On Wed, May 11, 2016 at 10:02:33PM +0200, Matthieu Bouron wrote:
> On Wed, May 11, 2016 at 9:04 PM, Reimar Döffinger
> wrote:
>
> >
> >
> > On 11.05.2016, at 20:37, Michael Niedermayer
> > wrote:
> >
> > > On Wed, May 11, 2016 at 06:39:20PM +0200, Matthieu Bouron wrote:
> > >> From: Matthieu Bou
On Wed, May 11, 2016 at 07:05:22PM -0300, James Almer wrote:
> On 5/11/2016 5:34 PM, Michael Niedermayer wrote:
> > config.h is needed for HAVE_*
> > config.h alone breaks build
> > some of the added includes may be unneccesary
> >
> > Signed-off-by: Michael Niedermayer
> > ---
> > libavutil/cpu
On Wed, May 11, 2016 at 07:03:23PM -0300, James Almer wrote:
> On 5/11/2016 5:34 PM, Michael Niedermayer wrote:
> > Signed-off-by: Michael Niedermayer
> > ---
> > libavutil/float_dsp-test.c |2 ++
> > 1 file changed, 2 insertions(+)
> >
> > diff --git a/libavutil/float_dsp-test.c b/libavutil
On 5/11/2016 7:26 PM, Michael Niedermayer wrote:
> On Wed, May 11, 2016 at 06:25:47PM -0300, James Almer wrote:
>> Signed-off-by: James Almer
>> ---
>> libavutil/aes-test.c | 36
>> libavutil/camellia-test.c | 7 ++-
>> libavutil/cast5-test.c| 7
On Wed, May 11, 2016 at 06:25:47PM -0300, James Almer wrote:
> Signed-off-by: James Almer
> ---
> libavutil/aes-test.c | 36
> libavutil/camellia-test.c | 7 ++-
> libavutil/cast5-test.c| 7 ++-
> libavutil/hash-test.c | 13 ++---
On 5/11/2016 5:34 PM, Michael Niedermayer wrote:
> Signed-off-by: Michael Niedermayer
> ---
> libavutil/float_dsp-test.c |2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/libavutil/float_dsp-test.c b/libavutil/float_dsp-test.c
> index 2603ee5..9963efb 100644
> --- a/libavutil/float_d
On 5/11/2016 5:34 PM, Michael Niedermayer wrote:
> config.h is needed for HAVE_*
> config.h alone breaks build
> some of the added includes may be unneccesary
>
> Signed-off-by: Michael Niedermayer
> ---
> libavutil/cpu-test.c | 33 -
> 1 file changed, 24 insert
Le 11 mai 2016 6:39 PM, "Matthieu Bouron" a
écrit :
>
> From: Matthieu Bouron
>
> ---
>
> Hello,
>
> Here are some benchmark on a rpi2 of the attached patch.
>
> ./ffmpeg -f lavfi -i
sine=440,aformat=sample_fmts=fltp,asetnsamples=4096,abench=start,aresample=48000,abench=stop
-t 1000 -f null -
>
>
Signed-off-by: James Almer
---
libavutil/aes-test.c | 36
libavutil/camellia-test.c | 7 ++-
libavutil/cast5-test.c| 7 ++-
libavutil/hash-test.c | 13 ++---
libavutil/murmur3-test.c | 6 --
libavutil/ripemd-test.c | 20 +
config.h is needed for HAVE_*
config.h alone breaks build
some of the added includes may be unneccesary
Signed-off-by: Michael Niedermayer
---
libavutil/cpu-test.c | 33 -
1 file changed, 24 insertions(+), 9 deletions(-)
diff --git a/libavutil/cpu-test.c b/liba
Signed-off-by: Michael Niedermayer
---
libavutil/float_dsp-test.c |2 ++
1 file changed, 2 insertions(+)
diff --git a/libavutil/float_dsp-test.c b/libavutil/float_dsp-test.c
index 2603ee5..9963efb 100644
--- a/libavutil/float_dsp-test.c
+++ b/libavutil/float_dsp-test.c
@@ -16,6 +16,8 @@
*
On Wed, 11 May 2016 19:57:19 +0200
Michael Niedermayer wrote:
> On Thu, May 05, 2016 at 01:45:55AM +0100, Derek Buitenhuis wrote:
> > I am not at all opposed to the idea. I just don't think the current way
> > of using internal headers, structs, and APIS (and a wrapper main()) should
> > at all
On Wed, May 11, 2016 at 9:04 PM, Reimar Döffinger
wrote:
>
>
> On 11.05.2016, at 20:37, Michael Niedermayer
> wrote:
>
> > On Wed, May 11, 2016 at 06:39:20PM +0200, Matthieu Bouron wrote:
> >> From: Matthieu Bouron
> >>
> >> ---
> >>
> >> Hello,
> >>
> >> Here are some benchmark on a rpi2 of th
On 11.05.2016, at 21:09, James Almer wrote:
> On 5/11/2016 4:04 PM, Reimar Döffinger wrote:
>> otherwise there's a risk that enabling the recently discussed
>> -ftree-vecorize for that file specifically would give better results.
>
> Tree vectorize is enabled for x86 and GCC >= 4.9 only. And ev
Le tridi 23 floréal, an CCXXIV, sebechlebsky...@gmail.com a écrit :
> From: Jan Sebechlebsky
>
> Signed-off-by: Jan Sebechlebsky
> ---
> libavformat/tee.c | 171
> --
> 1 file changed, 139 insertions(+), 32 deletions(-)
Just a few comments f
> > > Updated patch attached.
> >
> > Great, so far I found only these problematic files:
> >
> > https://www.datafilehost.com/d/871e7429
>
> Here are 6 more files (of course if you don't know how to fix something them
> decoder still LGTM - it can be fixed later):
>
> https://www.datafilehost
On 11.05.2016, at 17:00, Bruce Dawson
wrote:
> Microsoft is aware of the issue and is working on it. They might fix it.
> They might not, however, because regressions in non-standard behavior are
> not as serious. We will see.
>
> I'm committing a Chromium specific workaround because that allow
On 5/11/2016 4:04 PM, Reimar Döffinger wrote:
> otherwise there's a risk that enabling the recently discussed -ftree-vecorize
> for that file specifically would give better results.
Tree vectorize is enabled for x86 and GCC >= 4.9 only. And even then
it may not even remain like that much longer.
On 11.05.2016, at 14:03, Carl Eugen Hoyos wrote:
> 章军海 <243186085 qq.com> writes:
>
>> I wonder how you all test FATE on arm or arm64 hardware.
>
> To actually test fate on small devices like mobile phones,
> you have to use the configure option --target-exec and
> point it to ssh, you need
On Wed, May 11, 2016 at 08:16:49PM +0200, Christophe Gisquet wrote:
> 2016-05-01 15:33 GMT+02:00 Christophe Gisquet :
> > The loops are guaranteed to be at least multiples of 8, so this
> > unrolling is safe but allows exploiting execution ports.
> >
> > For int32 version: 68 -> 58c.
>
> Ping?
>
On 11.05.2016, at 20:37, Michael Niedermayer wrote:
> On Wed, May 11, 2016 at 06:39:20PM +0200, Matthieu Bouron wrote:
>> From: Matthieu Bouron
>>
>> ---
>>
>> Hello,
>>
>> Here are some benchmark on a rpi2 of the attached patch.
>>
>> ./ffmpeg -f lavfi -i
>> sine=440,aformat=sample_fmts=
On Wed, May 11, 2016 at 06:39:20PM +0200, Matthieu Bouron wrote:
> From: Matthieu Bouron
>
> ---
>
> Hello,
>
> Here are some benchmark on a rpi2 of the attached patch.
>
> ./ffmpeg -f lavfi -i
> sine=440,aformat=sample_fmts=fltp,asetnsamples=4096,abench=start,aresample=48000,abench=stop
> -
Hi,
Updated patch attached.
0001-avfilter-add-loudnorm.patch
Description: Binary data
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
2016-05-01 15:33 GMT+02:00 Christophe Gisquet :
> The loops are guaranteed to be at least multiples of 8, so this
> unrolling is safe but allows exploiting execution ports.
>
> For int32 version: 68 -> 58c.
Ping?
This was ok'ed by James irrespective of the auto-vectorization
discussion, but I don
On Thu, May 05, 2016 at 01:45:55AM +0100, Derek Buitenhuis wrote:
> On 5/5/2016 1:26 AM, Michael Niedermayer wrote:
> > It was possible to write and register external protocols, codecs,
> > (de)muxers,... in the distant past.
>
> Was there even a such thing as a public API at the time? ;)
There w
From: Matthieu Bouron
---
Hello,
Here are some benchmark on a rpi2 of the attached patch.
./ffmpeg -f lavfi -i
sine=440,aformat=sample_fmts=fltp,asetnsamples=4096,abench=start,aresample=48000,abench=stop
-t 1000 -f null -
With patch:avg=0.001159 speed=44,1x
Without patch: avg=0.001297 s
> > Updated patch attached.
>
> Great, so far I found only these problematic files:
>
> https://www.datafilehost.com/d/871e7429
Here are 6 more files (of course if you don't know how to fix something them
decoder still LGTM - it can be fixed later):
https://www.datafilehost.com/d/bc2e06a2
___
On Tue, May 10, 2016 at 05:45:59PM +0200, Matthieu Bouron wrote:
> From: Matthieu Bouron
>
> ---
> libavfilter/avfilter.c | 2 +-
> libavfilter/avfilter.h | 4 ++--
> libavfilter/framepool.c | 24
> libavfilter/framepool.h | 32
> lib
On Wed, May 11, 2016 at 10:51:32AM -0500, Dave Hess wrote:
> >you need something that gives more weight to large localized changes
> >likely
>
> Yes, good point. How about additional parameters to the function that
> specify a grid pattern and if any individual section in that grid meets the
> th
On Tue, May 10, 2016 at 05:46:00PM +0200, Matthieu Bouron wrote:
> From: Matthieu Bouron
>
> ---
> libavfilter/framepool.c | 106
>
> libavfilter/framepool.h | 36 +++-
> 2 files changed, 141 insertions(+), 1 deletion(-)
this someti
>you need something that gives more weight to large localized changes
>likely
Yes, good point. How about additional parameters to the function that
specify a grid pattern and if any individual section in that grid meets the
threshold, then it returns true? It could even have another parameter to
Hi,
I'll be working on tee muxer improvement during GSoC 2016 and I thought
maybe it is a good idea to ask about ideas which any of you might have
regarding what could be done in avformat/tee.
Currently, the tee muxer works in a simple way, incoming packets are
just iteratively fed to several
Microsoft is aware of the issue and is working on it. They might fix it.
They might not, however, because regressions in non-standard behavior are
not as serious. We will see.
I'm committing a Chromium specific workaround because that allows us to
test Microsoft's new optimizer - I want to be able
On Wed, May 11, 2016 at 4:19 PM, Bruce Dawson
wrote:
> The error is:
>
> wavdec.obj : error LNK2001: unresolved external symbol _ff_w64_guid_data
>
> This happens because the compiler does not recognize early enough
> that _ff_w64_guid_data is never actually needed, so it creates a reference
>
The error is:
wavdec.obj : error LNK2001: unresolved external symbol _ff_w64_guid_data
This happens because the compiler does not recognize early enough
that _ff_w64_guid_data is never actually needed, so it creates a reference
to it.
I am adding a temporary fix to Chromium's ffmpeg repo so
On Tue, May 10, 2016 at 01:22:53PM -0400, Bruce Dawson wrote:
> BTW, the reason I suggested the CONFIG_W64_DEMUXER patch that started this
> email thread is because the pre-release version of VS 2015 Update 3 can't
> handle that code as-is. Microsoft might change their compiler before the
> officia
On Tue, May 10, 2016 at 10:25:16AM -0500, Dave Hess wrote:
> I'm working on a post-process that will create thumbnails of a computer
> screen capture. Is there a frame change detection algorithm similar to
> "scene" that will detect changes from a previously detected frame or from x
> frames prior
> On 5/11/16, Paul B Mahol wrote:
> > Hi,
> >
> > patch attached.
> >
>
> Updated patch attached.
Great, so far I found only these problematic files:
https://www.datafilehost.com/d/871e7429
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http:/
On Wed, May 11, 2016 at 12:41:43PM +, Davinder Singh wrote:
> single patch
>
> On Sun, May 8, 2016 at 1:18 AM Davinder Singh wrote:
>
> > separated motion vector types (forward or backward) from frame picture
> > types as MVs are associated with picture types only in video coding.
> >
> > op
single patch
On Sun, May 8, 2016 at 1:18 AM Davinder Singh wrote:
> separated motion vector types (forward or backward) from frame picture
> types as MVs are associated with picture types only in video coding.
>
> option `mv` can have two values:
> forward predicted or backward predicted.
>
> op
章军海 <243186085 qq.com> writes:
> I wonder how you all test FATE on arm or arm64 hardware.
To actually test fate on small devices like mobile phones,
you have to use the configure option --target-exec and
point it to ssh, you need a running ssh daemon on your
device, this means a jailbroken iP
On Wed, May 11, 2016 at 03:39:45PM +0800, 章军海 wrote:
> Hi Michael:
>
> I have known how to use FATE test on MAC about x86 architecture code in
> FFMPEG,
> running make fate-rsync and then make fate in ffmpeg folder. While I really
> want to know
> how to finish FATE test for aarch
On 5/11/16, Paul B Mahol wrote:
> Hi,
>
> patch attached.
>
Updated patch attached.
0001-lavc-add-IFF-ANIM-decoder.patch.bz2
Description: BZip2 compressed data
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ff
Hi Michael:
I have known how to use FATE test on MAC about x86 architecture code in FFMPEG,
running make fate-rsync and then make fate in ffmpeg folder. While I really
want to know
how to finish FATE test for aarch64 architechture source code on hardware as
mine MX5 mobile
iphone
44 matches
Mail list logo
