44] <durandal_1707> michaelni: the hash you put into tickets are invalid
[20:44] <durandal_1707> invalid hashes are worse than no hash at all
[20:46] <nevcairiel> i just checked the truehd one, and it has the right hash
[20:48] <durandal_1707> mxf one is wrong
[20:57] <cone-106> ffmpeg.git 03Paul B Mahol 07master:b890198b669d: avfilter: add av_cold to init()
[20:58] <cone-106> ffmpeg.git 03Paul B Mahol 07master:c2d19430290d: fate: add phase test
[21:07] <michaelni> durandal_1707, hash fixed
[21:27] <durandal_1707> michaelni: those FFmpeg forums nobody use except noobs (expecting help that never comes) so should be removed from contact page
[21:58] <Mista-D> Can someone point to a location of default analyzedurtion. dealing with an older customized binary that won't accept "-analyzeduraiton 50M", so I want to increase the default in libav.
[22:03] <durandal_1707> libavformat/options_table.h
[22:26] <Mista-D> @durandal_1707: Thank you. It worked perfectly.
[23:20] <cone-106> ffmpeg.git 03Michael Niedermayer 07master:3e0c78bac63c: avcodec/imc: Check flcoeffs5
[00:00] --- Sat Sep 28 2013