[Ffmpeg-devel-irc] ffmpeg-devel.log.20140410

[burek]

[00:00] <lukaszmluki> Hi, I have question: is it possible that AVPacket.size may change for the certain stream or it is always the same?
[00:00] <wm4> of course it will change
[00:01] <lukaszmluki> inside one file, for the same stream, it may have different sizes, right?
[00:01] <wm4> yes
[00:01] <lukaszmluki> ok, thx
[00:03] <ubitux> lukaszmluki: ffprobe -show_entries packet=size -select_streams v:0 -of flat foobar.mp4
[00:04] <wm4> is ffprobe turing complete yet
[00:04] <ubitux> since it's the encoded video frames, packet sizes change every time, obviously
[00:04] <ubitux> wm4: eval should be
[00:04] <lukaszmluki> I think about audio packets, if thats matter.
[00:05] <ubitux> change 'v' into 'a'
[00:05] <lukaszmluki> oki
[00:05] <ubitux> and change 'video' into 'audio' in my next sentence
[00:05] <lukaszmluki> information that it is not constant is enough
[00:14] <cone-291> ffmpeg.git 03Michael Niedermayer 07master:f703dae7aa76: avcodec/cngdec: use av_mallocz_array()
[00:14] <cone-291> ffmpeg.git 03Michael Niedermayer 07master:968516cebf89: avformat/nut: add ff_nut_audio_extra_tags to demuxer too
[00:14] <cone-291> ffmpeg.git 03Michael Niedermayer 07master:849d40dce62d: avcodec/mpegvideo_enc: change some asserts to av_asserts
[01:34] <cone-291> ffmpeg.git 03Carl Eugen Hoyos 07master:bf87d582b1cb: Do not allow writing invalid wav channel layouts by default.
[01:34] <cone-291> ffmpeg.git 03Carl Eugen Hoyos 07master:2a3ff9fbbacf: Remove an unneeded include of avassert.h from amr.c.
[01:34] <cone-291> ffmpeg.git 03Carl Eugen Hoyos 07master:cf79f202757b: Fix a typo in amr.c.
[01:34] <cone-291> ffmpeg.git 03Michael Niedermayer 07master:421c116fc498: Merge remote-tracking branch 'cehoyos/master'
[01:38] <wm4> michaelni_: so what's the deal, are all enums now supposed to use fourcc values?
[01:40] <michaelni_> our codec ids do
[01:41] <michaelni_> i dont think anyone suggested to use fourccs for pixfmts
[01:44] <wm4> they seem to pop up in a number of places - even when not needed
[01:49] <michaelni> which one are you thinking of ?
[01:50] <michaelni> i mean which did pop up and isnt needed ?
[01:51] <wm4> e.g. the ones in libavdevice/avdevice.h 
[01:51] <lukaszmluki> probably my patch to avdevice.h
[01:51] <lukaszmluki> but that was Nicolas who wanted fourcc
[01:52] <wm4> but even the ones for codecids are not needed (though that damage is already done)
[01:52] <michaelni> damage ?
[01:53] <michaelni> fourccs should reduce the risk of conflicts
[01:53] <wm4> they make debugging harder etc.
[01:53] <wm4> they seem like an innecessary complication
[01:53] <wm4> *un
[02:00] <michaelni> iam not sure i agree, i mean a codec id of 123 tells me as much as 98765432 which is how they would look like when showen as literal numbers in a debuger
[02:01] <michaelni> ... and a good debuger should show the named identifer of an enum type
[02:01] <michaelni> either way
[02:09] <kierank> wm4: what about codecs in containers that lavc doesnt support
[02:09] <kierank> they need an identifier of some sort
[02:09] <wm4> kierank: codecids are never mixed with real fourccs
[02:09] <kierank> true
[03:06] <cone-291> ffmpeg.git 03Lukasz Marek 07master:fbf05759e38f: sws: fix warning about RGB_PACK24_B_OPERANDS redefinition
[04:03] <cone-291> ffmpeg.git 03Lukasz Marek 07master:e1c48e3593fb: lavf/utils: fix mem leak
[04:46] <cone-291> ffmpeg.git 03Reinhard Tartler 07release/2.2:a014b9614e61: Prepare for 10.1 Release
[04:46] <cone-291> ffmpeg.git 03Michael Niedermayer 07release/2.2:bd553941ae51: Merge commit 'a014b9614e6165fb8df9e756f46a95d516794678' into release/2.2
[04:53] <cone-291> ffmpeg.git 03Martin Storsjö 07release/2.2:e0aa76d38a02: golomb: Fix the implementation of get_se_golomb_long
[04:53] <cone-291> ffmpeg.git 03Michael Niedermayer 07release/2.2:b2583c2b6289: Merge commit 'e0aa76d38a02090245284fc157afb9074e9ff073' into release/2.2
[05:08] <cone-291> ffmpeg.git 03Michael Niedermayer 07release/2.2:a221c9bd7677: avcodec/libx264: move where x264opts is applied down so it isnt overridden by avctx & defaults
[05:08] <cone-291> ffmpeg.git 03Michael Niedermayer 07release/2.2:5cb2a1c3f0b9: avfilter/vf_pullup: fix gray8
[05:08] <cone-291> ffmpeg.git 03Michael Niedermayer 07release/2.2:02bae9f013e7: avcodec/h264: clear cur_pic structure instead of duplicating it in ff_h264_update_thread_context()
[05:08] <cone-291> ffmpeg.git 03Lukasz Marek 07release/2.2:fc8eb4c1f984: lavu/opt: validate range before dereference
[05:08] <cone-291> ffmpeg.git 03Michael Niedermayer 07release/2.2:f41622ecb4b3: avcodec/wma: use av_freep(), do not leave stale pointers in memory
[05:08] <cone-291> ffmpeg.git 03Michael Niedermayer 07release/2.2:544accc895f3: avformat/omadec: fix probetest failure
[05:08] <cone-291> ffmpeg.git 03Michael Niedermayer 07release/2.2:34592d04fbea: examples/avio_reading: fix null dereference on error
[05:09] <cone-291> ffmpeg.git 03Michael Niedermayer 07release/2.2:36cab9c408bd: avcodec/pthread_frame: fix missing unlock on error
[05:09] <cone-291> ffmpeg.git 03Michael Niedermayer 07release/2.2:b479b42b26b9: avcodec/msrle: check return code for success before use
[05:09] <cone-291> ffmpeg.git 03Michael Niedermayer 07release/2.2:352b0969e23c: swresample/resample: Limit filter length
[05:09] <cone-291> ffmpeg.git 03Michael Niedermayer 07release/2.2:d20ac551a861: swresample/dither: use av_malloc_array()
[05:09] <cone-291> ffmpeg.git 03Michael Niedermayer 07release/2.2:dde95268cc96: swresample/resample: use av_malloc_array() where appropriate
[05:09] <cone-291> ffmpeg.git 03Michael Niedermayer 07release/2.2:ce94955b3c0b: swscale/x86/swscale_template: loose hardcoded dstw_offset
[05:09] <cone-291> ffmpeg.git 03Anthoine Bourgeois 07release/2.2:dfddefa13afa: avcodec/dirac_arith: Fix build with PIC and stack-check options
[05:16] <cone-291> ffmpeg.git 03Michael Niedermayer 07release/2.2:e72c0a04664a: update for 2.2.1
[05:53] <cone-291> ffmpeg.git 03Lukasz Marek 07fatal: ambiguous argument 'refs/tags/n2.2.1': unknown revision or path not in the working tree.
[05:53] <cone-291> Use '--' to separate paths from revisions
[05:53] <cone-291> refs/tags/n2.2.1:HEAD: lavf/utils: fix mem leak
[07:43] <ubitux> heh koda is at it again
[07:47] <ubitux> > In a normal world one could envision ffmpeg and libav as two branches, one for the development/hacks/quickbugfix and the other for stable development.
[07:47] <ubitux> lol
[09:27] <t355u5> problem compiling 2.2.1 with x265:
[09:27] <t355u5> libavcodec/libx265.c: In function libx265_encode_init:
[09:27] <t355u5> libavcodec/libx265.c:124: error: x265_param has no member named bEnableVuiParametersPresentFlag
[09:27] <t355u5> libavcodec/libx265.c:125: error: x265_param has no member named bEnableAspectRatioIdc
[09:27] <t355u5> libavcodec/libx265.c:126: error: x265_param has no member named aspectRatioIdc
[09:27] <t355u5> libavcodec/libx265.c:127: error: x265_param has no member named sarWidth
[09:27] <t355u5> libavcodec/libx265.c:128: error: x265_param has no member named sarHeight
[09:29] <nevcairiel> 2.2.1 only works with the x265 version that was released when 2.2.1 was released, not with any newer versions
[09:29] <nevcairiel> thats how it is, and thats how it shall remain
[09:32] <t355u5> 2.2.1 was released today. and I'm using an x265 version from today
[09:32] <nevcairiel> well technically it should've said "when 2.2 was released"
[09:33] <nevcairiel> x265 changes its API once a week now, the release branch is not going to get updated for that. If you want to use x265, use ffmpeg git master
[09:34] <t355u5> which would be which version exactly? 
[09:34] <t355u5> I'm creating static builds for git master and for releases
[09:40] <relaxed> t355u5: it fails to build with git master too
[09:41] <nevcairiel> like i said, they keep changing it once a week these days, its not always perfectly up to date :p
[09:42] <nevcairiel> i saw a patch for the recent breakage on the mailing list however
[09:42] <t355u5> e1c48e3 built just fine
[09:43] <t355u5> I guess I will have to match the dates from 2.2 and x265 and try again
[09:45] <t355u5> thx for the info
[10:20] <t355u5> just tried with x265 from March 21st, since 2.2 was released on March 23rd - still the same error. does anyone have an idea which x265 is supposed to compile with 2.2?
[12:52] <ubitux> BBB: no need to ping koda, you are an uninvited troll according to him in the thread
[12:53] <ubitux> "It's hard to reply without reminding sad past topics, arousing trolls (a few have already took part uninvited)", "Much less the troll attempts in the other emails that got duly ignored."
[12:56] <BBB> ah, duly ignored
[12:56] <BBB> everyone that disagrees with me is an E V I L T R O L L!
[12:56] <BBB> there
[12:56] <BBB> I said it
[12:56] <BBB> thus it is true
[13:52] <Compn> ubitux : want to submit an api patch to libav-devel to call their bluff? :)
[13:52] <Compn> :P
[14:08] <cone-593> ffmpeg.git 03Derek Buitenhuis 07master:8945dcbb52b6: libx265: Use x265_param_parse to set the SAR
[14:25] <pross-au> sigh
[14:29] <ubitux> pross-au: he has commit access
[14:29] <pross-au> the logical deduction still holds
[14:30] <pross-au> join libav, get commit rights? no?
[14:33] <av500> fill out the form in triplicate
[14:33] <Daemon404> my biggest question is
[14:33] <Daemon404> why do you all keep replyoimg
[14:33] <Daemon404> replying*
[14:34] <ubitux> well we don't want to let him give the illusion that libav is trying to cooperate while we ignore it
[14:34] <ubitux> because that's actually what they do
[14:37] <Daemon404> wut
[14:38] <Daemon404> michaelni, found another jpeg that fails with avcodec btw
[14:38] <Daemon404> looking into it
[14:38] <Daemon404> not cmyk this time
[14:38] <pross-au> (i dunno abou you, but i found the repeated advice to stop merging helpful)
[15:03] <Compn> Daemon404 : probably a million jpegs that fail with avcodec
[15:03] <Compn> its a gigantic file format with so many crazy specs 
[15:03] <Compn> trillions of samples and broken software
[15:03] <Compn> :)
[15:09] <cone-593> ffmpeg.git 03Nicolas George 07master:ad8159e0fe2b: libavfilter: Add asettb filter for setting timebase for audio
[15:09] <cone-593> ffmpeg.git 03Michael Niedermayer 07master:476471495265: Merge remote-tracking branch 'qatar/master'
[17:48] <cone-593> ffmpeg.git 03Matt Oliver 07master:d1e6e5c88733: avcodec/x86: Exclude broken get_cabac under icl.
[17:49] <cone-593> ffmpeg.git 03Derek Buitenhuis 07master:fcc43e02d302: libx265: Use 16-bit SAR
[17:49] <cone-593> ffmpeg.git 03Derek Buitenhuis 07master:cdd21f1f03c5: libx265: Only set the SAR if it is valid
[20:05] <cone-593> ffmpeg.git 03Michael Niedermayer 07master:c486cfab792e: avcodec/imgconvert: fix nb_components and depth for PAL8 in get_pix_fmt_score()
[21:32] <cone-593> ffmpeg.git 03Anton Khirnov 07master:9a07fac678a8: mov: read hydrogenaudio replaygain information
[21:32] <cone-593> ffmpeg.git 03Michael Niedermayer 07master:6e573b77597c: Merge commit '9a07fac678a8540d076e635061bbaa4ed09a9431'
[21:32] <cone-593> ffmpeg.git 03Michael Niedermayer 07master:4d4bc89409fb: avformat/mov: merge mov_read_custom_metadata() and mov_read_custom()
[21:41] <cone-593> ffmpeg.git 03Derek Buitenhuis 07master:4a2226451e14: libx265: Use x265_param_parse to set the SAR
[21:41] <cone-593> ffmpeg.git 03Derek Buitenhuis 07master:cd0ac6f6e5c3: libx265: Use 16-bit SAR
[21:41] <cone-593> ffmpeg.git 03Derek Buitenhuis 07master:25fbfeed98d2: libx265: Only set the SAR if it is valid
[21:41] <cone-593> ffmpeg.git 03Michael Niedermayer 07master:a4662cd6b11e: Merge remote-tracking branch 'qatar/master'
[21:52] <jedir0x> I'm using the ffmpeg libraries on Android to render video.  On some devices when creating a 480x480 h.264 mp4 with aac audio the bottom 1/4 or so of the video is a large green bar.... does that mean anything to anyone? :)
[22:04] <iive> jedir0x: green means the area is full of zeroes.
[22:15] <jedir0x> i see
[22:15] <cone-593> ffmpeg.git 03James Almer 07master:09c16542085a: lavf/rtspcodes: add missing header include
[22:15] <cone-593> ffmpeg.git 03Michael Niedermayer 07master:6290ba005296: avfilter/settb: remove duplicate #include
[22:18] <Compn> jedir0x : try creating a window larger than your video res ? :)
[22:18] <Compn> but no idea
[22:20] <jedir0x> i was passing the wrong stuff to sws_scale :0
[22:20] <jedir0x> :) *
[22:20] <jedir0x> thanks for the tip
[22:21] <Compn> lol how did you figure it out from my guess ?
[22:27] <cone-593> ffmpeg.git 03Carl Eugen Hoyos 07master:ecb80fe78c58: Cosmetics: Remove an outdated comment from lavc/rawdec.c.
[22:37] <cone-593> ffmpeg.git 03Michael Niedermayer 07master:517cc81e939f: ffmpeg: stop protocols on interrupt when no transcoding has happened yet
[23:09] <cone-593> ffmpeg.git 03Anssi Hannula 07master:39b192b8e1c1: avformat/hls: do not call ffurl_seek/ffurl_size on crypto protocol
[00:00] --- Fri Apr 11 2014