On Mon, 3 Mar 2025, Krzysztof Pyrkosz via ffmpeg-devel wrote:
The idea is to split the 16 bit coefficients into lower and upper half,
invoke udot for the lower half, shift by 8, and follow by udot for the
upper half.
Benchmark on A78:
bgra_to_y_128_c: 682.0
We normally don't need else statements here; the common pattern
is to assign lower level SIMD implementations first, then
conditionally reassign higher level ones afterwards, if supported.
---
Sorry, I didn't notice this while reviewing the patch, until after
pushing it.
---
libswscale/aarch64/sw
On Mon, 3 Mar 2025, Krzysztof Pyrkosz via ffmpeg-devel wrote:
This patch replaces blocks of instructions performing rounding and
widening shifts with one-liners achieving the same result.
Before and after on A78
dmvr_8_12x20_neon: 86.2 ( 6.90x)
dmvr_8_20x12_
The functions are:
av_frame_side_data_get_nth_c()
av_frame_side_data_get_nth()
av_frame_get_nth_side_data()
They mimic the counterparts without the `_nth` suffix/infix, with an extra
argument, which specifies the relative position of the entry of a given
type.
Signed-off-by: Leandro Santiago
--
On Tue, Mar 04, 2025 at 01:47:10PM +0100, Andreas Rheinhardt wrote:
> Michael Niedermayer:
> > On Mon, Feb 03, 2025 at 09:54:57AM +0530, Viraaj Raulgaonkar wrote:
> >> Change the type of logctx from void* to AVFormatContext*. This is in
> >> preparation for the next commit.
> >> ---
> >> libavform
On Mon, Feb 03, 2025 at 09:54:58AM +0530, Viraaj Raulgaonkar wrote:
> If strict_std_compliance < FF_COMPLIANCE_STRICT, warn the user of the
> invalid sample rate instead of returning an error. In certain cases the
> sample rate can get decoded later on.
>
> Fixes Trac Ticket #11361.
> ---
> libav
On 3/4/2025 12:28 PM, Michael Niedermayer wrote:
On Tue, Mar 04, 2025 at 01:47:10PM +0100, Andreas Rheinhardt wrote:
Michael Niedermayer:
On Mon, Feb 03, 2025 at 09:54:57AM +0530, Viraaj Raulgaonkar wrote:
Change the type of logctx from void* to AVFormatContext*. This is in
preparation for the
Michael Niedermayer:
> On Mon, Feb 03, 2025 at 09:54:57AM +0530, Viraaj Raulgaonkar wrote:
>> Change the type of logctx from void* to AVFormatContext*. This is in
>> preparation for the next commit.
>> ---
>> libavformat/riff.h| 2 +-
>> libavformat/riffdec.c | 14 +++---
>> 2 files c
Signed-off-by: Kacper Michajłow
---
compat/w32pthreads.h | 30 ++
libavutil/thread.h | 2 ++
2 files changed, 32 insertions(+)
diff --git a/compat/w32pthreads.h b/compat/w32pthreads.h
index fd6428e29f..83b4819205 100644
--- a/compat/w32pthreads.h
+++ b/compat/w32pt
Signed-off-by: Kacper Michajłow
---
compat/w32pthreads.h | 30 ++
libavutil/thread.h | 2 ++
2 files changed, 32 insertions(+)
diff --git a/compat/w32pthreads.h b/compat/w32pthreads.h
index fd6428e29f..8d5b4729fa 100644
--- a/compat/w32pthreads.h
+++ b/compat/w32pt
On Tue, 4 Mar 2025 at 15:57, Kacper Michajlow wrote:
>
> On Tue, 4 Mar 2025 at 15:50, James Almer wrote:
> >
> > On 3/4/2025 9:52 AM, Kacper Michajłow wrote:
> > > Signed-off-by: Kacper Michajłow
> > > ---
> > > compat/w32pthreads.h | 30 ++
> > > libavutil/thread.
On Tue, Mar 4, 2025 at 4:02 AM Tomas Härdin wrote:
> So it's "US" even if something is broadcast in Sweden? T.35 has a whole
> list of countries, some of which don't exist any more, and some new
> ones are missing
Correct. So the T.35 country code doesn't actually have anything to
do with where
Andreas Rheinhardt:
> Patches attached
>
>
Will apply this patchset tomorrow unless there are objections.
- Andreas
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link abo
On Mon, 3 Mar 2025, 20:35 Michael Niedermayer,
wrote:
> Hi jb
>
> On Thu, Feb 27, 2025 at 12:11:39AM +0100, Jean-Baptiste Kempf wrote:
> > On Wed, 26 Feb 2025, at 02:51, Michael Niedermayer wrote:
> > > 1. I agree we need discussions, transparency and maybe IRC or some
> other
> > >audio/vide
>
> Nevertheless, the CC does issue a warning regarding unnecessarily
> offensive recent comments by Kieran on other topics, as well as for
> initially failing to provide background and context regarding Paul’s
> action on IRC.
>
Hi CC,
What comments are these? This is very vague as to what I've
Le 4 mars 2025 17:56:20 GMT+02:00, Kieran Kunhya via ffmpeg-devel
a écrit :
>>
>> Nevertheless, the CC does issue a warning regarding unnecessarily
>> offensive recent comments by Kieran on other topics, as well as for
>> initially failing to provide background and context regarding Paul’s
>> a
Hi Marth64
2nd mail to keep this clean and readable
On Tue, Mar 04, 2025 at 09:35:09AM -0600, Marth64 wrote:
> Hello,
>
> The FFmpeg Community Committee (CC) has received a complaint regarding
> alleged insults from Kieran against Yigithan and Michael.
>
> After careful review, the CC did not f
John Dorian:
> I discovered a crash here if log handler function tries to get the class
> name from "s"
>
> mpeg12dec.c:
> if (get_bits_left(&s->gb) < 0) {
> av_log(s, AV_LOG_ERROR, "overread %d\n", -get_bits_left(&s->gb));
> return AVERROR_INVALIDDATA;
> }
>
> And it seems
> -Original Message-
> From: ffmpeg-devel On Behalf Of
> epira...@gmail.com
> Sent: Dienstag, 4. März 2025 22:06
> To: FFmpeg development discussions and patches
> Subject: Re: [FFmpeg-devel] GSoC 2025 - OggKate subtitles support
>
>
>
> On 4 Mar 2025, at 20:44, Soft Works wrote:
>
On 3/4/25 22:06, Andreas Rheinhardt wrote:
I added a comment to #10 and modified #18 as described. I also changed
#21 to protect the macro in parentheses and simplified the FF_RC_OFFSET
macro in #31. Furthermore, there are now five more patches. All
attached. https://github.com/mkver/FFmpeg/tre
John Dorian:
> I discovered a crash here if log handler function tries to get the class
> name from "s"
>
> mpeg12dec.c:
> if (get_bits_left(&s->gb) < 0) {
> av_log(s, AV_LOG_ERROR, "overread %d\n", -get_bits_left(&s->gb));
> return AVERROR_INVALIDDATA;
> }
>
> And it seems
On 3/4/25 23:11, Andreas Rheinhardt wrote:
Patch attached
diff --git a/libavcodec/mpeg12dec.c b/libavcodec/mpeg12dec.c
index 843640edbf..14c028379e 100644
--- a/libavcodec/mpeg12dec.c
+++ b/libavcodec/mpeg12dec.c
@@ -1655,11 +1655,11 @@ static int mpeg_decode_slice(MpegEncContext *s, int
mb
Patch attached. Will apply it tomorrow unless there are objections.
- Andreas
From 9e2773034ba177861e32b746b5dbd8d0370661a5 Mon Sep 17 00:00:00 2001
From: Andreas Rheinhardt
Date: Tue, 4 Mar 2025 15:34:38 +0100
Subject: [PATCH] avformat/rtp_av1: Add necessary headers
Also use a void* instead of
On 3/4/2025 9:52 AM, Kacper Michajłow wrote:
Signed-off-by: Kacper Michajłow
---
compat/w32pthreads.h | 30 ++
libavutil/thread.h | 2 ++
2 files changed, 32 insertions(+)
diff --git a/compat/w32pthreads.h b/compat/w32pthreads.h
index fd6428e29f..83b4819205 10
On Tue, 4 Mar 2025 at 15:50, James Almer wrote:
>
> On 3/4/2025 9:52 AM, Kacper Michajłow wrote:
> > Signed-off-by: Kacper Michajłow
> > ---
> > compat/w32pthreads.h | 30 ++
> > libavutil/thread.h | 2 ++
> > 2 files changed, 32 insertions(+)
> >
> > diff --git
Hello,
The FFmpeg Community Committee (CC) has received a complaint regarding
alleged insults from Kieran against Yigithan and Michael.
After careful review, the CC did not find any recent comments from
Kieran that would constitute an insult against either of the alleged
targets. The CC did not e
Hi Marth64
first, thanks for publically explaining the decission
this allows me to comment.
On Tue, Mar 04, 2025 at 09:35:09AM -0600, Marth64 wrote:
> Hello,
>
> The FFmpeg Community Committee (CC) has received a complaint regarding
> alleged insults from Kieran against Yigithan and Michael.
>
Hi Kieran
On Tue, Mar 04, 2025 at 08:13:11AM -0600, Kieran Kunhya via ffmpeg-devel wrote:
[...]
> Hi Michael,
>
> Was all the STF discussion around the first application done in public?
No, jb contacted STF in the name of FFmpeg without any public discussion.
This crossed STFs own reaching out t
ode_slice = ff_nvdec_simple_decode_slice,
+.decode_slice = nvdec_vc1_decode_slice,
.frame_params = nvdec_vc1_frame_params,
.init = ff_nvdec_decode_init,
.uninit = ff_nvdec_decode_uninit,
---
base-commit: f76195ff656d6bea68feee
On Tue, Mar 4, 2025 at 6:05 PM Andreas Rheinhardt
wrote:
> Ramiro Polla:
> >
> > On 3/4/25 14:42, Andreas Rheinhardt wrote:
> >> (Mostly trivial) patches attached. A branch is at
> >> https://github.com/mkver/FFmpeg/tree/mpegvideo_misc
> >
> >
> > [PATCH 10/40] avcodec/mpegvideo_enc: Move default_
Ramiro Polla:
>
> On 3/4/25 14:42, Andreas Rheinhardt wrote:
>> (Mostly trivial) patches attached. A branch is at
>> https://github.com/mkver/FFmpeg/tree/mpegvideo_misc
>
>
> [PATCH 10/40] avcodec/mpegvideo_enc: Move default_mv_penalty to h261enc.c
>
>> diff --git a/libavcodec/h261enc.c b/libav
Don't error out, just ignore unknown codec numbers and pretend
decode succeeded. This is useful for older LucasArts titles
which stack a lot of different FOBJs with different codecs into
a single frame.
Signed-off-by: Manuel Lauss
---
v4: no changes
v3: no changes
v2: Mark frame as corrupt, sugg
This codec alternatingly skips and changes existing pixels.
A second 16bit parameter in the FOBJ header indicates how to do
the pixel changes: either by specifying a LUT in the codec datastream
or by adding a constant value to the pixel.
For ANIMv1 (Rebel Assault 1) the first 8bit parameter is used
On 4 Mar 2025, at 20:44, Soft Works wrote:
>> -Original Message-
>> From: ffmpeg-devel On Behalf Of
>> Gianluca Cannata
>> Sent: Dienstag, 4. März 2025 09:40
>> To: ffmpeg-devel@ffmpeg.org
>> Subject: [FFmpeg-devel] GSoC 2025 - OggKate subtitles support
>>
>> Good morning,
>>
>> I have
Hi,
Expanding on Devin's comment,
See the diagram I have created on page 3 of this deck showing the
digital caption embedding and this XDS stream visually:
https://docs.google.com/presentation/d/19_0EIol3xBy-ubyzHG-vw1y8OLOLOaJgHNlgkY2HsB4/edit#slide=id.g32866a10389_0_53
(Rest of the deck is a w
Ramiro Polla:
>
> On 3/4/25 23:11, Andreas Rheinhardt wrote:
>> Patch attached
>
>> diff --git a/libavcodec/mpeg12dec.c b/libavcodec/mpeg12dec.c
>> index 843640edbf..14c028379e 100644
>> --- a/libavcodec/mpeg12dec.c
>> +++ b/libavcodec/mpeg12dec.c
>> @@ -1655,11 +1655,11 @@ static int mpeg_decode
Gianluca Cannata (HE12025-03-04):
> I have to write a karaoke solution with FFmpeg libav libraries for a
> project of mine and I have found OggKate subtitles a good choice to
> implement such solution and I am thinking of writing a Kate demuxer and
> decoder along with a filter that will use OpenGL
Hi Nicolas,
I would like to add motion to subtitles like fading, scrolling, changing
colors and font size, eccetera.
Is it something that can be of public interest ?
If not, which mailing list is the most appropriate to ask support for
implementing such a thing ?
Sincerely
Gianluca
Il giorno
Good morning,
I have to write a karaoke solution with FFmpeg libav libraries for a
project of mine and I have found OggKate subtitles a good choice to
implement such solution and I am thinking of writing a Kate demuxer and
decoder along with a filter that will use OpenGL to render Kate subtitles
c
> ASS already supports all this. Have you tried using it for your project?
I have not tried yet. I will have.
> Top-posting is forbidden on this mailing-list. If you do not know what
it means look it up.
Sorry.
Thank you again for your support.
Sincerely
Gianluca
_
Gianluca Cannata (HE12025-03-04):
> I would like to add motion to subtitles like fading, scrolling, changing
> colors and font size, eccetera.
ASS already supports all this. Have you tried using it for your project?
> Il giorno mar 4 mar 2025 alle ore 11:46 Nicolas George ha
> scritto:
Top-post
> On Mar 2, 2025, at 05:39, Martin Storsjö wrote:
>
> On Thu, 20 Feb 2025, Zhao Zhili wrote:
>
>> From: Zhao Zhili
>>
>> clang does better than the assembly code before the patch, especially
>> for small size:
>>
>> hevc_idct_4x4_dc_8_c: 11.2 ( 1.00x)
>> he
mån 2025-03-03 klockan 08:55 -0500 skrev Devin Heitmueller:
> On Mon, Mar 3, 2025 at 7:25 AM Tomas Härdin wrote:
> >
> > Looking at CTA-708 at the moment and noticed this. In the future we
> > might want to make it possible for the user to set the country
> > code.
> > This patch makes finding us
mån 2025-03-03 klockan 14:12 +0100 skrev Andreas Rheinhardt:
> Tomas Härdin:
> > #include "libavutil/mem.h"
> > #include "atsc_a53.h"
> > +#include "itut35.h"
> > #include "get_bits.h"
> >
>
> Proper alphabetical ordering please. Here and in the other files.
A bit difficult since none of the
ons 2025-02-19 klockan 16:42 +0100 skrev Tomas Härdin:
> Hi
>
> Rather than continue the RFC thread for this, I decided to submit a
> patch that begins the process of documenting metadata keys.
>
> There's quite a lot of keys used in the codebase, and this patch does
> not cover all of them. But
> -Original Message-
> From: ffmpeg-devel On Behalf Of
> m.kaindl0...@gmail.com
> Sent: Tuesday, February 18, 2025 8:17 PM
> To: 'FFmpeg development discussions and patches' de...@ffmpeg.org>
> Subject: Re: [FFmpeg-devel] [PATCH] avfilter/dnn: add zero-shot image
> classification using
Tomas Härdin:
> * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301
> USA
> */
>
> -#include "libavutil/internal.h"
> -#include "libavutil/opt.h"
> -#include "libavutil/mem.h"
> -
> +#include "atsc_a53.h"
> #include "cbs.h"
> #include "cbs_h265.h"
> -#include "atsc_a
Tomas Härdin:
> +@item ref-frame-config @tab @tab @tab
> +@item reverb @tab S @tab argo_cvg @tab
The first is libvp9enc-only, the second argo_cvg only. IMO it is better
to document this at the documentation of the relevant encoder and demuxer.
- Andreas
__
The left and top parameters of an FOBJ are signed values.
Signed-off-by: Manuel Lauss
---
v4: revert v3, it arose due to a misunderstanding
v3: change the bytestream accessor to signed too
v2: no changes
libavcodec/sanm.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/li
On 4 Mar 2025, at 12:36, Andreas Rheinhardt wrote:
> Tomas Härdin:
>> +@item ref-frame-config @tab @tab @tab
>> +@item reverb @tab S @tab argo_cvg @tab
>
> The first is libvp9enc-only, the second argo_cvg only. IMO it is better
> to document this at the documentation of the relevant encoder and
On 3/4/25 14:42, Andreas Rheinhardt wrote:
(Mostly trivial) patches attached. A branch is at
https://github.com/mkver/FFmpeg/tree/mpegvideo_misc
[PATCH 10/40] avcodec/mpegvideo_enc: Move default_mv_penalty to h261enc.c
diff --git a/libavcodec/h261enc.c b/libavcodec/h261enc.c
index dabab9d8
Hi Michael,
Thanks for your thoughts.
CC did not ignore. One can find a fault in that thread no matter which
lens it is looked under. Our conclusion is aimed to be balanced.
Objectively, Kieran's point was valid - the topic started with focus
on the complaint of our environment having a high lear
I'm not sure I understand what documentation you'd like to add. The
behavior of ff_draw_init* is documented in drawutils.h, would you like
me to add a comment where these functions are called with no risk of
failing (as far as I can see, only qrencode.c)? My worry is that such
a comment can easily
Signed-off-by: James Almer
---
libavfilter/drawutils.c| 4
libavfilter/vsrc_testsrc.c | 19 +++
2 files changed, 23 insertions(+)
diff --git a/libavfilter/drawutils.c b/libavfilter/drawutils.c
index b86f666f1f..299eb0fd93 100644
--- a/libavfilter/drawutils.c
+++ b/libav
Patch attached
- Andreas
From 1095ce36d0e7fd26a290638ca6a505dc27c8 Mon Sep 17 00:00:00 2001
From: Andreas Rheinhardt
Date: Tue, 4 Mar 2025 23:06:38 +0100
Subject: [PATCH] avcodec/mpeg12dec: Use proper logcontext
Signed-off-by: Andreas Rheinhardt
---
libavcodec/mpeg12dec.c | 4 ++--
1 file
I discovered a crash here if log handler function tries to get the class
name from "s"
mpeg12dec.c:
if (get_bits_left(&s->gb) < 0) {
av_log(s, AV_LOG_ERROR, "overread %d\n", -get_bits_left(&s->gb));
return AVERROR_INVALIDDATA;
}
And it seems to be because MpegEncContext->cl
On Tue, 4 Mar 2025, 11:17 Michael Niedermayer,
wrote:
> Hi Kieran
>
> On Tue, Mar 04, 2025 at 08:13:11AM -0600, Kieran Kunhya via ffmpeg-devel
> wrote:
> [...]
> > Hi Michael,
> >
> > Was all the STF discussion around the first application done in public?
>
> No, jb contacted STF in the name of F
>
> Kierans reply later refered back to this:
> > OMG a GSoC contributor is complaining about how hard the contribution
>
That's a selective quotation of an incomplete sentence.
Kieran
>
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffm
> -Original Message-
> From: ffmpeg-devel On Behalf Of
> Gianluca Cannata
> Sent: Dienstag, 4. März 2025 09:40
> To: ffmpeg-devel@ffmpeg.org
> Subject: [FFmpeg-devel] GSoC 2025 - OggKate subtitles support
>
> Good morning,
>
> I have to write a karaoke solution with FFmpeg libav librarie
On 2023-08-23 4:52 p.m., Michael Niedermayer wrote:
On Fri, Aug 18, 2023 at 10:14:04PM -0400, Brad Smith wrote:
lsws/ppc/yuv2rgb_altivec: Fix build in non-VSX environments with Clang
Add a check for the existence of the vec_xl() function. Clang provides
the function even with VSX not enabled.
This allows switching it on conditionally and also for non float,
it may improve compression for RGB data that was paletted
or other synthetic images
Sponsored-by: Sovereign Tech Fund
Signed-off-by: Michael Niedermayer
---
libavcodec/ffv1.h | 1 +
libavcodec/ffv1dec.c | 8 ++
On Tue, Mar 04, 2025 at 10:12:40AM +0100, Tomas Härdin wrote:
[...]
> atsc_a53.c |2 +-
> vaapi_encode_h264.c | 18 --
> vaapi_encode_h265.c | 21 +
> vulkan_encode_h264.c | 17 +++--
> vulkan_encode_h265.c | 15 ++---
John Dorian:
> Yes, I have a custom log callback. I have multiple files open and I'd like
> to associate logged errors with the correct file, which I am doing through
> the objects that have an "opaque" member (not all, but most do). I don't
> think this is any abuse of the API as it does state tha
Hi Brad,
On Wed, Mar 5, 2025 at 2:24 AM Brad Smith
wrote:
> On 2023-08-23 4:52 p.m., Michael Niedermayer wrote:
> > On Fri, Aug 18, 2023 at 10:14:04PM -0400, Brad Smith wrote:
> >> lsws/ppc/yuv2rgb_altivec: Fix build in non-VSX environments with Clang
> >>
> >> Add a check for the existence of th
When playing out video to a Decklink SDI 4K card I've found that all/most
of the buffered frames get flushed instead of being displayed, essentially
shortening the video by the amount of pre-roll. I found a way to fix this
by adding a counter of outstanding frames, then delaying the call to
StopSch
This improves compression by 0.1% overall
I tried several other things but so far this is the best
compromise between complexity and compression
Sponsored-by: Sovereign Tech Fund
Signed-off-by: Michael Niedermayer
---
libavcodec/ffv1dec_template.c | 10 --
libavcodec/ffv1enc_template.c |
Patches attached
- Andreas
From 6c50b26de89f1bbcfe025aaf3b28c94b68376414 Mon Sep 17 00:00:00 2001
From: Andreas Rheinhardt
Date: Wed, 5 Mar 2025 00:50:49 +0100
Subject: [PATCH 1/3] avcodec/{h263,ituh263,msmpeg4,snow}dec: Use proper
logcontext
The logging functions here can be reached by codecs
On Fri, Jan 10, 2025 at 05:06:14AM +0100, Michael Niedermayer wrote:
> exr should not do a internal pixel format convert
>
> Sponsored-by: Sovereign Tech Fund
> Signed-off-by: Michael Niedermayer
will apply
[...]
--
Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB
Its n
Yes, I have a custom log callback. I have multiple files open and I'd like
to associate logged errors with the correct file, which I am doing through
the objects that have an "opaque" member (not all, but most do). I don't
think this is any abuse of the API as it does state that the first member
of
69 matches
Mail list logo
