Le 16/04/2025 à 15:26, Lynne a écrit :
We violated the spec, which, despite the actual command buffer pool
*not* being involved in any functions which require external synchronization
of the pool, *require* external synchronization even if only the
command buffers are used.
This also has the eff
Patches attached.
- Andreas
From 90f466ce00debdd8d4d557ffcd5ec2a7065d9719 Mon Sep 17 00:00:00 2001
From: Andreas Rheinhardt
Date: Fri, 11 Apr 2025 22:46:57 +0200
Subject: [PATCH 01/14] avcodec/magicyuvenc: Fix setting nb_slices
Do not derive it via av_cpu_count() in case AVCodecContext.slices
is
Michael Niedermayer:
> On Wed, Apr 16, 2025 at 11:17:30AM +0200, Andreas Rheinhardt wrote:
>> Patches attached.
>>
>> - Andreas
>
>> dvbsubenc.c |3 +++
>> 1 file changed, 3 insertions(+)
>> 27ce315dbaee02f8c92c12f8c9cd0c8c7edc54fb
>> 0001-avcodec-dvbsubenc-Sanity-check-num_rects.patch
>> F
The default behavior for VVenC (since v1.10.0) is to create an IDR with
leading pictures for the first picture in decoding order (POC 32). This
leads to FFmpeg generating an edit list with an empty entry, skipping
the leading pictures.
This patch fixes the calculation for the start_pts, while the
From: Zhao Zhili
Don't select sample with small dts when interleaved_read is disabled.
---
libavformat/mov.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libavformat/mov.c b/libavformat/mov.c
index 452690090c..cb81b6c4da 100644
--- a/libavformat/mov.c
+++ b/libavformat/mov
Some loop counters were initialized incorrectly.
This patch enhances the handling of loop iterations and residual parts.
---
libavcodec/loongarch/hevc_mc_bi_lsx.c | 126 --
libavcodec/loongarch/hevc_mc_uni_lsx.c | 325 +++--
libavcodec/loongarch/hevcdsp_lsx.c | 233
> -Original Message-
> From: ffmpeg-devel On Behalf Of
> Michael Niedermayer
> Sent: Mittwoch, 16. April 2025 23:25
> To: FFmpeg development discussions and patches de...@ffmpeg.org>
> Subject: Re: [FFmpeg-devel] [PATCH 1/3] avutil/dict2: Add
> AVDictionary2 with hash-based lookup
>
>
From: Guillaume Khayat
Improve performance (+17%) of ebur_128 filter using AVX2 and FMA instruction in
the body of the filter_frame function.
## Benchmark
Tested with hyperfine
hyperfine --warmup 2 "./ffmpeg_reference -i ~/test.wav -vn -af
ebur128=peak=none:framelog=quiet -f null -" "./ffmpe
And improve the vvencMsgLevel <-> AV_LOG_* mapping.
Signed-off-by: James Almer
---
libavcodec/libvvenc.c | 35 +++
1 file changed, 27 insertions(+), 8 deletions(-)
diff --git a/libavcodec/libvvenc.c b/libavcodec/libvvenc.c
index 29a49ba091..c71fd78f1e 100644
---
gkha...@spectre-music.com:
> From: Guillaume Khayat
>
> Improve performance (+17%) of ebur_128 filter using AVX2 and FMA instruction
> in the body of the filter_frame function.
>
> ## Benchmark
>
> Tested with hyperfine
>
> hyperfine --warmup 2 "./ffmpeg_reference -i ~/test.wav -vn -af
> ebu
From: IndecisiveTurtle
---
libavcodec/Makefile| 2 +-
libavcodec/vc2enc.c| 513 +
libavcodec/vc2enc_common.c | 376 +++
libavcodec/vc2enc_common.h | 196 ++
4 files changed, 581 insertions(+), 506 deletions
Hi
On Thu, Apr 17, 2025 at 07:57:46PM +0200, Nicolas George wrote:
> Michael Niedermayer (HE12025-04-17):
> > +AVMap *av_map_new(AVMapCompareFunc cmp_keyvalue, int cmp_flags,
> > AVMapCopyFunc copy, AVMapFreeFunc freef)
> > +{
> > +AVMap *s = av_mallocz(sizeof(*s));
> > +if (!s)
> > +
Hi
On Wed, Apr 16, 2025 at 10:12:19AM +, softworkz wrote:
> From: softworkz
>
> The key benefits are:
>
> - Different to other graph printing methods, this is outputting:
> - all graphs with runtime state
> (including auto-inserted filters)
> - each graph with its inputs and outputs
> -Original Message-
> From: ffmpeg-devel On Behalf Of
> Michael Niedermayer
> Sent: Donnerstag, 17. April 2025 01:41
> To: FFmpeg development discussions and patches de...@ffmpeg.org>
> Subject: Re: [FFmpeg-devel] [PATCH 2/3] doc/dict2: Add doc and api
> change for AVDictionary2
>
> H
From: softworkz
Signed-off-by: softworkz
---
fftools/textformat/tf_internal.h | 6 +-
1 file changed, 1 insertion(+), 5 deletions(-)
diff --git a/fftools/textformat/tf_internal.h b/fftools/textformat/tf_internal.h
index 7b326328cb..e145bc83bb 100644
--- a/fftools/textformat/tf_internal.h
+
From: softworkz
Signed-off-by: softworkz
---
ffbuild/common.mak | 28 ++-
fftools/Makefile | 3 +-
fftools/resources/.gitignore | 4 +
fftools/resources/Makefile | 27 +++
fftools/resources/graph.css | 353 +++
fftools/resources/gra
> -Original Message-
> From: ffmpeg-devel On Behalf Of
> softworkz .
> Sent: Mittwoch, 16. April 2025 13:33
> To: FFmpeg development discussions and patches de...@ffmpeg.org>
> Subject: Re: [FFmpeg-devel] [PATCH v2 02/10] fftools/textformat:
> Quality improvements
>
>
>
> > -Ori
> -Original Message-
> From: ffmpeg-devel On Behalf Of
> softworkz .
> Sent: Montag, 14. April 2025 17:45
> To: FFmpeg development discussions and patches de...@ffmpeg.org>
> Subject: Re: [FFmpeg-devel] SW's Patchsets Overview
>
> Hello,
>
>
> this is a follow-up to the list of patc
On 18/04/2025 01:55, IndecisiveTurtle wrote:
From: IndecisiveTurtle
Performance wise, encoding a 1080p 1-minute video is performed in about 2.5
minutes with the cpu encoder running on my Ryzen 5 4600H, while it takes about
30 seconds on my NVIDIA GTX 1650
Haar shader has a subgroup optimized
On 4/17/2025 8:55 PM, IndecisiveTurtle wrote:
From: IndecisiveTurtle
Performance wise, encoding a 1080p 1-minute video is performed in about 2.5
minutes with the cpu encoder running on my Ryzen 5 4600H, while it takes about
30 seconds on my NVIDIA GTX 1650
Haar shader has a subgroup optimize
On 4/17/2025 10:11 PM, Lynne wrote:
On 18/04/2025 01:55, IndecisiveTurtle wrote:
From: IndecisiveTurtle
Performance wise, encoding a 1080p 1-minute video is performed in
about 2.5 minutes with the cpu encoder running on my Ryzen 5 4600H,
while it takes about 30 seconds on my NVIDIA GTX 1650
Some loop counters were initialized incorrectly.
This patch enhances the handling of loop iterations and residual parts.
---
libavcodec/loongarch/hevc_mc_bi_lsx.c | 299 ++--
libavcodec/loongarch/hevc_mc_uni_lsx.c| 325 +-
libavcodec/loongarch/hevcdsp_in
On 4/17/2025 2:07 PM, Andreas Rheinhardt wrote:
gkha...@spectre-music.com:
From: Guillaume Khayat
Improve performance (+17%) of ebur_128 filter using AVX2 and FMA instruction in
the body of the filter_frame function.
## Benchmark
Tested with hyperfine
hyperfine --warmup 2 "./ffmpeg_referen
Michael Niedermayer (HE12025-04-17):
> +AVMap *av_map_new(AVMapCompareFunc cmp_keyvalue, int cmp_flags,
> AVMapCopyFunc copy, AVMapFreeFunc freef)
> +{
> +AVMap *s = av_mallocz(sizeof(*s));
> +if (!s)
> +return NULL;
> +
> +s->copy = copy;
> +s->freef = fre
From: softworkz
Signed-off-by: softworkz
---
fftools/textformat/avtextformat.c | 111 +++---
fftools/textformat/avtextformat.h | 6 +-
fftools/textformat/tf_default.c | 8 ++-
fftools/textformat/tf_ini.c | 2 +-
fftools/textformat/tf_json.c | 8 ++-
> -Original Message-
> From: ffmpeg-devel On Behalf Of
> Michael Niedermayer
> Sent: Donnerstag, 17. April 2025 20:42
> To: FFmpeg development discussions and patches de...@ffmpeg.org>
> Subject: Re: [FFmpeg-devel] [PATCH v2 09/10] fftools/graphprint: Add
> execution graph printing
>
>
From: softworkz
Signed-off-by: softworkz
---
fftools/textformat/avtextformat.c | 104 +++--
fftools/textformat/avtextformat.h | 16 ++---
fftools/textformat/avtextwriters.h | 11 ++-
fftools/textformat/tf_compact.c| 91 ++---
fftools/textform
Shortest cover letter for my longest-running FFmpeg patchset:
* Apply
* Build
* Add the "-sg" switch to any FFmpeg command line
* Press 'q' when you don't want to wait
SG = Show Graph
Documentation and examples can be found here:
https://github.com/softworkz/ffmpeg_output_apis/wiki
Versio
From: softworkz
The key benefits are:
- Different to other graph printing methods, this is outputting:
- all graphs with runtime state
(including auto-inserted filters)
- each graph with its inputs and outputs
- all filters with their in- and output pads
- all connections between all
From: softworkz
Signed-off-by: softworkz
---
fftools/textformat/avtextwriters.h | 2 +-
fftools/textformat/tf_compact.c| 32 ---
fftools/textformat/tf_default.c| 27 +++---
fftools/textformat/tf_flat.c | 25 +++--
fftools/textformat/tf_ini.c| 24 +++--
From: softworkz
to allow filtergraph printing to access the information.
Signed-off-by: softworkz
---
fftools/ffmpeg_filter.c | 190 +---
fftools/ffmpeg_filter.h | 234
2 files changed, 235 insertions(+), 189 deletions(-)
cr
From: softworkz
Signed-off-by: softworkz
---
doc/APIchanges | 3 +++
libavfilter/avfilter.c | 9 +
libavfilter/avfilter.h | 12
3 files changed, 24 insertions(+)
diff --git a/doc/APIchanges b/doc/APIchanges
index 22aa6fa5c7..4f60bbbe3d 100644
--- a/doc/APIchanges
From: softworkz
This function works analog to the avtext_print_string() which already
has a flags parameter.
Signed-off-by: softworkz
---
fftools/textformat/avtextformat.c | 21 +
fftools/textformat/avtextformat.h | 2 ++
2 files changed, 23 insertions(+)
diff --git a/fft
From: softworkz
remember this: -sg <= show-graph
Signed-off-by: softworkz
---
doc/ffmpeg.texi | 4 +
fftools/Makefile | 1 +
fftools/ffmpeg.c | 2 +-
fftools/ffmpeg.h | 1 +
fftools/ffmpeg_filter.c | 2 +-
fftools/ffmpeg_opt.c
> -Original Message-
> From: ffmpeg-devel On Behalf Of
> softworkz .
> Sent: Mittwoch, 16. April 2025 13:33
> To: FFmpeg development discussions and patches de...@ffmpeg.org>
> Subject: Re: [FFmpeg-devel] [PATCH v2 02/10] fftools/textformat:
> Quality improvements
>
>
>
> > -Ori
From: IndecisiveTurtle
Performance wise, encoding a 1080p 1-minute video is performed in about 2.5
minutes with the cpu encoder running on my Ryzen 5 4600H, while it takes about
30 seconds on my NVIDIA GTX 1650
Haar shader has a subgroup optimized variant that applies when configured
wavelet
On Mon, Apr 7, 2025 at 11:10 AM wrote:
> From: Thierry Foucu
>
> Make sure we are reading 16 bytes for the MD5
> ---
> libavformat/takdec.c | 5 -
> 1 file changed, 4 insertions(+), 1 deletion(-)
>
> diff --git a/libavformat/takdec.c b/libavformat/takdec.c
> index 21fff3fcbf..61b9f001c4 100
From: IndecisiveTurtle
Performance wise, encoding a 1080p 1-minute video is performed in about 2.5
minutes with the cpu encoder running on my Ryzen 5 4600H, while it takes about
30 seconds on my NVIDIA GTX 1650
Haar shader has a subgroup optimized variant that applies when configured
wavelet
Some loop counters were initialized incorrectly.
This patch enhances the handling of loop iterations and residual parts.
---
libavcodec/loongarch/hevc_mc_bi_lsx.c | 126 --
libavcodec/loongarch/hevc_mc_uni_lsx.c | 325 +++--
libavcodec/loongarch/hevcdsp_lsx.c | 233
On 4/17/2025 4:08 PM, Thierry Foucu wrote:
On Mon, Apr 7, 2025 at 11:10 AM wrote:
From: Thierry Foucu
Make sure we are reading 16 bytes for the MD5
---
libavformat/takdec.c | 5 -
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/libavformat/takdec.c b/libavformat/takdec.c
From: IndecisiveTurtle
---
libavcodec/vulkan/common.comp | 54 ---
1 file changed, 44 insertions(+), 10 deletions(-)
diff --git a/libavcodec/vulkan/common.comp b/libavcodec/vulkan/common.comp
index 10af9c0623..db216a2ac6 100644
--- a/libavcodec/vulkan/common.comp
From: IndecisiveTurtle
---
libavcodec/vc2enc.c| 2 +-
libavcodec/vc2enc_common.c | 18 +-
libavcodec/vc2enc_common.h | 2 +-
3 files changed, 11 insertions(+), 11 deletions(-)
diff --git a/libavcodec/vc2enc.c b/libavcodec/vc2enc.c
index 2e849eb09e..c0f542e116 100644
--
From: IndecisiveTurtle
Prevents compiler from mistaking it as a string
Also makes passing it to the GPU in a buffer easier
---
libavcodec/vc2enc_common.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libavcodec/vc2enc_common.h b/libavcodec/vc2enc_common.h
index eaaf5ac99c..
43 matches
Mail list logo
