[FFmpeg-devel] [PATCH] lavu: Adding hybrid-log gamma enum value and transfer function.
Hendrik Leppkes
h.leppkes at gmail.com
Thu Apr 21 23:03:33 CEST 2016
On Thu, Apr 21, 2016 at 10:57 PM, Neil Birkbeck <neil.birkbeck at gmail.com> wrote:
> The standard:
> http://www.arib.or.jp/english/html/overview/doc/2-STD-B67v1_0.pdf
>
> The choice of enum value of 18 is consistent with HEVC:
> http://phenix.it-sudparis.eu/jct/doc_end_user/current_document.php?id=10481
>
> and also with latest proposal for color trc in mkv:
> https://mailarchive.ietf.org/arch/search/?email_list=cellar&gbt=1&q=Colour+Format+proposal
>
> Signed-off-by: Neil Birkbeck <neil.birkbeck at gmail.com>
> ---
> libavutil/color_utils.c | 16 ++++++++++++++++
> libavutil/pixdesc.c | 2 +-
> libavutil/pixfmt.h | 1 +
> tests/ref/fate/color_utils | 19 +++++++++++++++++++
> 4 files changed, 37 insertions(+), 1 deletion(-)
>
> diff --git a/libavutil/color_utils.c b/libavutil/color_utils.c
> index 6dba46a..87e5f5e 100644
> --- a/libavutil/color_utils.c
> +++ b/libavutil/color_utils.c
> @@ -155,6 +155,18 @@ static double avpriv_trc_smpte_st428_1(double Lc)
> : pow(48.0 * Lc / 52.37, 1.0 / 2.6);
> }
>
> +
> +static double avpriv_trc_arib_std_b67(double Lc) {
> + // The function uses the definition from HEVC, which assumes that the peak
> + // white is input level = 1. (this is equivalent to scaling E = Lc * 12 and
> + // using the definition from the arib standard)
> + const double a = 0.17883277;
> + const double b = 0.28466892;
> + const double c = 0.55991073;
> + return (0.0 > Lc) ? 0.0 :
> + (Lc <= 1.0 / 12.0 ? sqrt(3.0 * Lc) : a * log(12.0 * Lc - b) + c);
> +}
> +
> avpriv_trc_function avpriv_get_trc_function_from_trc(enum AVColorTransferCharacteristic trc)
> {
> avpriv_trc_function func = NULL;
> @@ -209,6 +221,10 @@ avpriv_trc_function avpriv_get_trc_function_from_trc(enum AVColorTransferCharact
> func = avpriv_trc_smpte_st428_1;
> break;
>
> + case AVCOL_TRC_ARIB_STD_B67:
> + func = avpriv_trc_arib_std_b67;
> + break;
> +
> case AVCOL_TRC_RESERVED0:
> case AVCOL_TRC_UNSPECIFIED:
> case AVCOL_TRC_RESERVED:
> diff --git a/libavutil/pixdesc.c b/libavutil/pixdesc.c
> index 8a9475c..995244b 100644
> --- a/libavutil/pixdesc.c
> +++ b/libavutil/pixdesc.c
> @@ -2080,6 +2080,7 @@ static const char *color_transfer_names[AVCOL_TRC_NB] = {
> "bt470bg", "smpte170m", "smpte240m", "linear", "log100",
> "log316", "iec61966-2-4", "bt1361e", "iec61966-2-1",
> "bt2020-10", "bt2020-20", "smpte2084", "smpte428-1",
> + "arib-std-b67"
> };
>
> static const char *color_space_names[AVCOL_SPC_NB] = {
> @@ -2560,4 +2561,3 @@ int main(void){
> }
>
> #endif
> -
> diff --git a/libavutil/pixfmt.h b/libavutil/pixfmt.h
> index 546eb44..16bd5bc 100644
> --- a/libavutil/pixfmt.h
> +++ b/libavutil/pixfmt.h
> @@ -420,6 +420,7 @@ enum AVColorTransferCharacteristic {
> AVCOL_TRC_BT2020_12 = 15, ///< ITU-R BT2020 for 12 bit system
> AVCOL_TRC_SMPTEST2084 = 16, ///< SMPTE ST 2084 for 10, 12, 14 and 16 bit systems
> AVCOL_TRC_SMPTEST428_1 = 17, ///< SMPTE ST 428-1
> + AVCOL_TRC_ARIB_STD_B67 = 18, ///< Hybrid log-gamma
I find it odd that the name is seemingly unrelated to the comment, the
comment should probably elaborate more.
The commit message could probably also benefit from using the same name.
> AVCOL_TRC_NB, ///< Not part of ABI
> };
>
> diff --git a/tests/ref/fate/color_utils b/tests/ref/fate/color_utils
> index 6e80ebd..10f8055 100644
> --- a/tests/ref/fate/color_utils
> +++ b/tests/ref/fate/color_utils
> @@ -283,3 +283,22 @@ AVColorTransferCharacteristic=17 calling func(15123.456700) expected=39.174525
> AVColorTransferCharacteristic=17 calling func(19845.889230) expected=43.490646
> AVColorTransferCharacteristic=17 calling func(98678.423100) expected=80.593559
> AVColorTransferCharacteristic=17 calling func(99999.899998) expected=81.006971
> +AVColorTransferCharacteristic=18 calling func(-0.100000) expected=0.000000
> +AVColorTransferCharacteristic=18 calling func(-0.018054) expected=0.000000
> +AVColorTransferCharacteristic=18 calling func(-0.010000) expected=0.000000
> +AVColorTransferCharacteristic=18 calling func(-0.004490) expected=0.000000
> +AVColorTransferCharacteristic=18 calling func(0.000000) expected=0.000000
> +AVColorTransferCharacteristic=18 calling func(0.003162) expected=0.097400
> +AVColorTransferCharacteristic=18 calling func(0.005000) expected=0.122474
> +AVColorTransferCharacteristic=18 calling func(0.009000) expected=0.164317
> +AVColorTransferCharacteristic=18 calling func(0.015000) expected=0.212132
> +AVColorTransferCharacteristic=18 calling func(0.100000) expected=0.544089
> +AVColorTransferCharacteristic=18 calling func(1.000000) expected=1.000000
> +AVColorTransferCharacteristic=18 calling func(52.370000) expected=1.712092
> +AVColorTransferCharacteristic=18 calling func(125.098765) expected=1.867862
> +AVColorTransferCharacteristic=18 calling func(1999.111230) expected=2.363502
> +AVColorTransferCharacteristic=18 calling func(6945.443000) expected=2.586219
> +AVColorTransferCharacteristic=18 calling func(15123.456700) expected=2.725380
> +AVColorTransferCharacteristic=18 calling func(19845.889230) expected=2.773978
> +AVColorTransferCharacteristic=18 calling func(98678.423100) expected=3.060803
> +AVColorTransferCharacteristic=18 calling func(99999.899998) expected=3.063182
> --
> 2.8.0.rc3.226.g39d4020
>
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org
> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
More information about the ffmpeg-devel
mailing list