* [FFmpeg-devel] [PATCH] avfilter/vf_thumbnail: support high bit depth formats
@ 2025-06-12 14:47 Niklas Haas
2025-06-15 3:44 ` Andreas Rheinhardt
0 siblings, 1 reply; 3+ messages in thread
From: Niklas Haas @ 2025-06-12 14:47 UTC (permalink / raw)
To: ffmpeg-devel; +Cc: Niklas Haas
From: Niklas Haas <git@haasn.dev>
Implementation is the same as 8 bit formats, just right shifted to 8 bits.
It's worth pointing out that this also works for HDR formats (and even DV),
because the underlying implementation is just trying to minimize the histogram
difference. If anything, using a HDR format will result in a *more* accurate
detection, because HDR formats tend to be more perceptually uniform.
---
libavfilter/vf_thumbnail.c | 54 +++++++++++++++++++++++++++++++++++++-
1 file changed, 53 insertions(+), 1 deletion(-)
diff --git a/libavfilter/vf_thumbnail.c b/libavfilter/vf_thumbnail.c
index 90125a879c..f353f0495c 100644
--- a/libavfilter/vf_thumbnail.c
+++ b/libavfilter/vf_thumbnail.c
@@ -53,6 +53,7 @@ typedef struct ThumbContext {
int planewidth[4];
int planeheight[4];
+ int bitdepth;
} ThumbContext;
#define OFFSET(x) offsetof(ThumbContext, x)
@@ -193,7 +194,23 @@ static int do_slice(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
p += frame->linesize[0];
}
break;
- default:
+
+ case AV_PIX_FMT_YUV410P:
+ case AV_PIX_FMT_YUV411P:
+ case AV_PIX_FMT_YUV420P:
+ case AV_PIX_FMT_YUV422P:
+ case AV_PIX_FMT_YUV440P:
+ case AV_PIX_FMT_YUV444P:
+ case AV_PIX_FMT_YUVJ420P:
+ case AV_PIX_FMT_YUVJ422P:
+ case AV_PIX_FMT_YUVJ440P:
+ case AV_PIX_FMT_YUVJ444P:
+ case AV_PIX_FMT_YUVJ411P:
+ case AV_PIX_FMT_YUVA420P:
+ case AV_PIX_FMT_YUVA422P:
+ case AV_PIX_FMT_YUVA444P:
+ case AV_PIX_FMT_GBRP:
+ case AV_PIX_FMT_GBRAP:
for (int plane = 0; plane < 3; plane++) {
const int slice_start = (s->planeheight[plane] * jobnr) / nb_jobs;
const int slice_end = (s->planeheight[plane] * (jobnr+1)) / nb_jobs;
@@ -209,6 +226,25 @@ static int do_slice(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
}
}
break;
+
+ default:
+ for (int plane = 0; plane < 3; plane++) {
+ const int slice_start = (s->planeheight[plane] * jobnr) / nb_jobs;
+ const int slice_end = (s->planeheight[plane] * (jobnr+1)) / nb_jobs;
+ const int slice_offset = slice_start * frame->linesize[plane];
+ const uint16_t *p = (const uint16_t *) (frame->data[plane] + slice_offset);
+ const ptrdiff_t linesize = frame->linesize[plane] >> 1;
+ const int planewidth = s->planewidth[plane];
+ const int shift = s->bitdepth - 8;
+ int *hhist = hist + 256 * plane;
+
+ for (int j = slice_start; j < slice_end; j++) {
+ for (int i = 0; i < planewidth; i++)
+ hhist[(p[i] >> shift) & 0xFF]++;
+ p += linesize;
+ }
+ }
+ break;
}
return 0;
@@ -286,6 +322,7 @@ static int config_props(AVFilterLink *inlink)
s->planewidth[0] = s->planewidth[3] = inlink->w;
s->planeheight[1] = s->planeheight[2] = AV_CEIL_RSHIFT(inlink->h, desc->log2_chroma_h);
s->planeheight[0] = s->planeheight[3] = inlink->h;
+ s->bitdepth = desc->comp[0].depth;
return 0;
}
@@ -304,6 +341,21 @@ static const enum AVPixelFormat pix_fmts[] = {
AV_PIX_FMT_YUVJ411P,
AV_PIX_FMT_YUVA420P, AV_PIX_FMT_YUVA422P, AV_PIX_FMT_YUVA444P,
AV_PIX_FMT_GBRP, AV_PIX_FMT_GBRAP,
+
+ /* High bit depth formats */
+ AV_PIX_FMT_YUV420P9, AV_PIX_FMT_YUV422P9, AV_PIX_FMT_YUV444P9,
+ AV_PIX_FMT_YUV420P10, AV_PIX_FMT_YUV422P10, AV_PIX_FMT_YUV444P10,
+ AV_PIX_FMT_YUV420P12, AV_PIX_FMT_YUV422P12, AV_PIX_FMT_YUV444P12,
+ AV_PIX_FMT_YUV420P14, AV_PIX_FMT_YUV422P14, AV_PIX_FMT_YUV444P14,
+ AV_PIX_FMT_YUV420P16, AV_PIX_FMT_YUV422P16, AV_PIX_FMT_YUV444P16,
+ AV_PIX_FMT_YUVA420P9, AV_PIX_FMT_YUVA422P9, AV_PIX_FMT_YUVA444P9,
+ AV_PIX_FMT_YUVA420P10, AV_PIX_FMT_YUVA422P10, AV_PIX_FMT_YUVA444P10,
+ AV_PIX_FMT_YUVA422P12, AV_PIX_FMT_YUVA444P12,
+ AV_PIX_FMT_YUVA420P16, AV_PIX_FMT_YUVA422P16, AV_PIX_FMT_YUVA444P16,
+ AV_PIX_FMT_GBRP9, AV_PIX_FMT_GBRP10, AV_PIX_FMT_GBRP12,
+ AV_PIX_FMT_GBRAP12, AV_PIX_FMT_GBRP14, AV_PIX_FMT_GBRAP14,
+ AV_PIX_FMT_GBRP16, AV_PIX_FMT_GBRAP16,
+
AV_PIX_FMT_NONE
};
--
2.49.0
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [FFmpeg-devel] [PATCH] avfilter/vf_thumbnail: support high bit depth formats
2025-06-12 14:47 [FFmpeg-devel] [PATCH] avfilter/vf_thumbnail: support high bit depth formats Niklas Haas
@ 2025-06-15 3:44 ` Andreas Rheinhardt
2025-06-16 11:54 ` Niklas Haas
0 siblings, 1 reply; 3+ messages in thread
From: Andreas Rheinhardt @ 2025-06-15 3:44 UTC (permalink / raw)
To: ffmpeg-devel
Niklas Haas:
> From: Niklas Haas <git@haasn.dev>
>
> Implementation is the same as 8 bit formats, just right shifted to 8 bits.
>
> It's worth pointing out that this also works for HDR formats (and even DV),
> because the underlying implementation is just trying to minimize the histogram
> difference. If anything, using a HDR format will result in a *more* accurate
> detection, because HDR formats tend to be more perceptually uniform.
> ---
> libavfilter/vf_thumbnail.c | 54 +++++++++++++++++++++++++++++++++++++-
> 1 file changed, 53 insertions(+), 1 deletion(-)
>
> diff --git a/libavfilter/vf_thumbnail.c b/libavfilter/vf_thumbnail.c
> index 90125a879c..f353f0495c 100644
> --- a/libavfilter/vf_thumbnail.c
> +++ b/libavfilter/vf_thumbnail.c
> @@ -53,6 +53,7 @@ typedef struct ThumbContext {
>
> int planewidth[4];
> int planeheight[4];
> + int bitdepth;
> } ThumbContext;
>
> #define OFFSET(x) offsetof(ThumbContext, x)
> @@ -193,7 +194,23 @@ static int do_slice(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
> p += frame->linesize[0];
> }
> break;
> - default:
> +
> + case AV_PIX_FMT_YUV410P:
> + case AV_PIX_FMT_YUV411P:
> + case AV_PIX_FMT_YUV420P:
> + case AV_PIX_FMT_YUV422P:
> + case AV_PIX_FMT_YUV440P:
> + case AV_PIX_FMT_YUV444P:
> + case AV_PIX_FMT_YUVJ420P:
> + case AV_PIX_FMT_YUVJ422P:
> + case AV_PIX_FMT_YUVJ440P:
> + case AV_PIX_FMT_YUVJ444P:
> + case AV_PIX_FMT_YUVJ411P:
> + case AV_PIX_FMT_YUVA420P:
> + case AV_PIX_FMT_YUVA422P:
> + case AV_PIX_FMT_YUVA444P:
> + case AV_PIX_FMT_GBRP:
> + case AV_PIX_FMT_GBRAP:
> for (int plane = 0; plane < 3; plane++) {
> const int slice_start = (s->planeheight[plane] * jobnr) / nb_jobs;
> const int slice_end = (s->planeheight[plane] * (jobnr+1)) / nb_jobs;
> @@ -209,6 +226,25 @@ static int do_slice(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
> }
> }
> break;
> +
> + default:
> + for (int plane = 0; plane < 3; plane++) {
> + const int slice_start = (s->planeheight[plane] * jobnr) / nb_jobs;
> + const int slice_end = (s->planeheight[plane] * (jobnr+1)) / nb_jobs;
> + const int slice_offset = slice_start * frame->linesize[plane];
> + const uint16_t *p = (const uint16_t *) (frame->data[plane] + slice_offset);
> + const ptrdiff_t linesize = frame->linesize[plane] >> 1;
> + const int planewidth = s->planewidth[plane];
> + const int shift = s->bitdepth - 8;
> + int *hhist = hist + 256 * plane;
> +
> + for (int j = slice_start; j < slice_end; j++) {
> + for (int i = 0; i < planewidth; i++)
> + hhist[(p[i] >> shift) & 0xFF]++;
> + p += linesize;
> + }
> + }
> + break;
> }
>
> return 0;
> @@ -286,6 +322,7 @@ static int config_props(AVFilterLink *inlink)
> s->planewidth[0] = s->planewidth[3] = inlink->w;
> s->planeheight[1] = s->planeheight[2] = AV_CEIL_RSHIFT(inlink->h, desc->log2_chroma_h);
> s->planeheight[0] = s->planeheight[3] = inlink->h;
> + s->bitdepth = desc->comp[0].depth;
>
> return 0;
> }
> @@ -304,6 +341,21 @@ static const enum AVPixelFormat pix_fmts[] = {
> AV_PIX_FMT_YUVJ411P,
> AV_PIX_FMT_YUVA420P, AV_PIX_FMT_YUVA422P, AV_PIX_FMT_YUVA444P,
> AV_PIX_FMT_GBRP, AV_PIX_FMT_GBRAP,
> +
> + /* High bit depth formats */
> + AV_PIX_FMT_YUV420P9, AV_PIX_FMT_YUV422P9, AV_PIX_FMT_YUV444P9,
> + AV_PIX_FMT_YUV420P10, AV_PIX_FMT_YUV422P10, AV_PIX_FMT_YUV444P10,
> + AV_PIX_FMT_YUV420P12, AV_PIX_FMT_YUV422P12, AV_PIX_FMT_YUV444P12,
> + AV_PIX_FMT_YUV420P14, AV_PIX_FMT_YUV422P14, AV_PIX_FMT_YUV444P14,
> + AV_PIX_FMT_YUV420P16, AV_PIX_FMT_YUV422P16, AV_PIX_FMT_YUV444P16,
> + AV_PIX_FMT_YUVA420P9, AV_PIX_FMT_YUVA422P9, AV_PIX_FMT_YUVA444P9,
> + AV_PIX_FMT_YUVA420P10, AV_PIX_FMT_YUVA422P10, AV_PIX_FMT_YUVA444P10,
> + AV_PIX_FMT_YUVA422P12, AV_PIX_FMT_YUVA444P12,
> + AV_PIX_FMT_YUVA420P16, AV_PIX_FMT_YUVA422P16, AV_PIX_FMT_YUVA444P16,
> + AV_PIX_FMT_GBRP9, AV_PIX_FMT_GBRP10, AV_PIX_FMT_GBRP12,
> + AV_PIX_FMT_GBRAP12, AV_PIX_FMT_GBRP14, AV_PIX_FMT_GBRAP14,
> + AV_PIX_FMT_GBRP16, AV_PIX_FMT_GBRAP16,
> +
> AV_PIX_FMT_NONE
> };
>
Can't you use some AVPixFmtDescriptor magic to replace this huge list
with a query formats function?
- Andreas
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [FFmpeg-devel] [PATCH] avfilter/vf_thumbnail: support high bit depth formats
2025-06-15 3:44 ` Andreas Rheinhardt
@ 2025-06-16 11:54 ` Niklas Haas
0 siblings, 0 replies; 3+ messages in thread
From: Niklas Haas @ 2025-06-16 11:54 UTC (permalink / raw)
To: ffmpeg-devel
On Sun, 15 Jun 2025 05:44:43 +0200 Andreas Rheinhardt <andreas.rheinhardt@outlook.com> wrote:
> Niklas Haas:
> > From: Niklas Haas <git@haasn.dev>
> >
> > Implementation is the same as 8 bit formats, just right shifted to 8 bits.
> >
> > It's worth pointing out that this also works for HDR formats (and even DV),
> > because the underlying implementation is just trying to minimize the histogram
> > difference. If anything, using a HDR format will result in a *more* accurate
> > detection, because HDR formats tend to be more perceptually uniform.
> > ---
> > libavfilter/vf_thumbnail.c | 54 +++++++++++++++++++++++++++++++++++++-
> > 1 file changed, 53 insertions(+), 1 deletion(-)
> >
> > diff --git a/libavfilter/vf_thumbnail.c b/libavfilter/vf_thumbnail.c
> > index 90125a879c..f353f0495c 100644
> > --- a/libavfilter/vf_thumbnail.c
> > +++ b/libavfilter/vf_thumbnail.c
> > @@ -53,6 +53,7 @@ typedef struct ThumbContext {
> >
> > int planewidth[4];
> > int planeheight[4];
> > + int bitdepth;
> > } ThumbContext;
> >
> > #define OFFSET(x) offsetof(ThumbContext, x)
> > @@ -193,7 +194,23 @@ static int do_slice(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
> > p += frame->linesize[0];
> > }
> > break;
> > - default:
> > +
> > + case AV_PIX_FMT_YUV410P:
> > + case AV_PIX_FMT_YUV411P:
> > + case AV_PIX_FMT_YUV420P:
> > + case AV_PIX_FMT_YUV422P:
> > + case AV_PIX_FMT_YUV440P:
> > + case AV_PIX_FMT_YUV444P:
> > + case AV_PIX_FMT_YUVJ420P:
> > + case AV_PIX_FMT_YUVJ422P:
> > + case AV_PIX_FMT_YUVJ440P:
> > + case AV_PIX_FMT_YUVJ444P:
> > + case AV_PIX_FMT_YUVJ411P:
> > + case AV_PIX_FMT_YUVA420P:
> > + case AV_PIX_FMT_YUVA422P:
> > + case AV_PIX_FMT_YUVA444P:
> > + case AV_PIX_FMT_GBRP:
> > + case AV_PIX_FMT_GBRAP:
> > for (int plane = 0; plane < 3; plane++) {
> > const int slice_start = (s->planeheight[plane] * jobnr) / nb_jobs;
> > const int slice_end = (s->planeheight[plane] * (jobnr+1)) / nb_jobs;
> > @@ -209,6 +226,25 @@ static int do_slice(AVFilterContext *ctx, void *arg, int jobnr, int nb_jobs)
> > }
> > }
> > break;
> > +
> > + default:
> > + for (int plane = 0; plane < 3; plane++) {
> > + const int slice_start = (s->planeheight[plane] * jobnr) / nb_jobs;
> > + const int slice_end = (s->planeheight[plane] * (jobnr+1)) / nb_jobs;
> > + const int slice_offset = slice_start * frame->linesize[plane];
> > + const uint16_t *p = (const uint16_t *) (frame->data[plane] + slice_offset);
> > + const ptrdiff_t linesize = frame->linesize[plane] >> 1;
> > + const int planewidth = s->planewidth[plane];
> > + const int shift = s->bitdepth - 8;
> > + int *hhist = hist + 256 * plane;
> > +
> > + for (int j = slice_start; j < slice_end; j++) {
> > + for (int i = 0; i < planewidth; i++)
> > + hhist[(p[i] >> shift) & 0xFF]++;
> > + p += linesize;
> > + }
> > + }
> > + break;
> > }
> >
> > return 0;
> > @@ -286,6 +322,7 @@ static int config_props(AVFilterLink *inlink)
> > s->planewidth[0] = s->planewidth[3] = inlink->w;
> > s->planeheight[1] = s->planeheight[2] = AV_CEIL_RSHIFT(inlink->h, desc->log2_chroma_h);
> > s->planeheight[0] = s->planeheight[3] = inlink->h;
> > + s->bitdepth = desc->comp[0].depth;
> >
> > return 0;
> > }
> > @@ -304,6 +341,21 @@ static const enum AVPixelFormat pix_fmts[] = {
> > AV_PIX_FMT_YUVJ411P,
> > AV_PIX_FMT_YUVA420P, AV_PIX_FMT_YUVA422P, AV_PIX_FMT_YUVA444P,
> > AV_PIX_FMT_GBRP, AV_PIX_FMT_GBRAP,
> > +
> > + /* High bit depth formats */
> > + AV_PIX_FMT_YUV420P9, AV_PIX_FMT_YUV422P9, AV_PIX_FMT_YUV444P9,
> > + AV_PIX_FMT_YUV420P10, AV_PIX_FMT_YUV422P10, AV_PIX_FMT_YUV444P10,
> > + AV_PIX_FMT_YUV420P12, AV_PIX_FMT_YUV422P12, AV_PIX_FMT_YUV444P12,
> > + AV_PIX_FMT_YUV420P14, AV_PIX_FMT_YUV422P14, AV_PIX_FMT_YUV444P14,
> > + AV_PIX_FMT_YUV420P16, AV_PIX_FMT_YUV422P16, AV_PIX_FMT_YUV444P16,
> > + AV_PIX_FMT_YUVA420P9, AV_PIX_FMT_YUVA422P9, AV_PIX_FMT_YUVA444P9,
> > + AV_PIX_FMT_YUVA420P10, AV_PIX_FMT_YUVA422P10, AV_PIX_FMT_YUVA444P10,
> > + AV_PIX_FMT_YUVA422P12, AV_PIX_FMT_YUVA444P12,
> > + AV_PIX_FMT_YUVA420P16, AV_PIX_FMT_YUVA422P16, AV_PIX_FMT_YUVA444P16,
> > + AV_PIX_FMT_GBRP9, AV_PIX_FMT_GBRP10, AV_PIX_FMT_GBRP12,
> > + AV_PIX_FMT_GBRAP12, AV_PIX_FMT_GBRP14, AV_PIX_FMT_GBRAP14,
> > + AV_PIX_FMT_GBRP16, AV_PIX_FMT_GBRAP16,
> > +
> > AV_PIX_FMT_NONE
> > };
> >
>
> Can't you use some AVPixFmtDescriptor magic to replace this huge list
> with a query formats function?
Sure. I will submit that as a separate commit, I think.
>
> - Andreas
>
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel@ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
> To unsubscribe, visit link above, or email
> ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2025-06-16 11:54 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2025-06-12 14:47 [FFmpeg-devel] [PATCH] avfilter/vf_thumbnail: support high bit depth formats Niklas Haas
2025-06-15 3:44 ` Andreas Rheinhardt
2025-06-16 11:54 ` Niklas Haas
Git Inbox Mirror of the ffmpeg-devel mailing list - see https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
This inbox may be cloned and mirrored by anyone:
git clone --mirror https://master.gitmailbox.com/ffmpegdev/0 ffmpegdev/git/0.git
# If you have public-inbox 1.1+ installed, you may
# initialize and index your mirror using the following commands:
public-inbox-init -V2 ffmpegdev ffmpegdev/ https://master.gitmailbox.com/ffmpegdev \
ffmpegdev@gitmailbox.com
public-inbox-index ffmpegdev
Example config snippet for mirrors.
AGPL code for this site: git clone https://public-inbox.org/public-inbox.git