[FFmpeg-devel,v2] avfilter: add pu21 filter

Adds the PU21 encoding filter for high dynamic range images and video
quality assessment
Mantiuk, R., & Azimi, M. PU21: A novel perceptually uniform encoding for
adapting existing quality metrics for HDR
https://github.com/gfxdisp/pu21
---
 Changelog                |   1 +
 libavfilter/Makefile     |   1 +
 libavfilter/allfilters.c |   1 +
 libavfilter/version.h    |   4 +-
 libavfilter/vf_pu21.c    | 210 +++++++++++++++++++++++++++++++++++++++
 5 files changed, 215 insertions(+), 2 deletions(-)
 create mode 100644 libavfilter/vf_pu21.c

Thank you for your patch.  Some comments inline.

On 23/08/2024 10:07, Rajiv Harlalka wrote:
> Adds the PU21 encoding filter for high dynamic range images and video
> quality assessment
> Mantiuk, R., & Azimi, M. PU21: A novel perceptually uniform encoding for
> adapting existing quality metrics for HDR
> https://github.com/gfxdisp/pu21
> ---
>  Changelog                |   1 +
>  libavfilter/Makefile     |   1 +
>  libavfilter/allfilters.c |   1 +
>  libavfilter/version.h    |   4 +-
>  libavfilter/vf_pu21.c    | 210 +++++++++++++++++++++++++++++++++++++++
>  5 files changed, 215 insertions(+), 2 deletions(-)
>  create mode 100644 libavfilter/vf_pu21.c
> 
> diff --git a/Changelog b/Changelog
> index 571b27ad98..552aa6028a 100644
> --- a/Changelog
> +++ b/Changelog
> @@ -18,6 +18,7 @@ version <next>:
>  - D3D12VA HEVC encoder
>  - Cropping metadata parsing and writing in Matroska and MP4/MOV de/muxers
>  - Intel QSV-accelerated VVC decoding
> +- Addition of vf_pu21 encoding filter
>  
>  
>  version 7.0:
> diff --git a/libavfilter/Makefile b/libavfilter/Makefile
> index 63088e9286..102315f5bf 100644
> --- a/libavfilter/Makefile
> +++ b/libavfilter/Makefile
> @@ -121,6 +121,7 @@ OBJS-$(CONFIG_AXCORRELATE_FILTER)            += af_axcorrelate.o
>  OBJS-$(CONFIG_AZMQ_FILTER)                   += f_zmq.o
>  OBJS-$(CONFIG_BANDPASS_FILTER)               += af_biquads.o
>  OBJS-$(CONFIG_BANDREJECT_FILTER)             += af_biquads.o
> +OBJS-$(CONFIG_PU21_FILTER)                   += vf_pu21.o
>  OBJS-$(CONFIG_BASS_FILTER)                   += af_biquads.o
>  OBJS-$(CONFIG_BIQUAD_FILTER)                 += af_biquads.o
>  OBJS-$(CONFIG_BS2B_FILTER)                   += af_bs2b.o
> diff --git a/libavfilter/allfilters.c b/libavfilter/allfilters.c
> index 63600e9b58..ced076626c 100644
> --- a/libavfilter/allfilters.c
> +++ b/libavfilter/allfilters.c
> @@ -231,6 +231,7 @@ extern const AVFilter ff_vf_convolution_opencl;
>  extern const AVFilter ff_vf_convolve;
>  extern const AVFilter ff_vf_copy;
>  extern const AVFilter ff_vf_coreimage;
> +extern const AVFilter ff_vf_pu21;
>  extern const AVFilter ff_vf_corr;
>  extern const AVFilter ff_vf_cover_rect;
>  extern const AVFilter ff_vf_crop;
> diff --git a/libavfilter/version.h b/libavfilter/version.h
> index d8cd8a2cfb..7e0eb9af97 100644
> --- a/libavfilter/version.h
> +++ b/libavfilter/version.h
> @@ -31,8 +31,8 @@
>  
>  #include "version_major.h"
>  
> -#define LIBAVFILTER_VERSION_MINOR   2
> -#define LIBAVFILTER_VERSION_MICRO 102
> +#define LIBAVFILTER_VERSION_MINOR   3
> +#define LIBAVFILTER_VERSION_MICRO 100
>  
>  
>  #define LIBAVFILTER_VERSION_INT AV_VERSION_INT(LIBAVFILTER_VERSION_MAJOR, \
> diff --git a/libavfilter/vf_pu21.c b/libavfilter/vf_pu21.c
> new file mode 100644
> index 0000000000..c3b56f9c1b
> --- /dev/null
> +++ b/libavfilter/vf_pu21.c
> @@ -0,0 +1,210 @@
> +

Nit: empty line at top of file.

You should have a copyright disclaimer here.

> +#include "libavfilter/avfilter.h"
> +#include "libavutil/opt.h"
> +#include "libavutil/pixdesc.h"
> +#include "libavutil/mem.h"
> +#include "libavutil/imgutils.h"
> +#include "libavfilter/formats.h"
> +#include "libavutil/internal.h"
> +#include "video.h"
> +
> +static enum pu21_mode {

The static keyword here is ignored and a compilation warning.

> +  BANDING,
> +  BANDING_GLARE,
> +  PEAKS,
> +  PEAKS_GLARE
> +};
> +
> +typedef struct PU21Context {
> +  const AVClass* class;
> +  double L_min, L_max;
> +  int multiplier;
> +  int mode;
> +
> +  int depth;
> +  int nb_planes;
> +} PU21Context;
> +
> +const float pu21_params[4][7] = {
> +  // Reference: "PU21: A novel perceptually uniform encoding for adapting existing quality metrics for HDR"
> +  // Rafał K. Mantiuk and M. Azimi, Picture Coding Symposium 2021
> +  // https://github.com/gfxdisp/pu21
> +  {1.070275272f, 0.4088273932f, 0.153224308f, 0.2520326168f, 1.063512885f, 1.14115047f, 521.4527484f},  // BANDING
> +  {0.353487901f, 0.3734658629f, 8.277049286e-05f, 0.9062562627f, 0.09150303166f, 0.9099517204f, 596.3148142f},  // BANDING_GLARE
> +  {1.043882782f, 0.6459495343f, 0.3194584211f, 0.374025247f, 1.114783422f, 1.095360363f, 384.9217577f},  // PEAKS
> +  {816.885024f, 1479.463946f, 0.001253215609f, 0.9329636822f, 0.06746643971f, 1.573435413f, 419.6006374f}  // PEAKS_GLARE
> +};
> +
> +
> +#define OFFSET(x) offsetof(PU21Context, x)
> +#define FLAGS AV_OPT_FLAG_FILTERING_PARAM|AV_OPT_FLAG_VIDEO_PARAM
> +static const AVOption pu21_options[] = {
> +    { "L_min", "Set the minimum luminance value", OFFSET(L_min), AV_OPT_TYPE_DOUBLE, {.dbl = 0.005}, 0.005, 10000, FLAGS },
> +    { "L_max", "Set the maximum luminance value", OFFSET(L_max), AV_OPT_TYPE_DOUBLE, {.dbl = 10000}, 0.005, 10000, FLAGS },

L_min and L_max aren't used anywhere.

> +    { "multiplier", "Set the parameters for the encoding", OFFSET(multiplier), AV_OPT_TYPE_INT, {.i64 = 1}, 1, 10000, FLAGS},

This description could be more helpful.

> +    { "mode", "Set the mode of the PU21 encoding", OFFSET(mode), AV_OPT_TYPE_INT, {.i64 = BANDING_GLARE}, BANDING, PEAKS_GLARE, FLAGS},
> +    { NULL }
> +};
> +
> +AVFILTER_DEFINE_CLASS(pu21);
> +
> +static av_cold int pu21_init(AVFilterContext* ctx) {

Nit: FFmpeg uses K&R style, where a function's opening brace is on its own

Nit: FFmpeg uses four spaces for indentation, not two.

> +  PU21Context* pu21 = ctx->priv;
> +
> +  pu21->L_max = 300;
> +  pu21->L_min = 0.005;
> +  pu21->multiplier = 1;
> +  pu21->mode = BANDING_GLARE;
> +  return 0;
> +}
> +
> +static float apply_pu21(float pixel_val, PU21Context* pu21) {
> +  float Y, V;
> +  Y = pixel_val * (pu21->multiplier);
> +  V = fmax(pu21_params[pu21->mode][6] * (pow((pu21_params[pu21->mode][0] + pu21_params[pu21->mode][1] * pow(Y, pu21_params[pu21->mode][3])) / (1 + pu21_params[pu21->mode][2] * pow(Y, pu21_params[pu21->mode][3])), pu21_params[pu21->mode][4]) - pu21_params[pu21->mode][5]), 0);

Should be powf instead of pow to use the one defined in libavutil/libm.h

Nit: This line is very long (>250 chars!) Could it be broken into
multiple or otherwise made shorter?  For instance you could define a
temporary variable with a concise name for pu21_params[pu21->mode] which
is used a lot here.

> +  return V;
> +}
> +
> +static void pu21_encode8(AVFilterContext* ctx, AVFrame* in, AVFrame* out) {
> +  PU21Context* pu21 = ctx->priv;
> +  const int height = in->height;
> +  const int width = in->width;
> +
> +  int plane;
> +  for (plane = 0; plane < pu21->nb_planes; plane++) {
> +    const uint8_t* src8 = in->data[plane];
> +    uint8_t* dst8 = out->data[plane];
> +
> +    int  linesize = in->linesize[plane];
> +
> +    uint8_t pixel_val;
> +    for (int y = 0; y < height; y++) {
> +      for (int x = 0; x < width; x++) {
> +        pixel_val = src8[y * linesize + x];
> +        dst8[y * linesize + x] = (uint8_t)apply_pu21(pixel_val, pu21);
> +      }
> +    }
> +  }
> +}
> +
> +static void pu21_encode16(AVFilterContext* ctx, AVFrame* in, AVFrame* out) {
> +  PU21Context* pu21 = ctx->priv;
> +  const int height = in->height;
> +  const int width = in->width;
> +
> +  int plane;
> +  for (plane = 0; plane < pu21->nb_planes; plane++) {
> +    const uint16_t* src16 = (uint16_t*)in->data[plane];
> +    uint16_t* dst16 = (uint16_t*)out->data[plane];
> +
> +    int  linesize = in->linesize[plane] / 2;
> +
> +    uint16_t pixel_val;
> +    for (int y = 0; y < height; y++) {
> +      for (int x = 0; x < width; x++) {
> +        pixel_val = src16[y * linesize + x];
> +        dst16[y * linesize + x] = (uint16_t)apply_pu21(pixel_val, pu21);
> +      }
> +    }
> +  }
> +}
> +
> +static void pu21_encode32(AVFilterContext* ctx, AVFrame* in, AVFrame* out) {
> +  PU21Context* pu21 = ctx->priv;
> +  const int height = in->height;
> +  const int width = in->width;
> +
> +  int plane;
> +  for (plane = 0; plane < pu21->nb_planes; plane++) {
> +    const float* src32 = (float*)in->data[plane];
> +    float* dst32 = (float*)out->data[plane];
> +
> +    int  linesize = in->linesize[plane] / 4;
> +
> +    float pixel_val;
> +    for (int y = 0; y < height; y++) {
> +      for (int x = 0; x < width; x++) {
> +        pixel_val = src32[y * linesize + x];
> +        dst32[y * linesize + x] = (float)apply_pu21(pixel_val, pu21);
> +      }
> +    }
> +  }
> +}

There is a lot of repetition between the three functions above.  Could
you use a macro to make this a bit more concise?  Like in
libavfilter/vf_blend.c.

> +
> +static int filter_frame(AVFilterLink* inlink, AVFrame* input) {
> +  AVFilterContext* ctx = inlink->dst;
> +  PU21Context* pu21 = ctx->priv;
> +
> +  AVFilterLink* outlink = ctx->outputs[0];
> +  AVFrame* out;
> +
> +  if (av_frame_is_writable(input)) {
> +    out = input;
> +  }
> +  else {

Nit: usual style is `} else {`.  Same applies below.

> +    out = ff_get_video_buffer(outlink, outlink->w, outlink->h);
> +    if (!out) {
> +      av_frame_free(&input);
> +      return AVERROR(ENOMEM);
> +    }
> +    av_frame_copy_props(out, input);
> +  }
> +
> +  if (pu21->depth <= 8) {
> +    pu21_encode8(ctx, input, out);
> +  }
> +  else if (pu21->depth <= 16) {
> +    pu21_encode16(ctx, input, out);
> +  }
> +  else {
> +    pu21_encode32(ctx, input, out);
> +  }
> +
> +  if (out != input)
> +    av_frame_free(&input);
> +  return ff_filter_frame(outlink, out);
> +}
> +
> +static av_cold void uninit(AVFilterContext* ctx) {
> +  PU21Context* s = ctx->priv;

This variable is unused.

> +}
> +
> +static int config_input(AVFilterLink* inlink) {
> +  const AVPixFmtDescriptor* desc = av_pix_fmt_desc_get(inlink->format);
> +  PU21Context* s = inlink->dst->priv;
> +
> +  uninit(inlink->dst);
> +
> +  s->depth = desc->comp[0].depth;
> +  s->nb_planes = av_pix_fmt_count_planes(inlink->format);
> +
> +  return 0;
> +}
> +
> +static const AVFilterPad pu21_inputs[] = {
> +    {
> +        .name = "default",
> +        .type = AVMEDIA_TYPE_VIDEO,
> +        .config_props = config_input,
> +        .filter_frame = filter_frame,
> +    }
> +};
> +
> +static const AVFilterPad pu21_outputs[] = {
> +    {
> +        .name = "default",
> +        .type = AVMEDIA_TYPE_VIDEO,
> +    }
> +};
> +
> +const AVFilter ff_vf_pu21 = {
> +    .name = "pu21",
> +    .description = NULL_IF_CONFIG_SMALL("Convert HDR linear values to PU values using PU21 transform"),
> +    .priv_size = sizeof(PU21Context),
> +    .init = pu21_init,
> +    .uninit = uninit,
> +    FILTER_INPUTS(pu21_inputs),
> +    FILTER_OUTPUTS(pu21_outputs),
> +    .priv_class = &pu21_class,
> +    .flags = AVFILTER_FLAG_SUPPORT_TIMELINE_GENERIC,
> +};