On Thu, Jun 24, 2021 at 2:18 PM Pooja Venkatesan <po...@multicorewareinc.com>
wrote:
> From f6c1b86d99ab3a0dd57204f930e526a07e28fed7 Mon Sep 17 00:00:00 2001
> From: Praveen Karadugattu <praveenku...@multicorewareinc.com>
> Date: Sun, 20 Jun 2021 21:20:50 +0530
> Subject: [PATCH] Disable fall-back on traditional scenecut algorithm with
> --hist-scenecut
>
> ---
> doc/reST/cli.rst | 9 +++++++--
> source/CMakeLists.txt | 2 +-
> source/common/param.cpp | 8 +++++++-
> source/encoder/encoder.cpp | 17 ++++++++++++-----
> source/encoder/slicetype.cpp | 15 ++++++++++-----
> source/test/regression-tests.txt | 1 +
> source/x265.h | 4 ++++
> source/x265cli.h | 1 +
> 8 files changed, 43 insertions(+), 14 deletions(-)
>
> diff --git a/doc/reST/cli.rst b/doc/reST/cli.rst
> index 4154221af..280b734b3 100755
> --- a/doc/reST/cli.rst
> +++ b/doc/reST/cli.rst
> @@ -1470,7 +1470,8 @@ Slice decision options
> .. option:: --hist-scenecut, --no-hist-scenecut
>
> Indicates that scenecuts need to be detected using luma edge and chroma
> histograms.
> - :option:`--hist-scenecut` enables scenecut detection using the
> histograms and disables the default scene cut algorithm.
> + :option:`--hist-scenecut` enables scenecut detection using the
> histograms.
> + It also uses the intra and inter cost info to arrive at a scenecut
> decision from the default scenecut method.
> :option:`--no-hist-scenecut` disables histogram based scenecut algorithm.
>
> .. option:: --hist-threshold <0.0..1.0>
> @@ -1480,7 +1481,11 @@ Slice decision options
> greater than 0.2 against the previous frame as scenecut.
> Increasing the threshold reduces the number of scenecuts detected.
> Default 0.03.
> -
> +
> +.. option:: --disable-traditional-scenecut
> +
> + Indicates that the usage of traditional scenecut detection using intra
> and inter cost should be disabled when :option:`--hist-scenecut` is used.
> +
> .. option:: --radl <integer>
>
> Number of RADL pictures allowed infront of IDR. Requires closed gop
> interval.
> diff --git a/source/CMakeLists.txt b/source/CMakeLists.txt
> index b4e57b592..f4a9cb793 100755
> --- a/source/CMakeLists.txt
> +++ b/source/CMakeLists.txt
> @@ -29,7 +29,7 @@ option(NATIVE_BUILD "Target the build CPU" OFF)
> option(STATIC_LINK_CRT "Statically link C runtime for release builds" OFF)
> mark_as_advanced(FPROFILE_USE FPROFILE_GENERATE NATIVE_BUILD)
> # X265_BUILD must be incremented each time the public API is changed
> -set(X265_BUILD 200)
> +set(X265_BUILD 201)
> configure_file("${PROJECT_SOURCE_DIR}/x265.def.in"
> "${PROJECT_BINARY_DIR}/x265.def")
> configure_file("${PROJECT_SOURCE_DIR}/x265_config.h.in"
> diff --git a/source/common/param.cpp b/source/common/param.cpp
> index 8a27aaef3..3b84415c3 100755
> --- a/source/common/param.cpp
> +++ b/source/common/param.cpp
> @@ -172,6 +172,7 @@ void x265_param_default(x265_param* param)
> param->scenecutThreshold = 40; /* Magic number pulled in from x264 */
> param->edgeTransitionThreshold = 0.03;
> param->bHistBasedSceneCut = 0;
> + param->bDisableTradScdInHscd = 0;
> param->lookaheadSlices = 8;
> param->lookaheadThreads = 0;
> param->scenecutBias = 5.0;
> @@ -598,6 +599,7 @@ int x265_param_default_preset(x265_param* param, const
> char* preset, const char*
> param->lookaheadDepth = 0;
> param->scenecutThreshold = 0;
> param->bHistBasedSceneCut = 0;
> + param->bDisableTradScdInHscd = 0;
> param->rc.cuTree = 0;
> param->frameNumThreads = 1;
> }
> @@ -953,6 +955,7 @@ int x265_param_parse(x265_param* p, const char* name,
> const char* value)
> bError = false;
> p->scenecutThreshold = atoi(value);
> p->bHistBasedSceneCut = 0;
> + p->bDisableTradScdInHscd = 0;
> }
> }
> OPT("temporal-layers") p->bEnableTemporalSubLayers = atobool(value);
> @@ -1234,6 +1237,7 @@ int x265_param_parse(x265_param* p, const char*
> name, const char* value)
> }
> }
> OPT("hist-threshold") p->edgeTransitionThreshold = atof(value);
> + OPT("disable-traditional-scenecut") p->bDisableTradScdInHscd =
> atobool(value);
> OPT("rskip-edge-threshold") p->edgeVarThreshold =
> atoi(value)/100.0f;
> OPT("lookahead-threads") p->lookaheadThreads = atoi(value);
> OPT("opt-cu-delta-qp") p->bOptCUDeltaQP = atobool(value);
> @@ -2151,7 +2155,8 @@ char *x265_param2string(x265_param* p, int padx, int
> pady)
> s += sprintf(s, " rc-lookahead=%d", p->lookaheadDepth);
> s += sprintf(s, " lookahead-slices=%d", p->lookaheadSlices);
> s += sprintf(s, " scenecut=%d", p->scenecutThreshold);
> - s += sprintf(s, " hist-scenecut=%d", p->bHistBasedSceneCut);
> + BOOL(p->bHistBasedSceneCut, "hist-scenecut");
> + BOOL(p->bDisableTradScdInHscd, "disable-traditional-scenecut");
>
[AM] This will add invalid CLI option "no-disable-traditional-scenecut" to
info-sei. Please fix this.
> s += sprintf(s, " radl=%d", p->radl);
> BOOL(p->bEnableHRDConcatFlag, "splice");
> BOOL(p->bIntraRefresh, "intra-refresh");
> @@ -2467,6 +2472,7 @@ void x265_copy_params(x265_param* dst, x265_param*
> src)
> dst->lookaheadThreads = src->lookaheadThreads;
> dst->scenecutThreshold = src->scenecutThreshold;
> dst->bHistBasedSceneCut = src->bHistBasedSceneCut;
> + dst->bDisableTradScdInHscd = src->bDisableTradScdInHscd;
> dst->bIntraRefresh = src->bIntraRefresh;
> dst->maxCUSize = src->maxCUSize;
> dst->minCUSize = src->minCUSize;
> diff --git a/source/encoder/encoder.cpp b/source/encoder/encoder.cpp
> index c1e1cb46d..dfb872889 100644
> --- a/source/encoder/encoder.cpp
> +++ b/source/encoder/encoder.cpp
> @@ -3680,6 +3680,7 @@ void Encoder::configure(x265_param *p)
> p->keyframeMax = INT_MAX;
> p->scenecutThreshold = 0;
> p->bHistBasedSceneCut = 0;
> + p->bDisableTradScdInHscd = 0;
> }
> else if (p->keyframeMax <= 1)
> {
> @@ -3694,6 +3695,7 @@ void Encoder::configure(x265_param *p)
> p->bframes = 0;
> p->scenecutThreshold = 0;
> p->bHistBasedSceneCut = 0;
> + p->bDisableTradScdInHscd = 0;
> p->bFrameAdaptive = 0;
> p->rc.cuTree = 0;
> p->bEnableWeightedPred = 0;
> @@ -4421,12 +4423,17 @@ void Encoder::configure(x265_param *p)
> m_param->searchRange = m_param->hmeRange[2];
> }
>
> - if (p->bHistBasedSceneCut && !p->edgeTransitionThreshold)
> - {
> - p->edgeTransitionThreshold = 0.03;
> - x265_log(p, X265_LOG_WARNING, "using default threshold %.2lf for
> scene cut detection\n", p->edgeTransitionThreshold);
> - }
> + if (p->bHistBasedSceneCut && !p->edgeTransitionThreshold)
> + {
> + p->edgeTransitionThreshold = 0.03;
> + x265_log(p, X265_LOG_WARNING, "using default threshold %.2lf for
> scene cut detection.\n", p->edgeTransitionThreshold);
> + }
>
> + if (!p->bHistBasedSceneCut && p->bDisableTradScdInHscd)
> + {
> + p->bDisableTradScdInHscd = 0;
> + x265_log(p, X265_LOG_WARNING, "option
> --disable-traditional-scenecut requires --hist-scenecut to be enabled.\n");
> + }
> }
>
> void Encoder::readAnalysisFile(x265_analysis_data* analysis, int curPoc,
> const x265_picture* picIn, int paramBytes)
> diff --git a/source/encoder/slicetype.cpp b/source/encoder/slicetype.cpp
> index 0adb0d0db..ace3b5469 100644
> --- a/source/encoder/slicetype.cpp
> +++ b/source/encoder/slicetype.cpp
> @@ -2014,7 +2014,7 @@ void Lookahead::slicetypeAnalyse(Lowres **frames,
> bool bKeyframe)
> bool isScenecut = false;
>
> /* Temporal computations for scenecut detection */
> - if (m_param->bHistBasedSceneCut)
> + if (m_param->bHistBasedSceneCut && !m_param->bDisableTradScdInHscd)
> {
> for (int i = numFrames - 1; i > 0; i--)
> {
> @@ -2047,8 +2047,10 @@ void Lookahead::slicetypeAnalyse(Lowres **frames,
> bool bKeyframe)
> }
>
> /* When scenecut threshold is set, use scenecut detection for I frame
> placements */
> - if (!m_param->bHistBasedSceneCut || (m_param->bHistBasedSceneCut &&
> frames[1]->bScenecut))
> + if (!m_param->bHistBasedSceneCut || (m_param->bHistBasedSceneCut &&
> !m_param->bDisableTradScdInHscd && frames[1]->bScenecut))
> isScenecut = scenecut(frames, 0, 1, true, origNumFrames);
> + else if (m_param->bHistBasedSceneCut && frames[1]->bScenecut)
> + isScenecut = true;
>
> if (isScenecut && (m_param->bHistBasedSceneCut ||
> m_param->scenecutThreshold))
> {
> @@ -2061,14 +2063,17 @@ void Lookahead::slicetypeAnalyse(Lowres **frames,
> bool bKeyframe)
> m_extendGopBoundary = false;
> for (int i = m_param->bframes + 1; i < origNumFrames; i +=
> m_param->bframes + 1)
> {
> - if (!m_param->bHistBasedSceneCut ||
> (m_param->bHistBasedSceneCut && frames[i + 1]->bScenecut))
> + if (!m_param->bHistBasedSceneCut ||
> (m_param->bHistBasedSceneCut && !m_param->bDisableTradScdInHscd && frames[i
> + 1]->bScenecut))
> scenecut(frames, i, i + 1, true, origNumFrames);
>
> for (int j = i + 1; j <= X265_MIN(i + m_param->bframes + 1,
> origNumFrames); j++)
> {
> - if (frames[j]->bScenecut && scenecutInternal(frames, j -
> 1, j, true))
> + if (frames[j]->bScenecut)
> {
> - m_extendGopBoundary = true;
> + if (!m_param->bDisableTradScdInHscd)
> + m_extendGopBoundary = scenecutInternal(frames, j
> - 1, j, true);
> + else
> + m_extendGopBoundary = true;
> break;
> }
> }
> diff --git a/source/test/regression-tests.txt
> b/source/test/regression-tests.txt
> index 971c854df..828157cab 100644
> --- a/source/test/regression-tests.txt
> +++ b/source/test/regression-tests.txt
> @@ -160,6 +160,7 @@ Traffic_4096x2048_30p.y4m, --preset medium --frame-dup
> --dup-threshold 60 --hrd
> Kimono1_1920x1080_24_400.yuv,--preset superfast --qp 28 --zones 0,139,q=32
> sintel_trailer_2k_1920x1080_24.yuv, --preset medium --hist-scenecut
> --hist-threshold 0.02 --frame-dup --dup-threshold 60 --hrd --bitrate 10000
> --vbv-bufsize 15000 --vbv-maxrate 12000
> sintel_trailer_2k_1920x1080_24.yuv, --preset medium --hist-scenecut
> --hist-threshold 0.02
> +sintel_trailer_2k_1920x1080_24.yuv, --preset medium --hist-scenecut
> --hist-threshold 0.02 --disable-traditional-scenecut
> sintel_trailer_2k_1920x1080_24.yuv, --preset ultrafast --hist-scenecut
> --hist-threshold 0.02
> crowd_run_1920x1080_50.yuv, --preset faster --ctu 32 --rskip 2
> --rskip-edge-threshold 5
> crowd_run_1920x1080_50.yuv, --preset fast --ctu 64 --rskip 2
> --rskip-edge-threshold 5 --aq-mode 4
> diff --git a/source/x265.h b/source/x265.h
> index 324f3163f..3a65fdda6 100644
> --- a/source/x265.h
> +++ b/source/x265.h
> @@ -1963,6 +1963,10 @@ typedef struct x265_param
> /* Flag indicating whether the encoder should emit an End of Sequence
> * NAL at the end of every Coded Video Sequence. Default false */
> int bEnableEndOfSequence;
> +
> + /* Flag to turn off traditional scenecut detection in histogram based
> scenecut detection so that
> + * only spatial properties are used for scenecut detection. Default
> false */
> + int bDisableTradScdInHscd;
> } x265_param;
>
> /* x265_param_alloc:
> diff --git a/source/x265cli.h b/source/x265cli.h
> index 46a2b68ae..86136fbfb 100644
> --- a/source/x265cli.h
> +++ b/source/x265cli.h
> @@ -144,6 +144,7 @@ static const struct option long_options[] =
> { "hist-scenecut", no_argument, NULL, 0},
> { "no-hist-scenecut", no_argument, NULL, 0},
> { "hist-threshold", required_argument, NULL, 0},
> + { "disable-traditional-scenecut", no_argument, NULL, 0},
> { "fades", no_argument, NULL, 0 },
> { "no-fades", no_argument, NULL, 0 },
> { "scenecut-aware-qp", required_argument, NULL, 0 },
> --
> 2.24.0.windows.2
>
> Regards,
> *Pooja Venkatesan*,
> Video Codec Engineer,
> Media & AI analytics BU
>
> _______________________________________________
> x265-devel mailing list
> x265-devel@videolan.org
> https://mailman.videolan.org/listinfo/x265-devel
>
--
Regards,
*Aruna Matheswaran,*
Video Codec Engineer,
Media & AI analytics BU,
_______________________________________________
x265-devel mailing list
x265-devel@videolan.org
https://mailman.videolan.org/listinfo/x265-devel
