Hey all,
Does ffmpeg have an API for the real time transport protocol?
If yes, can you give links? I am unable to find.
Regards,
Abhilash Mhaisne
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg
On Tue, Oct 13, 2015 at 12:31:10AM -0400, Ganesh Ajjanagadde wrote:
> On Tue, Oct 13, 2015 at 12:26 AM, Ganesh Ajjanagadde wrote:
> > On Tue, Oct 13, 2015 at 12:16 AM, Carl Eugen Hoyos wrote:
> >> Ganesh Ajjanagadde mit.edu> writes:
> >>
> >>> Bench from libavfilter/astats on a 15 min clip.
> >>
Am 12.10.15 um 21:10 schrieb Clément Bœsch:
> On Fri, Oct 09, 2015 at 06:13:05PM +0200, Clément Bœsch wrote:
> [...]
You have 24 hours to make comments.
>>>
>>> 24 hours is IMHO too little. What is the hurry? Maybe at least 96 hours?
>>>
>>
>> People had time to comment on the thread already,
On Mon, 2015-10-05 at 14:25 +0200, Tobias Rapp wrote:
> On 05.10.2015 09:10, tim nicholson wrote:
> > On 04/10/15 13:07, Tomas Härdin wrote:
> >> On Mon, 2015-09-28 at 15:11 +0200, Tobias Rapp wrote:
> >>> [...]
> >>
> >> For me the most important thing is that anything dealing with MXF should
> >>
On Tue, Oct 13, 2015 at 12:18:58PM +0700, Muhammad Faiz wrote:
>
> From 893d4068adb9d3d2c118186bdc5645056f0ef172 Mon Sep 17 00:00:00 2001
> From: Muhammad Faiz
> Date: Tue, 13 Oct 2015 12:06:37 +0700
> Subject: [PATCH] avfilter/avf_showcqt: fix frame management
>
> follow frame writability rule
From a403f93a8fa20ce0d7345d9a00d75bc90fe7d73f Mon Sep 17 00:00:00 2001
From: Muhammad Faiz
Date: Tue, 13 Oct 2015 00:31:29 +0700
Subject: [PATCH] avcodec/fft: out of place permutation with av_fft_permute2
with optimization (more cache friendly)
also optimize av_fft_permute
machine specific code
On Tue, Oct 13, 2015 at 1:01 AM, Muhammad Faiz wrote:
>
>
I'm sorry, please ignore this (because of wrong subject line).
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
From 893d4068adb9d3d2c118186bdc5645056f0ef172 Mon Sep 17 00:00:00 2001
From: Muhammad Faiz
Date: Tue, 13 Oct 2015 12:06:37 +0700
Subject: [PATCH] avfilter/avf_showcqt: fix frame management
follow frame writability rule
reuse buffer
---
libavfilter/avf_showcqt.c | 61
On Tue, Oct 13, 2015 at 1:03 AM, Ganesh Ajjanagadde wrote:
> On Tue, Oct 13, 2015 at 12:44 AM, Carl Eugen Hoyos wrote:
>> Ganesh Ajjanagadde mit.edu> writes:
>>> On Tue, Oct 13, 2015 at 12:16 AM, Carl Eugen Hoyos wrote:
>>> > Ganesh Ajjanagadde mit.edu> writes:
>>> >
>>> >> Bench from libavfilt
On Tue, Oct 13, 2015 at 12:44 AM, Carl Eugen Hoyos wrote:
> Ganesh Ajjanagadde mit.edu> writes:
>> On Tue, Oct 13, 2015 at 12:16 AM, Carl Eugen Hoyos wrote:
>> > Ganesh Ajjanagadde mit.edu> writes:
>> >
>> >> Bench from libavfilter/astats on a 15 min clip.
>> >
>> > I believe that your test woul
Ganesh Ajjanagadde mit.edu> writes:
> On Tue, Oct 13, 2015 at 12:16 AM, Carl Eugen Hoyos wrote:
> > Ganesh Ajjanagadde mit.edu> writes:
> >
> >> Bench from libavfilter/astats on a 15 min clip.
> >
> > I believe that your test would indicate that the
> > old variant is faster or that no result can
On Tue, Oct 13, 2015 at 12:26 AM, Ganesh Ajjanagadde wrote:
> On Tue, Oct 13, 2015 at 12:16 AM, Carl Eugen Hoyos wrote:
>> Ganesh Ajjanagadde mit.edu> writes:
>>
>>> Bench from libavfilter/astats on a 15 min clip.
>>
>> I believe that your test would indicate that the
>> old variant is faster or
On Tue, Oct 13, 2015 at 12:16 AM, Carl Eugen Hoyos wrote:
> Ganesh Ajjanagadde mit.edu> writes:
>
>> Bench from libavfilter/astats on a 15 min clip.
>
> I believe that your test would indicate that the
> old variant is faster or that no result can be
> given which is what my tests show.
Look at
Ganesh Ajjanagadde mit.edu> writes:
> Bench from libavfilter/astats on a 15 min clip.
I believe that your test would indicate that the
old variant is faster or that no result can be
given which is what my tests show.
I am not sure if it makes sense to apply a patch
that is meant to improve s
Hi
you probably already saw it but just to be sure its not missed
mips optimizations where broken again today:
CC libavcodec/mips/aaccoder_mips.o
ffmpeg/libavcodec/mips/aaccoder_mips.c: In function ‘ff_aac_coder_init_mips’:
ffmpeg/libavcodec/mips/aaccoder_mips.c:2548: error: ‘AACEncOptions’
On 10/12/2015 3:15 PM, Andreas Cadhalpun wrote:
> On 11.10.2015 23:37, James Almer wrote:
>> On 10/11/2015 6:26 PM, Andreas Cadhalpun wrote:
>>> Anyway, please also remove the entry for av_ctz from APIchanges.
>>> It's not public API, so it never belonged there.
>>
>> Huh, even an APIChanges line?
On Mon, Oct 12, 2015 at 07:33:24PM -0500, Rodger Combs wrote:
> ---
> configure | 4
> doc/APIchanges| 3 +++
> libavutil/cpu.c | 4
> libavutil/cpu.h | 1 +
> libavutil/version.h | 2 +-
> libavutil/x86/cpu.c | 2 ++
> li
On Mon, Oct 12, 2015 at 08:15:10PM +0200, Andreas Cadhalpun wrote:
> On 11.10.2015 23:37, James Almer wrote:
> > On 10/11/2015 6:26 PM, Andreas Cadhalpun wrote:
> >> Anyway, please also remove the entry for av_ctz from APIchanges.
> >> It's not public API, so it never belonged there.
> >
> > Huh,
On Mon, Oct 12, 2015 at 07:33:32PM -0500, Rodger Combs wrote:
> ---
> tools/crypto_bench.c | 138
> +--
> 1 file changed, 135 insertions(+), 3 deletions(-)
>
> diff --git a/tools/crypto_bench.c b/tools/crypto_bench.c
> index ad20f95..0b1bfc8 100644
On Tue, Oct 13, 2015 at 12:00:07AM +0200, Michael Niedermayer wrote:
> On Mon, Oct 12, 2015 at 10:39:23PM +0200, Christophe Gisquet wrote:
> > Le 12 oct. 2015 21:37, "Michael Niedermayer" a
> > écrit :
> > >
> > > On Mon, Oct 12, 2015 at 07:37:43PM +0200, Christophe Gisquet wrote:
> > > > It was u
---
tools/crypto_bench.c | 51 +++
1 file changed, 43 insertions(+), 8 deletions(-)
diff --git a/tools/crypto_bench.c b/tools/crypto_bench.c
index 0b1bfc8..f84e2a7 100644
--- a/tools/crypto_bench.c
+++ b/tools/crypto_bench.c
@@ -32,6 +32,7 @@
#incl
---
tools/crypto_bench.c | 138 +--
1 file changed, 135 insertions(+), 3 deletions(-)
diff --git a/tools/crypto_bench.c b/tools/crypto_bench.c
index ad20f95..0b1bfc8 100644
--- a/tools/crypto_bench.c
+++ b/tools/crypto_bench.c
@@ -52,6 +52,7 @@ stat
---
tools/crypto_bench.c | 82
1 file changed, 82 insertions(+)
diff --git a/tools/crypto_bench.c b/tools/crypto_bench.c
index b3b24a6..ad20f95 100644
--- a/tools/crypto_bench.c
+++ b/tools/crypto_bench.c
@@ -118,6 +118,24 @@ static void run_la
---
tests/checkasm/Makefile | 2 +-
tests/checkasm/aes.c | 57 +++
tests/checkasm/checkasm.c | 1 +
tests/checkasm/checkasm.h | 1 +
4 files changed, 60 insertions(+), 1 deletion(-)
create mode 100644 tests/checkasm/aes.c
diff --git a/tests/c
---
libavutil/aes_internal.h | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/libavutil/aes_internal.h b/libavutil/aes_internal.h
index e5bf4bd..4944258 100644
--- a/libavutil/aes_internal.h
+++ b/libavutil/aes_internal.h
@@ -21,6 +21,7 @@
#ifndef AVUTIL_AES_INTERNAL_H
#d
---
libavutil/aes.c | 14 +-
1 file changed, 9 insertions(+), 5 deletions(-)
diff --git a/libavutil/aes.c b/libavutil/aes.c
index c1d860f..1fac4e8 100644
--- a/libavutil/aes.c
+++ b/libavutil/aes.c
@@ -280,7 +280,7 @@ int main(int argc, char **argv)
{ 0x10, 0xa5, 0x88, 0x69,
---
libavutil/aes.c | 4 ++
libavutil/aes_internal.h | 2 +
libavutil/x86/Makefile | 4 +-
libavutil/x86/aes.asm| 98
libavutil/x86/aes_init.c | 37 ++
5 files changed, 144 insertions(+), 1 deletion(-)
create mode
---
libavutil/aes.c | 16 +---
libavutil/aes_internal.h | 41 +
2 files changed, 42 insertions(+), 15 deletions(-)
create mode 100644 libavutil/aes_internal.h
diff --git a/libavutil/aes.c b/libavutil/aes.c
index 8d4..280e5fa 100644
---
configure | 4
doc/APIchanges| 3 +++
libavutil/cpu.c | 4
libavutil/cpu.h | 1 +
libavutil/version.h | 2 +-
libavutil/x86/cpu.c | 2 ++
libavutil/x86/cpu.h | 3 +++
libavutil/x86/x86inc.asm | 13 +++--
---
libavutil/aes.c | 47 +++
libavutil/aes_internal.h | 1 +
2 files changed, 32 insertions(+), 16 deletions(-)
diff --git a/libavutil/aes.c b/libavutil/aes.c
index 280e5fa..c1d860f 100644
--- a/libavutil/aes.c
+++ b/libavutil/aes.c
@@ -126,3
On Mon, Oct 12, 2015 at 07:37:46PM +0200, Christophe Gisquet wrote:
> When the input of a pass has 15 or 16 bits of precision (in particular
> the column pass), the addition of a bias to W4 may lead to overflows
> in the input to pmaddwd.
>
> This requires postponing the adding of the bias to afte
On Mon, Oct 12, 2015 at 07:37:45PM +0200, Christophe Gisquet wrote:
> omse goes from 0.03060703 (which fails for dct-test) to 0.01663750.
> This also actually improve the error of decoding the sample generated
> by fate-vsynth3-dnxhd1080i-10bit using simple_idct10 to FAANI, which
> goes (when resam
Hi,
On Mon, Oct 12, 2015 at 1:37 PM, Christophe Gisquet <
christophe.gisq...@gmail.com> wrote:
> This should be reused for a generic simple_idct10 function.
> Requires a bit of trickery to declare common constants in C.
> ---
> libavcodec/x86/constants.c| 28 +++
> libavcodec/x8
On Mon, Oct 12, 2015 at 07:37:44PM +0200, Christophe Gisquet wrote:
> This should be reused for a generic simple_idct10 function.
> Requires a bit of trickery to declare common constants in C.
> ---
> libavcodec/x86/constants.c| 28 +++
> libavcodec/x86/constants.h
On Mon, Oct 12, 2015 at 6:10 PM, Reynaldo H. Verdejo Pinochet
wrote:
> Signed-off-by: Reynaldo H. Verdejo Pinochet
> ---
> src/index | 33 +
> 1 file changed, 33 insertions(+)
>
> diff --git a/src/index b/src/index
> index 77946a4..4dec768 100644
> --- a/src/index
On Mon, 12 Oct 2015, Jaroslav Šnajdr wrote:
This fixes trac #4924, most likely not 100% correct.
Well, it is not, because file_start_time is the start timestamp of the
input, not the calculated start timestamp of the output...
I'd say if you want to use the concat demuxer this way, extend
On Mon, Oct 12, 2015 at 03:10:22PM -0700, Reynaldo H. Verdejo Pinochet wrote:
> Signed-off-by: Reynaldo H. Verdejo Pinochet
> ---
> src/index | 33 +
> 1 file changed, 33 insertions(+)
no objection from me
[...]
--
Michael GnuPG fingerprint: 9FF2128B147EF673
On Mon, Oct 12, 2015 at 07:37:42PM +0200, Christophe Gisquet wrote:
> Dequant or encoding were trying to reverse a scan that hadn't been
> applied...
> ---
> libavcodec/dnxhdenc.c | 7 ++-
> libavcodec/mpegvideo.h | 3 ++-
> libavcodec/mpegvideo_enc.c | 6 +++---
> 3 files changed, 11
Signed-off-by: Reynaldo H. Verdejo Pinochet
---
src/index | 33 +
1 file changed, 33 insertions(+)
diff --git a/src/index b/src/index
index 77946a4..4dec768 100644
--- a/src/index
+++ b/src/index
@@ -37,6 +37,39 @@
News
+ Telepoint & MediaHub are now s
On Mon, Oct 12, 2015 at 10:39:23PM +0200, Christophe Gisquet wrote:
> Le 12 oct. 2015 21:37, "Michael Niedermayer" a
> écrit :
> >
> > On Mon, Oct 12, 2015 at 07:37:43PM +0200, Christophe Gisquet wrote:
> > > It was useful to (accidentally?) spot an overflow in the column pass
> > > of the x86 sim
On Mon, Oct 12, 2015 at 8:50 AM, Rodger Combs wrote:
> diff --git a/tests/checkasm/checkasm.c b/tests/checkasm/checkasm.c
> @@ -117,6 +117,7 @@ static const struct {
> { "SSSE3","ssse3",AV_CPU_FLAG_SSSE3|AV_CPU_FLAG_ATOM },
> { "SSE4.1", "sse4", AV_CPU_FLAG_SSE4 },
> {
On Mon, Oct 12, 2015 at 4:57 PM, Ganesh Ajjanagadde wrote:
> On Mon, Oct 12, 2015 at 7:59 AM, Ganesh Ajjanagadde wrote:
>> On Mon, Oct 12, 2015 at 7:46 AM, Carl Eugen Hoyos wrote:
>>> Ganesh Ajjanagadde mit.edu> writes:
>>>
It is well known that fabs and fabsf are at least as fast and usua
On Mon, 12 Oct 2015, Bodecs Bela wrote:
Dear All,
currently stream specifiers may contain stream index, stream type,
stream id, program id, metadata key/value es usable config settings.
But you can not combine them. In some case, mainly working with
multiprogram mpeg-ts containers as input, th
On Mon, Oct 12, 2015 at 7:59 AM, Ganesh Ajjanagadde wrote:
> On Mon, Oct 12, 2015 at 7:46 AM, Carl Eugen Hoyos wrote:
>> Ganesh Ajjanagadde mit.edu> writes:
>>
>>> It is well known that fabs and fabsf are at least as fast and usually
>>> faster than the FFABS macro, at least on the gcc+glibc com
Le 12 oct. 2015 21:37, "Michael Niedermayer" a
écrit :
>
> On Mon, Oct 12, 2015 at 07:37:43PM +0200, Christophe Gisquet wrote:
> > It was useful to (accidentally?) spot an overflow in the column pass
> > of the x86 simple_idct10 implementation.
> > ---
> > tests/fate/vcodec.mak
Dear All,
currently stream specifiers may contain stream index, stream type,
stream id, program id, metadata key/value es usable config settings.
But you can not combine them. In some case, mainly working with
multiprogram mpeg-ts containers as input, this feature would be handy.
This patch makes
On Sun, Oct 11, 2015 at 05:11:03PM +0100, Ricardo Constantino wrote:
> Includes escapes that should now be supported and a few features not yet
> fully supported, like comments, regions, classes, ruby, and lang.
>
> All were tested with https://quuz.org/webvtt/ for validation, except
> regions bec
On Sun, Oct 11, 2015 at 05:11:02PM +0100, Ricardo Constantino wrote:
> Signed-off-by: Ricardo Constantino
> ---
> libavformat/webvttdec.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/libavformat/webvttdec.c b/libavformat/webvttdec.c
> index 43c2a63..47a3255 100644
>
On Sun, Oct 11, 2015 at 05:11:01PM +0100, Ricardo Constantino wrote:
> Bare ampersand characters are still accepted, even though out-of-spec.
> Also fixes adjacent tags not being parsed.
>
> Fixes trac #4915
>
> Signed-off-by: Ricardo Constantino
> ---
> libavcodec/webvttdec.c | 11 ++-
On Sun, Oct 11, 2015 at 8:27 PM, Ganesh Ajjanagadde wrote:
> On Sun, Oct 11, 2015 at 6:00 PM, Moritz Barsnick wrote:
>> On Sun, Oct 11, 2015 at 16:29:16 -0400, Ganesh Ajjanagadde wrote:
>>> + Microsoft Windows, the BSD's, Solaris, etc. under a wide variety of build
>>
On Mon, Oct 12, 2015 at 3:24 PM, Clément Bœsch wrote:
> On Mon, Oct 12, 2015 at 03:19:33PM -0400, Ganesh Ajjanagadde wrote:
>> On Mon, Oct 12, 2015 at 3:10 PM, Clément Bœsch wrote:
>> > On Fri, Oct 09, 2015 at 06:13:05PM +0200, Clément Bœsch wrote:
>> > [...]
>> >> > > You have 24 hours to make c
This fixes trac #4924, most likely not 100% correct.
---
libavformat/concatdec.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libavformat/concatdec.c b/libavformat/concatdec.c
index 832b7f4..1151b10 100644
--- a/libavformat/concatdec.c
+++ b/libavformat/concatdec.c
@@ -580,7
On Mon, Oct 12, 2015 at 07:37:43PM +0200, Christophe Gisquet wrote:
> It was useful to (accidentally?) spot an overflow in the column pass
> of the x86 simple_idct10 implementation.
> ---
> tests/fate/vcodec.mak | 6 +-
> tests/ref/vsynth/vsynth1-dnxhd-1080i-10bit
On Mon, Oct 12, 2015 at 03:19:33PM -0400, Ganesh Ajjanagadde wrote:
> On Mon, Oct 12, 2015 at 3:10 PM, Clément Bœsch wrote:
> > On Fri, Oct 09, 2015 at 06:13:05PM +0200, Clément Bœsch wrote:
> > [...]
> >> > > You have 24 hours to make comments.
> >> >
> >> > 24 hours is IMHO too little. What is t
On Mon, Oct 12, 2015 at 3:10 PM, Clément Bœsch wrote:
> On Fri, Oct 09, 2015 at 06:13:05PM +0200, Clément Bœsch wrote:
> [...]
>> > > You have 24 hours to make comments.
>> >
>> > 24 hours is IMHO too little. What is the hurry? Maybe at least 96 hours?
>> >
>>
>> People had time to comment on the
On Fri, Oct 09, 2015 at 06:13:05PM +0200, Clément Bœsch wrote:
[...]
> > > You have 24 hours to make comments.
> >
> > 24 hours is IMHO too little. What is the hurry? Maybe at least 96 hours?
> >
>
> People had time to comment on the thread already, but sure OK for 96
> hours.
>
Not sure if I
On Mon, Oct 12, 2015 at 9:54 AM, Rodger Combs wrote:
> +void checkasm_check_aes(void)
> +{
> +int i, j, d;
> +AVAES b;
> +uint8_t pt[32];
> +uint8_t temp[2][32];
> +uint8_t iv[2][16];
> +
> +for (d = 0; d <= 1; d++) {
> +for (i = 128; i <= 256; i += 64) {
> +
On Mon, 12 Oct 2015 20:50:58 +0200, Paul B Mahol wrote:
> ok
Pushed. Thanks.
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
On 10/12/15, Lou Logan wrote:
> Signed-off-by: Lou Logan
> ---
> doc/filters.texi | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/doc/filters.texi b/doc/filters.texi
> index 4abc786..34008e5 100644
> --- a/doc/filters.texi
> +++ b/doc/filters.texi
> @@ -7128,7 +7128,
Signed-off-by: Lou Logan
---
doc/filters.texi | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/doc/filters.texi b/doc/filters.texi
index 4abc786..34008e5 100644
--- a/doc/filters.texi
+++ b/doc/filters.texi
@@ -7128,7 +7128,7 @@ to create same output.
The filter accept the
On 11.10.2015 23:37, James Almer wrote:
> On 10/11/2015 6:26 PM, Andreas Cadhalpun wrote:
>> Anyway, please also remove the entry for av_ctz from APIchanges.
>> It's not public API, so it never belonged there.
>
> Huh, even an APIChanges line? Wonder if intmath.h was public at some
> point, or if
Hi all,
I wanted to test QSV encoding and decoding, which is included since
several months in FFmpeg, and have a few questions.
1. QSV encoding and decoding depends on libmfx, but no references are
present in the code to the library, so I have to suppose that the
required library is this one:
htt
From a403f93a8fa20ce0d7345d9a00d75bc90fe7d73f Mon Sep 17 00:00:00 2001
From: Muhammad Faiz
Date: Tue, 13 Oct 2015 00:31:29 +0700
Subject: [PATCH] avcodec/fft: out of place permutation with av_fft_permute2
with optimization (more cache friendly)
also optimize av_fft_permute
machine specific code
On 9/29/15, Paul B Mahol wrote:
> Signed-off-by: Paul B Mahol
> ---
> libavfilter/Makefile | 1 +
> libavfilter/allfilters.c | 1 +
> libavfilter/vf_setfps.c | 119
> +++
> 3 files changed, 121 insertions(+)
> create mode 100644 libavfilter/v
In particular for 10 and 12 bits.
---
libavcodec/dct-test.c | 2 ++
libavcodec/x86/dct-test.c | 10 ++
2 files changed, 12 insertions(+)
diff --git a/libavcodec/dct-test.c b/libavcodec/dct-test.c
index 56e1a62..72fe353 100644
--- a/libavcodec/dct-test.c
+++ b/libavcodec/dct-test.c
@@
On 12 frames of a 444p 12 bits DNxHR sequence, _put function:
C: 78902 decicycles in idct, 262071 runs, 73 skips
avx: 32478 decicycles in idct, 262045 runs, 99 skips
Difference between the 2:
stddev:0.39 PSNR:104.47 MAXDIFF:2
This is unavoidable and due to the scal
This is done by not passing clamping regs to the macro, and there,
deducing that this is a pure idct.
There are not many users (vf_spp is one), and the _add version doesn't
look like it will ever be used.
---
libavcodec/x86/idctdsp_init.c | 4 +-
libavcodec/x86/proresdsp.asm
Modeled from the prores version. Clips to [0;1023] and is bitexact.
Bitexactness requires to add offsets in different places compared to
prores or C, and makes the function approximately 2% slower.
For 16 frames of a DNxHD 4:2:2 10bits test sequence:
C:60861 decicycles in idct, 1048205 runs,
When the input of a pass has 15 or 16 bits of precision (in particular
the column pass), the addition of a bias to W4 may lead to overflows
in the input to pmaddwd.
This requires postponing the adding of the bias to after the first
butterfly. To do so, the fact that m15, unused although zeroed, is
This should be reused for a generic simple_idct10 function.
Requires a bit of trickery to declare common constants in C.
---
libavcodec/x86/constants.c| 28 +++
libavcodec/x86/constants.h| 16 ++
libavcodec/x86/proresdsp.asm | 263 +---
omse goes from 0.03060703 (which fails for dct-test) to 0.01663750.
This also actually improve the error of decoding the sample generated
by fate-vsynth3-dnxhd1080i-10bit using simple_idct10 to FAANI, which
goes (when resampled to yuv422p) from:
stddev:0.06 PSNR: 72.28 MAXDIFF:1
to identica
Changes since v3:
- proper fix (famous last words) for issues in encoder
- fate test that showed an overflow
- fix to previous code for said overflow
Patch 2 is somewhat independant from the others, and could be applied
separately.
Christophe Gisquet (9):
mpegvideo: dnxhdenc: permute 10bits con
It was useful to (accidentally?) spot an overflow in the column pass
of the x86 simple_idct10 implementation.
---
tests/fate/vcodec.mak | 6 +-
tests/ref/vsynth/vsynth1-dnxhd-1080i-10bit | 4
tests/ref/vsynth/vsynth2-dnxhd-1080i-10bit | 4
tests/ref/v
Dequant or encoding were trying to reverse a scan that hadn't been
applied...
---
libavcodec/dnxhdenc.c | 7 ++-
libavcodec/mpegvideo.h | 3 ++-
libavcodec/mpegvideo_enc.c | 6 +++---
3 files changed, 11 insertions(+), 5 deletions(-)
diff --git a/libavcodec/dnxhdenc.c b/libavcodec/dn
Changes since v3:
- proper fix (famous last words) for issues in encoder
- fate test that showed an overflow
- fix to previous code for said overflow
Patch 2 is somewhat independant from the others, and could be applied
separately.
Christophe Gisquet (9):
mpegvideo: dnxhdenc: permute 10bits con
On Mon, Oct 12, 2015 at 12:43 PM, Michael Niedermayer
wrote:
> On Sat, Oct 10, 2015 at 01:51:10PM -0400, Ganesh Ajjanagadde wrote:
>> Partially fixes Ticket 4727.
>>
>> -duration is not a safe expression, since duration can be INT_MIN.
>> One might ask how it can become INT_MIN.
>> Although it is
Hi,
2015-10-12 0:09 GMT+02:00 Michael Niedermayer :
> this seems to cause artifacts on the red/cyan edge in
> ./ffmpeg -f lavfi -i testsrc=s=1920x1080 -vframes 1 -pix_fmt yuv422p -vcodec
> dnxhd -flags +ildct -vb 185M test.mov
ok, ff_convert_matrix permutes the table, so the 8 bit case should us
On Sat, Oct 10, 2015 at 01:51:10PM -0400, Ganesh Ajjanagadde wrote:
> Partially fixes Ticket 4727.
>
> -duration is not a safe expression, since duration can be INT_MIN.
> One might ask how it can become INT_MIN.
> Although it is true that line 2574 is no longer reached with INT_MIN due
> to commi
On Mon, Oct 12, 2015 at 11:52 AM, Ganesh Ajjanagadde wrote:
> On Mon, Oct 12, 2015 at 11:43 AM, Michael Niedermayer
> wrote:
>> On Mon, Oct 12, 2015 at 11:22:17AM -0400, Ganesh Ajjanagadde wrote:
>>> On Mon, Oct 12, 2015 at 11:09 AM, Michael Niedermayer
>>> wrote:
>>> > On Fri, Oct 09, 2015 at 0
On Mon, Oct 12, 2015 at 11:47 AM, Michael Niedermayer
wrote:
> On Mon, Oct 12, 2015 at 11:16:00AM -0400, Ganesh Ajjanagadde wrote:
>> On Mon, Oct 12, 2015 at 10:54 AM, Ronald S. Bultje
>> wrote:
>> > Hi,
>> >
>> > On Mon, Oct 12, 2015 at 10:51 AM, Ganesh Ajjanagadde
>> > wrote:
>> >>
>> >> On T
On Mon, Oct 12, 2015 at 11:43 AM, Michael Niedermayer
wrote:
> On Mon, Oct 12, 2015 at 11:22:17AM -0400, Ganesh Ajjanagadde wrote:
>> On Mon, Oct 12, 2015 at 11:09 AM, Michael Niedermayer
>> wrote:
>> > On Fri, Oct 09, 2015 at 01:48:10PM -0400, Ganesh Ajjanagadde wrote:
>> >> res, absres are curr
On Mon, Oct 12, 2015 at 11:16:00AM -0400, Ganesh Ajjanagadde wrote:
> On Mon, Oct 12, 2015 at 10:54 AM, Ronald S. Bultje wrote:
> > Hi,
> >
> > On Mon, Oct 12, 2015 at 10:51 AM, Ganesh Ajjanagadde
> > wrote:
> >>
> >> On Thu, Oct 8, 2015 at 5:07 PM, Ganesh Ajjanagadde
> >> wrote:
> >> > On Sat,
On Mon, Oct 12, 2015 at 11:22:17AM -0400, Ganesh Ajjanagadde wrote:
> On Mon, Oct 12, 2015 at 11:09 AM, Michael Niedermayer
> wrote:
> > On Fri, Oct 09, 2015 at 01:48:10PM -0400, Ganesh Ajjanagadde wrote:
> >> res, absres are currently int's, which on most platforms is 32 bits.
> >> Unfortunately,
Hi,
On Fri, Oct 9, 2015 at 1:48 PM, Ganesh Ajjanagadde
wrote:
> res, absres are currently int's, which on most platforms is 32 bits.
> Unfortunately, data is untrusted, and on line 1267 res is manipulated
> with data. Thus, res can take on INT32_MIN/INT_MIN with crafted data,
> making FFABS on l
On Mon, Oct 12, 2015 at 11:23 AM, Ronald S. Bultje wrote:
> Hi,
>
> On Fri, Oct 9, 2015 at 1:48 PM, Ganesh Ajjanagadde
> wrote:
>>
>> res, absres are currently int's, which on most platforms is 32 bits.
>> Unfortunately, data is untrusted, and on line 1267 res is manipulated
>> with data. Thus, r
On Mon, 12 Oct 2015 11:16:00 -0400
Ganesh Ajjanagadde wrote:
> On Mon, Oct 12, 2015 at 10:54 AM, Ronald S. Bultje wrote:
> > Hi,
> >
> > On Mon, Oct 12, 2015 at 10:51 AM, Ganesh Ajjanagadde
> > wrote:
> >>
> >> On Thu, Oct 8, 2015 at 5:07 PM, Ganesh Ajjanagadde
> >> wrote:
> >> > On Sat, Oct 3
On Mon, Oct 12, 2015 at 10:30:01AM -0400, Derek Buitenhuis wrote:
> Signed-off-by: Derek Buitenhuis
> ---
> libavcodec/utils.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
should be ok i think
[...]
--
Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB
I ha
On Mon, Oct 12, 2015 at 11:23 AM, wm4 wrote:
> On Mon, 12 Oct 2015 11:16:00 -0400
> Ganesh Ajjanagadde wrote:
>
>> On Mon, Oct 12, 2015 at 10:54 AM, Ronald S. Bultje
>> wrote:
>> > Hi,
>> >
>> > On Mon, Oct 12, 2015 at 10:51 AM, Ganesh Ajjanagadde
>> > wrote:
>> >>
>> >> On Thu, Oct 8, 2015 at
On Mon, Oct 12, 2015 at 11:09 AM, Michael Niedermayer
wrote:
> On Fri, Oct 09, 2015 at 01:48:10PM -0400, Ganesh Ajjanagadde wrote:
>> res, absres are currently int's, which on most platforms is 32 bits.
>> Unfortunately, data is untrusted, and on line 1267 res is manipulated
>> with data. Thus, re
On Fri, Oct 09, 2015 at 01:48:10PM -0400, Ganesh Ajjanagadde wrote:
> res, absres are currently int's, which on most platforms is 32 bits.
> Unfortunately, data is untrusted, and on line 1267 res is manipulated
> with data. Thus, res can take on INT32_MIN/INT_MIN with crafted data,
> making FFABS o
On Mon, Oct 12, 2015 at 10:54 AM, Ronald S. Bultje wrote:
> Hi,
>
> On Mon, Oct 12, 2015 at 10:51 AM, Ganesh Ajjanagadde
> wrote:
>>
>> On Thu, Oct 8, 2015 at 5:07 PM, Ganesh Ajjanagadde
>> wrote:
>> > On Sat, Oct 3, 2015 at 8:17 AM, Ganesh Ajjanagadde
>> > wrote:
>> >> On Tue, Sep 29, 2015 at
Le nonidi 19 vendémiaire, an CCXXIV, Alex Agranovsky a écrit :
> Please see the modified patch attached.
Pushed, thanks.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
Hi,
On Mon, Oct 12, 2015 at 10:51 AM, Ganesh Ajjanagadde
wrote:
> On Thu, Oct 8, 2015 at 5:07 PM, Ganesh Ajjanagadde
> wrote:
> > On Sat, Oct 3, 2015 at 8:17 AM, Ganesh Ajjanagadde
> wrote:
> >> On Tue, Sep 29, 2015 at 10:49 AM, Ganesh Ajjanagadde
> wrote:
> >>> On Sun, Sep 27, 2015 at 9:39 P
Le nonidi 19 vendémiaire, an CCXXIV, Bodecs Bela a écrit :
> I have altered the code as you wrote.
Pushed, thanks.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http:/
On Thu, Oct 8, 2015 at 5:07 PM, Ganesh Ajjanagadde wrote:
> On Sat, Oct 3, 2015 at 8:17 AM, Ganesh Ajjanagadde wrote:
>> On Tue, Sep 29, 2015 at 10:49 AM, Ganesh Ajjanagadde
>> wrote:
>>> On Sun, Sep 27, 2015 at 9:39 PM, Ganesh Ajjanagadde
>>> wrote:
On Sun, Sep 27, 2015 at 9:18 PM, Mich
On Mon, Oct 12, 2015 at 9:53 AM, Michael Niedermayer
wrote:
> On Sun, Oct 11, 2015 at 08:39:24PM -0400, Ganesh Ajjanagadde wrote:
>> On Sat, Oct 10, 2015 at 9:47 AM, Nicolas George wrote:
>> > Le nonidi 19 vendémiaire, an CCXXIV, Clement Boesch a écrit :
>> >> You could change that last function
---
libavcodec/x86/vp9itxfm_16bpp.asm | 268 +-
1 file changed, 146 insertions(+), 122 deletions(-)
diff --git a/libavcodec/x86/vp9itxfm_16bpp.asm
b/libavcodec/x86/vp9itxfm_16bpp.asm
index 346c514..06b39ff 100644
--- a/libavcodec/x86/vp9itxfm_16bpp.asm
+++ b/l
Signed-off-by: Derek Buitenhuis
---
libavcodec/utils.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/libavcodec/utils.c b/libavcodec/utils.c
index 1751067..0e4f3c0 100644
--- a/libavcodec/utils.c
+++ b/libavcodec/utils.c
@@ -3266,7 +3266,7 @@ int av_lockmgr_register(int
---
libavcodec/x86/vp9itxfm.asm | 84 +-
libavcodec/x86/vp9itxfm_16bpp.asm | 95 ---
2 files changed, 91 insertions(+), 88 deletions(-)
diff --git a/libavcodec/x86/vp9itxfm.asm b/libavcodec/x86/vp9itxfm.asm
index a3e0f86..6
These aren't quite as helpful as the ones in 8bpp, since over there,
we can use pmulhrsw, but here the coefficients have too many bits to
be able to take advantage of pmulhrsw. However, we can still skip
cols for which all coefs are 0, and instead just zero the input data
for the row itx. This help
1 - 100 of 150 matches
Mail list logo
