[FFmpeg-devel] [PATCH] Update HDR10+ metadata structure.
James Almer
jamrial at gmail.com
Mon Jan 27 21:49:58 EET 2020
On 1/27/2020 4:08 PM, Mohammad Izadi wrote:
> From: Mohammad Izadi <izadi at google.com>
>
> Trying to read HDR10+ metadata from HEVC/SEI and pass it to packet side data in the follow-up CLs.
> ---
> libavutil/hdr_dynamic_metadata.c | 386 +++++++++++++++++++++++++++++++
> libavutil/hdr_dynamic_metadata.h | 51 +++-
> libavutil/version.h | 2 +-
> 3 files changed, 431 insertions(+), 8 deletions(-)
>
> diff --git a/libavutil/hdr_dynamic_metadata.c b/libavutil/hdr_dynamic_metadata.c
> index 0fa1ee82de..e321a22af8 100644
> --- a/libavutil/hdr_dynamic_metadata.c
> +++ b/libavutil/hdr_dynamic_metadata.c
> @@ -21,6 +21,19 @@
> #include "hdr_dynamic_metadata.h"
> #include "mem.h"
>
> +static const uint8_t usa_country_code = 0xB5;
> +static const uint16_t smpte_provider_code = 0x003C;
> +static const uint16_t smpte2094_40_provider_oriented_code = 0x0001;
> +static const uint16_t smpte2094_40_application_identifier = 0x04;
> +
> +static const int64_t luminance_den = 1;
> +static const int32_t peak_luminance_den = 15;
> +static const int64_t rgb_den = 100000;
> +static const int32_t fraction_pixel_den = 1000;
> +static const int32_t knee_point_den = 4095;
> +static const int32_t bezier_anchor_den = 1023;
> +static const int32_t saturation_weight_den = 8;
> +
> AVDynamicHDRPlus *av_dynamic_hdr_plus_alloc(size_t *size)
> {
> AVDynamicHDRPlus *hdr_plus = av_mallocz(sizeof(AVDynamicHDRPlus));
> @@ -33,6 +46,48 @@ AVDynamicHDRPlus *av_dynamic_hdr_plus_alloc(size_t *size)
> return hdr_plus;
> }
>
> +AVDynamicHDRPlus *av_dynamic_hdr_plus(GetBitContext *gb, size_t *size)
> +{
> + uint8_t country_code;
> + uint16_t provider_code;
> + uint16_t provider_oriented_code;
> + uint8_t application_identifier;
> + uint8_t application_version;
> + AVDynamicHDRPlus *hdr_plus;
> + int err;
> +
> + if (get_bits_left(gb) < 7)
> + return NULL;
> +
> + country_code = get_bits(gb, 8);
> + provider_code = get_bits(gb, 16);
> +
> + if (country_code != usa_country_code ||
> + provider_code != smpte_provider_code)
> + return NULL;
> +
> + // A/341 Amendment – 2094-40
> + provider_oriented_code = get_bits(gb, 16);
> + application_identifier = get_bits(gb, 8);
> + application_version = get_bits(gb, 8);
> + if (provider_oriented_code != smpte2094_40_provider_oriented_code ||
> + application_identifier != smpte2094_40_application_identifier)
> + return NULL;
> +
> + hdr_plus = av_dynamic_hdr_plus_alloc(size);
> + if (!hdr_plus)
> + return NULL;
> +
> + hdr_plus->application_version = application_version;
> + err = decode_itu_t_t35_to_dynamic_hdr_plus(gb, hdr_plus);
> + if (err < 0) {
> + av_freep(&hdr_plus);
> + return NULL;
> + }
> +
> + return hdr_plus;
> +}
> +
> AVDynamicHDRPlus *av_dynamic_hdr_plus_create_side_data(AVFrame *frame)
> {
> AVFrameSideData *side_data = av_frame_new_side_data(frame,
> @@ -43,5 +98,336 @@ AVDynamicHDRPlus *av_dynamic_hdr_plus_create_side_data(AVFrame *frame)
>
> memset(side_data->data, 0, sizeof(AVDynamicHDRPlus));
>
> +
> return (AVDynamicHDRPlus *)side_data->data;
> }
> +
> +int decode_itu_t_t35_to_dynamic_hdr_plus(GetBitContext *gb, AVDynamicHDRPlus *s)
> +{
> + int w, i, j;
> +
> + if (get_bits_left(gb) < 2)
> + return AVERROR_INVALIDDATA;
> + s->num_windows = get_bits(gb, 2);
> +
> + if (s->num_windows < 1 || s->num_windows > 3) {
> + return AVERROR_INVALIDDATA;
> + }
> +
> + if (get_bits_left(gb) < ((19 * 8 + 1) * (s->num_windows - 1)))
> + return AVERROR_INVALIDDATA;
> + for (w = 1; w < s->num_windows; w++) {
> + s->params[w].window_upper_left_corner_x.num = get_bits(gb, 16);
> + s->params[w].window_upper_left_corner_y.num = get_bits(gb, 16);
> + s->params[w].window_lower_right_corner_x.num = get_bits(gb, 16);
> + s->params[w].window_lower_right_corner_y.num = get_bits(gb, 16);
> + // The corners are set to absolute coordinates here. They should be
> + // converted to the relative coordinates (in [0, 1]) in the decoder.
> + s->params[w].window_upper_left_corner_x.den = 1;
> + s->params[w].window_upper_left_corner_y.den = 1;
> + s->params[w].window_lower_right_corner_x.den = 1;
> + s->params[w].window_lower_right_corner_y.den = 1;
> +
> + s->params[w].center_of_ellipse_x = get_bits(gb, 16);
> + s->params[w].center_of_ellipse_y = get_bits(gb, 16);
> + s->params[w].rotation_angle = get_bits(gb, 8);
> + s->params[w].semimajor_axis_internal_ellipse = get_bits(gb, 16);
> + s->params[w].semimajor_axis_external_ellipse = get_bits(gb, 16);
> + s->params[w].semiminor_axis_external_ellipse = get_bits(gb, 16);
> + s->params[w].overlap_process_option = get_bits1(gb);
> + }
> +
> + if (get_bits_left(gb) < 28)
> + return AVERROR(EINVAL);
> + s->targeted_system_display_maximum_luminance.num = get_bits(gb, 27);
> + s->targeted_system_display_maximum_luminance.den = luminance_den;
> + s->targeted_system_display_actual_peak_luminance_flag = get_bits1(gb);
> +
> + if (s->targeted_system_display_actual_peak_luminance_flag) {
> + int rows, cols;
> + if (get_bits_left(gb) < 10)
> + return AVERROR(EINVAL);
> + rows = get_bits(gb, 5);
> + cols = get_bits(gb, 5);
> + if (((rows < 2) || (rows > 25)) || ((cols < 2) || (cols > 25))) {
> + return AVERROR_INVALIDDATA;
> + }
> + s->num_rows_targeted_system_display_actual_peak_luminance = rows;
> + s->num_cols_targeted_system_display_actual_peak_luminance = cols;
> +
> + if (get_bits_left(gb) < (rows * cols * 4))
> + return AVERROR(EINVAL);
> +
> + for (i = 0; i < rows; i++) {
> + for (j = 0; j < cols; j++) {
> + s->targeted_system_display_actual_peak_luminance[i][j].num = get_bits(gb, 4);
> + s->targeted_system_display_actual_peak_luminance[i][j].den = peak_luminance_den;
> + }
> + }
> + }
> + for (w = 0; w < s->num_windows; w++) {
> + if (get_bits_left(gb) < (3 * 17 + 17 + 4))
> + return AVERROR(EINVAL);
> + for (i = 0; i < 3; i++) {
> + s->params[w].maxscl[i].num = get_bits(gb, 17);
> + s->params[w].maxscl[i].den = rgb_den;
> + }
> + s->params[w].average_maxrgb.num = get_bits(gb, 17);
> + s->params[w].average_maxrgb.den = rgb_den;
> + s->params[w].num_distribution_maxrgb_percentiles = get_bits(gb, 4);
> +
> + if (get_bits_left(gb) <
> + (s->params[w].num_distribution_maxrgb_percentiles * 24))
> + return AVERROR(EINVAL);
> + for (i = 0; i < s->params[w].num_distribution_maxrgb_percentiles; i++) {
> + s->params[w].distribution_maxrgb[i].percentage = get_bits(gb, 7);
> + s->params[w].distribution_maxrgb[i].percentile.num = get_bits(gb, 17);
> + s->params[w].distribution_maxrgb[i].percentile.den = rgb_den;
> + }
> +
> + if (get_bits_left(gb) < 10)
> + return AVERROR(EINVAL);
> + s->params[w].fraction_bright_pixels.num = get_bits(gb, 10);
> + s->params[w].fraction_bright_pixels.den = fraction_pixel_den;
> + }
> + if (get_bits_left(gb) < 1)
> + return AVERROR(EINVAL);
> + s->mastering_display_actual_peak_luminance_flag = get_bits1(gb);
> + if (s->mastering_display_actual_peak_luminance_flag) {
> + int rows, cols;
> + if (get_bits_left(gb) < 10)
> + return AVERROR(EINVAL);
> + rows = get_bits(gb, 5);
> + cols = get_bits(gb, 5);
> + if (((rows < 2) || (rows > 25)) || ((cols < 2) || (cols > 25))) {
> + return AVERROR_INVALIDDATA;
> + }
> + s->num_rows_mastering_display_actual_peak_luminance = rows;
> + s->num_cols_mastering_display_actual_peak_luminance = cols;
> +
> + if (get_bits_left(gb) < (rows * cols * 4))
> + return AVERROR(EINVAL);
> +
> + for (i = 0; i < rows; i++) {
> + for (j = 0; j < cols; j++) {
> + s->mastering_display_actual_peak_luminance[i][j].num = get_bits(gb, 4);
> + s->mastering_display_actual_peak_luminance[i][j].den = peak_luminance_den;
> + }
> + }
> + }
> +
> + for (w = 0; w < s->num_windows; w++) {
> + if (get_bits_left(gb) < 1)
> + return AVERROR(EINVAL);
> + s->params[w].tone_mapping_flag = get_bits1(gb);
> + if (s->params[w].tone_mapping_flag) {
> + if (get_bits_left(gb) < 28)
> + return AVERROR(EINVAL);
> + s->params[w].knee_point_x.num = get_bits(gb, 12);
> + s->params[w].knee_point_x.den = knee_point_den;
> + s->params[w].knee_point_y.num = get_bits(gb, 12);
> + s->params[w].knee_point_y.den = knee_point_den;
> + s->params[w].num_bezier_curve_anchors = get_bits(gb, 4);
> +
> + if (get_bits_left(gb) < (s->params[w].num_bezier_curve_anchors * 10))
> + return AVERROR(EINVAL);
> + for (i = 0; i < s->params[w].num_bezier_curve_anchors; i++) {
> + s->params[w].bezier_curve_anchors[i].num = get_bits(gb, 10);
> + s->params[w].bezier_curve_anchors[i].den = bezier_anchor_den;
> + }
> + }
> +
> + if (get_bits_left(gb) < 1)
> + return AVERROR(EINVAL);
> + s->params[w].color_saturation_mapping_flag = get_bits1(gb);
> + if (s->params[w].color_saturation_mapping_flag) {
> + if (get_bits_left(gb) < 6)
> + return AVERROR(EINVAL);
> + s->params[w].color_saturation_weight.num = get_bits(gb, 6);
> + s->params[w].color_saturation_weight.den = saturation_weight_den;
> + }
> + }
> +
> + skip_bits(gb, get_bits_left(gb));
> +
> + return 0;
> +}
> +
> +int get_encoded_dynamic_hdr_plus_buffer_size(const AVDynamicHDRPlus *s) {
> + int bit_count = 0;
> + int w, size;
> +
> + if (!s)
> + return 0;
> +
> + // 7 bytes for country code, provider code, and user identifier.
> + bit_count += 56;
> +
> + if (s->num_windows < 1 || s->num_windows > 3)
> + return 0;
> + // Count bits for window params.
> + bit_count += 2 + ((19 * 8 + 1) * (s->num_windows - 1));
> +
> + bit_count += 28;
> + if (s->targeted_system_display_actual_peak_luminance_flag) {
> + int rows, cols;
> + rows = s->num_rows_targeted_system_display_actual_peak_luminance;
> + cols = s->num_rows_targeted_system_display_actual_peak_luminance;
> + if (((rows < 2) || (rows > 25)) || ((cols < 2) || (cols > 25)))
> + return 0;
> +
> + bit_count += (10 + rows * cols * 4);
> + }
> + for (w = 0; w < s->num_windows; w++) {
> + bit_count += (3 * 17 + 17 + 4 + 10) +
> + (s->params[w].num_distribution_maxrgb_percentiles * 24);
> + }
> + bit_count++;
> +
> + if (s->mastering_display_actual_peak_luminance_flag) {
> + int rows, cols;
> + rows = s->num_rows_mastering_display_actual_peak_luminance;
> + cols = s->num_cols_mastering_display_actual_peak_luminance;
> + if (((rows < 2) || (rows > 25)) || ((cols < 2) || (cols > 25)))
> + return 0;
> +
> + bit_count += (10 + rows * cols * 4);
> + }
> +
> + for (w = 0; w < s->num_windows; w++) {
> + bit_count++;
> + if (s->params[w].tone_mapping_flag)
> + bit_count += (28 + s->params[w].num_bezier_curve_anchors * 10);
> +
> + bit_count++;
> + if (s->params[w].color_saturation_mapping_flag)
> + bit_count += 6;
> + }
> + size = bit_count / 8;
> + if (bit_count % 8 != 0)
> + size++;
> + return size;
> +}
> +
> +int encode_dynamic_hdr_plus_to_itu_t_t35(const AVDynamicHDRPlus *s, PutBitContext *pb)
> +{
> + int w, i, j;
> + if (!s)
> + return AVERROR(EINVAL);
> +
> + if (put_bits_left(pb) < get_encoded_dynamic_hdr_plus_buffer_size(s))
> + return AVERROR(EINVAL);
> + put_bits(pb, 8, usa_country_code);
> +
> + put_bits(pb, 16, smpte_provider_code);
> + put_bits(pb, 16, smpte2094_40_provider_oriented_code);
> + put_bits(pb, 8, smpte2094_40_application_identifier);
> + put_bits(pb, 8, s->application_version);
> +
> + put_bits(pb, 2, s->num_windows);
> +
> + for (w = 1; w < s->num_windows; w++) {
> + put_bits(pb, 16, s->params[w].window_upper_left_corner_x.num /
> + s->params[w].window_upper_left_corner_x.den);
> + put_bits(pb, 16, s->params[w].window_upper_left_corner_y.num /
> + s->params[w].window_upper_left_corner_y.den);
> + put_bits(pb, 16, s->params[w].window_lower_right_corner_x.num /
> + s->params[w].window_lower_right_corner_x.den);
> + put_bits(pb, 16, s->params[w].window_lower_right_corner_y.num /
> + s->params[w].window_lower_right_corner_y.den);
> + put_bits(pb, 16, s->params[w].center_of_ellipse_x);
> + put_bits(pb, 16, s->params[w].center_of_ellipse_y);
> + put_bits(pb, 8, s->params[w].rotation_angle);
> + put_bits(pb, 16, s->params[w].semimajor_axis_internal_ellipse);
> + put_bits(pb, 16, s->params[w].semimajor_axis_external_ellipse);
> + put_bits(pb, 16, s->params[w].semiminor_axis_external_ellipse);
> + put_bits(pb, 1, s->params[w].overlap_process_option);
> + }
> + put_bits(pb, 27,
> + s->targeted_system_display_maximum_luminance.num * luminance_den /
> + s->targeted_system_display_maximum_luminance.den);
> + put_bits(pb, 1, s->targeted_system_display_actual_peak_luminance_flag);
> + if (s->targeted_system_display_actual_peak_luminance_flag) {
> + int rows, cols;
> + rows = s->num_rows_targeted_system_display_actual_peak_luminance;
> + cols = s->num_cols_targeted_system_display_actual_peak_luminance;
> + put_bits(pb, 5, rows);
> + put_bits(pb, 5, cols);
> + for (i = 0; i < rows; i++) {
> + for (j = 0; j < cols; j++) {
> + put_bits(
> + pb, 4,
> + s->targeted_system_display_actual_peak_luminance[i][j].num *
> + peak_luminance_den /
> + s->targeted_system_display_actual_peak_luminance[i][j]
> + .den);
> + }
> + }
> + }
> + for (w = 0; w < s->num_windows; w++) {
> + for (i = 0; i < 3; i++) {
> + put_bits(pb, 17,
> + s->params[w].maxscl[i].num * rgb_den /
> + s->params[w].maxscl[i].den);
> + }
> + put_bits(pb, 17,
> + s->params[w].average_maxrgb.num * rgb_den /
> + s->params[w].average_maxrgb.den);
> + put_bits(pb, 4, s->params[w].num_distribution_maxrgb_percentiles);
> +
> + for (i = 0; i < s->params[w].num_distribution_maxrgb_percentiles; i++) {
> + put_bits(pb, 7, s->params[w].distribution_maxrgb[i].percentage);
> + put_bits(pb, 17,
> + s->params[w].distribution_maxrgb[i].percentile.num *
> + rgb_den /
> + s->params[w].distribution_maxrgb[i].percentile.den);
> + }
> + put_bits(pb, 10,
> + s->params[w].fraction_bright_pixels.num * fraction_pixel_den /
> + s->params[w].fraction_bright_pixels.den);
> + }
> + put_bits(pb, 1, s->mastering_display_actual_peak_luminance_flag);
> + if (s->mastering_display_actual_peak_luminance_flag) {
> + int rows, cols;
> + rows = s->num_rows_mastering_display_actual_peak_luminance;
> + cols = s->num_cols_mastering_display_actual_peak_luminance;
> + put_bits(pb, 5, rows);
> + put_bits(pb, 5, cols);
> + for (i = 0; i < rows; i++) {
> + for (j = 0; j < cols; j++) {
> + put_bits(
> + pb, 4,
> + s->mastering_display_actual_peak_luminance[i][j].num *
> + peak_luminance_den /
> + s->mastering_display_actual_peak_luminance[i][j].den);
> + }
> + }
> + }
> +
> + for (w = 0; w < s->num_windows; w++) {
> + put_bits(pb, 1, s->params[w].tone_mapping_flag);
> + if (s->params[w].tone_mapping_flag) {
> + put_bits(pb, 12,
> + s->params[w].knee_point_x.num * knee_point_den /
> + s->params[w].knee_point_x.den);
> + put_bits(pb, 12,
> + s->params[w].knee_point_y.num * knee_point_den /
> + s->params[w].knee_point_y.den);
> + put_bits(pb, 4, s->params[w].num_bezier_curve_anchors);
> + for (i = 0; i < s->params[w].num_bezier_curve_anchors; i++) {
> + put_bits(pb, 10,
> + s->params[w].bezier_curve_anchors[i].num *
> + bezier_anchor_den /
> + s->params[w].bezier_curve_anchors[i].den);
> + }
> + }
> + put_bits(pb, 1, s->params[w].color_saturation_mapping_flag);
> + if (s->params[w].color_saturation_mapping_flag)
> + put_bits(pb, 6,
> + s->params[w].color_saturation_weight.num *
> + saturation_weight_den /
> + s->params[w].color_saturation_weight.den);
> + }
> + flush_put_bits(pb);
> + return 0;
> +}
> diff --git a/libavutil/hdr_dynamic_metadata.h b/libavutil/hdr_dynamic_metadata.h
> index 2d72de56ae..4e075f2f58 100644
> --- a/libavutil/hdr_dynamic_metadata.h
> +++ b/libavutil/hdr_dynamic_metadata.h
> @@ -23,6 +23,8 @@
>
> #include "frame.h"
> #include "rational.h"
> +#include "libavcodec/get_bits.h"
> +#include "libavcodec/put_bits.h"
>
> /**
> * Option for overlapping elliptical pixel selectors in an image.
> @@ -241,11 +243,6 @@ typedef struct AVHDRPlusColorTransformParams {
> * the public ABI.
> */
> typedef struct AVDynamicHDRPlus {
> - /**
> - * Country code by Rec. ITU-T T.35 Annex A. The value shall be 0xB5.
> - */
> - uint8_t itu_t_t35_country_code;
This is an API break.
> -
> /**
> * Application version in the application defining document in ST-2094
> * suite. The value shall be set to 0.
> @@ -265,7 +262,7 @@ typedef struct AVDynamicHDRPlus {
>
> /**
> * The nominal maximum display luminance of the targeted system display,
> - * in units of 0.0001 candelas per square metre. The value shall be in
> + * in units of 1 candelas per square metre. The value shall be in
> * the range of 0 to 10000, inclusive.
> */
> AVRational targeted_system_display_maximum_luminance;
> @@ -322,14 +319,27 @@ typedef struct AVDynamicHDRPlus {
> AVRational mastering_display_actual_peak_luminance[25][25];
> } AVDynamicHDRPlus;
>
> +/**
> + * Allocate an AVDynamicHDRPlus structure and set its fields to
> + * the values decoded from the user data registered ITU-T T.35.
> + * The resulting struct can be freed using av_freep().
> + *
> + * @return An AVDynamicHDRPlus filled with decoded values or NULL
> + * on failure.
> + */
> +AVDynamicHDRPlus *av_dynamic_hdr_plus_alloc(size_t *size);
> +
> /**
> * Allocate an AVDynamicHDRPlus structure and set its fields to
> * default values. The resulting struct can be freed using av_freep().
> + * @param gb The input bit stream.
> + * @param size The size of allocated memory for the returned
> + * AVDynamicHDRPlus structure.
> *
> * @return An AVDynamicHDRPlus filled with default values or NULL
> * on failure.
> */
> -AVDynamicHDRPlus *av_dynamic_hdr_plus_alloc(size_t *size);
> +AVDynamicHDRPlus *av_dynamic_hdr_plus(GetBitContext *gb, size_t *size);
As is this. Also, GetBitContext and PutBitContext are internal structs
to libavcodec, and can't be used in public headers.
>
> /**
> * Allocate a complete AVDynamicHDRPlus and add it to the frame.
> @@ -340,4 +350,31 @@ AVDynamicHDRPlus *av_dynamic_hdr_plus_alloc(size_t *size);
> */
> AVDynamicHDRPlus *av_dynamic_hdr_plus_create_side_data(AVFrame *frame);
>
> +/**
> + * Decode the user data registered ITU-T T.35 to AVDynamicHDRPlus.
> + * @param gb The bit content to be decoded.
> + * @param s The decoded AVDynamicHDRPlus structure.
> + *
> + * @return 0 if succeed. Otherwise, returns the appropriate AVERROR.
> + */
> +int decode_itu_t_t35_to_dynamic_hdr_plus(GetBitContext *gb, AVDynamicHDRPlus *s);
> +
> +/**
> + * Get the size of buffer required to save the encoded bit stream of
> + * AVDynamicHDRPlus in the form of the user data registered ITU-T T.35.
> + * @param s The AVDynamicHDRPlus structure.
> + *
> + * @return The size of bit stream required for encoding. 0 if the data is invalid.
> + */
> +int get_encoded_dynamic_hdr_plus_buffer_size(const AVDynamicHDRPlus *s);
> +
> +/**
> + * Encode AVDynamicHDRPlus to the user data registered ITU-T T.35.
> + * @param s The AVDynamicHDRPlus structure to be encoded.
> + * @param pb The encoded bit stream.
> + *
> + * @return 0 if succeed. Otherwise, returns the appropriate AVERROR.
> + */
> +int encode_dynamic_hdr_plus_to_itu_t_t35(const AVDynamicHDRPlus *s, PutBitContext *pb);
> +
> #endif /* AVUTIL_HDR_DYNAMIC_METADATA_H */
> diff --git a/libavutil/version.h b/libavutil/version.h
> index af8f614aff..2bc1b98615 100644
> --- a/libavutil/version.h
> +++ b/libavutil/version.h
> @@ -79,7 +79,7 @@
> */
>
> #define LIBAVUTIL_VERSION_MAJOR 56
> -#define LIBAVUTIL_VERSION_MINOR 38
> +#define LIBAVUTIL_VERSION_MINOR 39
> #define LIBAVUTIL_VERSION_MICRO 100
>
> #define LIBAVUTIL_VERSION_INT AV_VERSION_INT(LIBAVUTIL_VERSION_MAJOR, \
>
More information about the ffmpeg-devel
mailing list