[FFmpeg-devel] [PATCH] libavcodec/libopenjpegenc: don't define OPJ_STATIC with openjpeg 2.1.1

[Kia]

This fixes bug #5694 and builds with libopenjpeg 2.1.1.

openjpeg commit 3ed5858902055d3500a6ab183f1395686921d026 hides all
symbols with __attribute__ ((visibility ("hidden"))) if OPJ_STATIC is
defined

Signed-off-by: Kia 
---
 libavcodec/libopenjpegdec.c | 1 +
 libavcodec/libopenjpegenc.c | 1 +
 2 files changed, 2 insertions(+)

diff --git a/libavcodec/libopenjpegdec.c b/libavcodec/libopenjpegdec.c
index 65167e6..efc0648 100644
--- a/libavcodec/libopenjpegdec.c
+++ b/libavcodec/libopenjpegdec.c
@@ -37,6 +37,7 @@
 #include "thread.h"
 
 #if HAVE_OPENJPEG_2_1_OPENJPEG_H
+#undef  OPJ_STATIC
 #  include 
 #elif HAVE_OPENJPEG_2_0_OPENJPEG_H
 #  include 
diff --git a/libavcodec/libopenjpegenc.c b/libavcodec/libopenjpegenc.c
index 1443551..2bc75e2 100644
--- a/libavcodec/libopenjpegenc.c
+++ b/libavcodec/libopenjpegenc.c
@@ -35,6 +35,7 @@
 #include "internal.h"
 
 #if HAVE_OPENJPEG_2_1_OPENJPEG_H
+#undef  OPJ_STATIC
 #  include 
 #elif HAVE_OPENJPEG_2_0_OPENJPEG_H
 #  include 
-- 
2.9.0

___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] [GSoC] Motion Interpolation

[Davinder Singh]

https://github.com/dsmudhar/FFmpeg/commits/dev

The Paper 2 algorithm is complete. It seems good. If I compare Paper 2
(which uses bilateral motion estimation) v/s motion vectors exported by
mEstimate filter:

$ tiny_psnr 60_source_2.yuv 60_mest-esa+obmc.yuv
stddev:1.43 PSNR: 45.02 MAXDIFF:  174 bytes:476928000/474163200

$ tiny_psnr 60_source_2.yuv 60_paper2_aobmc+cls.yuv
stddev:1.25 PSNR: 46.18 MAXDIFF:  187 bytes:476928000/474163200

Frame comparison: http://www.mediafire.com/?qe7sc4o0s4hgug5

Compared to simple OBMC which over-smooth edges, Objects clustering and
Adaptive OBMC makes the edges crisp but also introduce blocking artifacts
where MVs are bad (with default search window = 7). But I think it’s ESA’s
fault. The paper doesn’t specify which motion estimation method they used;
I have been using ESA. I think quality can be further improved with EPZS,
which I'm going to implement.

I also tried to tweak VS-BMC (Variable size block motion compensation)
which reduced the blocking artifacts in VS-BMC area. Had to do experiments
a lot, more to be done.

mEstimate filter (ESA) + Simple OBMC:
http://www.mediafire.com/?3b8j1zj1lsuw979
Paper 2 (full): http://www.mediafire.com/?npbw1iv6tmxwvyu


Regards,
DSM_
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] [PATCH] avformat/fivdec: cached keyframes before video or audio stream was created

[Steven Liu]

2016-07-25 10:29 GMT+08:00 XinZheng Zhang :

> Ping.
>
> On Thu, Jul 21, 2016 at 11:59 AM, XinZheng Zhang 
> wrote:
> > On Thu, Jul 21, 2016 at 5:03 AM, Moritz Barsnick 
> wrote:
> >> On Thu, Jul 21, 2016 at 00:06:44 +0800, zhangxinzheng wrote:
> >>
> >>> Subject: [FFmpeg-devel] [PATCH] avformat/fivdec: cached keyframes
> before video or audio stream was created
> >>^^^ flv
> >>
> >>> There has an error when seeking in a flv file, which key frames was
> >>> sorted before video frame.
> >>> This ensures that all the key frames was cached, and add to
> >>> corresponding stream when it was created.
> >>
> >> Thanks to the grammar, I don't understand the intent of the patch.
> >> (Don't get me wrong, I'm not scorning your English, but it needs to be
> >> improved.)
> >>
> >
> > In flvdec.c 420
> > case AMF_DATA_TYPE_OBJECT:
> > if ((vstream || astream) && key &&
> > ioc->seekable &&
> > !strcmp(KEYFRAMES_TAG, key) && depth == 1)
> > if (parse_keyframes_index(s, ioc, vstream ? vstream : astream,
> >   max_pos) < 0)
> >
> > If the first packet we got is object, both vstream and astream are NULL,
> > that may prevent keyframe index being parsed.
> >
> > This patch saves keyframe index before A/V streams are created.
> >
> > The issue was introduced since 09ae7b81ea.
> >
> > I have created a ticket #5729 with a sample attatched.
> >
> >>>  #define RESYNC_BUFFER_SIZE (1<<20)
> >>>
> >>> +
> >>> +typedef struct FLVKeyFrame {
> >>
> >> Why the extra newline?
> >>
> >
>


https://trac.ffmpeg.org/ticket/5729#comment:1
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] [PATCH] lavc: always build dnxhddata

[James Almer]

On 7/24/2016 8:13 PM, Michael Niedermayer wrote:
> On Sun, Jul 24, 2016 at 06:52:50PM -0300, James Almer wrote:
>> On 7/24/2016 6:14 PM, Michael Niedermayer wrote:
>>> On Sun, Jul 24, 2016 at 06:04:18PM -0300, James Almer wrote:
 On 7/24/2016 5:59 PM, Matthieu Bouron wrote:
> On Sun, Jul 24, 2016 at 10:53:51PM +0200, Michael Niedermayer wrote:
>> On Sun, Jul 24, 2016 at 10:40:58PM +0200, Michael Niedermayer wrote:
>>> On Sun, Jul 24, 2016 at 10:30:35PM +0200, Matthieu Bouron wrote:
 From: Matthieu Bouron 

 lavc/movenc rely on avpriv_dnxhd_parse_header_prefix declared by
 dnxhddata.h since e47981dab7fb7c9499b959cb0125b7281301969a.

 Fixes a missing symbol error in lavc/movenc if the dnxhd encoder is not
 enabled.
 ---
  libavcodec/Makefile | 5 +++--
  1 file changed, 3 insertions(+), 2 deletions(-)

 diff --git a/libavcodec/Makefile b/libavcodec/Makefile
 index abef19e..4fc4b09 100644
 --- a/libavcodec/Makefile
 +++ b/libavcodec/Makefile
 @@ -32,6 +32,7 @@ OBJS = allcodecs.o   
\
 codec_desc.o   
   \
 d3d11va.o  
   \
 dirac.o
   \
 +   dnxhddata.o
   \
 dv_profile.o   
   \
 imgconvert.o   
   \
 jni.o  
   \
 @@ -241,8 +242,8 @@ OBJS-$(CONFIG_DIRAC_DECODER)   += 
 diracdec.o dirac.o diracdsp.o diractab
dirac_arith.o mpeg12data.o 
 dirac_dwt.o \
dirac_vlc.o
  OBJS-$(CONFIG_DFA_DECODER) += dfa.o
 -OBJS-$(CONFIG_DNXHD_DECODER)   += dnxhddec.o dnxhddata.o
 -OBJS-$(CONFIG_DNXHD_ENCODER)   += dnxhdenc.o dnxhddata.o
 +OBJS-$(CONFIG_DNXHD_DECODER)   += dnxhddec.o
 +OBJS-$(CONFIG_DNXHD_ENCODER)   += dnxhdenc.o
>>>
>>> why dont you add dnxhddata.o to the demuxer needing it ?
>>
>> patch ok, i missed that this was a dependancy between the 2 libs
>> thought it was within a lib
>
> Pushed. Thanks.
>
> Matthieu

 Please revert this change and instead add a 
 "OBJS-$(CONFIG_MOV_MUXER) += dnxhddata.o" line to the libavformat
 dependencies section.
>>>
>>> doesnt this break the ABI of libavcodec ?
>>
>> No more than it already is, assuming this can be considered a
>> mistake in how avpriv functions are made available.
> 
> yes
> 
> 
>>
>> dnxhddata.c contains like three avpriv functions, the oldest
>> from 2013, and the file has never been built unconditionally.
>> Any ffmpeg build without dnxhd de/encoder, mxf muxer or dnxhd
>> demuxer from the past few years hasn't built it.
>> There are other files, like mpeg4audio.c, that are also not
>> built unconditionally and also have avpriv functions.
> 
> yes
> but tradition doesnt make it correct
> i think this hasnt hit anyone because its a problem for mainstream
> distros that care about ABI and they dont disable alot of stuff
> People disabling alot of stuff are probably building for a
> specifici use case, some embeded system some debuging case whatever
> they likely arent affected or concerned by the build libs all
> changing ABI together
> 
> 
>>
>>> either avpriv_dnxhd_parse_header_prefix is part of the ABI of
>>> libavcodec or its not
>>
> 
>> If avpriv functions are meant to be part of the ABI and their
>> presence not depend on configure time options, then that's
>> something that will have to be changed across the codebase
>> in general.
> 
> yes
> 
> They have to be part of the ABI though because only the ABI can be
> used from outside and they are used from outside the lib
> 
> if you want an hypothetical example where the current code breaks
> 1. a distro maintainer rebuilding the libs with some things disabled
> 2. a user installing a new package that pulls in a new libavcodec
>(could be a video player) but not libavformat
> 3. everything using libavformat breaks due to missing symbbols
> 
> removing a function from the ABI requires a major version and soname
> bump
> 
> we could document that changing the configure options affect the ABI
> but that would be very unfriendly to distro maintainers

A change in configure options can affect even public API right now.
See avutil's opencl and lzo modules.
We should probably change them to behave like 

Re: [FFmpeg-devel] [PATCH] avformat/fivdec: cached keyframes before video or audio stream was created

[XinZheng Zhang]

Ping.

On Thu, Jul 21, 2016 at 11:59 AM, XinZheng Zhang  wrote:
> On Thu, Jul 21, 2016 at 5:03 AM, Moritz Barsnick  wrote:
>> On Thu, Jul 21, 2016 at 00:06:44 +0800, zhangxinzheng wrote:
>>
>>> Subject: [FFmpeg-devel] [PATCH] avformat/fivdec: cached keyframes before 
>>> video or audio stream was created
>>^^^ flv
>>
>>> There has an error when seeking in a flv file, which key frames was
>>> sorted before video frame.
>>> This ensures that all the key frames was cached, and add to
>>> corresponding stream when it was created.
>>
>> Thanks to the grammar, I don't understand the intent of the patch.
>> (Don't get me wrong, I'm not scorning your English, but it needs to be
>> improved.)
>>
>
> In flvdec.c 420
> case AMF_DATA_TYPE_OBJECT:
> if ((vstream || astream) && key &&
> ioc->seekable &&
> !strcmp(KEYFRAMES_TAG, key) && depth == 1)
> if (parse_keyframes_index(s, ioc, vstream ? vstream : astream,
>   max_pos) < 0)
>
> If the first packet we got is object, both vstream and astream are NULL,
> that may prevent keyframe index being parsed.
>
> This patch saves keyframe index before A/V streams are created.
>
> The issue was introduced since 09ae7b81ea.
>
> I have created a ticket #5729 with a sample attatched.
>
>>>  #define RESYNC_BUFFER_SIZE (1<<20)
>>>
>>> +
>>> +typedef struct FLVKeyFrame {
>>
>> Why the extra newline?
>>
>
> The extra newline has been removed in new patch.
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] Outreachy 2016 december

[Michael Niedermayer]

Hi

On Sat, Jul 23, 2016 at 03:36:53PM -0700, Yayoi Ukai wrote:
> On Jul 22, 2016 10:18 AM, "Michael Niedermayer" 
> wrote:
> >
> > On Wed, Jul 06, 2016 at 11:25:21PM +0200, Michael Niedermayer wrote:
> > > Hi all
> > >
> > > The next Outreachy round starts soon
> > > FFmpeg has till august 22 IIUC (https://www.gnome.org/outreachy/)
> > > to express interrest to participate.
> > >
> > > We need an admin and backup admins.
> >
> > ping!
> 
> Hi everyone!
> 
> I'm happy to help with the administrative tasks of this project. As far as
> Outreachy goes, I will not be able to fully commit as a participant. I can
> help with fundraising, recruitment, and any administrarative tasks that you
> would like me to assist with. I would be happy to contribute code, but I
> would be limited to 10 hours per week working on small coding projects. I
> attempted to familiarise myself with Trac, but there are still some aspects
> of it I need guidance with. If there are any documentation available that I
> am not aware of, I would be happy to review it.

about trac in general there is:
https://trac.ffmpeg.org/wiki/TracGuide

but simply looking at some existing trac wiki page (by clicking on
"Edit this page" but not changing it just looking at whats there
for various constructs) should explain the trac wiki well enough
to create new pages ...

you can also use the text from "Edit this page" to copy wiki text
to a new page like copying parts of the last outreachy page to a new

If you like you surely can create the next outreachy page from the
previous Outreachy and GSoC page
It would be needed to check that the mentors and admins listed
previously are still able to spend te time & available and that
their projects are still available too.

Also people should be suggested to help improve the page. add projects,
volunteer as mentors, and so on ...

[...]

thanks
-- 
Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Awnsering whenever a program halts or runs forever is
On a turing machine, in general impossible (turings halting problem).
On any real computer, always possible as a real computer has a finite number
of states N, and will either halt in less than N cycles or never halt.


signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] [PATCH] lavc: always build dnxhddata

[Michael Niedermayer]

On Sun, Jul 24, 2016 at 06:52:50PM -0300, James Almer wrote:
> On 7/24/2016 6:14 PM, Michael Niedermayer wrote:
> > On Sun, Jul 24, 2016 at 06:04:18PM -0300, James Almer wrote:
> >> On 7/24/2016 5:59 PM, Matthieu Bouron wrote:
> >>> On Sun, Jul 24, 2016 at 10:53:51PM +0200, Michael Niedermayer wrote:
>  On Sun, Jul 24, 2016 at 10:40:58PM +0200, Michael Niedermayer wrote:
> > On Sun, Jul 24, 2016 at 10:30:35PM +0200, Matthieu Bouron wrote:
> >> From: Matthieu Bouron 
> >>
> >> lavc/movenc rely on avpriv_dnxhd_parse_header_prefix declared by
> >> dnxhddata.h since e47981dab7fb7c9499b959cb0125b7281301969a.
> >>
> >> Fixes a missing symbol error in lavc/movenc if the dnxhd encoder is not
> >> enabled.
> >> ---
> >>  libavcodec/Makefile | 5 +++--
> >>  1 file changed, 3 insertions(+), 2 deletions(-)
> >>
> >> diff --git a/libavcodec/Makefile b/libavcodec/Makefile
> >> index abef19e..4fc4b09 100644
> >> --- a/libavcodec/Makefile
> >> +++ b/libavcodec/Makefile
> >> @@ -32,6 +32,7 @@ OBJS = allcodecs.o   
> >>\
> >> codec_desc.o   
> >>   \
> >> d3d11va.o  
> >>   \
> >> dirac.o
> >>   \
> >> +   dnxhddata.o
> >>   \
> >> dv_profile.o   
> >>   \
> >> imgconvert.o   
> >>   \
> >> jni.o  
> >>   \
> >> @@ -241,8 +242,8 @@ OBJS-$(CONFIG_DIRAC_DECODER)   += 
> >> diracdec.o dirac.o diracdsp.o diractab
> >>dirac_arith.o mpeg12data.o 
> >> dirac_dwt.o \
> >>dirac_vlc.o
> >>  OBJS-$(CONFIG_DFA_DECODER) += dfa.o
> >> -OBJS-$(CONFIG_DNXHD_DECODER)   += dnxhddec.o dnxhddata.o
> >> -OBJS-$(CONFIG_DNXHD_ENCODER)   += dnxhdenc.o dnxhddata.o
> >> +OBJS-$(CONFIG_DNXHD_DECODER)   += dnxhddec.o
> >> +OBJS-$(CONFIG_DNXHD_ENCODER)   += dnxhdenc.o
> >
> > why dont you add dnxhddata.o to the demuxer needing it ?
> 
>  patch ok, i missed that this was a dependancy between the 2 libs
>  thought it was within a lib
> >>>
> >>> Pushed. Thanks.
> >>>
> >>> Matthieu
> >>
> >> Please revert this change and instead add a 
> >> "OBJS-$(CONFIG_MOV_MUXER) += dnxhddata.o" line to the libavformat
> >> dependencies section.
> > 
> > doesnt this break the ABI of libavcodec ?
> 
> No more than it already is, assuming this can be considered a
> mistake in how avpriv functions are made available.

yes


> 
> dnxhddata.c contains like three avpriv functions, the oldest
> from 2013, and the file has never been built unconditionally.
> Any ffmpeg build without dnxhd de/encoder, mxf muxer or dnxhd
> demuxer from the past few years hasn't built it.
> There are other files, like mpeg4audio.c, that are also not
> built unconditionally and also have avpriv functions.

yes
but tradition doesnt make it correct
i think this hasnt hit anyone because its a problem for mainstream
distros that care about ABI and they dont disable alot of stuff
People disabling alot of stuff are probably building for a
specifici use case, some embeded system some debuging case whatever
they likely arent affected or concerned by the build libs all
changing ABI together


> 
> > either avpriv_dnxhd_parse_header_prefix is part of the ABI of
> > libavcodec or its not
> 

> If avpriv functions are meant to be part of the ABI and their
> presence not depend on configure time options, then that's
> something that will have to be changed across the codebase
> in general.

yes

They have to be part of the ABI though because only the ABI can be
used from outside and they are used from outside the lib

if you want an hypothetical example where the current code breaks
1. a distro maintainer rebuilding the libs with some things disabled
2. a user installing a new package that pulls in a new libavcodec
   (could be a video player) but not libavformat
3. everything using libavformat breaks due to missing symbbols

removing a function from the ABI requires a major version and soname
bump

we could document that changing the configure options affect the ABI
but that would be very unfriendly to distro maintainers

maybe we can just enable the conditional parts when shared libs
are enabled. This without much thinking seems a possible solution
that would if its enough be better than fully unconditional.
But if we go this way someone should think about if this is really
enough first ...


> 

Re: [FFmpeg-devel] [PATCH] lavc: always build dnxhddata

[James Almer]

On 7/24/2016 6:14 PM, Michael Niedermayer wrote:
> On Sun, Jul 24, 2016 at 06:04:18PM -0300, James Almer wrote:
>> On 7/24/2016 5:59 PM, Matthieu Bouron wrote:
>>> On Sun, Jul 24, 2016 at 10:53:51PM +0200, Michael Niedermayer wrote:
 On Sun, Jul 24, 2016 at 10:40:58PM +0200, Michael Niedermayer wrote:
> On Sun, Jul 24, 2016 at 10:30:35PM +0200, Matthieu Bouron wrote:
>> From: Matthieu Bouron 
>>
>> lavc/movenc rely on avpriv_dnxhd_parse_header_prefix declared by
>> dnxhddata.h since e47981dab7fb7c9499b959cb0125b7281301969a.
>>
>> Fixes a missing symbol error in lavc/movenc if the dnxhd encoder is not
>> enabled.
>> ---
>>  libavcodec/Makefile | 5 +++--
>>  1 file changed, 3 insertions(+), 2 deletions(-)
>>
>> diff --git a/libavcodec/Makefile b/libavcodec/Makefile
>> index abef19e..4fc4b09 100644
>> --- a/libavcodec/Makefile
>> +++ b/libavcodec/Makefile
>> @@ -32,6 +32,7 @@ OBJS = allcodecs.o 
>>  \
>> codec_desc.o 
>> \
>> d3d11va.o
>> \
>> dirac.o  
>> \
>> +   dnxhddata.o  
>> \
>> dv_profile.o 
>> \
>> imgconvert.o 
>> \
>> jni.o
>> \
>> @@ -241,8 +242,8 @@ OBJS-$(CONFIG_DIRAC_DECODER)   += diracdec.o 
>> dirac.o diracdsp.o diractab
>>dirac_arith.o mpeg12data.o 
>> dirac_dwt.o \
>>dirac_vlc.o
>>  OBJS-$(CONFIG_DFA_DECODER) += dfa.o
>> -OBJS-$(CONFIG_DNXHD_DECODER)   += dnxhddec.o dnxhddata.o
>> -OBJS-$(CONFIG_DNXHD_ENCODER)   += dnxhdenc.o dnxhddata.o
>> +OBJS-$(CONFIG_DNXHD_DECODER)   += dnxhddec.o
>> +OBJS-$(CONFIG_DNXHD_ENCODER)   += dnxhdenc.o
>
> why dont you add dnxhddata.o to the demuxer needing it ?

 patch ok, i missed that this was a dependancy between the 2 libs
 thought it was within a lib
>>>
>>> Pushed. Thanks.
>>>
>>> Matthieu
>>
>> Please revert this change and instead add a 
>> "OBJS-$(CONFIG_MOV_MUXER) += dnxhddata.o" line to the libavformat
>> dependencies section.
> 
> doesnt this break the ABI of libavcodec ?

No more than it already is, assuming this can be considered a
mistake in how avpriv functions are made available.

dnxhddata.c contains like three avpriv functions, the oldest
from 2013, and the file has never been built unconditionally.
Any ffmpeg build without dnxhd de/encoder, mxf muxer or dnxhd
demuxer from the past few years hasn't built it.
There are other files, like mpeg4audio.c, that are also not
built unconditionally and also have avpriv functions.

> either avpriv_dnxhd_parse_header_prefix is part of the ABI of
> libavcodec or its not

If avpriv functions are meant to be part of the ABI and their
presence not depend on configure time options, then that's
something that will have to be changed across the codebase
in general.

Looking at this function closely now, it could have been added
to dnxhddata.h instead. No need for avpriv_ prefix and to
forcefully build an object file full of data tables that it
doesn't need.

> 
> [...]
> 
> 
> 
> ___
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
> 

___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

[FFmpeg-devel] [PATCH] lavf/movenc: add missing dependency on dnxhddata

[Matthieu Bouron]

From: Matthieu Bouron 

Fixes a missing symbol error in lavc/movenc if the dnxhd encoder is not
enabled.
---

New patch attached. If it's accepted, I'll revert the previous one and push the
new one.

---
 libavcodec/Makefile | 1 +
 1 file changed, 1 insertion(+)

diff --git a/libavcodec/Makefile b/libavcodec/Makefile
index abef19e..2c26b77 100644
--- a/libavcodec/Makefile
+++ b/libavcodec/Makefile
@@ -809,6 +809,7 @@ OBJS-$(CONFIG_MATROSKA_AUDIO_MUXER)+= mpeg4audio.o 
vorbis_data.o\
   flac.o flacdata.o
 OBJS-$(CONFIG_MATROSKA_MUXER)  += flac.o flacdata.o vorbis_data.o
 OBJS-$(CONFIG_MOV_DEMUXER) += ac3tab.o
+OBJS-$(CONFIG_MOV_MUXER)   += dnxhddata.o
 OBJS-$(CONFIG_MP2_MUXER)   += mpegaudiodata.o mpegaudiodecheader.o
 OBJS-$(CONFIG_MP3_MUXER)   += mpegaudiodata.o mpegaudiodecheader.o
 OBJS-$(CONFIG_MPEGTS_MUXER)+= mpeg4audio.o
-- 
2.9.0

___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] [PATCH] lavc: always build dnxhddata

[Michael Niedermayer]

On Sun, Jul 24, 2016 at 06:04:18PM -0300, James Almer wrote:
> On 7/24/2016 5:59 PM, Matthieu Bouron wrote:
> > On Sun, Jul 24, 2016 at 10:53:51PM +0200, Michael Niedermayer wrote:
> >> On Sun, Jul 24, 2016 at 10:40:58PM +0200, Michael Niedermayer wrote:
> >>> On Sun, Jul 24, 2016 at 10:30:35PM +0200, Matthieu Bouron wrote:
>  From: Matthieu Bouron 
> 
>  lavc/movenc rely on avpriv_dnxhd_parse_header_prefix declared by
>  dnxhddata.h since e47981dab7fb7c9499b959cb0125b7281301969a.
> 
>  Fixes a missing symbol error in lavc/movenc if the dnxhd encoder is not
>  enabled.
>  ---
>   libavcodec/Makefile | 5 +++--
>   1 file changed, 3 insertions(+), 2 deletions(-)
> 
>  diff --git a/libavcodec/Makefile b/libavcodec/Makefile
>  index abef19e..4fc4b09 100644
>  --- a/libavcodec/Makefile
>  +++ b/libavcodec/Makefile
>  @@ -32,6 +32,7 @@ OBJS = allcodecs.o 
>   \
>  codec_desc.o 
>  \
>  d3d11va.o
>  \
>  dirac.o  
>  \
>  +   dnxhddata.o  
>  \
>  dv_profile.o 
>  \
>  imgconvert.o 
>  \
>  jni.o
>  \
>  @@ -241,8 +242,8 @@ OBJS-$(CONFIG_DIRAC_DECODER)   += diracdec.o 
>  dirac.o diracdsp.o diractab
> dirac_arith.o mpeg12data.o 
>  dirac_dwt.o \
> dirac_vlc.o
>   OBJS-$(CONFIG_DFA_DECODER) += dfa.o
>  -OBJS-$(CONFIG_DNXHD_DECODER)   += dnxhddec.o dnxhddata.o
>  -OBJS-$(CONFIG_DNXHD_ENCODER)   += dnxhdenc.o dnxhddata.o
>  +OBJS-$(CONFIG_DNXHD_DECODER)   += dnxhddec.o
>  +OBJS-$(CONFIG_DNXHD_ENCODER)   += dnxhdenc.o
> >>>
> >>> why dont you add dnxhddata.o to the demuxer needing it ?
> >>
> >> patch ok, i missed that this was a dependancy between the 2 libs
> >> thought it was within a lib
> > 
> > Pushed. Thanks.
> > 
> > Matthieu
> 
> Please revert this change and instead add a 
> "OBJS-$(CONFIG_MOV_MUXER) += dnxhddata.o" line to the libavformat
> dependencies section.

doesnt this break the ABI of libavcodec ?
either avpriv_dnxhd_parse_header_prefix is part of the ABI of
libavcodec or its not

[...]
-- 
Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Never trust a computer, one day, it may think you are the virus. -- Compn


signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

[FFmpeg-devel] [PATCH] lavc/ffjni: replace ff_jni_{attach, detach} with ff_jni_get_env

[Matthieu Bouron]

From: Matthieu Bouron 

If a JNI environment is not already attached to the thread where the
MediaCodec calls are made the current implementation will attach /
detach an environment for each MediaCodec call wasting some CPU time.

ff_jni_get_env replaces ff_jni_{attach,detach} by permanently attaching
an environment (if it is not already the case) to the current thread.
The environment will be automatically detached at the thread destruction
using a pthread_key callback.

Saves around 5% of CPU time (out of 20%) while decoding a stream with
MediaCodec.
---
 libavcodec/ffjni.c  |  43 +
 libavcodec/ffjni.h  |  15 +--
 libavcodec/mediacodec.c |  14 +--
 libavcodec/mediacodec_surface.c |  14 +--
 libavcodec/mediacodec_wrapper.c | 200 
 5 files changed, 74 insertions(+), 212 deletions(-)

diff --git a/libavcodec/ffjni.c b/libavcodec/ffjni.c
index 82ee5d3..13eabb0 100644
--- a/libavcodec/ffjni.c
+++ b/libavcodec/ffjni.c
@@ -31,25 +31,42 @@
 #include "jni.h"
 #include "ffjni.h"
 
-static JavaVM *java_vm = NULL;
+static JavaVM *java_vm;
+static pthread_key_t current_env;
+static pthread_once_t once = PTHREAD_ONCE_INIT;
 static pthread_mutex_t lock = PTHREAD_MUTEX_INITIALIZER;
 
-JNIEnv *ff_jni_attach_env(int *attached, void *log_ctx)
+static void jni_detach_env(void *data)
+{
+if (java_vm) {
+(*java_vm)->DetachCurrentThread(java_vm);
+}
+}
+
+static void jni_create_pthread_key(void)
+{
+pthread_key_create(_env, jni_detach_env);
+}
+
+JNIEnv *ff_jni_get_env(void *log_ctx)
 {
 int ret = 0;
 JNIEnv *env = NULL;
 
-*attached = 0;
-
 pthread_mutex_lock();
 if (java_vm == NULL) {
 java_vm = av_jni_get_java_vm(log_ctx);
 }
-pthread_mutex_unlock();
 
 if (!java_vm) {
 av_log(log_ctx, AV_LOG_ERROR, "No Java virtual machine has been 
registered\n");
-return NULL;
+goto done;
+}
+
+pthread_once(, jni_create_pthread_key);
+
+if ((env = pthread_getspecific(current_env)) != NULL) {
+goto done;
 }
 
 ret = (*java_vm)->GetEnv(java_vm, (void **), JNI_VERSION_1_6);
@@ -59,7 +76,7 @@ JNIEnv *ff_jni_attach_env(int *attached, void *log_ctx)
 av_log(log_ctx, AV_LOG_ERROR, "Failed to attach the JNI 
environment to the current thread\n");
 env = NULL;
 } else {
-*attached = 1;
+pthread_setspecific(current_env, env);
 }
 break;
 case JNI_OK:
@@ -72,19 +89,11 @@ JNIEnv *ff_jni_attach_env(int *attached, void *log_ctx)
 break;
 }
 
+done:
+pthread_mutex_unlock();
 return env;
 }
 
-int ff_jni_detach_env(void *log_ctx)
-{
-if (java_vm == NULL) {
-av_log(log_ctx, AV_LOG_ERROR, "No Java virtual machine has been 
registered\n");
-return AVERROR(EINVAL);
-}
-
-return (*java_vm)->DetachCurrentThread(java_vm);
-}
-
 char *ff_jni_jstring_to_utf_chars(JNIEnv *env, jstring string, void *log_ctx)
 {
 char *ret = NULL;
diff --git a/libavcodec/ffjni.h b/libavcodec/ffjni.h
index 990c7b9..721a246 100644
--- a/libavcodec/ffjni.h
+++ b/libavcodec/ffjni.h
@@ -26,7 +26,10 @@
 #include 
 
 /*
- * Attach a JNI environment to the current thread.
+ * Attach permanently a JNI environment to the current thread and retrieve it.
+ *
+ * If successfully attached, the JNI environment will be automatically detached
+ * at thread destruction.
  *
  * @param attached pointer to an integer that will be set to 1 if the
  * environment has been attached to the current thread or 0 if it is
@@ -34,15 +37,7 @@
  * @param log_ctx context used for logging, can be NULL
  * @return the JNI environment on success, NULL otherwise
  */
-JNIEnv *ff_jni_attach_env(int *attached, void *log_ctx);
-
-/*
- * Detach the JNI environment from the current thread.
- *
- * @param log_ctx context used for logging, can be NULL
- * @return 0 on success, < 0 otherwise
- */
-int ff_jni_detach_env(void *log_ctx);
+JNIEnv *ff_jni_get_env(void *log_ctx);
 
 /*
  * Convert a jstring to its utf characters equivalent.
diff --git a/libavcodec/mediacodec.c b/libavcodec/mediacodec.c
index 5b79798..cd8766f 100644
--- a/libavcodec/mediacodec.c
+++ b/libavcodec/mediacodec.c
@@ -43,9 +43,8 @@ int av_mediacodec_default_init(AVCodecContext *avctx, 
AVMediaCodecContext *ctx,
 {
 int ret = 0;
 JNIEnv *env = NULL;
-int attached = 0;
 
-env = ff_jni_attach_env(, avctx);
+env = ff_jni_get_env(avctx);
 if (!env) {
 return AVERROR_EXTERNAL;
 }
@@ -58,17 +57,12 @@ int av_mediacodec_default_init(AVCodecContext *avctx, 
AVMediaCodecContext *ctx,
 ret = AVERROR_EXTERNAL;
 }
 
-if (attached) {
-ff_jni_detach_env(avctx);
-}
-
 return ret;
 }
 
 void av_mediacodec_default_free(AVCodecContext *avctx)
 {
 JNIEnv *env = NULL;
-int attached = 0;
 
 AVMediaCodecContext *ctx = avctx->hwaccel_context;
 
@@ -76,7 

Re: [FFmpeg-devel] [PATCH] lavc: always build dnxhddata

[James Almer]

On 7/24/2016 5:59 PM, Matthieu Bouron wrote:
> On Sun, Jul 24, 2016 at 10:53:51PM +0200, Michael Niedermayer wrote:
>> On Sun, Jul 24, 2016 at 10:40:58PM +0200, Michael Niedermayer wrote:
>>> On Sun, Jul 24, 2016 at 10:30:35PM +0200, Matthieu Bouron wrote:
 From: Matthieu Bouron 

 lavc/movenc rely on avpriv_dnxhd_parse_header_prefix declared by
 dnxhddata.h since e47981dab7fb7c9499b959cb0125b7281301969a.

 Fixes a missing symbol error in lavc/movenc if the dnxhd encoder is not
 enabled.
 ---
  libavcodec/Makefile | 5 +++--
  1 file changed, 3 insertions(+), 2 deletions(-)

 diff --git a/libavcodec/Makefile b/libavcodec/Makefile
 index abef19e..4fc4b09 100644
 --- a/libavcodec/Makefile
 +++ b/libavcodec/Makefile
 @@ -32,6 +32,7 @@ OBJS = allcodecs.o   
\
 codec_desc.o \
 d3d11va.o\
 dirac.o  \
 +   dnxhddata.o  \
 dv_profile.o \
 imgconvert.o \
 jni.o\
 @@ -241,8 +242,8 @@ OBJS-$(CONFIG_DIRAC_DECODER)   += diracdec.o 
 dirac.o diracdsp.o diractab
dirac_arith.o mpeg12data.o 
 dirac_dwt.o \
dirac_vlc.o
  OBJS-$(CONFIG_DFA_DECODER) += dfa.o
 -OBJS-$(CONFIG_DNXHD_DECODER)   += dnxhddec.o dnxhddata.o
 -OBJS-$(CONFIG_DNXHD_ENCODER)   += dnxhdenc.o dnxhddata.o
 +OBJS-$(CONFIG_DNXHD_DECODER)   += dnxhddec.o
 +OBJS-$(CONFIG_DNXHD_ENCODER)   += dnxhdenc.o
>>>
>>> why dont you add dnxhddata.o to the demuxer needing it ?
>>
>> patch ok, i missed that this was a dependancy between the 2 libs
>> thought it was within a lib
> 
> Pushed. Thanks.
> 
> Matthieu

Please revert this change and instead add a 
"OBJS-$(CONFIG_MOV_MUXER) += dnxhddata.o" line to the libavformat
dependencies section.

___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] [PATCH] lavc: always build dnxhddata

[James Almer]

On 7/24/2016 5:53 PM, Michael Niedermayer wrote:
> On Sun, Jul 24, 2016 at 10:40:58PM +0200, Michael Niedermayer wrote:
>> On Sun, Jul 24, 2016 at 10:30:35PM +0200, Matthieu Bouron wrote:
>>> From: Matthieu Bouron 
>>>
>>> lavc/movenc rely on avpriv_dnxhd_parse_header_prefix declared by
>>> dnxhddata.h since e47981dab7fb7c9499b959cb0125b7281301969a.
>>>
>>> Fixes a missing symbol error in lavc/movenc if the dnxhd encoder is not
>>> enabled.
>>> ---
>>>  libavcodec/Makefile | 5 +++--
>>>  1 file changed, 3 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/libavcodec/Makefile b/libavcodec/Makefile
>>> index abef19e..4fc4b09 100644
>>> --- a/libavcodec/Makefile
>>> +++ b/libavcodec/Makefile
>>> @@ -32,6 +32,7 @@ OBJS = allcodecs.o
>>>   \
>>> codec_desc.o \
>>> d3d11va.o\
>>> dirac.o  \
>>> +   dnxhddata.o  \
>>> dv_profile.o \
>>> imgconvert.o \
>>> jni.o\
>>> @@ -241,8 +242,8 @@ OBJS-$(CONFIG_DIRAC_DECODER)   += diracdec.o 
>>> dirac.o diracdsp.o diractab
>>>dirac_arith.o mpeg12data.o 
>>> dirac_dwt.o \
>>>dirac_vlc.o
>>>  OBJS-$(CONFIG_DFA_DECODER) += dfa.o
>>> -OBJS-$(CONFIG_DNXHD_DECODER)   += dnxhddec.o dnxhddata.o
>>> -OBJS-$(CONFIG_DNXHD_ENCODER)   += dnxhdenc.o dnxhddata.o
>>> +OBJS-$(CONFIG_DNXHD_DECODER)   += dnxhddec.o
>>> +OBJS-$(CONFIG_DNXHD_ENCODER)   += dnxhdenc.o
>>
>> why dont you add dnxhddata.o to the demuxer needing it ?
> 
> patch ok, i missed that this was a dependancy between the 2 libs
> thought it was within a lib

It can be done as you originally said. See the line for mfx
muxer under the "libavformat dependencies" section.

Just add one for mov muxer.

> 
> [...]
> 
> 
> 
> ___
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
> 

___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] [PATCH] lavc: always build dnxhddata

[Michael Niedermayer]

On Sun, Jul 24, 2016 at 10:40:58PM +0200, Michael Niedermayer wrote:
> On Sun, Jul 24, 2016 at 10:30:35PM +0200, Matthieu Bouron wrote:
> > From: Matthieu Bouron 
> > 
> > lavc/movenc rely on avpriv_dnxhd_parse_header_prefix declared by
> > dnxhddata.h since e47981dab7fb7c9499b959cb0125b7281301969a.
> > 
> > Fixes a missing symbol error in lavc/movenc if the dnxhd encoder is not
> > enabled.
> > ---
> >  libavcodec/Makefile | 5 +++--
> >  1 file changed, 3 insertions(+), 2 deletions(-)
> > 
> > diff --git a/libavcodec/Makefile b/libavcodec/Makefile
> > index abef19e..4fc4b09 100644
> > --- a/libavcodec/Makefile
> > +++ b/libavcodec/Makefile
> > @@ -32,6 +32,7 @@ OBJS = allcodecs.o
> >   \
> > codec_desc.o \
> > d3d11va.o\
> > dirac.o  \
> > +   dnxhddata.o  \
> > dv_profile.o \
> > imgconvert.o \
> > jni.o\
> > @@ -241,8 +242,8 @@ OBJS-$(CONFIG_DIRAC_DECODER)   += diracdec.o 
> > dirac.o diracdsp.o diractab
> >dirac_arith.o mpeg12data.o 
> > dirac_dwt.o \
> >dirac_vlc.o
> >  OBJS-$(CONFIG_DFA_DECODER) += dfa.o
> > -OBJS-$(CONFIG_DNXHD_DECODER)   += dnxhddec.o dnxhddata.o
> > -OBJS-$(CONFIG_DNXHD_ENCODER)   += dnxhdenc.o dnxhddata.o
> > +OBJS-$(CONFIG_DNXHD_DECODER)   += dnxhddec.o
> > +OBJS-$(CONFIG_DNXHD_ENCODER)   += dnxhdenc.o
> 
> why dont you add dnxhddata.o to the demuxer needing it ?

patch ok, i missed that this was a dependancy between the 2 libs
thought it was within a lib

[...]

-- 
Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Concerning the gods, I have no means of knowing whether they exist or not
or of what sort they may be, because of the obscurity of the subject, and
the brevity of human life -- Protagoras


signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] [PATCH] lavc: always build dnxhddata

[Michael Niedermayer]

On Sun, Jul 24, 2016 at 10:30:35PM +0200, Matthieu Bouron wrote:
> From: Matthieu Bouron 
> 
> lavc/movenc rely on avpriv_dnxhd_parse_header_prefix declared by
> dnxhddata.h since e47981dab7fb7c9499b959cb0125b7281301969a.
> 
> Fixes a missing symbol error in lavc/movenc if the dnxhd encoder is not
> enabled.
> ---
>  libavcodec/Makefile | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/libavcodec/Makefile b/libavcodec/Makefile
> index abef19e..4fc4b09 100644
> --- a/libavcodec/Makefile
> +++ b/libavcodec/Makefile
> @@ -32,6 +32,7 @@ OBJS = allcodecs.o  
> \
> codec_desc.o \
> d3d11va.o\
> dirac.o  \
> +   dnxhddata.o  \
> dv_profile.o \
> imgconvert.o \
> jni.o\
> @@ -241,8 +242,8 @@ OBJS-$(CONFIG_DIRAC_DECODER)   += diracdec.o 
> dirac.o diracdsp.o diractab
>dirac_arith.o mpeg12data.o 
> dirac_dwt.o \
>dirac_vlc.o
>  OBJS-$(CONFIG_DFA_DECODER) += dfa.o
> -OBJS-$(CONFIG_DNXHD_DECODER)   += dnxhddec.o dnxhddata.o
> -OBJS-$(CONFIG_DNXHD_ENCODER)   += dnxhdenc.o dnxhddata.o
> +OBJS-$(CONFIG_DNXHD_DECODER)   += dnxhddec.o
> +OBJS-$(CONFIG_DNXHD_ENCODER)   += dnxhdenc.o

why dont you add dnxhddata.o to the demuxer needing it ?

[...]
-- 
Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Opposition brings concord. Out of discord comes the fairest harmony.
-- Heraclitus


signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] A question regarding dangerous call inside libavformat\utils.c::has_decode_delay_been_guessed()

[Ivan Uskov]

Hello Michael,

Sunday, July 24, 2016, 11:25:29 PM, you wrote:

MN> On Sun, Jul 24, 2016 at 11:18:38PM +0300, Ivan Uskov wrote:
>> Hello Michael,
>> 
>> Sunday, July 24, 2016, 11:11:32 PM, you wrote:
>> 
>> MN> On Sun, Jul 24, 2016 at 09:55:21PM +0300, Ivan Uskov wrote:
>> >> Hello All,
>> >> 
>> >> I have discovered the following issue:
>> >> Latest builds of ffmpeg crashes into the h264.c when *hardware* qsv-based 
>> >> h264 decoder uses.
>> >> The crash does appear inside the
>> >> 
>> >> int avpriv_h264_has_num_reorder_frames(AVCodecContext *avctx)
>> >> {
>> >> H264Context *h = avctx->priv_data;
>> >> return h && h->ps.sps ? h->ps.sps->num_reorder_frames : 0;
>> >> }
>> >> 
>> >> It is obvious, that casting to H264Context cannot be used for qsv decoder
>> >> because there is QSVH2645Context. Similar issue will appear for CUVID
>> >> decoder case (CuvidContext uses), Android MediaCodec H.264 decoder
>> >> (MediaCodecH264DecContext uses), possible another cases existing.
>> >> 
>> >> The caller function is
>> >> 
>> >> static int has_decode_delay_been_guessed(AVStream *st)
>> >> {
>> >> if (st->codecpar->codec_id != AV_CODEC_ID_H264) return 1;
>> >> if (!st->info) // if we have left find_stream_info then 
>> >> nb_decoded_frames won't increase anymore for stream copy
>> >> return 1;
>> >> #if CONFIG_H264_DECODER
>> >> if (st->internal->avctx->has_b_frames &&
>> >>avpriv_h264_has_num_reorder_frames(st->internal->avctx) == 
>> >> st->internal->avctx->has_b_frames)
>> >> return 1;
>> >> #endif
>> >> if (st->internal->avctx->has_b_frames<3)
>> >> return st->nb_decoded_frames >= 7;
>> >> else if (st->internal->avctx->has_b_frames<4)
>> >> return st->nb_decoded_frames >= 18;
>> >> else
>> >> return st->nb_decoded_frames >= 20;
>> >> }
>> >> ...which called by update_initial_timestamps()
>> >> 
>> >> Have anybody the idea how it can be correctly fixed?
>> >> Looks like has_decode_delay_been_guessed() should be corrected.
>> 
>> MN> this code is not new, this is in git since 4 years
>> MN> commit bafa1c7f383d6c1c0f3d207395fe6a574092b7ac
>> MN> Date:   Mon Jul 2 23:16:59 2012 +020
>> 
>> MN> why does it cause a problem now ?
>> 
>> MN> why does libavformat use hw h264 decoding ?
>> I do not know that is reason why I'm asking.
>> I just run ffmpeg like
>> ffmpeg -c:v h264_qsv -i ./Ducks.Take.Off.720p.QHD.CRF24.x264-CtrlHD.mkv 
>> -vcodec h264_qsv -y ./result.mp4
>> ..and it crashes inside avpriv_h264_has_num_reorder_frames(AVCodecContext 
>> *avctx)
>> 
>> When I replace
>> #if CONFIG_H264_DECODER
>> to
>> #if 0
>> ...all working fine.
>> 
>> I can not say when exactly the issue appeared, about several weeks ago all
>> worked correct. But current build is broken at this place.

MN> can you use git bisect to figure out when this issue appeared?
Ok, I'll try, it will take some time.
MN> was the decoder used in libavformat previously also the hw instead of
MN> sw decoder ?
There are no appropriate methods in hw decoder to be called from libavformat.
So hw decoder exactly was never used.


MN> [...]



-- 
Best regards,
 Ivanmailto:ivan.us...@nablet.com

___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

[FFmpeg-devel] [PATCH] lavc: always build dnxhddata

[Matthieu Bouron]

From: Matthieu Bouron 

lavc/movenc rely on avpriv_dnxhd_parse_header_prefix declared by
dnxhddata.h since e47981dab7fb7c9499b959cb0125b7281301969a.

Fixes a missing symbol error in lavc/movenc if the dnxhd encoder is not
enabled.
---
 libavcodec/Makefile | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/libavcodec/Makefile b/libavcodec/Makefile
index abef19e..4fc4b09 100644
--- a/libavcodec/Makefile
+++ b/libavcodec/Makefile
@@ -32,6 +32,7 @@ OBJS = allcodecs.o
  \
codec_desc.o \
d3d11va.o\
dirac.o  \
+   dnxhddata.o  \
dv_profile.o \
imgconvert.o \
jni.o\
@@ -241,8 +242,8 @@ OBJS-$(CONFIG_DIRAC_DECODER)   += diracdec.o 
dirac.o diracdsp.o diractab
   dirac_arith.o mpeg12data.o 
dirac_dwt.o \
   dirac_vlc.o
 OBJS-$(CONFIG_DFA_DECODER) += dfa.o
-OBJS-$(CONFIG_DNXHD_DECODER)   += dnxhddec.o dnxhddata.o
-OBJS-$(CONFIG_DNXHD_ENCODER)   += dnxhdenc.o dnxhddata.o
+OBJS-$(CONFIG_DNXHD_DECODER)   += dnxhddec.o
+OBJS-$(CONFIG_DNXHD_ENCODER)   += dnxhdenc.o
 OBJS-$(CONFIG_DPX_DECODER) += dpx.o
 OBJS-$(CONFIG_DPX_ENCODER) += dpxenc.o
 OBJS-$(CONFIG_DSD_LSBF_DECODER)+= dsddec.o dsd.o
-- 
2.9.0

___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] A question regarding dangerous call inside libavformat\utils.c::has_decode_delay_been_guessed()

[Michael Niedermayer]

On Sun, Jul 24, 2016 at 11:18:38PM +0300, Ivan Uskov wrote:
> Hello Michael,
> 
> Sunday, July 24, 2016, 11:11:32 PM, you wrote:
> 
> MN> On Sun, Jul 24, 2016 at 09:55:21PM +0300, Ivan Uskov wrote:
> >> Hello All,
> >> 
> >> I have discovered the following issue:
> >> Latest builds of ffmpeg crashes into the h264.c when *hardware* qsv-based 
> >> h264 decoder uses.
> >> The crash does appear inside the
> >> 
> >> int avpriv_h264_has_num_reorder_frames(AVCodecContext *avctx)
> >> {
> >> H264Context *h = avctx->priv_data;
> >> return h && h->ps.sps ? h->ps.sps->num_reorder_frames : 0;
> >> }
> >> 
> >> It is obvious, that casting to H264Context cannot be used for qsv decoder
> >> because there is QSVH2645Context. Similar issue will appear for CUVID
> >> decoder case (CuvidContext uses), Android MediaCodec H.264 decoder
> >> (MediaCodecH264DecContext uses), possible another cases existing.
> >> 
> >> The caller function is
> >> 
> >> static int has_decode_delay_been_guessed(AVStream *st)
> >> {
> >> if (st->codecpar->codec_id != AV_CODEC_ID_H264) return 1;
> >> if (!st->info) // if we have left find_stream_info then 
> >> nb_decoded_frames won't increase anymore for stream copy
> >> return 1;
> >> #if CONFIG_H264_DECODER
> >> if (st->internal->avctx->has_b_frames &&
> >>avpriv_h264_has_num_reorder_frames(st->internal->avctx) == 
> >> st->internal->avctx->has_b_frames)
> >> return 1;
> >> #endif
> >> if (st->internal->avctx->has_b_frames<3)
> >> return st->nb_decoded_frames >= 7;
> >> else if (st->internal->avctx->has_b_frames<4)
> >> return st->nb_decoded_frames >= 18;
> >> else
> >> return st->nb_decoded_frames >= 20;
> >> }
> >> ...which called by update_initial_timestamps()
> >> 
> >> Have anybody the idea how it can be correctly fixed?
> >> Looks like has_decode_delay_been_guessed() should be corrected.
> 
> MN> this code is not new, this is in git since 4 years
> MN> commit bafa1c7f383d6c1c0f3d207395fe6a574092b7ac
> MN> Date:   Mon Jul 2 23:16:59 2012 +020
> 
> MN> why does it cause a problem now ?
> 
> MN> why does libavformat use hw h264 decoding ?
> I do not know that is reason why I'm asking.
> I just run ffmpeg like
> ffmpeg -c:v h264_qsv -i ./Ducks.Take.Off.720p.QHD.CRF24.x264-CtrlHD.mkv 
> -vcodec h264_qsv -y ./result.mp4
> ..and it crashes inside avpriv_h264_has_num_reorder_frames(AVCodecContext 
> *avctx)
> 
> When I replace
> #if CONFIG_H264_DECODER
> to
> #if 0
> ...all working fine.
> 
> I can not say when exactly the issue appeared, about several weeks ago all
> worked correct. But current build is broken at this place.

can you use git bisect to figure out when this issue appeared?

was the decoder used in libavformat previously also the hw instead of
sw decoder ?


[...]
-- 
Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Dictatorship: All citizens are under surveillance, all their steps and
actions recorded, for the politicians to enforce control.
Democracy: All politicians are under surveillance, all their steps and
actions recorded, for the citizens to enforce control.


signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] A question regarding dangerous call inside libavformat\utils.c::has_decode_delay_been_guessed()

[Ivan Uskov]

Hello Michael,

Sunday, July 24, 2016, 11:11:32 PM, you wrote:

MN> On Sun, Jul 24, 2016 at 09:55:21PM +0300, Ivan Uskov wrote:
>> Hello All,
>> 
>> I have discovered the following issue:
>> Latest builds of ffmpeg crashes into the h264.c when *hardware* qsv-based 
>> h264 decoder uses.
>> The crash does appear inside the
>> 
>> int avpriv_h264_has_num_reorder_frames(AVCodecContext *avctx)
>> {
>> H264Context *h = avctx->priv_data;
>> return h && h->ps.sps ? h->ps.sps->num_reorder_frames : 0;
>> }
>> 
>> It is obvious, that casting to H264Context cannot be used for qsv decoder
>> because there is QSVH2645Context. Similar issue will appear for CUVID
>> decoder case (CuvidContext uses), Android MediaCodec H.264 decoder
>> (MediaCodecH264DecContext uses), possible another cases existing.
>> 
>> The caller function is
>> 
>> static int has_decode_delay_been_guessed(AVStream *st)
>> {
>> if (st->codecpar->codec_id != AV_CODEC_ID_H264) return 1;
>> if (!st->info) // if we have left find_stream_info then 
>> nb_decoded_frames won't increase anymore for stream copy
>> return 1;
>> #if CONFIG_H264_DECODER
>> if (st->internal->avctx->has_b_frames &&
>>avpriv_h264_has_num_reorder_frames(st->internal->avctx) == 
>> st->internal->avctx->has_b_frames)
>> return 1;
>> #endif
>> if (st->internal->avctx->has_b_frames<3)
>> return st->nb_decoded_frames >= 7;
>> else if (st->internal->avctx->has_b_frames<4)
>> return st->nb_decoded_frames >= 18;
>> else
>> return st->nb_decoded_frames >= 20;
>> }
>> ...which called by update_initial_timestamps()
>> 
>> Have anybody the idea how it can be correctly fixed?
>> Looks like has_decode_delay_been_guessed() should be corrected.

MN> this code is not new, this is in git since 4 years
MN> commit bafa1c7f383d6c1c0f3d207395fe6a574092b7ac
MN> Date:   Mon Jul 2 23:16:59 2012 +020

MN> why does it cause a problem now ?

MN> why does libavformat use hw h264 decoding ?
I do not know that is reason why I'm asking.
I just run ffmpeg like
ffmpeg -c:v h264_qsv -i ./Ducks.Take.Off.720p.QHD.CRF24.x264-CtrlHD.mkv -vcodec 
h264_qsv -y ./result.mp4
..and it crashes inside avpriv_h264_has_num_reorder_frames(AVCodecContext 
*avctx)

When I replace
#if CONFIG_H264_DECODER
to
#if 0
...all working fine.

I can not say when exactly the issue appeared, about several weeks ago all
worked correct. But current build is broken at this place.


MN> this seems like a bad idea to me, as libavcodec would use
MN> hw decoding too potentially resulting in multiple hw decoders
MN> running at the same time for no good reason.

MN> the one in libavformat is just used to parse a few header values
MN> ideally that should be done in the AVParser for h264 ...

MN> [...]




-- 
Best regards,
 Ivanmailto:ivan.us...@nablet.com

___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] A question regarding dangerous call inside libavformat\utils.c::has_decode_delay_been_guessed()

[Michael Niedermayer]

On Sun, Jul 24, 2016 at 09:55:21PM +0300, Ivan Uskov wrote:
> Hello All,
> 
> I have discovered the following issue:
> Latest builds of ffmpeg crashes into the h264.c when *hardware* qsv-based 
> h264 decoder uses.
> The crash does appear inside the
> 
> int avpriv_h264_has_num_reorder_frames(AVCodecContext *avctx)
> {
> H264Context *h = avctx->priv_data;
> return h && h->ps.sps ? h->ps.sps->num_reorder_frames : 0;
> }
> 
> It is obvious, that casting to H264Context cannot be used for qsv decoder
> because there is QSVH2645Context. Similar issue will appear for CUVID
> decoder case (CuvidContext uses), Android MediaCodec H.264 decoder
> (MediaCodecH264DecContext uses), possible another cases existing.
> 
> The caller function is
> 
> static int has_decode_delay_been_guessed(AVStream *st)
> {
> if (st->codecpar->codec_id != AV_CODEC_ID_H264) return 1;
> if (!st->info) // if we have left find_stream_info then nb_decoded_frames 
> won't increase anymore for stream copy
> return 1;
> #if CONFIG_H264_DECODER
> if (st->internal->avctx->has_b_frames &&
>avpriv_h264_has_num_reorder_frames(st->internal->avctx) == 
> st->internal->avctx->has_b_frames)
> return 1;
> #endif
> if (st->internal->avctx->has_b_frames<3)
> return st->nb_decoded_frames >= 7;
> else if (st->internal->avctx->has_b_frames<4)
> return st->nb_decoded_frames >= 18;
> else
> return st->nb_decoded_frames >= 20;
> }
> ...which called by update_initial_timestamps()
> 
> Have anybody the idea how it can be correctly fixed?
> Looks like has_decode_delay_been_guessed() should be corrected.

this code is not new, this is in git since 4 years
commit bafa1c7f383d6c1c0f3d207395fe6a574092b7ac
Date:   Mon Jul 2 23:16:59 2012 +020

why does it cause a problem now ?

why does libavformat use hw h264 decoding ?
this seems like a bad idea to me, as libavcodec would use
hw decoding too potentially resulting in multiple hw decoders
running at the same time for no good reason.

the one in libavformat is just used to parse a few header values
ideally that should be done in the AVParser for h264 ...

[...]

-- 
Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Why not whip the teacher when the pupil misbehaves? -- Diogenes of Sinope


signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] [PATCH v2] avcodec/bsf: Set EOF flag only if pkt == NULL

[Jan Sebechlebsky]

On 07/22/2016 10:14 PM, Nicolas George wrote:

Le quintidi 5 thermidor, an CCXXIV, sebechlebsky...@gmail.com a écrit :

From: Jan Sebechlebsky 

Set BSF EOF flag only if pkt == NULL in av_bsf_send_packet().

Signed-off-by: Jan Sebechlebsky 
---
  I agree, it seems cleaner that way.

  Thanks,
  please apply this version of patch then and ignore
  the patch changing the comment.

  Regards,
  Jan

  libavcodec/bsf.c | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/libavcodec/bsf.c b/libavcodec/bsf.c
index 88b7f29..9b9ada7 100644
--- a/libavcodec/bsf.c
+++ b/libavcodec/bsf.c
@@ -172,7 +172,7 @@ int av_bsf_init(AVBSFContext *ctx)
  
  int av_bsf_send_packet(AVBSFContext *ctx, AVPacket *pkt)

  {
-if (!pkt || !pkt->data) {
+if (!pkt) {

It we make the case where pkt != NULL but no data nor side data forbidden, I
would suggest to detect it.

I know that others disagree, but I think an assert is the best solution for
that: if the caller pass a forbidden value, it can expect an undefined
behaviour, and an assert failure is the most sympathetic undefined
behaviours for developers.
I gave this a second thought, wouldn't it be better to simply ignore pkt 
without payload? So after caller would send empty packet using 
av_bsf_send_packet, he would get AVERROR(EAGAIN) from the next 
av_bsf_receive_packet call (from the definition in documentation 
AVERROR(EAGAIN) means "more data needed" when returned by 
av_bsf_receive_packet).
However, if you think it is better to reserve packet without payload for 
some future use, I won't object.


Regards,
Jan
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] ffbox0 server hosting

[Clément Bœsch]

On Sat, Jul 23, 2016 at 11:12:40AM +0200, Michael Niedermayer wrote:
> On Tue, Jul 19, 2016 at 07:32:01PM +0200, Clément Bœsch wrote:
> > Hi,
> > 
> > As stated in previous meeting¹ I'd like to stop paying for ffbox0. This is
> > not exactly an issue today for me, but I feel like it could be avoided.
> > There was a need for a fast solution a while ago, but it's now long past
> > that time (it was about one year ago), and I'd like to be discharged from
> > this.
> > 
> > So on 19th October 2016 (3 months from now), I will close the ffbox0
> > subscription.
> > 
> > Note that this is not only a money problem, it comes with a hosting
> > responsibility I'd like to get rid of; I'm thinking notably about the
> > abuse complaints I got wrt rtmpdump (because yes, this box is not only for
> > the FFmpeg project, but also all kind of other projects I have absolutely
> > 0 interest in funding, hosting and taking responsibility for myself).
> 
> raz (and me but really raz did 99% of the work) has moved things from
> ffbox0 to ffbox0-bg.
> 

Thank you very much to you and Nikolay.

> About monthly ffbox0 fees, if you want them refunded please submit
> a refund request to the ffmpeg-devel ML (see similar ones on the ML
> about refunding of travell costs). All refund requests should be easy
> search, detect and discussable on the ML ...
> 

It's about 50€/months, for one year. I'm fine with that, I don't need a
refund. Thank you.

> If you wish to shutdown ffbox0 in 3 months, would you have time to
> go over the other free server & hosting offers we had? It would make
> sense for us probably to keep the number of servers we have available
> non decreasing so we can move things "easily" in case future need
> arises ...

I haven't followed at all the different offers, I don't feel like poking
random people, sorry.

> Thanks alot for providing ffbox0 to FFmpeg, MPlayer, rtmpdump, ...
> when it was needed

Sure, you're welcome

-- 
Clément B.


signature.asc
Description: PGP signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

[FFmpeg-devel] A question regarding dangerous call inside libavformat\utils.c::has_decode_delay_been_guessed()

[Ivan Uskov]

Hello All,

I have discovered the following issue:
Latest builds of ffmpeg crashes into the h264.c when *hardware* qsv-based h264 
decoder uses.
The crash does appear inside the

int avpriv_h264_has_num_reorder_frames(AVCodecContext *avctx)
{
H264Context *h = avctx->priv_data;
return h && h->ps.sps ? h->ps.sps->num_reorder_frames : 0;
}

It is obvious, that casting to H264Context cannot be used for qsv decoder
because there is QSVH2645Context. Similar issue will appear for CUVID
decoder case (CuvidContext uses), Android MediaCodec H.264 decoder
(MediaCodecH264DecContext uses), possible another cases existing.

The caller function is

static int has_decode_delay_been_guessed(AVStream *st)
{
if (st->codecpar->codec_id != AV_CODEC_ID_H264) return 1;
if (!st->info) // if we have left find_stream_info then nb_decoded_frames 
won't increase anymore for stream copy
return 1;
#if CONFIG_H264_DECODER
if (st->internal->avctx->has_b_frames &&
   avpriv_h264_has_num_reorder_frames(st->internal->avctx) == 
st->internal->avctx->has_b_frames)
return 1;
#endif
if (st->internal->avctx->has_b_frames<3)
return st->nb_decoded_frames >= 7;
else if (st->internal->avctx->has_b_frames<4)
return st->nb_decoded_frames >= 18;
else
return st->nb_decoded_frames >= 20;
}
...which called by update_initial_timestamps()

Have anybody the idea how it can be correctly fixed?
Looks like has_decode_delay_been_guessed() should be corrected.
  

-- 
Best regards,
 Ivan  mailto:ivan.us...@nablet.com

___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] [PATCH]lavc/h264_ps: Be more verbose when truncating likely oversized pps.

[Carl Eugen Hoyos]

Michael Niedermayer  niedermayer.cc> writes:

> > Subject: [PATCH] lavc/h264_ps: Be more verbose when truncating likely
> >  oversized PPS.
> 
> should be ok

Patch applied.

Thank you, Carl Eugen

___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] [PATCH] lavc: add mpeg4 mediacodec decoder

[Matthieu Bouron]

On Sun, Jul 24, 2016 at 11:37:08AM +0200, Thomas Volkert wrote:
> From: Thomas Volkert 
> 
> ---
>  configure|   1 +
>  libavcodec/Makefile  |   1 +
>  libavcodec/allcodecs.c   |   1 +
>  libavcodec/mediacodecdec_mpeg4.c | 239 
> +++
>  4 files changed, 242 insertions(+)
>  create mode 100644 libavcodec/mediacodecdec_mpeg4.c

Hi,

Thanks for the patchset, I will review it and test it during the upcoming days.

Matthieu

[...]
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] [PATCH] libavcodec/qsvdec.c: Restoring decoding functionality after unsuccessful merge from libav.

[Hendrik Leppkes]

On Sun, Jul 24, 2016 at 3:43 PM, Ivan Uskov  wrote:
> Hello Mark,
>
> Saturday, July 23, 2016, 11:42:46 PM, you wrote:
>
> MT> On 23/07/16 20:33, Ivan Uskov wrote:
>>> If  you are use qsv, I would like to recommend to roll-back to version 
>>> before
>>> d30cf57a7b2097b565db02ecfffbdc9c16423d0e
>>> Really  the  d30cf57a7b2097b565db02ecfffbdc9c16423d0e is useless and only
>>> makes code complex and work slow, the sync variable is not mandatory to be
>>> allocated  on  heap  at  all.  libav  guys  did a big mistake when have 
>>> added
>>> such "feature".
>
> MT> Are you sure that works correctly with all streams, maybe one with some
> MT> sort of delay?  It looks like you can add multiple things to the fifo in
> MT> one call and but then only ever remove one, so using the stack to hold
> MT> them across invocations is not appropriate (though failure may be very 
> subtle).
> Yes, I'm sure.
> There are following arguments:
> 1. Intel's samples by itself use general variable for the 'sync' arg, not
> heap allocated.
> 2. Intel documentation says nothing special for this arg.
> 3. I did several qsv-based transcoders for windows and linux no any issues
> without this questionable trick.
> 4. A fifo object used previously to keep 'sync' handles already keeps data on
> heap :-)
>
> There only one known issue into initial implementation, the fifo size may be 
> not
> enough an there is no check for overflow. So at some cases wrapper may
> overwrite sync entries in fifo and some decoded frames may be lost. But it is
> not related with discussed problem.
>
> MT> (Note that I'm not sufficiently familiar with Media SDK to confidently
> MT> assert anything about how this code should behave, so please don't read 
> too much into what I say.)
>
> MT> In general I think if you are happy that the previous implementation in
> MT> ffmpeg was correct then please do revert, because the code as it is now 
> is completely broken.
> Ok, will do.
>

OT:
Ivan, can you *please* fix your quoting style to use standard quotes
as everyone else (ie. single > for every level)? Reading your replies
is practically impossible.

- Hendrik
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] [PATCH] qsvdec.c: Revert "Merge commit '3c53627ac17fc6bdea5029be57da1e03b32d265d'"

[Michael Niedermayer]

On Sun, Jul 24, 2016 at 06:47:37PM +0300, Ivan Uskov wrote:
> Hello All,
> 
> This reverts commit d30cf57a7b2097b565db02ecfffbdc9c16423d0e, reversing
> changes made to acc155ac55baa95d1c16c0364b02244bc04d83a8.
> The commit  d30cf57a7b2097b565db02ecfffbdc9c16423d0e provided irrelevant
> code complexity and decoding slowdown. But main disadvantage of this commit 
> is decoder crash.
> So it should be reverted.
> Please review.
> 
> -- 
> Best regards,
>  Ivan  mailto:ivan.us...@nablet.com

>  qsvdec.c |   34 ++
>  1 file changed, 6 insertions(+), 28 deletions(-)
> 2dafbabfa4e364cb9fdd39d2e85eb89108a2710b  
> 0001-Revert-Merge-commit-3c53627ac17fc6bdea5029be57da1e03.patch
> From 1e2ba6de0095e63c7f6ef66b5516c3e4789af7f3 Mon Sep 17 00:00:00 2001
> From: Ivan Uskov 
> Date: Sun, 24 Jul 2016 09:59:42 -0400
> Subject: [PATCH] Revert "Merge commit
>  '3c53627ac17fc6bdea5029be57da1e03b32d265d'" This reverts commit
>  d30cf57a7b2097b565db02ecfffbdc9c16423d0e, reversing changes made to
>  acc155ac55baa95d1c16c0364b02244bc04d83a8. The commit
>  d30cf57a7b2097b565db02ecfffbdc9c16423d0e provided irrelevant code complexity
>  and decoding slowdown. But main disadvantage of this commit is decoder crash.
>  So it should be reverted.

applied

thanks

[...]

-- 
Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Dictatorship: All citizens are under surveillance, all their steps and
actions recorded, for the politicians to enforce control.
Democracy: All politicians are under surveillance, all their steps and
actions recorded, for the citizens to enforce control.


signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

[FFmpeg-devel] [PATCH] af_hdcd: Add counter for cdt expirations

[Burt P]

Adds a counter for when the "code detect timer" expired without
finding a valid packet.

Signed-off-by: Burt P 
---
 libavfilter/af_hdcd.c | 15 +--
 1 file changed, 13 insertions(+), 2 deletions(-)

diff --git a/libavfilter/af_hdcd.c b/libavfilter/af_hdcd.c
index 72560db..f265ed9 100644
--- a/libavfilter/af_hdcd.c
+++ b/libavfilter/af_hdcd.c
@@ -846,6 +846,7 @@ typedef struct {
  * steps of 0.5, but no value below -6.0 dB should appear. */
 int gain_counts[16]; /* for cursiosity, mostly */
 int max_gain;
+int count_sustain_expired;/* occurences of code detect timer expiring 
without detecting a code */
 
 AVFilterContext *fctx; /* filter context for logging errors */
 } hdcd_state_t;
@@ -895,6 +896,8 @@ static void hdcd_reset(hdcd_state_t *state, unsigned rate)
 state->count_transient_filter = 0;
 for(i = 0; i < 16; i++) state->gain_counts[i] = 0;
 state->max_gain = 0;
+
+state->count_sustain_expired = 0;
 }
 
 /* update the user info/counters */
@@ -978,8 +981,11 @@ static int hdcd_integrate(hdcd_state_t *state, int *flag, 
const int32_t *samples
 
 static int hdcd_scan(hdcd_state_t *state, const int32_t *samples, int max, int 
stride)
 {
+int cdt_active = 0;
+/* code detect timer */
 int result;
 if (state->sustain > 0) {
+cdt_active = 1;
 if (state->sustain <= max) {
 state->control = 0;
 max = state->sustain;
@@ -992,11 +998,15 @@ static int hdcd_scan(hdcd_state_t *state, const int32_t 
*samples, int max, int s
 int consumed = hdcd_integrate(state, , samples, max - result, 
stride);
 result += consumed;
 if (flag > 0) {
+/* reset timer if code detected in channel */
 state->sustain = state->sustain_reset;
 break;
 }
 samples += consumed * stride;
 }
+/* code detect timer expired */
+if (cdt_active && state->sustain == 0)
+state->count_sustain_expired++;
 return result;
 }
 
@@ -1198,12 +1208,13 @@ static av_cold void uninit(AVFilterContext *ctx)
 hdcd_state_t *state = >state[i];
 av_log(ctx, AV_LOG_VERBOSE, "Channel %d: counter A: %d, B: %d, C: 
%d\n", i,
 state->code_counterA, state->code_counterB, 
state->code_counterC);
-av_log(ctx, AV_LOG_VERBOSE, "Channel %d: pe: %d, tf: %d, almost_A: %d, 
checkfail_B: %d, unmatched_C: %d\n", i,
+av_log(ctx, AV_LOG_VERBOSE, "Channel %d: pe: %d, tf: %d, almost_A: %d, 
checkfail_B: %d, unmatched_C: %d, cdt_expired: %d\n", i,
 state->count_peak_extend,
 state->count_transient_filter,
 state->code_counterA_almost,
 state->code_counterB_checkfails,
-state->code_counterC_unmatched);
+state->code_counterC_unmatched,
+state->count_sustain_expired);
 for (j = 0; j <= state->max_gain; j++) {
 av_log(ctx, AV_LOG_VERBOSE, "Channel %d: tg %0.1f: %d\n", i, 
GAINTOFLOAT(j), state->gain_counts[j]);
 }
-- 
2.7.4

___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] [PATCH] avcodec/qsv: remove MFX_EXTBUFF_CODING_OPTION3

[Ivan Uskov]

Hello Zeranoe,

Saturday, June 18, 2016, 7:33:12 AM, you wrote:

zgc> From: Kyle Schwarz 

zgc> 4th generation Intel CPUs don't support MFX_EXTBUFF_CODING_OPTION3.

zgc> This patch fixes bug #5324.
zgc> ---
zgc>  libavcodec/qsvenc.c | 18 --
zgc>  1 file changed, 18 deletions(-)

zgc> diff --git a/libavcodec/qsvenc.c b/libavcodec/qsvenc.c
zgc> index 132cf47..a561df8 100644
zgc> --- a/libavcodec/qsvenc.c
zgc> +++ b/libavcodec/qsvenc.c
zgc> @@ -132,9 +132,6 @@ static void dump_video_param(AVCodecContext *avctx, 
QSVEncContext *q,
zgc>  #if QSV_HAVE_CO2
zgc>  mfxExtCodingOption2 *co2 = (mfxExtCodingOption2*)coding_opts[1];
zgc>  #endif
zgc> -#if QSV_HAVE_CO3
zgc> -mfxExtCodingOption3 *co3 = (mfxExtCodingOption3*)coding_opts[2];
zgc> -#endif
zgc>  
zgc>  av_log(avctx, AV_LOG_VERBOSE, "profile: %s; level: %"PRIu16"\n",
zgc> print_profile(info->CodecProfile), info->CodecLevel);
zgc> @@ -186,12 +183,6 @@ static void dump_video_param(AVCodecContext *avctx, 
QSVEncContext *q,
zgc> info->ICQQuality, co2->LookAheadDepth);
zgc>  }
zgc>  #endif
zgc> -#if QSV_HAVE_QVBR
-else if (info->>RateControlMethod == MFX_RATECONTROL_QVBR) {
zgc> -av_log(avctx, AV_LOG_VERBOSE, "QVBRQuality: %"PRIu16"\n",
zgc> -   co3->QVBRQuality);
zgc> -}
zgc> -#endif
zgc>  
zgc>  av_log(avctx, AV_LOG_VERBOSE, "NumSlice: %"PRIu16"; NumRefFrame: 
%"PRIu16"\n",
zgc> info->NumSlice, info->NumRefFrame);
zgc> @@ -577,12 +568,6 @@ static int qsv_retrieve_enc_params(AVCodecContext 
*avctx, QSVEncContext *q)
zgc>  .Header.BufferSz = sizeof(co2),
zgc>  };
zgc>  #endif
zgc> -#if QSV_HAVE_CO3
zgc> -mfxExtCodingOption3 co3 = {
zgc> -.Header.BufferId = MFX_EXTBUFF_CODING_OPTION3,
zgc> -.Header.BufferSz = sizeof(co3),
zgc> -};
zgc> -#endif
zgc>  
zgc>  mfxExtBuffer *ext_buffers[] = {
zgc>  (mfxExtBuffer*),
zgc> @@ -590,9 +575,6 @@ static int qsv_retrieve_enc_params(AVCodecContext 
*avctx, QSVEncContext *q)
zgc>  #if QSV_HAVE_CO2
zgc>  (mfxExtBuffer*),
zgc>  #endif
zgc> -#if QSV_HAVE_CO3
zgc> -(mfxExtBuffer*),
zgc> -#endif
zgc>  };
zgc>  
zgc>  int need_pps = avctx->codec_id != AV_CODEC_ID_MPEG2VIDEO;
First, I'm not sure that "4th generation Intel CPUs don't support 
MFX_EXTBUFF_CODING_OPTION3".
Do you have any reference which says it? The documentation says:
"This structure is available since SDK API 1.11.", i.e. it depended by API
level but not by CPU generation.
For any case I believe it is wrong to remove code if some issues are appearing
on old CPUs. Especially if we already have got a pre-processor condition to 
disable
MFX_EXTBUFF_CODING_OPTION3 by correct way. If you have got some
constructive suggestions to improve disabling MFX_EXTBUFF_CODING_OPTION3 by
condition (possible in real-time code, not by preprocessor) please provide.
But stupid disabling of a feature is not the good solution.

-- 
Best regards,
 Ivanmailto:ivan.us...@nablet.com

___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

[FFmpeg-devel] [PATCH] qsvdec.c: Revert "Merge commit '3c53627ac17fc6bdea5029be57da1e03b32d265d'"

[Ivan Uskov]

Hello All,

This reverts commit d30cf57a7b2097b565db02ecfffbdc9c16423d0e, reversing
changes made to acc155ac55baa95d1c16c0364b02244bc04d83a8.
The commit  d30cf57a7b2097b565db02ecfffbdc9c16423d0e provided irrelevant
code complexity and decoding slowdown. But main disadvantage of this commit is 
decoder crash.
So it should be reverted.
Please review.

-- 
Best regards,
 Ivan  mailto:ivan.us...@nablet.com

0001-Revert-Merge-commit-3c53627ac17fc6bdea5029be57da1e03.patch
Description: Binary data
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] libavcodec : add psd image file decoder

[Martin Vignali]

2016-07-24 16:36 GMT+02:00 Moritz Barsnick :

> On Sun, Jul 24, 2016 at 15:21:00 +0200, Martin Vignali wrote:
> > Subject: [PATCH] libavcodec : add decoder for .psd image file.
> [...]
> > +- Psd Decoder
> [...]
> > +@item PSD  @tab   @tab X
> > +@tab Photoshop
> [...]
> > +.long_name = NULL_IF_CONFIG_SMALL("Photoshop file"),
>
> ".psd", "psd", "Psd", "Photoshop file". I think some consistency would
> be nice. E.g. use "Photoshop PSD file" as long name, "PSD" in the
> descriptions (and not the name of the extension with dot).
>
> But that's just cosmetics...
>
> > +if ((s->channel_count < 1)||(s->channel_count > 56)) {
> [...]
> > +if ((s->height < 1)||(s->height > 3)) {
> [...]
> > +if ((s->width < 1)||(s->width > 3)) {
>
> You could leave whitespace around the operators, for readability.
>
> > +avpriv_report_missing_feature(avctx, "channel depth
> unsupported for rgb %d", s->channel_depth);
>
> For the sake of "grammar", I would say:
> avpriv_report_missing_feature(avctx, "channel depth %d
> unsupported for rgb", s->channel_depth);
> (Same with channel_count and with grayscale below.)
>
> > +/* reorganize uncompress data. each channel is store one after the
> other */
>   ^E   ^stored
>
> > +if ((b[4] == 0)&&(b[5] == 1)) {/* version 1 is PSD, version 2 is
> PSB */
> [...]
> > +if ((AV_RL32(b+6) == 0)&&(AV_RL16(b+10) == 0))/* reserved must be 0
> */
> [...]
> > +if ((color_mode <= 9)&&(color_mode != 5)&&(color_mode != 6))
> Whitespace, as mentioned above.
>
>
> Correct locally.
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] libavcodec : add psd image file decoder

[Moritz Barsnick]

On Sun, Jul 24, 2016 at 15:21:00 +0200, Martin Vignali wrote:
> Subject: [PATCH] libavcodec : add decoder for .psd image file.
[...]
> +- Psd Decoder
[...]
> +@item PSD  @tab   @tab X
> +@tab Photoshop
[...]
> +.long_name = NULL_IF_CONFIG_SMALL("Photoshop file"),

".psd", "psd", "Psd", "Photoshop file". I think some consistency would
be nice. E.g. use "Photoshop PSD file" as long name, "PSD" in the
descriptions (and not the name of the extension with dot).

But that's just cosmetics...

> +if ((s->channel_count < 1)||(s->channel_count > 56)) {
[...]
> +if ((s->height < 1)||(s->height > 3)) {
[...]
> +if ((s->width < 1)||(s->width > 3)) {

You could leave whitespace around the operators, for readability.

> +avpriv_report_missing_feature(avctx, "channel depth 
> unsupported for rgb %d", s->channel_depth);

For the sake of "grammar", I would say:
avpriv_report_missing_feature(avctx, "channel depth %d 
unsupported for rgb", s->channel_depth);
(Same with channel_count and with grayscale below.)

> +/* reorganize uncompress data. each channel is store one after the other 
> */
  ^E   ^stored 

> +if ((b[4] == 0)&&(b[5] == 1)) {/* version 1 is PSD, version 2 is PSB */
[...]
> +if ((AV_RL32(b+6) == 0)&&(AV_RL16(b+10) == 0))/* reserved must be 0 */
[...]
> +if ((color_mode <= 9)&&(color_mode != 5)&&(color_mode != 6))
Whitespace, as mentioned above.

Moritz
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] Questions about Video Memory in qsv decoders

[Ivan Uskov]

Hello Michael,

Sunday, July 24, 2016, 12:16:59 AM, you wrote:

MN> On Sat, Jul 23, 2016 at 11:19:59PM +0300, Ivan Uskov wrote:
>> Hello 张玉晓,
>> 
>> Friday, July 22, 2016, 4:10:36 AM, you wrote:
>> 
>> 张> I have a question when learning ffmpeg qsv decoder and Intel media sdk.
>> 
>> 张> The intel media sdk suggest to use Video Memory while doing Hardware
>> 张> decoding, use System Memory while doing Software decoding.
>> 
>> 张> FFmpeg only used System Memory to decode with qsv (Intel msdk) in file
>> 张> libavcodec/qsvdec.c. Is there any problem when use Video Memory? 
>> 
>> As qsv modules maintainer I can confirm, yes, there are problems.
>> 
>> First (objective) problem that in common case there are 3 qsv components
>> should be linked by a pipeline through GPU memory: 
>> *qsv video decoder
>> *qsv vpp filter
>> *qsv video encoder
>> But current ffmpeg architecture have restrictions to share one common qsv
>> session between libavcodec and libavfilter.
>> 
>> Second  (subjective) problem that nobody in ffmpeg community is interested to
>> advance  QSV  codecs. Most of  patches at this year provided by me or by my
>> colleagues (including tries to solve issue with GPU memory) were rejected or
>> ignored silently. Moreover, if you will try to clone a latest ffmpeg you 
>> found qsv
>> decoder broken, it crashes at any try of use. The patch to fix it was posted
>> by me at April 25 but was ignored.

MN> which patches should be applied to fix the issues ?
MN> can you repost them ?
Currently I believe that broken merge from libav should be not improved/fixed
by some way but just reverted because it provides nothing helpful, only
complexity. I will provide new patch soon.

MN> [...]



-- 
Best regards,
 Ivanmailto:ivan.us...@nablet.com

___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] [PATCH] libavcodec/qsvdec.c: Restoring decoding functionality after unsuccessful merge from libav.

[Ivan Uskov]

Hello Mark,

Saturday, July 23, 2016, 11:42:46 PM, you wrote:

MT> On 23/07/16 20:33, Ivan Uskov wrote:
>> If  you are use qsv, I would like to recommend to roll-back to version before
>> d30cf57a7b2097b565db02ecfffbdc9c16423d0e
>> Really  the  d30cf57a7b2097b565db02ecfffbdc9c16423d0e is useless and only
>> makes code complex and work slow, the sync variable is not mandatory to be
>> allocated  on  heap  at  all.  libav  guys  did a big mistake when have added
>> such "feature".

MT> Are you sure that works correctly with all streams, maybe one with some
MT> sort of delay?  It looks like you can add multiple things to the fifo in
MT> one call and but then only ever remove one, so using the stack to hold
MT> them across invocations is not appropriate (though failure may be very 
subtle).
Yes, I'm sure.
There are following arguments:
1. Intel's samples by itself use general variable for the 'sync' arg, not
heap allocated.
2. Intel documentation says nothing special for this arg.
3. I did several qsv-based transcoders for windows and linux no any issues
without this questionable trick.
4. A fifo object used previously to keep 'sync' handles already keeps data on
heap :-)

There only one known issue into initial implementation, the fifo size may be not
enough an there is no check for overflow. So at some cases wrapper may
overwrite sync entries in fifo and some decoded frames may be lost. But it is
not related with discussed problem.

MT> (Note that I'm not sufficiently familiar with Media SDK to confidently
MT> assert anything about how this code should behave, so please don't read too 
much into what I say.)

MT> In general I think if you are happy that the previous implementation in
MT> ffmpeg was correct then please do revert, because the code as it is now is 
completely broken.
Ok, will do.

-- 
Best regards,
 Ivanmailto:ivan.us...@nablet.com

___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] libavcodec : add psd image file decoder

[Martin Vignali]

> breaks fate
>
> make: *** [fate-filter-metadata-ebur128] Error 1
> --- -   2016-07-23 18:50:11.633752058 +0200
> +++ tests/data/fate/probe-format-roundup14142016-07-23
> 18:50:10.913635588 +0200
> @@ -1 +1 @@
> -mpeg
> +psd_pipe
> Test probe-format-roundup1414 failed. Look at
> tests/data/fate/probe-format-roundup1414.err for details.
> make: *** [fate-probe-format-roundup1414] Error 1
> --- -   2016-07-23 18:50:11.657384876 +0200
> +++ tests/data/fate/probe-format-roundup997 2016-07-23
> 18:50:10.917635588 +0200
> @@ -1 +1 @@
> -mpeg
> +psd_pipe
> Test probe-format-roundup997 failed. Look at
> tests/data/fate/probe-format-roundup997.err for details.
> make: *** [fate-probe-format-roundup997] Error 1
> --- -   2016-07-23 18:50:11.667958765 +0200
> +++ tests/data/fate/probe-format-roundup13832016-07-23
> 18:50:10.917635588 +0200
> @@ -1 +1 @@
> -mp3
> +psd_pipe
> Test probe-format-roundup1383 failed. Look at
> tests/data/fate/probe-format-roundup1383.err for details.
> make: *** [fate-probe-format-roundup1383] Error 1
> make: Target `fate' not remade because of errors.
>
>
Hello,

Corrected patch in attach who fix the psd_probe function.

Comments welcome

Martin
From 2b368b8592cf0c278da0954e41705ddfc0f342b6 Mon Sep 17 00:00:00 2001
From: Martin Vignali 
Date: Sun, 24 Jul 2016 15:17:18 +0200
Subject: [PATCH] libavcodec : add decoder for .psd image file.

Decode the Image Data Section (who contain merge picture).
Support RGB/A and Grayscale/A in 8bits and 16 bits by channel.
Support uncompress and rle compression in Image Data Section
---
 Changelog|   1 +
 doc/general.texi |   2 +
 libavcodec/Makefile  |   1 +
 libavcodec/allcodecs.c   |   1 +
 libavcodec/avcodec.h |   1 +
 libavcodec/codec_desc.c  |   7 +
 libavcodec/psd.c | 395 +++
 libavformat/Makefile |   1 +
 libavformat/allformats.c |   1 +
 libavformat/img2.c   |   1 +
 libavformat/img2dec.c|  35 +
 11 files changed, 446 insertions(+)
 create mode 100644 libavcodec/psd.c

diff --git a/Changelog b/Changelog
index 99cdb80..97b11f1 100644
--- a/Changelog
+++ b/Changelog
@@ -3,6 +3,7 @@ releases are sorted from youngest to oldest.
 
 version :
 
+- Psd Decoder
 
 version 3.1:
 - DXVA2-accelerated HEVC Main10 decoding
diff --git a/doc/general.texi b/doc/general.texi
index 4db209f..06775f4 100644
--- a/doc/general.texi
+++ b/doc/general.texi
@@ -577,6 +577,8 @@ following image formats are supported:
 @item PNG  @tab X @tab X
 @item PPM  @tab X @tab X
 @tab Portable PixelMap image
+@item PSD  @tab   @tab X
+@tab Photoshop
 @item PTX  @tab   @tab X
 @tab V.Flash PTX format
 @item SGI  @tab X @tab X
diff --git a/libavcodec/Makefile b/libavcodec/Makefile
index fd0d1f0..913b4a8 100644
--- a/libavcodec/Makefile
+++ b/libavcodec/Makefile
@@ -456,6 +456,7 @@ OBJS-$(CONFIG_PRORES_LGPL_DECODER) += proresdec_lgpl.o proresdsp.o proresdat
 OBJS-$(CONFIG_PRORES_ENCODER)  += proresenc_anatoliy.o
 OBJS-$(CONFIG_PRORES_AW_ENCODER)   += proresenc_anatoliy.o
 OBJS-$(CONFIG_PRORES_KS_ENCODER)   += proresenc_kostya.o proresdata.o
+OBJS-$(CONFIG_PSD_DECODER) += psd.o
 OBJS-$(CONFIG_PTX_DECODER) += ptx.o
 OBJS-$(CONFIG_QCELP_DECODER)   += qcelpdec.o \
   celp_filters.o acelp_vectors.o \
diff --git a/libavcodec/allcodecs.c b/libavcodec/allcodecs.c
index 54efaad..458f603 100644
--- a/libavcodec/allcodecs.c
+++ b/libavcodec/allcodecs.c
@@ -278,6 +278,7 @@ void avcodec_register_all(void)
 REGISTER_ENCODER(PRORES_AW, prores_aw);
 REGISTER_ENCODER(PRORES_KS, prores_ks);
 REGISTER_DECODER(PRORES_LGPL,   prores_lgpl);
+REGISTER_DECODER(PSD,   psd);
 REGISTER_DECODER(PTX,   ptx);
 REGISTER_DECODER(QDRAW, qdraw);
 REGISTER_DECODER(QPEG,  qpeg);
diff --git a/libavcodec/avcodec.h b/libavcodec/avcodec.h
index 39713ed..75a9970 100644
--- a/libavcodec/avcodec.h
+++ b/libavcodec/avcodec.h
@@ -409,6 +409,7 @@ enum AVCodecID {
 AV_CODEC_ID_MAGICYUV,
 AV_CODEC_ID_SHEERVIDEO,
 AV_CODEC_ID_YLC,
+AV_CODEC_ID_PSD,
 
 /* various PCM "codecs" */
 AV_CODEC_ID_FIRST_AUDIO = 0x1, ///< A dummy id pointing at the start of audio codecs
diff --git a/libavcodec/codec_desc.c b/libavcodec/codec_desc.c
index 9d94b72..427553c 100644
--- a/libavcodec/codec_desc.c
+++ b/libavcodec/codec_desc.c
@@ -1425,6 +1425,13 @@ static const AVCodecDescriptor codec_descriptors[] = {
 .props = AV_CODEC_PROP_INTRA_ONLY | AV_CODEC_PROP_LOSSLESS,
 },
 {
+.id= AV_CODEC_ID_PSD,
+.type  = AVMEDIA_TYPE_VIDEO,
+.name  = "psd",
+.long_name = NULL_IF_CONFIG_SMALL("Photoshop file"),
+.props = AV_CODEC_PROP_INTRA_ONLY | 

[FFmpeg-devel] [PATCH 4/4] lavc: add vp8 mediacodec decoder

[Thomas Volkert]

From: Thomas Volkert 

---
 Changelog  |  1 +
 configure  |  1 +
 libavcodec/Makefile|  1 +
 libavcodec/allcodecs.c |  1 +
 libavcodec/mediacodecdec_vp8.c | 89 ++
 libavcodec/version.h   |  2 +-
 6 files changed, 94 insertions(+), 1 deletion(-)
 create mode 100644 libavcodec/mediacodecdec_vp8.c

diff --git a/Changelog b/Changelog
index e35224c..6eaec1b 100644
--- a/Changelog
+++ b/Changelog
@@ -9,6 +9,7 @@ version :
 - VP8 in Ogg muxing
 - MediaCodec MPEG-4 decoding
 - MediaCodec H.263 decoding
+- MediaCodec VP8 decoding
 
 
 version 3.1:
diff --git a/configure b/configure
index 18cc5f2..445d975 100755
--- a/configure
+++ b/configure
@@ -2655,6 +2655,7 @@ vc1_vdpau_hwaccel_deps="vdpau"
 vc1_vdpau_hwaccel_select="vc1_decoder"
 vp8_cuvid_hwaccel_deps="cuda cuvid CUVIDVP9PICPARAMS"
 vp9_cuvid_hwaccel_deps="cuda cuvid CUVIDVP9PICPARAMS"
+vp8_mediacodec_decoder_deps="mediacodec"
 vp9_d3d11va_hwaccel_deps="d3d11va DXVA_PicParams_VP9"
 vp9_d3d11va_hwaccel_select="vp9_decoder"
 vp9_dxva2_hwaccel_deps="dxva2 DXVA_PicParams_VP9"
diff --git a/libavcodec/Makefile b/libavcodec/Makefile
index 0bc0211..466c5a4 100644
--- a/libavcodec/Makefile
+++ b/libavcodec/Makefile
@@ -593,6 +593,7 @@ OBJS-$(CONFIG_VP6_DECODER) += vp6.o vp56.o 
vp56data.o \
 OBJS-$(CONFIG_VP7_DECODER) += vp8.o vp56rac.o
 OBJS-$(CONFIG_VP8_DECODER) += vp8.o vp56rac.o
 OBJS-$(CONFIG_VP8_CUVID_DECODER)   += cuvid.o
+OBJS-$(CONFIG_VP8_MEDIACODEC_DECODER)  += mediacodecdec_vp8.o
 OBJS-$(CONFIG_VP9_DECODER) += vp9.o vp9dsp.o vp56rac.o 
vp9dsp_8bpp.o \
   vp9dsp_10bpp.o vp9dsp_12bpp.o
 OBJS-$(CONFIG_VP9_CUVID_DECODER)   += cuvid.o
diff --git a/libavcodec/allcodecs.c b/libavcodec/allcodecs.c
index db2f25d..b750c0e 100644
--- a/libavcodec/allcodecs.c
+++ b/libavcodec/allcodecs.c
@@ -353,6 +353,7 @@ void avcodec_register_all(void)
 REGISTER_DECODER(VP6F,  vp6f);
 REGISTER_DECODER(VP7,   vp7);
 REGISTER_DECODER(VP8,   vp8);
+REGISTER_DECODER(VP8_MEDIACODEC,vp8_mediacodec);
 REGISTER_DECODER(VP9,   vp9);
 REGISTER_DECODER(VQA,   vqa);
 REGISTER_DECODER(WEBP,  webp);
diff --git a/libavcodec/mediacodecdec_vp8.c b/libavcodec/mediacodecdec_vp8.c
new file mode 100644
index 000..6beac70
--- /dev/null
+++ b/libavcodec/mediacodecdec_vp8.c
@@ -0,0 +1,89 @@
+/*
+ * Android MediaCodec VP8 decoder
+ *
+ * Copyright (c) 2016 Thomas Volkert 
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include "libavutil/avassert.h"
+#include "libavutil/common.h"
+#include "libavutil/fifo.h"
+
+#include "avcodec.h"
+#include "internal.h"
+#include "mediacodecdec.h"
+#include "mediacodec_wrapper.h"
+
+#define CODEC_MIME "video/x-vnd.on2.vp8"
+
+static av_cold int mediacodec_decoder_init_vp8(AVCodecContext *avctx)
+{
+int ret;
+
+FFAMediaFormat *format = NULL;
+MediaCodecDecContext *s = avctx->priv_data;
+
+format = ff_AMediaFormat_new();
+if (!format) {
+av_log(avctx, AV_LOG_ERROR, "Failed to create media format\n");
+ret = AVERROR_EXTERNAL;
+goto done;
+}
+
+ff_AMediaFormat_setString(format, "mime", CODEC_MIME);
+ff_AMediaFormat_setInt32(format, "width", avctx->width);
+ff_AMediaFormat_setInt32(format, "height", avctx->height);
+
+if ((ret = ff_mediacodec_dec_init(avctx, s, CODEC_MIME, format)) < 0) {
+goto done;
+}
+
+av_log(avctx, AV_LOG_INFO, "MediaCodec started successfully, ret = %d\n", 
ret);
+
+s->fifo = av_fifo_alloc(sizeof(AVPacket));
+if (!s->fifo) {
+ret = AVERROR(ENOMEM);
+goto done;
+}
+
+av_init_packet(>output_pkt);
+
+done:
+if (format) {
+ff_AMediaFormat_delete(format);
+}
+
+if (ret < 0) {
+   ff_mediacodec_decoder_close(avctx);
+}
+
+return ret;
+}
+
+AVCodec ff_vp8_mediacodec_decoder = {
+.name   = "vp8_mediacodec",
+.long_name  = NULL_IF_CONFIG_SMALL("VP8 Android MediaCodec decoder"),
+.type   = AVMEDIA_TYPE_VIDEO,
+.id  

[FFmpeg-devel] [PATCH 1/4] lavc: add mpeg4 mediacodec decoder

[Thomas Volkert]

From: Thomas Volkert 

---
 Changelog|   1 +
 configure|   1 +
 libavcodec/Makefile  |   1 +
 libavcodec/allcodecs.c   |   1 +
 libavcodec/mediacodecdec_mpeg4.c | 239 +++
 libavcodec/version.h |   2 +-
 6 files changed, 244 insertions(+), 1 deletion(-)
 create mode 100644 libavcodec/mediacodecdec_mpeg4.c

diff --git a/Changelog b/Changelog
index 2bd18ec..b8bbdb9 100644
--- a/Changelog
+++ b/Changelog
@@ -7,6 +7,7 @@ version :
 - Changed metadata print option to accept general urls
 - Alias muxer for Ogg Video (.ogv)
 - VP8 in Ogg muxing
+- MediaCodec MPEG-4 decoding
 
 
 version 3.1:
diff --git a/configure b/configure
index 1b41303..9004b06 100755
--- a/configure
+++ b/configure
@@ -2621,6 +2621,7 @@ mpeg2_videotoolbox_hwaccel_select="mpeg2video_decoder"
 mpeg2_xvmc_hwaccel_deps="xvmc"
 mpeg2_xvmc_hwaccel_select="mpeg2video_decoder"
 mpeg4_crystalhd_decoder_select="crystalhd"
+mpeg4_mediacodec_decoder_deps="mediacodec"
 mpeg4_mmal_decoder_deps="mmal"
 mpeg4_mmal_decoder_select="mmal"
 mpeg4_mmal_hwaccel_deps="mmal"
diff --git a/libavcodec/Makefile b/libavcodec/Makefile
index abef19e..642cf2a 100644
--- a/libavcodec/Makefile
+++ b/libavcodec/Makefile
@@ -317,6 +317,7 @@ OBJS-$(CONFIG_H264_DECODER)+= h264.o 
h264_cabac.o h264_cavlc.o \
   h2645_parse.o
 OBJS-$(CONFIG_H264_CUVID_DECODER)  += cuvid.o
 OBJS-$(CONFIG_H264_MEDIACODEC_DECODER) += mediacodecdec_h264.o
+OBJS-$(CONFIG_MPEG4_MEDIACODEC_DECODER) += mediacodecdec_mpeg4.o
 OBJS-$(CONFIG_H264_MMAL_DECODER)   += mmaldec.o
 OBJS-$(CONFIG_H264_NVENC_ENCODER)  += nvenc_h264.o
 OBJS-$(CONFIG_NVENC_ENCODER)   += nvenc_h264.o
diff --git a/libavcodec/allcodecs.c b/libavcodec/allcodecs.c
index 951e199..2d98694 100644
--- a/libavcodec/allcodecs.c
+++ b/libavcodec/allcodecs.c
@@ -239,6 +239,7 @@ void avcodec_register_all(void)
 REGISTER_ENCDEC (MPEG2VIDEO,mpeg2video);
 REGISTER_ENCDEC (MPEG4, mpeg4);
 REGISTER_DECODER(MPEG4_CRYSTALHD,   mpeg4_crystalhd);
+REGISTER_DECODER(MPEG4_MEDIACODEC,  mpeg4_mediacodec);
 REGISTER_DECODER(MPEG4_MMAL,mpeg4_mmal);
 #if FF_API_VDPAU
 REGISTER_DECODER(MPEG4_VDPAU,   mpeg4_vdpau);
diff --git a/libavcodec/mediacodecdec_mpeg4.c b/libavcodec/mediacodecdec_mpeg4.c
new file mode 100644
index 000..7c4559b
--- /dev/null
+++ b/libavcodec/mediacodecdec_mpeg4.c
@@ -0,0 +1,239 @@
+/*
+ * Android MediaCodec MPEG 4 decoder
+ *
+ * Copyright (c) 2016 Thomas Volkert 
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include "libavutil/avassert.h"
+#include "libavutil/common.h"
+#include "libavutil/fifo.h"
+
+#include "avcodec.h"
+#include "internal.h"
+#include "mediacodecdec.h"
+#include "mediacodec_wrapper.h"
+
+#define CODEC_MIME "video/mp4v-es"
+
+typedef struct MediaCodecMPEG4DecContext {
+
+MediaCodecDecContext *ctx;
+
+AVBSFContext *bsf;
+
+AVFifoBuffer *fifo;
+
+AVPacket filtered_pkt;
+
+} MediaCodecMPEG4DecContext;
+
+static av_cold int mediacodec_decode_close(AVCodecContext *avctx)
+{
+MediaCodecMPEG4DecContext *s = avctx->priv_data;
+
+ff_mediacodec_dec_close(avctx, s->ctx);
+s->ctx = NULL;
+
+av_fifo_free(s->fifo);
+
+av_bsf_free(>bsf);
+av_packet_unref(>filtered_pkt);
+
+return 0;
+}
+
+static av_cold int mediacodec_decode_init(AVCodecContext *avctx)
+{
+int ret;
+
+FFAMediaFormat *format = NULL;
+MediaCodecMPEG4DecContext *s = avctx->priv_data;
+
+format = ff_AMediaFormat_new();
+if (!format) {
+av_log(avctx, AV_LOG_ERROR, "Failed to create media format\n");
+ret = AVERROR_EXTERNAL;
+goto done;
+}
+
+ff_AMediaFormat_setString(format, "mime", CODEC_MIME);
+ff_AMediaFormat_setInt32(format, "width", avctx->width);
+ff_AMediaFormat_setInt32(format, "height", avctx->height);
+
+s->ctx = av_mallocz(sizeof(*s->ctx));
+if (!s->ctx) {
+av_log(avctx, AV_LOG_ERROR, "Failed to allocate 
MediaCodecDecContext\n");
+ret = AVERROR(ENOMEM);
+goto done;
+}
+
+if ((ret = ff_mediacodec_dec_init(avctx, s->ctx, 

[FFmpeg-devel] [PATCH 3/4] lavc: add h.263 mediacodec decoder

[Thomas Volkert]

From: Thomas Volkert 

---
 Changelog   |  1 +
 configure   |  1 +
 libavcodec/Makefile |  1 +
 libavcodec/allcodecs.c  |  1 +
 libavcodec/mediacodecdec_h263.c | 90 +
 libavcodec/version.h|  2 +-
 6 files changed, 95 insertions(+), 1 deletion(-)
 create mode 100644 libavcodec/mediacodecdec_h263.c

diff --git a/Changelog b/Changelog
index b8bbdb9..e35224c 100644
--- a/Changelog
+++ b/Changelog
@@ -8,6 +8,7 @@ version :
 - Alias muxer for Ogg Video (.ogv)
 - VP8 in Ogg muxing
 - MediaCodec MPEG-4 decoding
+- MediaCodec H.263 decoding
 
 
 version 3.1:
diff --git a/configure b/configure
index 9004b06..18cc5f2 100755
--- a/configure
+++ b/configure
@@ -2548,6 +2548,7 @@ videotoolbox_hwaccel_deps="videotoolbox pthreads"
 videotoolbox_hwaccel_extralibs="-framework QuartzCore"
 xvmc_deps="X11_extensions_XvMClib_h"
 
+h263_mediacodec_decoder_deps="mediacodec"
 h263_vaapi_hwaccel_deps="vaapi"
 h263_vaapi_hwaccel_select="h263_decoder"
 h263_videotoolbox_hwaccel_deps="videotoolbox"
diff --git a/libavcodec/Makefile b/libavcodec/Makefile
index 642cf2a..0bc0211 100644
--- a/libavcodec/Makefile
+++ b/libavcodec/Makefile
@@ -309,6 +309,7 @@ OBJS-$(CONFIG_H263_DECODER)+= h263dec.o h263.o 
ituh263dec.o\
   intelh263dec.o h263data.o
 OBJS-$(CONFIG_H263_ENCODER)+= mpeg4videoenc.o mpeg4video.o  \
   h263.o ituh263enc.o flvenc.o 
h263data.o
+OBJS-$(CONFIG_H263_MEDIACODEC_DECODER) += mediacodecdec_h263.o
 OBJS-$(CONFIG_H264_DECODER)+= h264.o h264_cabac.o h264_cavlc.o \
   h264_direct.o h264_loopfilter.o  \
   h264_mb.o h264_picture.o h264_ps.o \
diff --git a/libavcodec/allcodecs.c b/libavcodec/allcodecs.c
index 2d98694..db2f25d 100644
--- a/libavcodec/allcodecs.c
+++ b/libavcodec/allcodecs.c
@@ -193,6 +193,7 @@ void avcodec_register_all(void)
 REGISTER_ENCDEC (H263,  h263);
 REGISTER_DECODER(H263I, h263i);
 REGISTER_ENCDEC (H263P, h263p);
+REGISTER_DECODER(H263_MEDIACODEC,   h263_mediacodec);
 REGISTER_DECODER(H264,  h264);
 REGISTER_DECODER(H264_CRYSTALHD,h264_crystalhd);
 REGISTER_DECODER(H264_MEDIACODEC,   h264_mediacodec);
diff --git a/libavcodec/mediacodecdec_h263.c b/libavcodec/mediacodecdec_h263.c
new file mode 100644
index 000..5f2c412
--- /dev/null
+++ b/libavcodec/mediacodecdec_h263.c
@@ -0,0 +1,90 @@
+/*
+ * Android MediaCodec H.263 decoder
+ *
+ * Copyright (c) 2016 Thomas Volkert 
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include "libavutil/avassert.h"
+#include "libavutil/common.h"
+#include "libavutil/fifo.h"
+
+#include "avcodec.h"
+#include "internal.h"
+#include "mediacodecdec.h"
+#include "mediacodec_wrapper.h"
+
+#define CODEC_MIME "video/3gpp"
+
+static av_cold int mediacodec_decoder_init_h263(AVCodecContext *avctx)
+{
+int ret;
+
+FFAMediaFormat *format = NULL;
+MediaCodecDecContext *s = avctx->priv_data;
+
+format = ff_AMediaFormat_new();
+if (!format) {
+av_log(avctx, AV_LOG_ERROR, "Failed to create media format\n");
+ret = AVERROR_EXTERNAL;
+goto done;
+}
+
+ff_AMediaFormat_setString(format, "mime", CODEC_MIME);
+ff_AMediaFormat_setInt32(format, "width", avctx->width);
+ff_AMediaFormat_setInt32(format, "height", avctx->height);
+
+if ((ret = ff_mediacodec_dec_init(avctx, s, CODEC_MIME, format)) < 0) {
+goto done;
+}
+
+av_log(avctx, AV_LOG_INFO, "MediaCodec started successfully, ret = %d\n", 
ret);
+
+s->fifo = av_fifo_alloc(sizeof(AVPacket));
+if (!s->fifo) {
+ret = AVERROR(ENOMEM);
+goto done;
+}
+
+av_init_packet(>output_pkt);
+
+done:
+if (format) {
+ff_AMediaFormat_delete(format);
+}
+
+if (ret < 0) {
+   ff_mediacodec_decoder_close(avctx);
+}
+
+return ret;
+}
+
+AVCodec ff_h263_mediacodec_decoder = {
+.name   = "h263_mediacodec",
+.long_name  = NULL_IF_CONFIG_SMALL("H.263 

[FFmpeg-devel] [PATCH 2/4] lavc: share more code between h.264 and mpeg4 decoder

[Thomas Volkert]

From: Thomas Volkert 

---
 libavcodec/mediacodecdec.c   | 140 ++-
 libavcodec/mediacodecdec.h   |  12 +++
 libavcodec/mediacodecdec_h264.c  | 175 +++
 libavcodec/mediacodecdec_mpeg4.c | 154 +++---
 4 files changed, 170 insertions(+), 311 deletions(-)

diff --git a/libavcodec/mediacodecdec.c b/libavcodec/mediacodecdec.c
index df60104..d780ca4 100644
--- a/libavcodec/mediacodecdec.c
+++ b/libavcodec/mediacodecdec.c
@@ -23,6 +23,7 @@
 #include 
 #include 
 
+#include "libavutil/avassert.h"
 #include "libavutil/atomic.h"
 #include "libavutil/common.h"
 #include "libavutil/mem.h"
@@ -152,6 +153,8 @@ static void ff_mediacodec_dec_unref(MediaCodecDecContext *s)
 return;
 
 if (!avpriv_atomic_int_add_and_fetch(>refcount, -1)) {
+av_freep(>codec_name);
+
 if (s->codec) {
 ff_AMediaCodec_delete(s->codec);
 s->codec = NULL;
@@ -167,8 +170,17 @@ static void ff_mediacodec_dec_unref(MediaCodecDecContext 
*s)
 s->surface = NULL;
 }
 
-av_freep(>codec_name);
-av_freep();
+if (s->bsf) {
+av_bsf_free(>bsf);
+   s->bsf = NULL;
+}
+
+if (s->fifo) {
+   av_fifo_free(s->fifo);
+   s->fifo = NULL;
+}
+
+av_packet_unref(>output_pkt);
 }
 }
 
@@ -752,6 +764,130 @@ int ff_mediacodec_dec_is_flushing(AVCodecContext *avctx, 
MediaCodecDecContext *s
 return s->flushing;
 }
 
+void ff_mediacodec_decoder_flush(AVCodecContext *avctx)
+{
+   MediaCodecDecContext *s = avctx->priv_data;
+
+while (av_fifo_size(s->fifo)) {
+AVPacket pkt;
+av_fifo_generic_read(s->fifo, , sizeof(pkt), NULL);
+av_packet_unref();
+}
+av_fifo_reset(s->fifo);
+
+av_packet_unref(>output_pkt);
+
+ff_mediacodec_dec_flush(avctx, s);
+}
+
+int ff_mediacodec_decoder_decode(AVCodecContext *avctx, void *data,
+ int *got_frame, AVPacket *avpkt)
+{
+   MediaCodecDecContext *s = avctx->priv_data;
+AVFrame *frame= data;
+int ret;
+
+/* buffer the input packet */
+if (avpkt->size) {
+AVPacket input_pkt = { 0 };
+
+if (av_fifo_space(s->fifo) < sizeof(input_pkt)) {
+ret = av_fifo_realloc2(s->fifo,
+   av_fifo_size(s->fifo) + sizeof(input_pkt));
+if (ret < 0)
+return ret;
+}
+
+ret = av_packet_ref(_pkt, avpkt);
+if (ret < 0)
+return ret;
+av_fifo_generic_write(s->fifo, _pkt, sizeof(input_pkt), NULL);
+}
+
+/*
+ * MediaCodec.flush() discards both input and output buffers, thus we
+ * need to delay the call to this function until the user has released or
+ * renderered the frames he retains.
+ *
+ * After we have buffered an input packet, check if the codec is in the
+ * flushing state. If it is, we need to call ff_mediacodec_dec_flush.
+ *
+ * ff_mediacodec_dec_flush returns 0 if the flush cannot be performed on
+ * the codec (because the user retains frames). The codec stays in the
+ * flushing state.
+ *
+ * ff_mediacodec_dec_flush returns 1 if the flush can actually be
+ * performed on the codec. The codec leaves the flushing state and can
+ * process again packets.
+ *
+ * ff_mediacodec_dec_flush returns a negative value if an error has
+ * occurred.
+ *
+ */
+if (ff_mediacodec_dec_is_flushing(avctx, s)) {
+if (!ff_mediacodec_dec_flush(avctx, s)) {
+return avpkt->size;
+}
+}
+
+/* process buffered data */
+while (!*got_frame) {
+/* prepare the input data -- execute BSF if needed */
+if (s->output_pkt.size <= 0) {
+av_packet_unref(>output_pkt);
+
+/* no more data */
+if (av_fifo_size(s->fifo) < sizeof(AVPacket)) {
+return avpkt->size ? avpkt->size :
+ff_mediacodec_dec_decode(avctx, s, frame, got_frame, 
avpkt);
+}
+
+if (s->bsf) {
+AVPacket unfiltered_pkt = { 0 };
+
+av_fifo_generic_read(s->fifo, _pkt, 
sizeof(unfiltered_pkt), NULL);
+
+ret = av_bsf_send_packet(s->bsf, _pkt);
+   if (ret < 0) {
+   return ret;
+   }
+
+   ret = av_bsf_receive_packet(s->bsf, 
>output_pkt);
+   if (ret == AVERROR(EAGAIN)) {
+   goto done;
+   }
+
+   /* h264_mp4toannexb is used here and does not 
requires flushing */
+   av_assert0(ret != AVERROR_EOF);
+
+   if (ret < 0) {
+  

Re: [FFmpeg-devel] [PATCH 3/4] af_hdcd: Add counter for cdt expirations

[Moritz Barsnick]

Nitpicking:

> Adds a counter for when the code detect timer expired without
   ^ detects

> +int count_sustain_expired;/* occurances of code detect timer 
> expiring without detecting a code */
^ occurrences

> +if (cdt_active && state->sustain == 0) state->count_sustain_expired++;

I think, according to ffmpeg style, you're supposed to put the
instructions on a separate line.

Moritz
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] [PATCH] lavc: add mpeg4 mediacodec decoder

[Michael Niedermayer]

On Sun, Jul 24, 2016 at 11:37:08AM +0200, Thomas Volkert wrote:
> From: Thomas Volkert 
> 
> ---
>  configure|   1 +
>  libavcodec/Makefile  |   1 +
>  libavcodec/allcodecs.c   |   1 +
>  libavcodec/mediacodecdec_mpeg4.c | 239 
> +++
>  4 files changed, 242 insertions(+)
>  create mode 100644 libavcodec/mediacodecdec_mpeg4.c

confirmed that this fixes the build issue

if it works its probably ok, but maybye wait a bit so others have a
chance to comment

thx

[..]

thanks

-- 
Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Frequently ignored answer#1 FFmpeg bugs should be sent to our bugtracker. User
questions about the command line tools should be sent to the ffmpeg-user ML.
And questions about how to use libav* should be sent to the libav-user ML.


signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] [PATCH v2] tests/fate: add dnxhr encoding tests

[Michael Niedermayer]

On Sat, Jul 23, 2016 at 06:10:33PM -0700, Mark Reid wrote:
> added sws_flags flags and tested against x86_32
> 
> ---
>  tests/fate/seek.mak   |  2 ++
>  tests/fate/vcodec.mak | 26 -
>  tests/ref/seek/vsynth_lena-dnxhd-4k-hr-lb | 40 
> +++
>  tests/ref/vsynth/vsynth1-dnxhd-2k-hr-hq   |  4 +++
>  tests/ref/vsynth/vsynth1-dnxhd-4k-hr-lb   |  4 +++
>  tests/ref/vsynth/vsynth1-dnxhd-720p-hr-lb |  4 +++
>  tests/ref/vsynth/vsynth1-dnxhd-uhd-hr-sq  |  4 +++
>  tests/ref/vsynth/vsynth2-dnxhd-2k-hr-hq   |  4 +++
>  tests/ref/vsynth/vsynth2-dnxhd-4k-hr-lb   |  4 +++
>  tests/ref/vsynth/vsynth2-dnxhd-720p-hr-lb |  4 +++
>  tests/ref/vsynth/vsynth2-dnxhd-uhd-hr-sq  |  4 +++
>  tests/ref/vsynth/vsynth3-dnxhd-2k-hr-hq   |  4 +++
>  tests/ref/vsynth/vsynth3-dnxhd-4k-hr-lb   |  4 +++
>  tests/ref/vsynth/vsynth3-dnxhd-720p-hr-lb |  4 +++
>  tests/ref/vsynth/vsynth3-dnxhd-uhd-hr-sq  |  4 +++
>  tests/ref/vsynth/vsynth_lena-dnxhd-2k-hr-hq   |  4 +++
>  tests/ref/vsynth/vsynth_lena-dnxhd-4k-hr-lb   |  4 +++
>  tests/ref/vsynth/vsynth_lena-dnxhd-720p-hr-lb |  4 +++
>  tests/ref/vsynth/vsynth_lena-dnxhd-uhd-hr-sq  |  4 +++
>  19 files changed, 131 insertions(+), 1 deletion(-)
>  create mode 100644 tests/ref/seek/vsynth_lena-dnxhd-4k-hr-lb
>  create mode 100644 tests/ref/vsynth/vsynth1-dnxhd-2k-hr-hq
>  create mode 100644 tests/ref/vsynth/vsynth1-dnxhd-4k-hr-lb
>  create mode 100644 tests/ref/vsynth/vsynth1-dnxhd-720p-hr-lb
>  create mode 100644 tests/ref/vsynth/vsynth1-dnxhd-uhd-hr-sq
>  create mode 100644 tests/ref/vsynth/vsynth2-dnxhd-2k-hr-hq
>  create mode 100644 tests/ref/vsynth/vsynth2-dnxhd-4k-hr-lb
>  create mode 100644 tests/ref/vsynth/vsynth2-dnxhd-720p-hr-lb
>  create mode 100644 tests/ref/vsynth/vsynth2-dnxhd-uhd-hr-sq
>  create mode 100644 tests/ref/vsynth/vsynth3-dnxhd-2k-hr-hq
>  create mode 100644 tests/ref/vsynth/vsynth3-dnxhd-4k-hr-lb
>  create mode 100644 tests/ref/vsynth/vsynth3-dnxhd-720p-hr-lb
>  create mode 100644 tests/ref/vsynth/vsynth3-dnxhd-uhd-hr-sq
>  create mode 100644 tests/ref/vsynth/vsynth_lena-dnxhd-2k-hr-hq
>  create mode 100644 tests/ref/vsynth/vsynth_lena-dnxhd-4k-hr-lb
>  create mode 100644 tests/ref/vsynth/vsynth_lena-dnxhd-720p-hr-lb
>  create mode 100644 tests/ref/vsynth/vsynth_lena-dnxhd-uhd-hr-sq

applied

thanks

[...]
-- 
Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Everything should be made as simple as possible, but not simpler.
-- Albert Einstein


signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

Re: [FFmpeg-devel] [PATCH] lavfi/curves: do not automatically insert points at x=0 and x=1

[Clément Bœsch]

On Sun, Jul 17, 2016 at 05:03:35PM +0200, Clément Bœsch wrote:
> There is actually a need for the origin and end point not to be defined.
> We can not automatically insert them with the y value of the first and
> last point as it will influence the curves in a wrong way.
> 
> Fixes #5397
> ---
>  Changelog   |  1 +
>  doc/filters.texi| 10 ++
>  libavfilter/version.h   |  2 +-
>  libavfilter/vf_curves.c | 85 
> -
>  4 files changed, 47 insertions(+), 51 deletions(-)
> 

Applied.

...with a few more bonus (curves plotting and 10-bits support)

-- 
Clément B.


signature.asc
Description: PGP signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel

[FFmpeg-devel] [PATCH] lavc: add mpeg4 mediacodec decoder

[Thomas Volkert]

From: Thomas Volkert 

---
 configure|   1 +
 libavcodec/Makefile  |   1 +
 libavcodec/allcodecs.c   |   1 +
 libavcodec/mediacodecdec_mpeg4.c | 239 +++
 4 files changed, 242 insertions(+)
 create mode 100644 libavcodec/mediacodecdec_mpeg4.c

diff --git a/configure b/configure
index 1b41303..9004b06 100755
--- a/configure
+++ b/configure
@@ -2621,6 +2621,7 @@ mpeg2_videotoolbox_hwaccel_select="mpeg2video_decoder"
 mpeg2_xvmc_hwaccel_deps="xvmc"
 mpeg2_xvmc_hwaccel_select="mpeg2video_decoder"
 mpeg4_crystalhd_decoder_select="crystalhd"
+mpeg4_mediacodec_decoder_deps="mediacodec"
 mpeg4_mmal_decoder_deps="mmal"
 mpeg4_mmal_decoder_select="mmal"
 mpeg4_mmal_hwaccel_deps="mmal"
diff --git a/libavcodec/Makefile b/libavcodec/Makefile
index abef19e..642cf2a 100644
--- a/libavcodec/Makefile
+++ b/libavcodec/Makefile
@@ -317,6 +317,7 @@ OBJS-$(CONFIG_H264_DECODER)+= h264.o 
h264_cabac.o h264_cavlc.o \
   h2645_parse.o
 OBJS-$(CONFIG_H264_CUVID_DECODER)  += cuvid.o
 OBJS-$(CONFIG_H264_MEDIACODEC_DECODER) += mediacodecdec_h264.o
+OBJS-$(CONFIG_MPEG4_MEDIACODEC_DECODER) += mediacodecdec_mpeg4.o
 OBJS-$(CONFIG_H264_MMAL_DECODER)   += mmaldec.o
 OBJS-$(CONFIG_H264_NVENC_ENCODER)  += nvenc_h264.o
 OBJS-$(CONFIG_NVENC_ENCODER)   += nvenc_h264.o
diff --git a/libavcodec/allcodecs.c b/libavcodec/allcodecs.c
index 951e199..2d98694 100644
--- a/libavcodec/allcodecs.c
+++ b/libavcodec/allcodecs.c
@@ -239,6 +239,7 @@ void avcodec_register_all(void)
 REGISTER_ENCDEC (MPEG2VIDEO,mpeg2video);
 REGISTER_ENCDEC (MPEG4, mpeg4);
 REGISTER_DECODER(MPEG4_CRYSTALHD,   mpeg4_crystalhd);
+REGISTER_DECODER(MPEG4_MEDIACODEC,  mpeg4_mediacodec);
 REGISTER_DECODER(MPEG4_MMAL,mpeg4_mmal);
 #if FF_API_VDPAU
 REGISTER_DECODER(MPEG4_VDPAU,   mpeg4_vdpau);
diff --git a/libavcodec/mediacodecdec_mpeg4.c b/libavcodec/mediacodecdec_mpeg4.c
new file mode 100644
index 000..7c4559b
--- /dev/null
+++ b/libavcodec/mediacodecdec_mpeg4.c
@@ -0,0 +1,239 @@
+/*
+ * Android MediaCodec MPEG 4 decoder
+ *
+ * Copyright (c) 2016 Thomas Volkert 
+ *
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include "libavutil/avassert.h"
+#include "libavutil/common.h"
+#include "libavutil/fifo.h"
+
+#include "avcodec.h"
+#include "internal.h"
+#include "mediacodecdec.h"
+#include "mediacodec_wrapper.h"
+
+#define CODEC_MIME "video/mp4v-es"
+
+typedef struct MediaCodecMPEG4DecContext {
+
+MediaCodecDecContext *ctx;
+
+AVBSFContext *bsf;
+
+AVFifoBuffer *fifo;
+
+AVPacket filtered_pkt;
+
+} MediaCodecMPEG4DecContext;
+
+static av_cold int mediacodec_decode_close(AVCodecContext *avctx)
+{
+MediaCodecMPEG4DecContext *s = avctx->priv_data;
+
+ff_mediacodec_dec_close(avctx, s->ctx);
+s->ctx = NULL;
+
+av_fifo_free(s->fifo);
+
+av_bsf_free(>bsf);
+av_packet_unref(>filtered_pkt);
+
+return 0;
+}
+
+static av_cold int mediacodec_decode_init(AVCodecContext *avctx)
+{
+int ret;
+
+FFAMediaFormat *format = NULL;
+MediaCodecMPEG4DecContext *s = avctx->priv_data;
+
+format = ff_AMediaFormat_new();
+if (!format) {
+av_log(avctx, AV_LOG_ERROR, "Failed to create media format\n");
+ret = AVERROR_EXTERNAL;
+goto done;
+}
+
+ff_AMediaFormat_setString(format, "mime", CODEC_MIME);
+ff_AMediaFormat_setInt32(format, "width", avctx->width);
+ff_AMediaFormat_setInt32(format, "height", avctx->height);
+
+s->ctx = av_mallocz(sizeof(*s->ctx));
+if (!s->ctx) {
+av_log(avctx, AV_LOG_ERROR, "Failed to allocate 
MediaCodecDecContext\n");
+ret = AVERROR(ENOMEM);
+goto done;
+}
+
+if ((ret = ff_mediacodec_dec_init(avctx, s->ctx, CODEC_MIME, format)) < 0) 
{
+s->ctx = NULL;
+goto done;
+}
+
+av_log(avctx, AV_LOG_INFO, "MediaCodec started successfully, ret = %d\n", 
ret);
+
+s->fifo = av_fifo_alloc(sizeof(AVPacket));
+if (!s->fifo) {
+ret = AVERROR(ENOMEM);
+goto done;
+}
+
+const AVBitStreamFilter *bsf = av_bsf_get_by_name("mpeg4_unpack_bframes");
+