[FFmpeg-devel] [PR] doc/filters: add entry for mestimate_d3d12 (PR #21610)

[FFmpeg-devel] [PR] doc/filters: add entry for mestimate_d3d12 (PR #21610)

[Steven Xiao via ffmpeg-devel]

PR #21610 opened by Steven Xiao (younengxiao)
URL: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/21610
Patch URL: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/21610.patch

Add documentation for the filter "mestimate_d3d12" to the filters documentation.

This is the follow-up for commit https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/21217/commits



>From aac5406afceab57e2afe0dfd23d31b569a7aae32 Mon Sep 17 00:00:00 2001
From: stevxiao <steven.xiao@amd.com>
Date: Thu, 29 Jan 2026 10:14:25 -0500
Subject: [PATCH] doc/filters: add entry for mestimate_d3d12

Add documentation for the filter "mestimate_d3d12" to the filters documentation.

This is the follow-up for commit https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/21217/commits

Signed-off-by: younengxiao <steven.xiao@amd.com>
---
 doc/filters.texi | 26 ++++++++++++++++++++++++++
 1 file changed, 26 insertions(+)

diff --git a/doc/filters.texi b/doc/filters.texi
index 0f64b4a3fa..9aecea7d88 100644
--- a/doc/filters.texi
+++ b/doc/filters.texi
@@ -18096,6 +18096,32 @@ Macroblock size. Default @code{16}.
 Search parameter. Default @code{7}.
 @end table
 
+@section mestimate_d3d12
+
+Estimate and export motion vectors using D3D12 hardware-accelerated motion estimation.
+This filter uses GPU hardware motion estimation capabilities available in
+DirectX 12 Video APIs to achieve significant performance improvements
+over the software-based @code{mestimate} filter.
+
+Motion vectors are stored in frame side data to be used by other filters.
+
+This filter requires the input to be in @code{d3d12} hardware pixel format.
+The filter uses quarter-pixel precision for motion vector estimation.
+
+This filter accepts the following options:
+@table @option
+@item mb_size
+Macroblock size. Only @code{8} and @code{16} are supported.
+Default @code{16}.
+@end table
+
+@subsection Examples
+
+Estimate motion vectors using D3D12 hardware acceleration with 8x8 blocks:
+@example
+ffmpeg -hwaccel d3d12va -hwaccel_output_format d3d12 -i input.mp4 -vf mestimate_d3d12=mb_size=8 -f null -
+@end example
+
 @section midequalizer
 
 Apply Midway Image Equalization effect using two video streams.
-- 
2.52.0

_______________________________________________
ffmpeg-devel mailing list -- ffmpeg-devel@ffmpeg.org
To unsubscribe send an email to ffmpeg-devel-leave@ffmpeg.org