As recommended by Nicholas George. Uses refcounting for planar formats similar
to this change for alphamerge
https://patchwork.ffmpeg.org/project/ffmpeg/patch/20250807195800.77462-1-jestrada.l...@gmail.com/
---
libavfilter/vf_extractplanes.c | 41 +-
1 file changed
Addressed and attached the updated patch. I sent my previous email
incorrectly. My bad.
On Thu, Aug 7, 2025 at 1:13 PM James Almer wrote:
> On 8/7/2025 4:58 PM, Jorge Estrada wrote:
> > Use reference plane when handling planar formats. Should address
> > https://code.ffmpeg.or
Addressed. Uses av_buffer_replace now
---
libavfilter/vf_alphamerge.c | 24 +---
1 file changed, 17 insertions(+), 7 deletions(-)
diff --git a/libavfilter/vf_alphamerge.c b/libavfilter/vf_alphamerge.c
index f5779484a9..69c0dd04ea 100644
--- a/libavfilter/vf_alphamerge.c
+++ b
Use reference plane when handling planar formats. Should address
https://code.ffmpeg.org/FFmpeg/FFmpeg/issues/20153
Example:
ffmpeg -f lavfi -i "color=red:s=640x480:d=5,format=yuva420p" \
-f lavfi -i
"color=black:s=640x480:d=5,format=yuv420p,geq=lum='255*gt(150,hypot(X-W/2,Y-H/2))':a=0"
\
formats in alphamerge or keep anything CUDA related separated into
alphamerge_cuda which would basically be doing the same thing?
On Thu, Aug 7, 2025 at 3:15 AM Nicolas George wrote:
> Jorge Estrada (HE12025-08-06):
> > This patch adds the alphamerge_cuda video filter
>
> Thank
This patch adds the alphamerge_cuda video filter
Example usage:
ffmpeg -f lavfi -i "color=c=red:s=1280x720:d=5,format=yuva420p" \
-f lavfi -i
"color=c=black:s=1280x720:d=5,geq=lum='255*gt(W/4,hypot(X-W/2,Y-H/2))'" \
-f lavfi -i "color=c=blue:s=1280x720:d=5" \
-filter_complex \
"[0:v]hwupload_cuda
ping
On Sat, Jul 19, 2025 at 3:32 PM Jorge Estrada
wrote:
> When converting from yuva420p to yuv420p the resize loop would try to
> access an output plane that did not exist.
> Added a check for the output planes existence.
>
> Example cmd: ffmpeg -hwaccel cuda -hwaccel_output
-c:v h264_nvenc -f null -
Signed-off-by: Jorge Estrada
---
libavfilter/vf_scale_npp.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libavfilter/vf_scale_npp.c b/libavfilter/vf_scale_npp.c
index 1b1b7b9fc9..5387bf9c7a 100644
--- a/libavfilter/vf_scale_npp.c
+++ b/libavfilter
e_npp=format=yuv420p[video],[video][image]overlay_cuda=enable='between(t,30,60)'"
-c:v h264_nvenc -c:a copy -y overlay-output-gpu.mp4
Signed-off-by: Jorge Estrada
---
libavfilter/vf_overlay_cuda.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/libavfilter/vf_overlay_
Changes look good to me. I appreciate the fixes. Thanks!
On Thu, Jul 3, 2025 at 3:49 PM Timo Rothenpieler
wrote:
> On 03.07.2025 23:05, Jorge Estrada wrote:
> > ping
>
> Completely forgot about this, sorry.
>
> Patch does not compile, cause vector_types.h does not exist.
ping
On Thu, Jun 19, 2025 at 7:43 PM Jorge Estrada
wrote:
> Thanks for the feedback. Addressed your points and attached an updated
> patch
>
> - Simplified the documentation to just point to the pad filter for usage
> but kept the examples
> - Made the adjustments to vf_pad_
Thanks for the feedback. Addressed your points and attached an updated patch
- Simplified the documentation to just point to the pad filter for usage
but kept the examples
- Made the adjustments to vf_pad_cuda.c
- Updated vf_pad_cuda.cu. Should be easier to add 10-12 bit support down
the line now.
From: Jorge Estrada <--global>
This patch adds the pad_cuda video filter. A filter similar to the existing pad
filter but accelerated by CUDA.
The filter shares the same options as the software pad filter.
Example usage:
ffmpeg -hwaccel cuda -hwaccel_output_format cuda -i input.m
Sounds reasonable. I'll look into rewriting this and will re-submit a CUDA
variant.
On Mon, Jun 16, 2025 at 4:08 AM Timo Rothenpieler
wrote:
> On 16/06/2025 03:38, Jorge Estrada wrote:
> > Got it. I don't mind writing a CUDA variant of this and submitting it
> > separat
Got it. I don't mind writing a CUDA variant of this and submitting it
separately. What's the reasoning behind preferring a plain CUDA filter over
an NPP based one?
On Sun, Jun 15, 2025 at 5:37 PM Timo Rothenpieler
wrote:
> On 16.06.2025 02:14, Jorge Estrada wrote:
> >
This patch adds the pad_npp video filter. A filter similar to the existing pad
filter but accelerated by NPP.
The filter shares the same options as the software pad filter.
Example usage:
ffmpeg -hwaccel cuda -hwaccel_output_format cuda -i input.mp4 -vf
"pad_npp=w=iw+100:h=ih+100:x=-1:y=-1:colo
16 matches
Mail list logo
