[FFmpeg-devel] [PATCH] lavc/hevc_parse: Don't take a HEVCContext

2016-04-25 Thread Derek Buitenhuis
It's not even used anymore, and the checks are no longer functionally important. Signed-off-by: Derek Buitenhuis <derek.buitenh...@gmail.com> --- libavcodec/hevc.c| 2 +- libavcodec/hevc.h| 4 ++-- libavcodec/hevc_parse.c | 11 +-- libavcodec/hevc_parser.c | 2

Re: [FFmpeg-devel] [PATCH] lavf/concatdec: remove unrelated change during codecpar merge.

2016-04-24 Thread Derek Buitenhuis
On 4/24/2016 3:39 PM, Nicolas George wrote: > Clearing the extradata is not related to the codecpar change, > and it breaks if auto_convert is disabled. > > Fix trac ticket #5461. > > Signed-off-by: Nicolas George > --- > libavformat/concatdec.c | 5 - > 1 file changed, 5

Re: [FFmpeg-devel] [PATCH] lavf/concatdec: remove unrelated change during codecpar merge.

2016-04-24 Thread Derek Buitenhuis
On 4/24/2016 4:26 PM, Nicolas George wrote: > That may be true, but the fix is incorrect, and should not have sneaked > during a completely unrelated merge. For the sake of tracking changes, I > want to revert this, and then look for a more correct fix. The fix was deemed OK at the time by a few

Re: [FFmpeg-devel] [PATCH] libavcodec/qsvdec_h2645.c Bug fixed: wrong ticks_per_frame.

2016-04-25 Thread Derek Buitenhuis
On 4/25/2016 2:14 PM, Ivan Uskov wrote: > The attached patch does fixes the issue of frames duplication when > elementary h.264 stream decodes by qsvdec. Could you perhaps elaborate in the commit message, and a code comment, on why it must be 2? Where does that value come from, etc. -

Re: [FFmpeg-devel] [PATCH] lavc/hevc_parse: Don't take a HEVCContext

2016-04-25 Thread Derek Buitenhuis
On 4/25/2016 7:50 PM, John Warburton wrote: > Is it possible that this patch, particularly to libavcodec/hevc.h, is > causing my compilation error today, cross-compiling using mingw-w64 > and gcc-5.3.0 from GNU/Linux to Windows 64-bit? Yes it looks as if qsv also called that, and I missed it. I

[FFmpeg-devel] [PATCH] qsvenc_hevc: Fix usage of ff_hevc_extract_rbsp

2016-04-25 Thread Derek Buitenhuis
Signed-off-by: Derek Buitenhuis <derek.buitenh...@gmail.com> --- grep reveals no other usages. --- libavcodec/qsvenc_hevc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/libavcodec/qsvenc_hevc.c b/libavcodec/qsvenc_hevc.c index 30fde72..e0bead9 100644 --- a/liba

Re: [FFmpeg-devel] [PATCH] qsvenc_hevc: Fix usage of ff_hevc_extract_rbsp

2016-04-25 Thread Derek Buitenhuis
On 4/25/2016 11:37 PM, Michael Niedermayer wrote: > i cant easily test this but does this need a > av_freep of skipped_bytes_pos somewhere ? > > otherwise it should be ok I cannot easily test it either, unfortunately. I'm not sure why it would, though (without having looked very hard). - Derek

Re: [FFmpeg-devel] CONFIG_W64_DEMUXER and dead-code elimination

2016-04-25 Thread Derek Buitenhuis
On 4/25/2016 4:49 PM, Carl Eugen Hoyos wrote: > Can you confirm that this is Windows-only, ie that lto > works fine on Linux with the Intel compiler? > (It did work fine when I last tested.) FWIW, I've had issues with LTO on Linux, with clang and gold, due to (lack of) DCE. - Derek

Re: [FFmpeg-devel] [PATCH] configure: Don't require nonfree for nvenc

2016-04-27 Thread Derek Buitenhuis
On 4/27/2016 2:17 PM, Ricardo Constantino wrote: > configure already checks for minimum major version 6 which is the > first one using MIT license and fails if lower. Gotcha. - Derek ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org

Re: [FFmpeg-devel] [PATCH] Bug fixed: wrong ticks_per_frame. For H.264 stream ticks_per_frame should be 2

2016-04-27 Thread Derek Buitenhuis
On 4/27/2016 9:50 AM, Ivan Uskov wrote: > Since it was added to commit message too, it is not the issue, right? > I left the "bug fixed" because it was really the bug which force decoder to > produce doubled frames output. I don't think it is worth debating, really. I'll push this patch today

Re: [FFmpeg-devel] [PATCH] configure: Don't require nonfree for nvenc

2016-04-27 Thread Derek Buitenhuis
On 4/23/2016 6:00 PM, Timo Rothenpieler wrote: > As the nvEncodeApi.h header is now MIT licensed, this can be dropped. > The loaded CUDA and NVENC libraries are part of the nvidia driver, and > thus count as system libraries. > --- > configure | 1 - > 1 file changed, 1 deletion(-) I presume

Re: [FFmpeg-devel] [PATCH] vf_colorspace: add floyd-steinberg dithering option to full conversion.

2016-04-26 Thread Derek Buitenhuis
On 4/26/2016 5:41 PM, Ronald S. Bultje wrote: > --- > doc/filters.texi | 13 > libavfilter/colorspacedsp.c | 9 +++ > libavfilter/colorspacedsp.h | 6 ++ > libavfilter/colorspacedsp_template.c | 128 > +++ >

Re: [FFmpeg-devel] [PATCH] Bug fixed: wrong ticks_per_frame. For H.264 stream ticks_per_frame should be 2

2016-04-26 Thread Derek Buitenhuis
On 4/26/2016 4:45 PM, wm4 wrote: > I can see that this code is run only for h264, and I can see that you > set the field to 2. The added comment adds no new information and is > useless. The ticks_per_frame docu literally says "should be 2 for h.264" or something. I only thought it should be in

Re: [FFmpeg-devel] [PATCH] lavf/concatdec: remove unrelated change during codecpar merge.

2016-04-25 Thread Derek Buitenhuis
On 4/25/2016 1:09 PM, Nicolas George wrote: > Indeed, I had made a note to look into it, if only to silence the warnings, > and it seems the new API does the right thing. That will take a little time, > though, it would probably be better to have the fix immediately. I'm not opposed to it, as

Re: [FFmpeg-devel] Remove Derek Buitenhuis from MAINTAINERS

2016-05-19 Thread Derek Buitenhuis
On 5/19/2016 2:12 PM, Michael Niedermayer wrote: > if derek still wants to leave in 2 weeks then so be it, its his choice > but i really hope things can be resolved in a way that everyone > stays and works together and is happy I will wait 2 weeks. - Derek

Re: [FFmpeg-devel] [PATCH]lavc/libx265: Support gray encoding

2016-05-12 Thread Derek Buitenhuis
On 5/12/2016 10:43 AM, Carl Eugen Hoyos wrote: > I am also attaching the patch I just sent to x265-devel that allows to > test the feature. Would this not need a version requirement bump in configure, if the feature is new? - Derek ___ ffmpeg-devel

Re: [FFmpeg-devel] [PATCH]lavc/libx265: Support gray encoding

2016-05-12 Thread Derek Buitenhuis
On 5/12/2016 4:18 PM, Carl Eugen Hoyos wrote: > The feature is not new (and the necessary define in the header > was always there) but it did not work (at least for some > revisions). > How should I proceed? Hmm. I'm not really a fan of bumping up the required x265 version, since it has been

Re: [FFmpeg-devel] [PATCH]lavf/rawenc: Add a G.729 muxer

2016-05-18 Thread Derek Buitenhuis
On 5/18/2016 2:30 PM, Carl Eugen Hoyos wrote: > Since you indicated lately that you are so terribly interested > in the tone here: > Please stop making such comments! It's a legitimate criticism. You refuse to follow the FFmpeg developer rules of using git format-patch. - Derek

Re: [FFmpeg-devel] [PATCH]lavf/rawenc: Add a G.729 muxer

2016-05-18 Thread Derek Buitenhuis
On 5/18/2016 2:32 PM, Carl Eugen Hoyos wrote: > I should probably add that this is an interesting comment > coming from somebody who breaks FFmpeg with nearly every > commit and not only absolutely refuses to work on fixes but > even blocks fixes if others want to help him. Frankly, this is

Re: [FFmpeg-devel] [PATCH]lavf/rawenc: Add a G.729 muxer

2016-05-18 Thread Derek Buitenhuis
On 5/18/2016 1:10 PM, Paul B Mahol wrote: > Please attach actual patch with commit log message. I would agree, but for some reason Carl feels he is exempt from the FFmpeg developer rules. - Derek ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org

Re: [FFmpeg-devel] [FFmpeg-cvslog] Merge commit '2758cdedfb7ac61f8b5e4861f99218b6fd43491d'

2016-05-04 Thread Derek Buitenhuis
On 5/4/2016 11:58 PM, Thierry Foucu wrote: >> Can i ask why it cannot be part of the public API? See answer below. >> It was really practical just to register new prefix, specially, when you >> have to create X number of them. It's not harder to do via AVIO callbacks in actual code using the

Re: [FFmpeg-devel] [PATCH] lavd/avfoundation: use AVCodecParameters

2016-05-05 Thread Derek Buitenhuis
On 5/5/2016 2:08 PM, Rick Kern wrote: > Fixes "Could not find codec parameters for stream" error (#5494) > > Signed-off-by: Rick Kern > --- > libavdevice/avfoundation.m | 26 +- > 1 file changed, 13 insertions(+), 13 deletions(-) Patch looks good to

Re: [FFmpeg-devel] [PATCH] set exact ref frame in AVMotionVector

2016-05-06 Thread Derek Buitenhuis
On 5/6/2016 10:54 AM, Davinder Singh wrote: > change the confusing "source" var to "ref" of AVMotionVector. This would be a backwards incompatible API break, no? - Derek ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org

Re: [FFmpeg-devel] [PATCH] avcodec: Add AVClass to AVCodecParameters

2016-05-10 Thread Derek Buitenhuis
On 5/10/2016 12:43 PM, Michael Niedermayer wrote: > This patch adds an AVClass field to AVCodecParameters like we use in > other public structures. > i fail to see how this patch is related to your reply I apologize. I thought it was the start of an attempt to address the "issue" from the other

Re: [FFmpeg-devel] [PATCH] avcodec: Add AVClass to AVCodecParameters

2016-05-10 Thread Derek Buitenhuis
On 5/10/2016 2:53 PM, wm4 wrote: > Indeed, av_opt calls on structs which don't support it compile without > warning, but mysteriously crash at runtime. That's a violation of this > principle. And no, it doesn't mean that every struct should have an > AVClass, because that would be insane. Is

Re: [FFmpeg-devel] [FFmpeg-cvslog] Merge commit '2758cdedfb7ac61f8b5e4861f99218b6fd43491d'

2016-05-04 Thread Derek Buitenhuis
On 5/4/2016 9:08 PM, Thierry Foucu wrote: > I noticed that this commit remove the function ffurl_register_protocol > which allows people to register their own protocol when using libavformat. This is intended. It was never, ever a public API on purpose, and it was never supported or allowed. >

Re: [FFmpeg-devel] [PATCH] avcodec: Add AVClass to AVCodecParameters

2016-05-10 Thread Derek Buitenhuis
On 5/10/2016 7:24 AM, Hendrik Leppkes wrote: > I don't like this, the struct is pretty cleanly defined and unlikely > to be extended much over time. > Most other structs have AVOptions so the CLI can interact with it, but > this struct is not meant to be modified by users, its just a direct > line

Re: [FFmpeg-devel] [PATCH] libnut: Pass stream to ff_parse_specific_params

2016-04-21 Thread Derek Buitenhuis
On 4/21/2016 1:52 PM, Derek Buitenhuis wrote: > Signed-off-by: Derek Buitenhuis <der...@vimeo.com> This should be my gmail. Fixed locally. ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] [PATCH] Add support for spherical uuid in mov

2016-04-21 Thread Derek Buitenhuis
On 4/21/2016 9:13 AM, Hendrik Leppkes wrote: > One developer from Libav is working on defining side-data for this > spherical data. If you want to push this forward I suggest you join > those efforts. > ie. this patch for mov: https://patches.libav.org/patch/60300/ (part > of a longer series)

[FFmpeg-devel] [PATCH] libnut: Pass stream to ff_parse_specific_params

2016-04-21 Thread Derek Buitenhuis
This function bo longer takes an AVCodecContext. Fixes ticket #5430. Signed-off-by: Derek Buitenhuis <der...@vimeo.com> --- Since I'm sending this patch anyway, I had a few questions: 1. Where the heck is libnut's official repo. Google only turns up random GitHub forks, and

Re: [FFmpeg-devel] [PATCH] Implement hdcd filtering

2016-04-15 Thread Derek Buitenhuis
On 3/22/2016 11:22 AM, Benjamin St wrote: > This patch applies filtering/decoding for hdcds(see ticket #4441) . The > filter is heavily based on > https://github.com/kode54/foo_hdcd/. (Is this ok? Copyright?) > > Discuss, Review Hi, Sorry to be late to the party (I just saw this pushed.) Why

[FFmpeg-devel] Hostile Attitudes

2016-05-08 Thread Derek Buitenhuis
Hi, I've been doing merges for a while now, and I am happy to continue doing them too, but, I grow quite tired of certain individuals doing nothing but, for lack of a better term, defecate on my efforts. It's really quite demotivating and immediately sours my day. I am not going to finger point,

Re: [FFmpeg-devel] [PATCH] fate: Remove duplicate wmv8_x8intra.wmv test

2016-05-08 Thread Derek Buitenhuis
On 5/8/2016 5:49 PM, Michael Niedermayer wrote: > --disable-optimizations "fixes" the test on mips So is this a compiler bug? - Derek ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] [PATCH] fate: Remove duplicate wmv8_x8intra.wmv test

2016-05-07 Thread Derek Buitenhuis
On 5/8/2016 12:42 AM, Michael Niedermayer wrote: > Also temporary enable the test so we get updated fate failure statistics > > Note, this does not work on all platforms, it fails on MIPS > and ml archives indicate it failed on x86 openbsd with some compilers as well > > Signed-off-by: Michael

Re: [FFmpeg-devel] [PATCH] avcodec: add properties for lossless to AVCodecParameters

2016-05-09 Thread Derek Buitenhuis
On 5/9/2016 1:29 PM, Michael Niedermayer wrote: > Also theres a deeper problem, > User want, benefit from and sometimes need, "every single piece of > information a decoder might output.", well not "every" of course > but its hard to draw lines what may be needed and what not. They can decode a

Re: [FFmpeg-devel] [PATCH] avcodec: add properties for lossless to AVCodecParameters

2016-05-09 Thread Derek Buitenhuis
On 5/9/2016 8:36 AM, wm4 wrote: > I'd argue that AVCodecParameters should contain only container > properties, or we'd end up with duplicating most of AVCodecContext in > it. Agreed. IMO the correct solution for user applications is to decode a frame. It's not reasonable to expect a *demuxer* to

[FFmpeg-devel] [PATCH] mjpegdec: Properly fail on malloc failure

2016-05-09 Thread Derek Buitenhuis
Signed-off-by: Derek Buitenhuis <derek.buitenh...@gmail.com> --- libavcodec/mjpegdec.c | 56 +++ 1 file changed, 30 insertions(+), 26 deletions(-) diff --git a/libavcodec/mjpegdec.c b/libavcodec/mjpegdec.c index 7d38fc2..32e8cee

Re: [FFmpeg-devel] [PATCH] avcodec: add properties for lossless to AVCodecParameters

2016-05-09 Thread Derek Buitenhuis
On 5/9/2016 1:48 PM, Carl Eugen Hoyos wrote: > This should get applied asap imo. NAK! This patch has at least 3 people dissenting and it doesn't actually "fix" anything except cosmetic output on the console, which is NOT meant to be parsed not even stable. > Arguing that the frame should be

Re: [FFmpeg-devel] [PATCH] avcodec: add properties for lossless to AVCodecParameters

2016-05-09 Thread Derek Buitenhuis
On 5/9/2016 1:52 PM, Carl Eugen Hoyos wrote: > So can you answer why the codec aspect ratio is provided but > other information is not? Most because e.g. some *conatiners* need this info. No containers, to my knowledge, need this 'lossless' info. > Do you agree that the answer is that decoders

Re: [FFmpeg-devel] [PATCH] avcodec: add properties for lossless to AVCodecParameters

2016-05-09 Thread Derek Buitenhuis
On 5/9/2016 2:19 PM, Paul B Mahol wrote: > Isn't st->codec deprecated? It is, but the old behavior should be maintained for the duration of its deprecation period. It's a bug not to add it back to st->codec in my opinion. - Derek ___ ffmpeg-devel

Re: [FFmpeg-devel] [PATCH] avcodec: add properties for lossless to AVCodecParameters

2016-05-09 Thread Derek Buitenhuis
On 5/9/2016 2:22 PM, Paul B Mahol wrote: > Once st->codec is gone, how would this lossless info be gathered back? As myself and others have said above: decode a frame. It is not a demuxer's job to provide info gleaned from decoding. Either avformat should be properly separate as it (kind of) is

Re: [FFmpeg-devel] [PATCH] avcodec: add properties for lossless to AVCodecParameters

2016-05-09 Thread Derek Buitenhuis
On 5/9/2016 2:28 PM, Hendrik Leppkes wrote: > And before people argue that avformat does this anyway today - one of > the hopes is to make it stop doing that for many "simple" codecs where > this is just not necessary, and say a parser could extract all the > important information with much less

[FFmpeg-devel] [PATCH] avcodec: Mark some codecs with threadsafe init as such

2017-02-07 Thread Derek Buitenhuis
Signed-off-by: Derek Buitenhuis <derek.buitenh...@gmail.com> --- I wrote this in Nov 2015, and it has resided on my GitHub fork in a branch since then, since I never ot around to finishing it. I noticed a spike in interest in the 'feature', so I rebase and am forwarding this, i

Re: [FFmpeg-devel] [PATCH] mov: Remove ancient heuristic hack

2016-08-25 Thread Derek Buitenhuis
On 8/24/2016 10:54 PM, Michael Niedermayer wrote: > IIRC the removed code tried to detect a reorder delay that is not > possible in a valid file due to the profile constraints. Aka dts and > pts are too far appart for the largest amount of buffers allowed in > any codec. Basing this on timestamps

Re: [FFmpeg-devel] [PATCH] mov: Remove ancient heuristic hack

2016-08-25 Thread Derek Buitenhuis
On 8/25/2016 3:40 PM, Michael Niedermayer wrote: > but its probably best to remove in a seperate patch so if it breaks > something bisect would immedeatly point to which of the 2 changes > caused it Sounds good. If you think this patch is OK, please push it with this part of the commit message

[FFmpeg-devel] [PATCH] vf_fps: Don't flush a cached frame if it should have been dropped

2016-10-04 Thread Derek Buitenhuis
This fixes downconverting framerates to multiples. For example, prior to this patch, converting 900 frames at 60 fps to 30 fps would output 451 frames instead of the correct 450. Signed-off-by: Derek Buitenhuis <derek.buitenh...@gmail.com> --- DISCLAIMER: I don't know libavfilter ver

Re: [FFmpeg-devel] [PATCH] vf_fps: Don't flush a cached frame if it should have been dropped

2016-10-04 Thread Derek Buitenhuis
On 10/4/2016 10:21 PM, Michael Niedermayer wrote: >> This "break" may actually be more correct output, but I am not familiar >> enough >> with lavfi or vf_fps to say. Removing the last frame in case where it should >> have >> been removed was the entire point of this patch. > > the change to

Re: [FFmpeg-devel] [PATCH] vf_fps: Don't flush a cached frame if it should have been dropped

2016-10-04 Thread Derek Buitenhuis
On 10/4/2016 9:58 PM, Michael Niedermayer wrote: > breaks fate > and the change to fate looks wrong, the last frame is lost > make fate-filter-fps > TESTfilter-fps > --- ./tests/ref/fate/filter-fps 2016-10-04 14:46:19.642736770 +0200 > +++ tests/data/fate/filter-fps 2016-10-04

[FFmpeg-devel] [PATCH 1/2 v2] mov: Remove ancient heuristic hack

2016-08-25 Thread Derek Buitenhuis
This breaks files with legitimate single-entry edit lists, and the hack, introduced in f03a081df09f9c4798a17d7e24446ed47924b11b, has no link to any known sample in its commit message. Signed-off-by: Derek Buitenhuis <derek.buitenh...@gmail.com> --- libavformat/isom.h | 1 - libavformat

Re: [FFmpeg-devel] [PATCH] mov: Remove ancient heuristic hack

2016-08-25 Thread Derek Buitenhuis
On 8/25/2016 4:52 PM, Michael Niedermayer wrote: >> the patch removes all uses of wrong_dts, the field should be >> > removed too > oops i forgot cc-ing you, iam not used to reply-all on the ML OK. I thought it was used in the FLV demuxer too, but it seems it has it's own copy inside the FLV

Re: [FFmpeg-devel] [PATCH] mov: Remove ancient heuristic hack

2016-08-25 Thread Derek Buitenhuis
On 8/25/2016 5:34 PM, James Almer wrote: > On 8/25/2016 12:45 PM, Carl Eugen Hoyos wrote: >> 2016-08-24 16:55 GMT+02:00 Derek Buitenhuis <derek.buitenh...@gmail.com>: >>> This breaks files with legitimate single-entry edit lists, >>

Re: [FFmpeg-devel] [PATCH 1/2 v2] mov: Remove ancient heuristic hack

2016-08-25 Thread Derek Buitenhuis
On 8/25/2016 5:18 PM, Michael Niedermayer wrote: > ok if i push with the changes requested in: > https://lists.ffmpeg.org/pipermail/ffmpeg-devel/2016-August/198303.html > ? (seems another reply with forgotten CC) You can change 'hack' to 'incorrect workaround' if you wish. However, removing the

Re: [FFmpeg-devel] [PATCH 1/2 v2] mov: Remove ancient heuristic hack

2016-08-25 Thread Derek Buitenhuis
On 8/25/2016 5:49 PM, Derek Buitenhuis wrote: > However, removing the part about the sample being missing in > the commit message would be wrong, because it is, in fact, not > linked in the commit message. I must step out now, so please push whenever some consensus is reached.

[FFmpeg-devel] [PATCH 2/2 v2] mov: Remove old b-frame/video delay heuristic

2016-08-25 Thread Derek Buitenhuis
This was added before edts support existed, and is no longer valid. Signed-off-by: Derek Buitenhuis <derek.buitenh...@gmail.com> --- libavformat/mov.c | 3 --- 1 file changed, 3 deletions(-) diff --git a/libavformat/mov.c b/libavformat/mov.c index 54c63ad..bc11fb8 100644 --- a/libavformat

[FFmpeg-devel] [PATCH] mov: Remove ancient heuristic hack

2016-08-24 Thread Derek Buitenhuis
quot; them at the expense of breaking legitimate files. Signed-off-by: Derek Buitenhuis <derek.buitenh...@gmail.com> --- Let's get this out of the way first: I'm not 'returning to FFmpeg'. I do not intend to be involved in the community again, nor its overall direction and discussions, short of a m

Re: [FFmpeg-devel] [PATCH] mov: Remove ancient heuristic hack

2016-08-24 Thread Derek Buitenhuis
On 8/24/2016 6:50 PM, Carl Eugen Hoyos wrote: >> This breaks files with legitimate single-entry edit lists, >> and the hack, introduced in f03a081df09f9c4798a17d7e24446ed47924b11b, >> has no link to any known sample in its commit message, nor > > I suspect the commit message links to this sample:

Re: [FFmpeg-devel] [PATCH] mov: Remove ancient heuristic hack

2016-08-24 Thread Derek Buitenhuis
On 8/24/2016 7:00 PM, Carl Eugen Hoyos wrote: > Does this patch also fix the issue? > http://ffmpeg.org/pipermail/ffmpeg-devel/2016-August/197817.html It doesn't. The patch I sent fixes a problem with a current workaround in the mov demuxer. Edit lists are just one way it could have triggered it.

Re: [FFmpeg-devel] [PATCH] lavf/mov.c: Fallback to finding non-keyframe in fix_index, if keyframe search fails.

2016-11-06 Thread Derek Buitenhuis
On 11/6/2016 12:40 PM, Michael Niedermayer wrote: > IMO we should support it if its possible without large hacks I don't really have an opinion myself, so to speak. I was just noting it. - Derek ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org

[FFmpeg-devel] [PATCH] doc/libx26[45]: Add documentation for forced-idr

2016-11-06 Thread Derek Buitenhuis
Signed-off-by: Derek Buitenhuis <derek.buitenh...@gmail.com> --- As request by Michael. --- doc/encoders.texi | 8 1 file changed, 8 insertions(+) diff --git a/doc/encoders.texi b/doc/encoders.texi index 5a60e7e..ba60e0a 100644 --- a/doc/encoders.texi +++ b/doc/encoders.texi @@ -

Re: [FFmpeg-devel] [PATCH] lavf/mov.c: Fallback to finding non-keyframe in fix_index, if keyframe search fails.

2016-11-05 Thread Derek Buitenhuis
On 11/4/2016 9:53 PM, Sasi Inguva wrote: > So, if I am not wrong, the only case when that would happen is, if edit > list media time was < 0 . And that would be a corrupt MOV file, unless edit > list media time = -1. > I tried to create such a file with AtomInspector but it doesn't let me. FWIW,

Re: [FFmpeg-devel] [PATCH] avcodec/libx264: fix forced_idr logic

2016-11-11 Thread Derek Buitenhuis
On 11/11/2016 2:30 PM, Michael Niedermayer wrote: > if noone adds suport fo the 3rd constant then the patch should > likely be applied, unless i miss something This patch LGTM. When I added this option in c981b1145a857c8f962c93b8eecb1c613b20ffe9, the type was INT, and the default was -1, and it

Re: [FFmpeg-devel] [PATCH] avcodec/libx264: fix forced_idr logic

2016-11-11 Thread Derek Buitenhuis
On 11/11/2016 3:26 PM, Derek Buitenhuis wrote: > This patch LGTM. > > When I added this option in c981b1145a857c8f962c93b8eecb1c613b20ffe9, the > type was > INT, and the default was -1, and it was correct. It was later broken in > fb99ef0bd39a1859d0e65c6c16caa8e17dd3cfbe.

Re: [FFmpeg-devel] [PATCH] lavf/mov.c: Fallback to finding non-keyframe in fix_index, if keyframe search fails.

2016-11-10 Thread Derek Buitenhuis
On 11/9/2016 6:51 PM, Sasi Inguva wrote: > Signed-off-by: Sasi Inguva > --- > libavformat/mov.c | 41 > tests/fate/mov.mak | 6 +- > tests/ref/fate/mp4-init-nonkeyframe | 120 > > 3 files

Re: [FFmpeg-devel] [PATCH] lavc/utils.c: Make sure skip_samples never goes negative.

2016-11-10 Thread Derek Buitenhuis
On 11/4/2016 10:31 PM, Sasi Inguva wrote: > -avctx->internal->skip_samples -= frame->nb_samples; > +avctx->internal->skip_samples = FFMAX(0, > avctx->internal->skip_samples - frame->nb_samples); LGTM. - Derek ___ ffmpeg-devel

Re: [FFmpeg-devel] [PATCH] lavf/mov: Add support for edit list parsing.

2016-10-21 Thread Derek Buitenhuis
> In YouTube we have long been receiving MOV files from users, which have > non-trivial edit lists (Those edit lists which are not just used to offset > video start from audio start) and multiple edit lists. Recently the uploads > of such files has increased with the introduction of apps that

Re: [FFmpeg-devel] [PATCH] lavf/mov: Add support for edit list parsing.

2016-10-21 Thread Derek Buitenhuis
On 10/21/2016 4:32 PM, Derek Buitenhuis wrote: > * Audio packets. Especially audio packets with a large number of samples. > It's extremely likely that edits will not fall on packet boundaries, and > depending on the number of samples per packet, audio sync i

Re: [FFmpeg-devel] [PATCH] lavf/mov: Add support for edit list parsing.

2016-10-21 Thread Derek Buitenhuis
On 10/21/2016 6:47 PM, Sasi Inguva wrote: > * Audio packets. Especially audio packets with a large number of > samples. > It's extremely likely that edits will not fall on packet > boundaries, and > depending on the number of samples per packet, audio sync issues >

Re: [FFmpeg-devel] [PATCH] lavf/mov.c: Set skip_samples according to first edit list, when -ignore_editlist is set.

2016-11-14 Thread Derek Buitenhuis
On 11/12/2016 6:02 AM, Sasi Inguva wrote: > +/* Adjust skip_samples correctly when ignore_editlist is set, to support > gapless playback */ > +if (mov->ignore_editlist && sc->num_non_empty_elst <= 1 && > +st->codecpar->codec_id == AV_CODEC_ID_AAC && > sc->first_elist_start_time >

Re: [FFmpeg-devel] [PATCH] avcodec/libx264: fix forced_idr logic

2016-11-14 Thread Derek Buitenhuis
On 11/11/2016 4:01 PM, Timo Rothenpieler wrote: > -1 and 0 both mean false now, and I left in the option to pass -1 to > stay compatible with possible 3rd parties who pass it. Well that's certainly non-obvious... > So changing to default to 0 doesn't really matter, and I decided to keep > the

Re: [FFmpeg-devel] [PATCH] lavf/mov.c: Use the correct timescale when seeking for audio.

2016-11-01 Thread Derek Buitenhuis
On 10/31/2016 1:48 PM, Rostislav Pehlivanov wrote: > The tests pass on my machine, pushed after Derek said it looks correct. Hi Sasi, I notice there's still one difference after this patch. Could you explain if this is intended: -"duration": 3072, -

Re: [FFmpeg-devel] [PATCH] lavf/mov.c: Use the correct timescale when seeking for audio.

2016-11-01 Thread Derek Buitenhuis
On 11/1/2016 5:01 PM, Sasi Inguva wrote: > When we add up (last pts + 0.023 ) it amounts to that . We fix the last > packet duration according to edit list segment duration in our edit list > patch. Here last pts = 1324.652 . I figured this was the case. IIRC, the OGG code works like this

Re: [FFmpeg-devel] [PATCH] lavf/mov.c: Use the correct timescale when seeking for audio.

2016-10-31 Thread Derek Buitenhuis
On 10/27/2016 10:48 PM, Sasi Inguva wrote: > gentle ping. > > Thanks! A ping from me, too. - Derek ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] [PATCH] lavf/mov.c: Use the first sidx for tracks without sidx.

2016-11-03 Thread Derek Buitenhuis
On 11/3/2016 4:56 PM, Sasi Inguva wrote: >> I am just converting the *reference* stream duration from that streams' > timescale to the timescale of the stream I am setting to. > > Indeed, I misread this (midnight emails FTL; I read time_scale 3 times instead of 2). Anyway, the rest seems OK

Re: [FFmpeg-devel] [PATCH] lavf/mov.c: Use the first sidx for tracks without sidx.

2016-11-02 Thread Derek Buitenhuis
On 10/26/2016 7:31 PM, Sasi Inguva wrote: > According to spec ISO_IEC_15444_12 "For any media stream for which no segment > index is present, referred to as non‐indexed stream, the media stream > associated with the first Segment Index box in the segment serves as a > reference stream in a

Re: [FFmpeg-devel] [PATCH] lavf/mov: Add support for edit list parsing.

2016-11-02 Thread Derek Buitenhuis
On 10/24/2016 6:43 AM, Sasi Inguva wrote: > Just sent a patch, correcting a bug in the edit list code. PTAL. Hi again, it's me. Back with more files this patch set broke :). This time it's very easy to reproduce: $ ffmpeg -i sample.mp4 -f null - This works before your patchset, and also if

Re: [FFmpeg-devel] [PATCH] lavf/mov.c: Fallback to finding non-keyframe in fix_index, if keyframe search fails.

2016-11-04 Thread Derek Buitenhuis
On 11/3/2016 10:36 PM, Sasi Inguva wrote: > +if (index == -1) { > +av_log(mov->fc, AV_LOG_ERROR, > + "st: %d edit list %"PRId64" Cannot find an index > entry before timestamp: %"PRId64"\n", > + st->index, edit_list_index,

[FFmpeg-devel] [PATCH] libx265: Add option to force IDR frames

2016-11-02 Thread Derek Buitenhuis
This is in the same the same vein as c981b1145a857c8f962c93b8eecb1c613b20ffe9. Signed-off-by: Derek Buitenhuis <derek.buitenh...@gmail.com> --- libavcodec/libx265.c | 5 - libavcodec/version.h | 2 +- 2 files changed, 5 insertions(+), 2 deletions(-) diff --git a/libavcodec/libx2

[FFmpeg-devel] [PATCH] h264_slice: Wait for refs to be available before we use them in error concealment

2016-12-07 Thread Derek Buitenhuis
This could happen when there was a frame number gap and frame threading was used. This fixes #5458. Debugging-by: Ronald S. Bultje <rsbul...@gmail.com> Debugging-by: Justin Ruggles <justin.rugg...@gmail.com> Signed-off-by: Derek Buitenhuis <derek.buitenh...@gmail.com> --

Re: [FFmpeg-devel] [PATCH] h264_slice: Wait for refs to be available before we use them in error concealment

2016-12-07 Thread Derek Buitenhuis
On 12/7/2016 6:43 PM, Carl Eugen Hoyos wrote: > 2016-12-07 19:06 GMT+01:00 Derek Buitenhuis <derek.buitenh...@gmail.com>: >> This could happen when there was a frame number >> gap and frame threading was used. >> >> This fixes #5458. > > Does this patch ha

[FFmpeg-devel] [PATCH v3] h264_slice: Wait for refs to be available before we use them in error concealment

2016-12-08 Thread Derek Buitenhuis
This could happen when there was a frame number gap and frame threading was used. This fixes #5458. Debugging-by: Ronald S. Bultje <rsbul...@gmail.com> Debugging-by: Justin Ruggles <justin.rugg...@gmail.com> Signed-off-by: Derek Buitenhuis <derek.buitenh...@gmail.com> --- Extr

[FFmpeg-devel] [PATCH 0/2 v2] h264_slice race condition fixes

2016-12-08 Thread Derek Buitenhuis
!=). The 2nd patch in the series is the same as before. Derek Buitenhuis (2): h264_slice: Add forgotten progress update for the 2nd field in a slice h264_slice: Wait for refs to be available before we use them in error concealment libavcodec/h264_slice.c | 5 + 1 file changed, 5 insertions

[FFmpeg-devel] [PATCH 2/2] h264_slice: Wait for refs to be available before we use them in error concealment

2016-12-08 Thread Derek Buitenhuis
This could happen when there was a frame number gap and frame threading was used. This fixes #5458. Debugging-by: Ronald S. Bultje <rsbul...@gmail.com> Debugging-by: Justin Ruggles <justin.rugg...@gmail.com> Signed-off-by: Derek Buitenhuis <derek.buitenh...@gmail.com&g

[FFmpeg-devel] [PATCH 1/2] h264_slice: Add forgotten progress update for the 2nd field in a slice

2016-12-08 Thread Derek Buitenhuis
Signed-off-by: Derek Buitenhuis <derek.buitenh...@gmail.com> --- libavcodec/h264_slice.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/libavcodec/h264_slice.c b/libavcodec/h264_slice.c index 1f2c06521e..a8a8731138 100644 --- a/libavcodec/h264_slice.c +++ b/libavcodec/h264_s

Re: [FFmpeg-devel] [PATCH 1/2] h264_slice: Add forgotten progress update for the 2nd field in a slice

2016-12-08 Thread Derek Buitenhuis
On 12/8/2016 3:46 PM, Carl Eugen Hoyos wrote: > Does this have a speed impact for valid h264 streams? Nope. - Derek ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] [PATCH] h264_slice: Wait for refs to be available before we use them in error concealment

2016-12-07 Thread Derek Buitenhuis
On 12/7/2016 8:10 PM, Michael Niedermayer wrote: > this deadlocks on some samples > ill try to find one that i can share I have a vague idea maybe why, but I'd need a sample, yes. - Derek ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org

[FFmpeg-devel] [PATCH] fate: Add h264 test for frame num gaps

2016-12-09 Thread Derek Buitenhuis
Signed-off-by: Derek Buitenhuis <derek.buitenh...@gmail.com> --- Sample is here: https://trac.ffmpeg.org/attachment/ticket/5458/nondeterministic_cut.h264 --- tests/fate/h264.mak | 2 ++ tests/ref/fate/h264-missing-frame | 35 +++ 2 files chang

Re: [FFmpeg-devel] [PATCH v3] h264_slice: Wait for refs to be available before we use them in error concealment

2016-12-09 Thread Derek Buitenhuis
On 12/9/2016 2:29 PM, Michael Niedermayer wrote: > this looks reasonable Pushed. - Derek ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

[FFmpeg-devel] [PATCH] libavcodec/tests: Add avpacket test to .gitignore

2016-12-09 Thread Derek Buitenhuis
Signed-off-by: Derek Buitenhuis <derek.buitenh...@gmail.com> --- Happened to notice this when adding a FATE test. --- libavcodec/tests/.gitignore | 1 + 1 file changed, 1 insertion(+) diff --git a/libavcodec/tests/.gitignore b/libavcodec/tests/.gitignore index d8ab947abe..0ab029696d

[FFmpeg-devel] [PATCH] hevc: Mark as having threadsafe init

2017-01-10 Thread Derek Buitenhuis
Signed-off-by: Derek Buitenhuis <derek.buitenh...@gmail.com> --- It looks like it already is threadsafe, amazingly? I don't see any table inits, somehow. --- libavcodec/hevc.c | 1 + 1 file changed, 1 insertion(+) diff --git a/libavcodec/hevc.c b/libavcodec/hevc.c index 7c563a3..607a8da

Re: [FFmpeg-devel] libavcodec memory usage

2017-01-03 Thread Derek Buitenhuis
On 1/3/2017 9:57 PM, Ronald S. Bultje wrote: > Hm, right... OK, yeah, I understand the issue, but I somehow doubt there's > an easy way to resolve that that isn't disgustingly hacky on our side. If you're not very fond of happiness or joy, you could run a separate 64-bit process to decode and

[FFmpeg-devel] [PATCH] webm_dash_manifest_demuxer: Fix initialization range for files with cues at the front

2017-04-21 Thread Derek Buitenhuis
The WebM DASH spec states: The Initialization Segment shall not contain Clusters or Cues. The Segment Index corresponds to the Cues. Previously, it included the cues if they were at the front. Signed-off-by: Derek Buitenhuis <derek.buitenh...@gmail.com> --- libavformat/matrosk

[FFmpeg-devel] [PATCH 2/2] webmdashenc: Validate the 'streams' adaptation sets parameter

2017-04-20 Thread Derek Buitenhuis
It should not be a value larger than the number of streams we have, or it will cause invalid reads and/or SIGSEGV. Signed-off-by: Derek Buitenhuis <derek.buitenh...@gmail.com> --- libavformat/webmdashenc.c | 7 ++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/libav

[FFmpeg-devel] [PATCH 0/2] webmdashenc crash fixes

2017-04-20 Thread Derek Buitenhuis
This whole file is sloppy... I found these on my very first mux with it, ever, by accident. Derek Buitenhuis (2): webmdashenc: Require the 'adaptation_sets' option to be ste webmdashenc: Validate the 'streams' adaptation sets parameter libavformat/webmdashenc.c | 11 ++- 1 file

[FFmpeg-devel] [PATCH 1/2] webmdashenc: Require the 'adaptation_sets' option to be ste

2017-04-20 Thread Derek Buitenhuis
This seems to be non-optional, and if the muxer is run without it, strlen() is run on NULL, causing a segfault. Signed-off-by: Derek Buitenhuis <derek.buitenh...@gmail.com> --- libavformat/webmdashenc.c | 4 1 file changed, 4 insertions(+) diff --git a/libavformat/webmdashe

Re: [FFmpeg-devel] [PATCH] webm_dash_manifest_demuxer: Fix UB in cue timestamp string code and make it actually work

2017-04-20 Thread Derek Buitenhuis
On 4/20/2017 3:58 PM, Michael Niedermayer wrote: > On Thu, Apr 20, 2017 at 03:02:54PM +0100, Derek Buitenhuis wrote: >> The original author apparently never read the documentation for snprintf, >> or even tested that the output was correct. > This statement does not belong i

Re: [FFmpeg-devel] [PATCH 2/2] webmdashenc: Validate the 'streams' adaptation sets parameter

2017-04-20 Thread Derek Buitenhuis
On 4/20/2017 3:47 PM, Michael Niedermayer wrote: > breaks fate v2 sent. - Derek ___ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

[FFmpeg-devel] [PATCH 2/2 v2] webmdashenc: Validate the 'streams' adaptation sets parameter

2017-04-20 Thread Derek Buitenhuis
It should not be a value larger than the number of streams we have, or it will cause invalid reads and/or SIGSEGV. Signed-off-by: Derek Buitenhuis <derek.buitenh...@gmail.com> --- libavformat/webmdashenc.c | 6 +- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/libav

[FFmpeg-devel] [PATCH] webm_dash_manifest_demuxer: Fix UB in cue timestamp string code and make it actually work

2017-04-20 Thread Derek Buitenhuis
-by: Derek Buitenhuis <derek.buitenh...@gmail.com> --- libavformat/matroskadec.c | 14 +++--- 1 file changed, 11 insertions(+), 3 deletions(-) diff --git a/libavformat/matroskadec.c b/libavformat/matroskadec.c index 9adca8d..320d8bf 100644 --- a/libavformat/matroskadec.c +++ b/libav

Re: [FFmpeg-devel] [PATCH 1/2] webmdashenc: Require the 'adaptation_sets' option to be ste

2017-04-20 Thread Derek Buitenhuis
On 4/20/2017 4:58 PM, James Almer wrote: > EINVAL IMO, since it's an user option. AVERROR_INVALIDDATA was defined as AVERROR(EINVAL) for years, I guess I forgot it changed to something else... git blame puts it at 72af5d8a020e4dca1118b3ede67be983b33e27c5 in 2010... My memory was seriously

Re: [FFmpeg-devel] [PATCH] webm_dash_manifest_demuxer: Fix initialization range for files with cues at the front

2017-04-23 Thread Derek Buitenhuis
On 4/21/2017 4:40 PM, Derek Buitenhuis wrote: > The WebM DASH spec states: > The Initialization Segment shall not contain Clusters or Cues. > The Segment Index corresponds to the Cues. > > Previously, it included the cues if they were at the front. > > Signed-off

<    1   2   3   4   5   6   7   8   9   >