wrapper/FFmpeg
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

avcodec/h265_metadata: Add options to set width/height after crop

Browse source 
It's a common usecase to request a video size after crop. Before
this patch, user must know the video size before crop, then set
crop_right/crop_bottom accordingly. Since HEVC can have different
CTU size, it's not easy to get/deduce the video size before crop.
With the new width/height options, there is no such requirement.

Signed-off-by: Zhao Zhili <zhilizhao@tencent.com>
This commit is contained in:
Zhao Zhili 2024-05-23 15:08:54 +08:00
parent d55f5cba7b
commit 680b3cee1f
3 changed files with 103 additions and 4 deletions
Download patch file Download diff file Expand all files Collapse all files

4
doc/bitstream_filters.texi
View file

@ -456,6 +456,10 @@ will replace the current ones if the stream is already cropped.
These fields are set in pixels. Note that some sizes may not be
representable if the chroma is subsampled (H.265 section 7.4.3.2.1).
@item width
@item height
Set width and height after crop.
@item level
Set the level in the VPS and SPS. See H.265 section A.4 and tables
A.6 and A.7.

101
libavcodec/bsf/h265_metadata.c
View file

@ -59,6 +59,8 @@ typedef struct H265MetadataContext {
int crop_right;
int crop_top;
int crop_bottom;
int width;
int height;
int level;
int level_guess;
@ -188,12 +190,94 @@ static int h265_metadata_update_vps(AVBSFContext *bsf,
return 0;
}
static int h265_metadata_deduce_crop(AVBSFContext *bsf, const H265RawSPS *sps,
int *crop_left, int *crop_right,
int *crop_top, int *crop_bottom)
{
const H265MetadataContext *ctx = bsf->priv_data;
int left = ctx->crop_left;
int right = ctx->crop_right;
int top = ctx->crop_top;
int bottom = ctx->crop_bottom;
if (ctx->width > 0) {
if (ctx->width > sps->pic_width_in_luma_samples) {
av_log(bsf, AV_LOG_ERROR,
"The width option value %d is larger than picture width %d\n",
ctx->width, sps->pic_width_in_luma_samples);
return AVERROR(EINVAL);
}
if (left < 0) {
if (right > 0)
left = sps->pic_width_in_luma_samples - ctx->width - right;
else
left = 0;
}
if (right < 0)
right = sps->pic_width_in_luma_samples - ctx->width - left;
if (left < 0 || right < 0 || (left + right + ctx->width) !=
sps->pic_width_in_luma_samples) {
av_log(bsf, AV_LOG_ERROR,
"Invalid value for crop_left %d, crop_right %d, width after "
"crop %d, with picture width %d\n",
ctx->crop_left, ctx->crop_right, ctx->width,
sps->pic_width_in_luma_samples);
return AVERROR(EINVAL);
}
}
if (ctx->height > 0) {
if (ctx->height > sps->pic_height_in_luma_samples) {
av_log(bsf, AV_LOG_ERROR,
"The height option value %d is larger than picture height %d\n",
ctx->height, sps->pic_height_in_luma_samples);
return AVERROR(EINVAL);
}
if (top < 0) {
if (bottom > 0)
top = sps->pic_height_in_luma_samples - ctx->height - bottom;
else
top = 0;
}
if (bottom < 0)
bottom = sps->pic_height_in_luma_samples - ctx->height - top;
if (top < 0 || bottom < 0 || (top + bottom + ctx->height) !=
sps->pic_height_in_luma_samples) {
av_log(bsf, AV_LOG_ERROR,
"Invalid value for crop_top %d, crop_bottom %d, height after "
"crop %d, with picture height %d\n",
ctx->crop_top, ctx->crop_bottom, ctx->height,
sps->pic_height_in_luma_samples);
return AVERROR(EINVAL);
}
}
*crop_left = left;
*crop_right = right;
*crop_top = top;
*crop_bottom = bottom;
return 0;
}
static int h265_metadata_update_sps(AVBSFContext *bsf,
H265RawSPS *sps)
{
H265MetadataContext *ctx = bsf->priv_data;
int need_vui = 0;
int crop_unit_x, crop_unit_y;
/* Use local variables to avoid modifying context fields in case of video
* resolution changed. Crop doesn't work well with resolution change, this
* is the best we can do.
*/
int crop_left, crop_right, crop_top, crop_bottom;
int ret;
if (ctx->sample_aspect_ratio.num && ctx->sample_aspect_ratio.den) {
int num, den, i;
@ -290,6 +374,11 @@ static int h265_metadata_update_sps(AVBSFContext *bsf,
}
}
ret = h265_metadata_deduce_crop(bsf, sps, &crop_left, &crop_right,
&crop_top, &crop_bottom);
if (ret < 0)
return ret;
if (sps->separate_colour_plane_flag || sps->chroma_format_idc == 0) {
crop_unit_x = 1;
crop_unit_y = 1;
@ -298,14 +387,14 @@ static int h265_metadata_update_sps(AVBSFContext *bsf,
crop_unit_y = 1 + (sps->chroma_format_idc < 2);
}
#define CROP(border, unit) do { \
if (ctx->crop_ ## border >= 0) { \
if (ctx->crop_ ## border % unit != 0) { \
if (crop_ ## border >= 0) { \
if (crop_ ## border % unit != 0) { \
av_log(bsf, AV_LOG_ERROR, "Invalid value for crop_%s: " \
"must be a multiple of %d.\n", #border, unit); \
return AVERROR(EINVAL); \
} \
sps->conf_win_ ## border ## _offset = \
ctx->crop_ ## border / unit; \
crop_ ## border / unit; \
sps->conformance_window_flag = 1; \
} \
} while (0)
@ -454,6 +543,12 @@ static const AVOption h265_metadata_options[] = {
{ "crop_bottom", "Set bottom border crop offset",
OFFSET(crop_bottom), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, HEVC_MAX_HEIGHT, FLAGS },
{ "width", "Set width after crop",
OFFSET(width), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, HEVC_MAX_WIDTH, FLAGS },
{ "height", "Set height after crop",
OFFSET(height), AV_OPT_TYPE_INT,
{ .i64 = -1 }, -1, HEVC_MAX_HEIGHT, FLAGS },
{ "level", "Set level (tables A.6 and A.7)",
OFFSET(level), AV_OPT_TYPE_INT,

2
libavcodec/version.h
View file

@ -30,7 +30,7 @@
#include "version_major.h"
#define LIBAVCODEC_VERSION_MINOR 6
#define LIBAVCODEC_VERSION_MICRO 100
#define LIBAVCODEC_VERSION_MICRO 101
#define LIBAVCODEC_VERSION_INT AV_VERSION_INT(LIBAVCODEC_VERSION_MAJOR, \
LIBAVCODEC_VERSION_MINOR, \