# HG changeset patch
# User gopi Satykrishna Akisetty
# Date 1571121977 -19800
# Tue Oct 15 12:16:17 2019 +0530
# Node ID deaecadc43060ba37a85d9724a1a306a86433432
# Parent 37648fca915b389bafe923d8443818359e80ebf2
Improved scenecut detection
This patch does the following.
1. identifies scenecuts using sad of edge and chroma histogram based
thresholding.
2. Add option "--hist-scenecut" and "--hist-threshold' to enable improved
scenecut method for slice type decisions,rate control and a threshold for
determining scene-cuts.
3. Removed duplicate edgefilter code and created a global function for use in
scene cut detection and aq in Lookahead.
diff -r 37648fca915b -r deaecadc4306 doc/reST/cli.rst
--- a/doc/reST/cli.rst Fri Oct 11 12:45:52 2019 +0530
+++ b/doc/reST/cli.rst Tue Oct 15 12:16:17 2019 +0530
@@ -1427,6 +1427,18 @@
intra cost of a frame used in scenecut detection. For example, a value
of 5 indicates,
if the inter cost of a frame is greater than or equal to 95 percent of
the intra cost of the frame,
then detect this frame as scenecut. Values between 5 and 15 are
recommended. Default 5.
+
+.. option:: --hist-scenecut, --no-hist-scenecut
+
+ indicates that I-frames need to be inserted using edge and color
histogram based scenecut algorithm.
+ option: `--hist-scencut` 1 enables adaptive I frame placement using
this method and disables the default scene cut algorithm.
+ option:`--no-hist-scenecut` adaptive I frame placement.
+
+.. option:: --hist-threshold <0.0..2.0>
+
+ This value represents the threshold for SAD for edge histograms used in
scenecut detection. This requires hist-scenecut to be enabled.
+ For example, a value of 0.2 indicates that a frame with SAD value
greater than 0.2 against the previous frame as scenecut.
+ Values between 0.0 and 2.0 are recommended. Default 0.1.
.. option:: --radl <integer>
diff -r 37648fca915b -r deaecadc4306 source/CMakeLists.txt
--- a/source/CMakeLists.txt Fri Oct 11 12:45:52 2019 +0530
+++ b/source/CMakeLists.txt Tue Oct 15 12:16:17 2019 +0530
@@ -29,7 +29,7 @@
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 180)
+set(X265_BUILD 181)
configure_file("${PROJECT_SOURCE_DIR}/x265.def.in"
"${PROJECT_BINARY_DIR}/x265.def")
configure_file("${PROJECT_SOURCE_DIR}/x265_config.h.in"
diff -r 37648fca915b -r deaecadc4306 source/common/CMakeLists.txt
--- a/source/common/CMakeLists.txt Fri Oct 11 12:45:52 2019 +0530
+++ b/source/common/CMakeLists.txt Tue Oct 15 12:16:17 2019 +0530
@@ -151,4 +151,5 @@
predict.cpp predict.h
scalinglist.cpp scalinglist.h
quant.cpp quant.h contexts.h
- deblock.cpp deblock.h)
+ deblock.cpp deblock.h
+ scenecut.h scenecut.cpp)
diff -r 37648fca915b -r deaecadc4306 source/common/common.h
--- a/source/common/common.h Fri Oct 11 12:45:52 2019 +0530
+++ b/source/common/common.h Tue Oct 15 12:16:17 2019 +0530
@@ -129,14 +129,20 @@
typedef uint64_t sum2_t;
typedef uint64_t pixel4;
typedef int64_t ssum2_t;
+#define HISTOGRAM_SIZE 1024
+#define SHIFT 1
#else
typedef uint8_t pixel;
typedef uint16_t sum_t;
typedef uint32_t sum2_t;
typedef uint32_t pixel4;
typedef int32_t ssum2_t; // Signed sum
+#define HISTOGRAM_SIZE 256
+#define SHIFT 0
#endif // if HIGH_BIT_DEPTH
+#define PI 3.14159265
+
#if X265_DEPTH < 10
typedef uint32_t sse_t;
#else
diff -r 37648fca915b -r deaecadc4306 source/common/param.cpp
--- a/source/common/param.cpp Fri Oct 11 12:45:52 2019 +0530
+++ b/source/common/param.cpp Tue Oct 15 12:16:17 2019 +0530
@@ -167,6 +167,8 @@
param->bFrameAdaptive = X265_B_ADAPT_TRELLIS;
param->bBPyramid = 1;
param->scenecutThreshold = 40; /* Magic number pulled in from x264 */
+ param->edgeTransitionThreshold = 0.01;
+ param->bHistbasedScenecut = false;
param->lookaheadSlices = 8;
param->lookaheadThreads = 0;
param->scenecutBias = 5.0;
@@ -567,6 +569,7 @@
param->bframes = 0;
param->lookaheadDepth = 0;
param->scenecutThreshold = 0;
+ param->bHistbasedScenecut = false;
param->rc.cuTree = 0;
param->frameNumThreads = 1;
}
@@ -609,7 +612,7 @@
return 0;
}
-static int x265_atobool(const char* str, bool& bError)
+static bool x265_atobool(const char* str, bool& bError)
{
if (!strcmp(str, "1") ||
!strcmp(str, "true") ||
@@ -920,6 +923,7 @@
{
bError = false;
p->scenecutThreshold = atoi(value);
+ p->bHistbasedScenecut = false;
}
}
OPT("temporal-layers") p->bEnableTemporalSubLayers = atobool(value);
@@ -1186,6 +1190,31 @@
OPT("opt-ref-list-length-pps") p->bOptRefListLengthPPS =
atobool(value);
OPT("multi-pass-opt-rps") p->bMultiPassOptRPS = atobool(value);
OPT("scenecut-bias") p->scenecutBias = atof(value);
+ OPT("hist-scenecut")
+ {
+ p->bHistbasedScenecut = atobool(value);
+
+ if (bError)
+ {
+ bError = false;
+ p->bHistbasedScenecut = false;
+ }
+
+ if (p->bHistbasedScenecut) {
+ bError = false;
+ p->scenecutThreshold = 0;
+ }
+
+ }
+ OPT("hist-threshold") {
+ p->edgeTransitionThreshold = atof(value);
+ if (bError)
+ {
+ bError = false;
+ p->edgeTransitionThreshold = 0.01;
+ x265_log(p, X265_LOG_INFO, "using default threshold %.2lf for
scene cut detection\n", p->edgeTransitionThreshold);
+ }
+ }
OPT("lookahead-threads") p->lookaheadThreads = atoi(value);
OPT("opt-cu-delta-qp") p->bOptCUDeltaQP = atobool(value);
OPT("multi-pass-opt-analysis") p->analysisMultiPassRefine =
atobool(value);
@@ -1623,8 +1652,14 @@
"Valid Logging level -1:none 0:error 1:warning 2:info 3:debug
4:full");
CHECK(param->scenecutThreshold < 0,
"scenecutThreshold must be greater than 0");
- CHECK(param->scenecutBias < 0 || 100 < param->scenecutBias,
- "scenecut-bias must be between 0 and 100");
+ if (param->scenecutThreshold) {
+ CHECK(param->scenecutBias < 0 || 100 < param->scenecutBias,
+ "scenecut-bias must be between 0 and 100");
+ }
+ else if (param->bHistbasedScenecut) {
+ CHECK(param->edgeTransitionThreshold < 0.0 || 2.0 <
param->edgeTransitionThreshold,
+ "hist-threshold must be between 0.0 and 2.0");
+ }
CHECK(param->radl < 0 || param->radl > param->bframes,
"radl must be between 0 and bframes");
CHECK(param->rdPenalty < 0 || param->rdPenalty > 2,
@@ -1780,10 +1815,20 @@
x265_log(param, X265_LOG_INFO, "ME / range / subpel / merge :
%s / %d / %d / %d\n",
x265_motion_est_names[param->searchMethod], param->searchRange,
param->subpelRefine, param->maxNumMergeCand);
- if (param->keyframeMax != INT_MAX || param->scenecutThreshold)
- x265_log(param, X265_LOG_INFO, "Keyframe min / max / scenecut / bias:
%d / %d / %d / %.2lf\n", param->keyframeMin, param->keyframeMax,
param->scenecutThreshold, param->scenecutBias * 100);
+ if (param->scenecutThreshold && param->keyframeMax != INT_MAX)
+ param->edgeTransitionThreshold = 0.0;
+ else if (param->bHistbasedScenecut && param->keyframeMax != INT_MAX)
+ param->scenecutBias = 0.0;
+ else if (param->keyframeMax != INT_MAX) {
+ param->edgeTransitionThreshold = 0.0;
+ param->scenecutBias = 0.0;
+ }
+
+ if (param->keyframeMax == INT_MAX)
+ x265_log(param, X265_LOG_INFO, "Keyframe min / max / scenecut :
disabled\n");
else
- x265_log(param, X265_LOG_INFO, "Keyframe min / max / scenecut :
disabled\n");
+ x265_log(param, X265_LOG_INFO, "Keyframe min / max / scenecut / bias /
threshold : %d / %d / %d / %.2lf / %.2lf\n", param->keyframeMin,
param->keyframeMax, ( param->bHistbasedScenecut || param->scenecutThreshold ),
param->scenecutBias * 100, param->edgeTransitionThreshold);
+
if (param->cbQpOffset || param->crQpOffset)
x265_log(param, X265_LOG_INFO, "Cb/Cr QP Offset :
%d / %d\n", param->cbQpOffset, param->crQpOffset);
@@ -1949,6 +1994,8 @@
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);
+ s += sprintf(s, " hist-threshold=%.2f", p->edgeTransitionThreshold);
s += sprintf(s, " radl=%d", p->radl);
BOOL(p->bEnableHRDConcatFlag, "splice");
BOOL(p->bIntraRefresh, "intra-refresh");
@@ -2096,6 +2143,8 @@
BOOL(p->bOptRefListLengthPPS, "opt-ref-list-length-pps");
BOOL(p->bMultiPassOptRPS, "multi-pass-opt-rps");
s += sprintf(s, " scenecut-bias=%.2f", p->scenecutBias);
+ s += sprintf(s, " hist-threshold=%.2f", p->edgeTransitionThreshold);
+
BOOL(p->bOptCUDeltaQP, "opt-cu-delta-qp");
BOOL(p->bAQMotion, "aq-motion");
BOOL(p->bEmitHDRSEI, "hdr");
@@ -2246,6 +2295,7 @@
dst->lookaheadSlices = src->lookaheadSlices;
dst->lookaheadThreads = src->lookaheadThreads;
dst->scenecutThreshold = src->scenecutThreshold;
+ dst->bHistbasedScenecut = src->bHistbasedScenecut;
dst->bIntraRefresh = src->bIntraRefresh;
dst->maxCUSize = src->maxCUSize;
dst->minCUSize = src->minCUSize;
@@ -2403,6 +2453,7 @@
dst->bOptRefListLengthPPS = src->bOptRefListLengthPPS;
dst->bMultiPassOptRPS = src->bMultiPassOptRPS;
dst->scenecutBias = src->scenecutBias;
+ dst->edgeTransitionThreshold = src->edgeTransitionThreshold;
dst->gopLookahead = src->lookaheadDepth;
dst->bOptCUDeltaQP = src->bOptCUDeltaQP;
dst->analysisMultiPassDistortion = src->analysisMultiPassDistortion;
diff -r 37648fca915b -r deaecadc4306 source/common/scenecut.cpp
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/source/common/scenecut.cpp Tue Oct 15 12:16:17 2019 +0530
@@ -0,0 +1,674 @@
+#include <iostream>
+#include <sstream>
+#include <vector>
+#include <algorithm>
+#include <math.h>
+#include "encoder.h"
+#include "scenecut.h"
+#include "slicetype.h"
+
+using namespace std;
+using namespace X265_NS;
+
+#define EDGE_PLANE_COUNT 1
+
+namespace X265_NS {
+
+void computeEdge(pixel * edgePic, pixel *refPic, pixel * edgeTheta, intptr_t
stride, int height, int width)
+ {
+ float gradientH = 0, gradientV = 0, radians = 0, theta = 0;
+ float gradientMagnitude = 0;
+ pixel blackPixel = 0;
+
+ //Applying Sobel filter
+ for (int rowNum = 0; rowNum < height; rowNum++)
+ {
+ for (int colNum = 0; colNum < width; colNum++)
+ {
+ edgeTheta[(rowNum*stride) + colNum] = 0;
+ if ((rowNum != 0) && (colNum != 0) && (rowNum != height - 1)
&& (colNum != width - 1)) //Ignoring the border pixels of the picture
+ {
+ /*Horizontal and vertical gradients
+ [ -3 0 3 ] [-3 -10 -3 ]
+ gH = [ -10 0 10] gV = [ 0 0 0 ]
+ [ -3 0 3 ] [ 3 10 3 ]*/
+
+ const intptr_t rowOne = (rowNum - 1)*stride, colOne =
colNum - 1;
+ const intptr_t rowTwo = rowNum * stride, colTwo = colNum;
+ const intptr_t rowThree = (rowNum + 1)*stride, colThree =
colNum + 1;
+ const intptr_t index = (rowNum*stride) + colNum;
+
+ gradientH = (float)(-3 * refPic[rowOne + colOne] + 3 *
refPic[rowOne + colThree] - 10 * refPic[rowTwo + colOne] + 10 * refPic[rowTwo +
colThree] - 3 * refPic[rowThree + colOne] + 3 * refPic[rowThree + colThree]);
+ gradientV = (float)(-3 * refPic[rowOne + colOne] - 10 *
refPic[rowOne + colTwo] - 3 * refPic[rowOne + colThree] + 3 * refPic[rowThree +
colOne] + 10 * refPic[rowThree + colTwo] + 3 * refPic[rowThree + colThree]);
+
+ gradientMagnitude = sqrtf(gradientH * gradientH +
gradientV * gradientV);
+ radians = atan2(gradientV, gradientH);
+ theta = (float)((radians * 180) / PI);
+ if (theta < 0)
+ theta = 180 + theta;
+ edgeTheta[(rowNum*stride) + colNum] = (pixel)theta;
+
+ edgePic[index] = (pixel)(gradientMagnitude >=
edge_threshold ? whitePixel : blackPixel);
+ }
+ }
+ }
+ }
+
+ Histogram::Histogram()
+ {
+ memset(frequency_distribution, 0, HISTOGRAM_SIZE * sizeof(int32_t));
+ }
+
+ Histogram::Histogram(Histogram const& hist) {
+ memset(frequency_distribution, 0, HISTOGRAM_SIZE * sizeof(int32_t));
+ for (int i = 0; i < HISTOGRAM_SIZE; i++) {
+ frequency_distribution[i] = hist.frequency_distribution[i];
+ }
+ }
+
+ Histogram & Histogram::operator=(Histogram const& hist)
+ {
+ memset(frequency_distribution, 0, HISTOGRAM_SIZE * sizeof(int32_t));
+ memcpy(frequency_distribution, hist.frequency_distribution,
sizeof(int32_t)*HISTOGRAM_SIZE);
+ return *this;
+ }
+
+ Histogram::~Histogram() {}
+
+ YuvHistogram::YuvHistogram() {}
+
+ void YuvHistogram::initHistograms(int32_t planecount) {
+ this->plane_count = planecount;
+ m_isalloc = false;
+ bisUpdated = false;
+ param = NULL;
+
+ yuv_hist = edge_hist = NULL;
+ plane_sizes = plane_heights = plane_widths = NULL;
+ edgePic = edgeThetaPic = NULL;
+
+ plane_sizes = X265_MALLOC(int32_t, plane_count);
+ plane_heights = X265_MALLOC(int32_t, plane_count);
+ plane_widths = X265_MALLOC(int32_t, plane_count);
+
+ if (!plane_sizes || !plane_heights || !plane_widths) {
+ x265_log(param, X265_LOG_ERROR, "unable to allocate memory for
plane dimensions\n");
+ m_isalloc &= false;
+ }
+ else {
+ memset(plane_sizes, 0, plane_count * sizeof(int32_t));
+ memset(plane_heights, 0, plane_count * sizeof(int32_t));
+ memset(plane_widths, 0, plane_count * sizeof(int32_t));
+ m_isalloc &= true;
+ }
+
+ yuv_hist = X265_MALLOC(Histogram, plane_count);
+ edge_hist = X265_MALLOC(Histogram, plane_count);
+
+ if (!yuv_hist || !edge_hist) {
+ m_isalloc &= false;
+ x265_log(param, X265_LOG_ERROR, "unable to allocate memory for
histograms\n");
+ }
+
+ }
+
+ void YuvHistogram::initHistograms(x265_param *p) {
+ param = p;
+ plane_count = x265_cli_csps[param->internalCsp].planes;
+ bisUpdated = false;
+ m_isalloc = false;
+
+ yuv_hist = edge_hist = NULL;
+ plane_sizes = plane_heights = plane_widths = NULL;
+ edgePic = edgeThetaPic = NULL;
+
+ plane_sizes = X265_MALLOC(int32_t, plane_count);
+ plane_heights = X265_MALLOC(int32_t, plane_count);
+ plane_widths = X265_MALLOC(int32_t, plane_count);
+
+ if (!plane_sizes || !plane_heights || !plane_widths) {
+ x265_log(param, X265_LOG_ERROR, "unable to allocate memory for
plane dimensions\n");
+ m_isalloc &= false;
+ }
+ else {
+ memset(plane_sizes, 0, plane_count * sizeof(int32_t));
+ memset(plane_heights, 0, plane_count * sizeof(int32_t));
+ memset(plane_widths, 0, plane_count * sizeof(int32_t));
+ m_isalloc &= true;
+ }
+
+ yuv_hist = X265_MALLOC(Histogram, plane_count);
+ edge_hist = X265_MALLOC(Histogram, plane_count);
+
+ if (!yuv_hist || !edge_hist) {
+ m_isalloc &= false;
+ x265_log(param, X265_LOG_ERROR, "unable to allocate memory for
histograms\n");
+ }
+
+ }
+
+ bool YuvHistogram::allocHistogramBuffers() {
+ //allocate memory for edge filter output and histograms
+ bool isalloc = true;
+
+ edgePic = X265_MALLOC(pixel*, plane_count);
+ edgeThetaPic = X265_MALLOC(pixel*, plane_count);
+
+ if (!edgePic || !edgeThetaPic) {
+ isalloc &= false;
+ x265_log(param, X265_LOG_ERROR, "unable to allocate memory for
edge buffers\n");
+ return isalloc;
+ }
+
+ for (int i = 0; i < plane_count; i++) {
+
+ edgePic[i] = edgeThetaPic[i] = NULL;
+ edgePic[i] = X265_MALLOC(pixel, plane_sizes[i]);
+ edgeThetaPic[i] = X265_MALLOC(pixel, plane_sizes[i]);
+
+ if (edgePic[i] && edgeThetaPic[i]) {
+ memset(edgePic[i], 0, plane_sizes[i] * sizeof(pixel));
+ memset(edgeThetaPic[i], 0, plane_sizes[i] * sizeof(pixel));
+ isalloc &= true;
+ }
+ else
+ isalloc &= false;
+ }
+ return isalloc;
+ }
+
+ YuvHistogram::~YuvHistogram()
+ {
+ freeHistogramBuffers(); //change implementation based on allocation
changes
+ }
+
+ YuvHistogram::YuvHistogram(YuvHistogram const& hist) {
+
+ maxuv_hist = hist.maxuv_hist;
+ plane_count = hist.plane_count;
+ bisUpdated = hist.bisUpdated;
+ param = hist.param;
+ memcpy(plane_sizes, hist.plane_sizes, plane_count * sizeof(int32_t));
+ memcpy(plane_heights, hist.plane_heights, plane_count *
sizeof(int32_t));
+ memcpy(plane_widths, hist.plane_widths, plane_count * sizeof(int32_t));
+ memcpy(yuv_hist, hist.yuv_hist, plane_count * sizeof(Histogram));
+ memcpy(edge_hist, hist.edge_hist, plane_count * sizeof(Histogram));
+
+ if (!m_isalloc) {
+ m_isalloc = false;
+ m_isalloc = allocHistogramBuffers();
+ }
+
+ if (m_isalloc) {
+ for (int i = 0; i < plane_count; i++) {
+ if (edgePic[i] && edgeThetaPic[i]) {
+ memcpy(edgePic[i], hist.edgePic[i], plane_sizes[i] *
sizeof(pixel));
+ memcpy(edgeThetaPic[i], hist.edgeThetaPic[i],
plane_sizes[i] * sizeof(pixel));
+ }
+ }
+ }
+
+ }
+
+ YuvHistogram & YuvHistogram ::operator=(const YuvHistogram & copy_hist)
+ {
+ maxuv_hist = copy_hist.maxuv_hist;
+ plane_count = copy_hist.plane_count;
+ bisUpdated = copy_hist.bisUpdated;
+ param = copy_hist.param;
+ memcpy(plane_sizes, copy_hist.plane_sizes, plane_count *
sizeof(int32_t));
+ memcpy(plane_heights, copy_hist.plane_heights, plane_count *
sizeof(int32_t));
+ memcpy(plane_widths, copy_hist.plane_widths, plane_count *
sizeof(int32_t));
+ memcpy(yuv_hist, copy_hist.yuv_hist, plane_count * sizeof(Histogram));
+ memcpy(edge_hist, copy_hist.edge_hist, plane_count *
sizeof(Histogram));
+
+ if (!m_isalloc) {
+ m_isalloc = false;
+ m_isalloc = allocHistogramBuffers();
+ }
+
+ if (m_isalloc) {
+ for (int i = 0; i < plane_count; i++) {
+ if (edgePic[i] && edgeThetaPic[i]) {
+ memcpy(edgePic[i], copy_hist.edgePic[i], plane_sizes[i] *
sizeof(pixel));
+ memcpy(edgeThetaPic[i], copy_hist.edgeThetaPic[i],
plane_sizes[i] * sizeof(pixel));
+ }
+ }
+ }
+
+ return *this;
+ }
+
+ void YuvHistogram::initFrameDimensions(x265_picture & pic) {
+
+ for (int i = 0; i < plane_count; i++) {
+ plane_widths[i] = pic.width;
+ plane_heights[i] = pic.height >>
x265_cli_csps[pic.colorSpace].height[i];
+ plane_sizes[i] = plane_widths[i] * plane_heights[i];
+ }
+ }
+
+ void YuvHistogram::freeHistogramBuffers() {
+ //de allocate memory for histograms and edge filtered output
+ if (edgePic && edgeThetaPic) {
+ for (int i = 0; i < plane_count; i++) {
+ if (edgePic[i] && edgeThetaPic[i]) {
+ X265_FREE_ZERO(edgePic[i]);
+ X265_FREE_ZERO(edgeThetaPic[i]);
+ }
+ }
+ X265_FREE_ZERO(edgePic);
+ X265_FREE_ZERO(edgeThetaPic);
+ }
+
+ if (plane_sizes && plane_heights && plane_widths) {
+ X265_FREE_ZERO(plane_sizes);
+ X265_FREE_ZERO(plane_heights);
+ X265_FREE_ZERO(plane_widths);
+ }
+
+ if (yuv_hist && edge_hist) {
+ X265_FREE_ZERO(yuv_hist);
+ X265_FREE_ZERO(edge_hist);
+ }
+ }
+
+ bool YuvHistogram::edgeFilter(x265_picture *frame) {
+
+ if (!m_isalloc) {
+ initFrameDimensions(*frame);
+ m_isalloc = allocHistogramBuffers();
+ }
+
+ if (m_isalloc) {
+ for (int idx = 0; idx < EDGE_PLANE_COUNT; idx++) {
+
+ memset(edgePic[idx], 0, sizeof(pixel) * plane_sizes[idx]);
+ memset(edgeThetaPic[idx], 0, sizeof(pixel) * plane_sizes[idx]);
+
+ pixel *src = (pixel*)frame->planes[idx];
+ pixel *edge_pic = edgePic[idx];
+ pixel *ref_pic = src;
+ pixel *edge_theta = edgeThetaPic[idx];
+
+ assert(edge_pic != NULL);
+ assert(ref_pic != NULL);
+ memcpy(edge_pic, src, plane_sizes[idx] * sizeof(pixel));
+ memcpy(ref_pic, src, plane_sizes[idx] * sizeof(pixel));
+
+ computeEdge(edge_pic, ref_pic, edge_theta,plane_widths[idx],
plane_heights[idx], plane_widths[idx]);
+ }
+ return true;
+ }
+ else {
+ return false;
+ }
+ }
+
+ bool YuvHistogram::computeHistograms(x265_picture &cur_frame) {
+
+ bool bsuccess = false;
+ bsuccess = computeLumaEdgeHistogram(cur_frame);
+ if (bsuccess) {
+ if (plane_count > 1) {
+ bsuccess &= computeChromaHistogram(cur_frame);
+ }
+ return bsuccess;
+ }
+ else {
+ return bsuccess;
+ }
+
+ }
+
+ bool YuvHistogram::computeLumaEdgeHistogram(x265_picture &frame) {
+
+ pixel pixel_val = 0;
+
+ memset(edge_hist[0].frequency_distribution, 0, HISTOGRAM_SIZE *
sizeof(int32_t));
+
+ int size = frame.height*(frame.stride[0] >> SHIFT);
+
+ for (int i = 0; i < size; i++) {
+ pixel_val = edgePic[0][i];
+ edge_hist[0].frequency_distribution[pixel_val]++;
+ }
+ return true;
+ }
+
+ bool YuvHistogram::computeChromaHistogram(x265_picture &frame) {
+ /*u hist calculation*/
+ pixel pixel_val = 0;
+ int32_t pixel_ucount = 0, pixel_vcount = 0;
+
+ int u_height = (frame.height >>
x265_cli_csps[frame.colorSpace].height[1]);
+ int size = u_height * (frame.stride[1] >> SHIFT);
+ memset(yuv_hist[1].frequency_distribution, 0, HISTOGRAM_SIZE *
sizeof(int32_t));
+
+ for (int i = 0; i < size; i++) {
+ pixel_val = *((pixel *)frame.planes[1] + i);
+ yuv_hist[1].frequency_distribution[pixel_val]++;
+ pixel_ucount++;
+ }
+
+ /*v hist calculation for independent uv planes */
+
+ if (plane_count == 3) {
+ pixel_val = 0;
+ int v_height = (frame.height >>
x265_cli_csps[frame.colorSpace].height[2]);
+ size = v_height * (frame.stride[2] >> SHIFT);
+ memset(yuv_hist[2].frequency_distribution, 0, HISTOGRAM_SIZE *
sizeof(int32_t));
+
+ for (int i = 0; i < size; i++) {
+ pixel_val = *((pixel *)frame.planes[2] + i);
+ yuv_hist[2].frequency_distribution[pixel_val]++;
+ pixel_vcount++;
+ }
+
+ for (int i = 0; i < HISTOGRAM_SIZE; i++) {
+ maxuv_hist.frequency_distribution[i] =
max(yuv_hist[1].frequency_distribution[i],
yuv_hist[2].frequency_distribution[i]);
+ }
+
+ }
+ else {
+ maxuv_hist = yuv_hist[1]; //for two planes scenario
+ }
+
+ return true;
+ }
+
+ bool YuvHistogram::isUpdated() {
+ return bisUpdated;
+ }
+
+ void YuvHistogram::setUpdateFlag(bool flag) {
+ bisUpdated = flag;
+ }
+
+ bool YuvHistogram::getUpdateFlag() {
+ return bisUpdated;
+ }
+
+ SadYuv::SadYuv() { }
+
+ void SadYuv::initSadYuv(int planecount) {
+ this->plane_count = planecount;
+ sad_yuv = NULL;
+ psad_yuv = NULL;
+ sad_yuv = X265_MALLOC(int32_t, plane_count);
+ psad_yuv = X265_MALLOC(double, plane_count);
+ if (sad_yuv && psad_yuv) {
+ memset(sad_yuv, 0, plane_count * sizeof(int32_t));
+ memset(psad_yuv, 0, sizeof(double) * plane_count);
+ }
+ }
+
+ SadYuv & SadYuv::operator=(SadYuv const& sad_val) {
+ this->plane_count = sad_val.plane_count;
+ if (!sad_yuv && !psad_yuv) {
+ sad_yuv = NULL;
+ psad_yuv = NULL;
+ sad_yuv = X265_MALLOC(int32_t, plane_count);
+ psad_yuv = X265_MALLOC(double, plane_count);
+ if (sad_yuv && psad_yuv) {
+ memcpy(sad_yuv, sad_val.sad_yuv, plane_count *
sizeof(int32_t));
+ memcpy(psad_yuv, sad_val.psad_yuv, sizeof(double) *
plane_count);
+ }
+ }
+ else {
+ if (sad_yuv)
+ memcpy(sad_yuv, sad_val.sad_yuv, plane_count *
sizeof(int32_t));
+ if (psad_yuv)
+ memcpy(psad_yuv, sad_val.psad_yuv, sizeof(double) *
plane_count);
+ }
+ return *this;
+ }
+
+ SadYuv::~SadYuv() {
+ if (sad_yuv && psad_yuv) {
+ X265_FREE(sad_yuv);
+ X265_FREE(psad_yuv);
+ }
+ }
+
+ int sad_stats::frames_scanned=0;
+ int sad_stats::line_number=0;
+
+ sad_stats::sad_stats(int planecount, double threshold) {
+ this->plane_count = planecount;
+ calculateThresholds(threshold);
+ allocateBuffers();
+ }
+
+ void sad_stats::calculateThresholds(double threshold) {
+ edge_hist_threshold = threshold;
+ strength_factor = 2.0;
+ chroma_hist_threshold = threshold * 10.0;
+ scaled_edge_threshold = edge_hist_threshold * strength_factor;
+ scaled_chroma_threshold = chroma_hist_threshold * strength_factor;
+ }
+
+ void sad_stats::init() {
+ bscene_cut = NULL;
+ bdrop_frame = NULL;
+ sad_vals = NULL;
+ maxuv_sad_vals = NULL;
+ edge_sad_vals = NULL;
+ prev_hist = NULL;
+ }
+
+ sad_stats::~sad_stats() {
+ releaseBuffers();
+ }
+
+ void sad_stats::allocateBuffers() {
+
+ init();
+ sad_vals = new SadYuv[DUP_BUFFER]();
+ maxuv_sad_vals = new SadYuv[DUP_BUFFER]();
+ edge_sad_vals = new SadYuv[DUP_BUFFER]();
+ prev_hist = new YuvHistogram();
+ prev_hist->initHistograms(plane_count);
+
+ for (int i = 0; i < DUP_BUFFER; i++) {
+ sad_vals[i].initSadYuv(plane_count);
+ maxuv_sad_vals[i].initSadYuv(plane_count);
+ edge_sad_vals[i].initSadYuv(plane_count);
+ }
+
+ bscene_cut = new bool[DUP_BUFFER];
+ bdrop_frame = new bool[DUP_BUFFER];
+
+ if (!sad_vals || !maxuv_sad_vals || !edge_sad_vals || !bscene_cut ||
!bdrop_frame) {
+ x265_log(NULL, X265_LOG_ERROR, "Heap Error !");
+ exit(101);
+ }
+ else {
+ memset(bscene_cut, false, 2 * sizeof(bool));
+ memset(bdrop_frame, false, 2 * sizeof(bool));
+ }
+
+ }
+
+ void sad_stats::releaseBuffers() {
+ if (sad_vals && maxuv_sad_vals && edge_sad_vals && bscene_cut &&
bdrop_frame && prev_hist) {
+ delete[] sad_vals;
+ delete[] maxuv_sad_vals;
+ delete[] edge_sad_vals;
+ delete[] bscene_cut;
+ delete[] bdrop_frame;
+ delete prev_hist;
+ }
+ }
+
+ bool sad_stats::computeSadValue(YuvHistogram *input_frames, int32_t*
plane_sizes) {
+
+ int32_t *yuv_sad_val = NULL, *edge_sad_val = NULL, *maxuv_sad_val =
NULL;
+
+ double *maxuv_normalized_sad = NULL, *yuv_norm_sad = NULL,
*edge_normalized_sads = NULL;
+
+ YuvHistogram * ref_hist = NULL, *cur_hist = NULL;
+
+ /*inorder to process frames as per poc's updated by frame duplication
*/
+ if (frames_scanned > 0) {
+
+ if (!input_frames[0].isUpdated() && input_frames[1].isUpdated()) {
+ ref_hist = prev_hist;
+ cur_hist = input_frames + 1;
+
+ yuv_sad_val = sad_vals[1].sad_yuv,
+ edge_sad_val = edge_sad_vals[1].sad_yuv,
+ maxuv_sad_val = maxuv_sad_vals[1].sad_yuv;
+ maxuv_normalized_sad = maxuv_sad_vals[1].psad_yuv,
+ yuv_norm_sad = sad_vals[1].psad_yuv,
+ edge_normalized_sads = edge_sad_vals[1].psad_yuv;
+ input_frames[1].setUpdateFlag(false);
+
+ }
+ else if (input_frames[0].isUpdated() &&
input_frames[1].isUpdated()) {
+ ref_hist = prev_hist;
+ cur_hist = input_frames;
+
+ yuv_sad_val = sad_vals[0].sad_yuv,
+ edge_sad_val = edge_sad_vals[0].sad_yuv,
+ maxuv_sad_val = maxuv_sad_vals[0].sad_yuv;
+ maxuv_normalized_sad = maxuv_sad_vals[0].psad_yuv,
+ yuv_norm_sad = sad_vals[0].psad_yuv,
+ edge_normalized_sads = edge_sad_vals[0].psad_yuv;
+ input_frames[0].setUpdateFlag(false);
+
+ }
+ else if (input_frames[0].isUpdated() &&
!input_frames[1].isUpdated()) {
+ ref_hist = prev_hist;
+ cur_hist = input_frames;
+
+ yuv_sad_val = sad_vals[0].sad_yuv,
+ edge_sad_val = edge_sad_vals[0].sad_yuv,
+ maxuv_sad_val = maxuv_sad_vals[0].sad_yuv;
+ maxuv_normalized_sad = maxuv_sad_vals[0].psad_yuv,
+ yuv_norm_sad = sad_vals[0].psad_yuv,
+ edge_normalized_sads = edge_sad_vals[0].psad_yuv;
+ input_frames[0].setUpdateFlag(false);
+ }
+ else {
+ return true;
+ }
+ }
+ else {
+ cur_hist = input_frames;
+
+ yuv_sad_val = sad_vals[0].sad_yuv,
+ edge_sad_val = edge_sad_vals[0].sad_yuv,
+ maxuv_sad_val = maxuv_sad_vals[0].sad_yuv;
+
+ maxuv_normalized_sad = maxuv_sad_vals[0].psad_yuv,
+ yuv_norm_sad = sad_vals[0].psad_yuv,
+ edge_normalized_sads = edge_sad_vals[0].psad_yuv;
+ input_frames[0].setUpdateFlag(false);
+ }
+
+ if (frames_scanned == 0) { //first frame is scenecut by default no sad
computation for the same.
+
+ maxuv_sad_val[0] = 0;
+ maxuv_normalized_sad[0] = 0.0;
+ memset(yuv_sad_val, 0 , plane_count * sizeof(int32_t));
+ memset(edge_sad_val, 0, plane_count * sizeof(int32_t));
+ memset(edge_normalized_sads, 0, plane_count * sizeof(double));
+ memset(yuv_norm_sad, 0, plane_count * sizeof(double));
+
+ }
+ else {
+ int32_t freq_diff[3];
+ int32_t maxuv_freq_diff[1];
+ int32_t edge_freq_diff[3];
+ double color_probability_diff[3], edge_probability_diff[3];
+
+ memset(yuv_sad_val, 0, plane_count*sizeof(int32_t));
+ memset(edge_sad_val, 0, plane_count*sizeof(int32_t));
+
+ memset(yuv_norm_sad, 0, plane_count * sizeof(double));
+ memset(edge_normalized_sads, 0, plane_count * sizeof(double));
+ memset(color_probability_diff, 0, plane_count * sizeof(double));
+ memset(edge_probability_diff, 0, plane_count * sizeof(double));
+
+ maxuv_normalized_sad[0] = 0.0;
+ maxuv_sad_val[0] = 0;
+
+ memset(freq_diff, 0, 3 * sizeof(int32_t));
+ memset(maxuv_freq_diff, 0, sizeof(int32_t));
+ memset(edge_freq_diff, 0, 3 * sizeof(int32_t));
+
+ for (int i = 0; i < plane_count; i++) {
+ {
+ for (int j = 0; j < HISTOGRAM_SIZE; j++) {
+
+ if (i == 0 && plane_count >= 1) {
+ maxuv_freq_diff[i] =
(abs(cur_hist->maxuv_hist.frequency_distribution[j] -
ref_hist->maxuv_hist.frequency_distribution[j]));
+ maxuv_sad_val[i] += maxuv_freq_diff[i];
+ maxuv_normalized_sad[i] +=
(double)maxuv_freq_diff[i] / plane_sizes[i];
+ edge_freq_diff[i] =
abs(cur_hist->edge_hist[i].frequency_distribution[j] -
ref_hist->edge_hist[i].frequency_distribution[j]);
+ edge_probability_diff[i] =
double(edge_freq_diff[i]) / plane_sizes[i];
+ edge_sad_val[i] += edge_freq_diff[i];
+ edge_normalized_sads[i] +=
edge_probability_diff[i];
+ }
+ else {
+ freq_diff[i] =
abs(cur_hist->yuv_hist[i].frequency_distribution[j] -
ref_hist->yuv_hist[i].frequency_distribution[j]);
+ color_probability_diff[i] = (double)freq_diff[i] /
plane_sizes[i];
+ yuv_sad_val[i] += freq_diff[i];
+ yuv_norm_sad[i] += color_probability_diff[i];
+ }
+
+ }
+
+ }
+ }
+
+ }
+
+ *prev_hist = *cur_hist;
+
+ frames_scanned++;
+
+ return true;
+ }
+
+ void sad_stats::findSceneCuts(x265_picture * picList, bool& bdup) {
+
+ if (frames_scanned == 1) {
+ //for first frame
+ bscene_cut[0] = true;
+ bdrop_frame[0] = false;
+ picList->analysisData.bScenecut = (int)getSceneCutflag(0);
+ bdup = getDropflag(0);
+ picList->analysisData.edgeSadValue =
edge_sad_vals[0].psad_yuv[0];
+ picList->analysisData.chromaSadValue =
maxuv_sad_vals[0].psad_yuv[0];
+ }
+ else {
+ bscene_cut[1] = bdrop_frame[1] = false;
+ if (edge_sad_vals[1].psad_yuv[0] == 0) {
+ bdrop_frame[1] = true;
+ }
+ else if (edge_sad_vals[1].psad_yuv[0] > edge_hist_threshold ||
maxuv_sad_vals[1].psad_yuv[0] >= chroma_hist_threshold) {
+ bscene_cut[1] = true;
+ bdrop_frame[1] = false;
+ }
+ else if (edge_sad_vals[1].psad_yuv[0] > scaled_edge_threshold
|| maxuv_sad_vals[1].psad_yuv[0] >= scaled_chroma_threshold) {
+ bscene_cut[1] = true;
+ bdrop_frame[1] = false;
+ }
+ picList->analysisData.bScenecut = (int)getSceneCutflag(1);
+ bdup = getDropflag(1);
+ picList->analysisData.edgeSadValue =
edge_sad_vals[1].psad_yuv[0];
+ picList->analysisData.chromaSadValue =
maxuv_sad_vals[1].psad_yuv[0];
+ }
+ }
+
+ bool sad_stats::getDropflag(int i) {
+ return bdrop_frame[i];
+ }
+
+ bool sad_stats::getSceneCutflag(int i) {
+ return bscene_cut[i];
+ }
+
+}
\ No newline at end of file
diff -r 37648fca915b -r deaecadc4306 source/common/scenecut.h
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/source/common/scenecut.h Tue Oct 15 12:16:17 2019 +0530
@@ -0,0 +1,147 @@
+#ifndef SCENECUT_H
+#define SCENECUT_H
+
+#include <string>
+#include <iostream>
+#include <sstream>
+#include <vector>
+#include <algorithm>
+#include <math.h>
+
+#include "yuv.h"
+#include "common.h"
+
+#ifdef HIGH_BIT_DEPTH
+#define edge_threshold 1023.0
+#define whitePixel 1023.0
+#else
+#define edge_threshold 255.0
+#define pixel whitePixel 255.0
+#endif
+
+using namespace std;
+
+namespace X265_NS {
+
+ class Histogram {
+
+ public:
+ int32_t frequency_distribution[HISTOGRAM_SIZE];
+
+ Histogram();
+
+ Histogram(Histogram const& hist);
+
+ Histogram & operator=(Histogram const& hist);
+
+ ~Histogram();
+
+ };
+
+ class YuvHistogram {
+ public:
+ Histogram *yuv_hist;
+ Histogram *edge_hist;
+ int32_t *plane_sizes;
+ int32_t *plane_heights;
+ int32_t *plane_widths;
+
+ Histogram maxuv_hist;
+ int32_t plane_count;
+ bool bisUpdated;
+
+ pixel** edgePic;
+ pixel** edgeThetaPic;
+
+ x265_param * param; /*for handling various color spaces*/
+ bool m_isalloc;
+
+ YuvHistogram();
+
+ void initHistograms(int32_t plane_count);
+
+ void initHistograms(x265_param *p);
+
+ bool allocHistogramBuffers();
+
+ YuvHistogram(YuvHistogram const& hist);
+
+ YuvHistogram & operator=(const YuvHistogram & copy_hist);
+
+ ~YuvHistogram();
+
+ void initFrameDimensions(x265_picture & pic);
+
+ void freeHistogramBuffers();
+
+ bool edgeFilter(x265_picture *frame);
+
+ bool computeHistograms(x265_picture &cur_frame);
+
+ bool computeLumaEdgeHistogram(x265_picture &frame);
+
+ bool computeChromaHistogram(x265_picture &frame);
+
+ bool isUpdated();
+
+ void setUpdateFlag(bool flag);
+
+ bool getUpdateFlag();
+
+ };
+
+ struct SadYuv {
+ int32_t *sad_yuv;
+ double *psad_yuv;
+ int plane_count;
+ ~SadYuv();
+ SadYuv();
+ void initSadYuv(int plane_count);
+ SadYuv & operator=(SadYuv const& sad_val);
+};
+
+ class sad_stats {
+
+ bool *bscene_cut;
+ bool *bdrop_frame;
+ SadYuv * sad_vals;
+ SadYuv * maxuv_sad_vals;
+ SadYuv * edge_sad_vals;
+ int plane_count;
+ static int line_number;
+ static int frames_scanned;
+ YuvHistogram *prev_hist;
+ double edge_hist_threshold;
+ double chroma_hist_threshold;
+ double scaled_chroma_threshold;
+ double scaled_edge_threshold;
+ double strength_factor;
+
+ public:
+ sad_stats(int plane_count, double threshold);
+
+ ~sad_stats();
+
+ void init();
+
+ void allocateBuffers();
+
+ void releaseBuffers();
+
+ void calculateThresholds(double threshold);
+
+ bool computeSadValue(YuvHistogram *frames, int32_t* plane_sizes);
+
+ void findSceneCuts(x265_picture * piclist,bool & bdup);
+
+ bool getDropflag(int i);
+
+ bool getSceneCutflag(int i);
+
+ };
+
+void computeEdge(pixel * edgePic, pixel *refPic, pixel * edgeTheta, intptr_t
stride, int height, int width);
+
+}
+
+#endif
diff -r 37648fca915b -r deaecadc4306 source/encoder/api.cpp
--- a/source/encoder/api.cpp Fri Oct 11 12:45:52 2019 +0530
+++ b/source/encoder/api.cpp Tue Oct 15 12:16:17 2019 +0530
@@ -31,6 +31,7 @@
#include "nal.h"
#include "bitcost.h"
#include "svt.h"
+#include "scenecut.h"
#if ENABLE_LIBVMAF
#include "libvmaf.h"
@@ -117,7 +118,10 @@
x265_log(param, X265_LOG_INFO, "build info %s\n", PFX(build_info_str));
encoder = new Encoder;
-
+ encoder->m_sad_stats = new
sad_stats(x265_cli_csps[p->internalCsp].planes,param->edgeTransitionThreshold);
+ encoder->m_hist_of_adj_frames = new YuvHistogram[2];
+ encoder->m_hist_of_adj_frames[0].initHistograms(p);
+ encoder->m_hist_of_adj_frames[1].initHistograms(p);
#ifdef SVT_HEVC
if (param->bEnableSvtHevc)
@@ -809,6 +813,7 @@
CHECKED_MALLOC_ZERO(interData->ref, int32_t,
analysis->numCUsInFrame * X265_MAX_PRED_MODE_PER_CTU * numDir);
}
analysis->interData = interData;
+ analysis->bScenecut = false;
return;
@@ -924,6 +929,7 @@
pic->rpu.payloadSize = 0;
pic->rpu.payload = NULL;
pic->picStruct = 0;
+ pic->bufUpdated = false;
if ((param->analysisSave || param->analysisLoad) || (param->bAnalysisType
== AVC_INFO))
{
@@ -933,7 +939,9 @@
uint32_t numCUsInFrame = widthInCU * heightInCU;
pic->analysisData.numCUsInFrame = numCUsInFrame;
pic->analysisData.numPartitions = param->num4x4Partitions;
+ pic->analysisData.bScenecut = false;
}
+
}
void x265_picture_free(x265_picture *p)
@@ -955,7 +963,8 @@
{
if (param && param->rc.zonefileCount) {
for (int i = 0; i < param->rc.zonefileCount; i++)
- x265_free(param->rc.zones[i].zoneParam);
+ if(param->rc.zones[i].zoneParam)
+ x265_free(param->rc.zones[i].zoneParam);
}
if (param && (param->rc.zoneCount || param->rc.zonefileCount))
x265_free(param->rc.zones);
diff -r 37648fca915b -r deaecadc4306 source/encoder/encoder.cpp
--- a/source/encoder/encoder.cpp Fri Oct 11 12:45:52 2019 +0530
+++ b/source/encoder/encoder.cpp Tue Oct 15 12:16:17 2019 +0530
@@ -119,6 +119,9 @@
m_frameEncoder[i] = NULL;
for (uint32_t i = 0; i < DUP_BUFFER; i++)
m_dupBuffer[i] = NULL;
+
+ m_hist_of_adj_frames = NULL;
+ m_sad_stats = NULL;
MotionEstimate::initScales();
#if ENABLE_HDR10_PLUS
@@ -162,7 +165,9 @@
int rows = (p->sourceHeight + p->maxCUSize - 1) >>
g_log2Size[p->maxCUSize];
int cols = (p->sourceWidth + p->maxCUSize - 1) >>
g_log2Size[p->maxCUSize];
- if (m_param->bEnableFrameDuplication)
+
+
+ if (m_param->bEnableFrameDuplication || m_param->bHistbasedScenecut)
{
size_t framesize = 0;
int pixelbytes = p->sourceBitDepth > 8 ? 2 : 1;
@@ -184,6 +189,7 @@
m_dupBuffer[i]->dupPlane = NULL;
m_dupBuffer[i]->dupPlane = X265_MALLOC(char, framesize);
m_dupBuffer[i]->dupPic->planes[0] = m_dupBuffer[i]->dupPlane;
+ m_dupBuffer[i]->bufUpdated = false;
m_dupBuffer[i]->bOccupied = false;
m_dupBuffer[i]->bDup = false;
}
@@ -820,7 +826,7 @@
m_exportedPic = NULL;
}
- if (m_param->bEnableFrameDuplication)
+ if (m_param->bEnableFrameDuplication || m_param->bHistbasedScenecut)
{
for (uint32_t i = 0; i < DUP_BUFFER; i++)
{
@@ -1280,6 +1286,33 @@
return psnrWeight = (psnrY * 6 + psnrU + psnrV) / 8;
}
+void Encoder::updateSceneCutAndFrameDuplicateFlags() {
+ /* SCD computation and drop flag*/
+ for (int i = 0; i < DUP_BUFFER; i++) {
+ if (m_dupBuffer[i]->bufUpdated) {
+ m_hist_of_adj_frames[i].setUpdateFlag(true);
+ m_hist_of_adj_frames[i].edgeFilter(m_dupBuffer[i]->dupPic);
+ m_hist_of_adj_frames[i].computeHistograms(*m_dupBuffer[i]->dupPic);
+ m_sad_stats->computeSadValue(m_hist_of_adj_frames,
m_hist_of_adj_frames->plane_sizes);
+ m_sad_stats->findSceneCuts(m_dupBuffer[i]->dupPic,
m_dupBuffer[i]->bDup);
+
+ if (m_dupBuffer[i]->dupPic->analysisData.bScenecut) {
+ x265_log(m_param, X265_LOG_DEBUG, "scene cut at %d edge hist
sad: %0.4lf maxuv hist sad: %0.4lf\n",
+
m_dupBuffer[i]->dupPic->poc,m_dupBuffer[i]->dupPic->analysisData.edgeSadValue,m_dupBuffer[i]->dupPic->analysisData.chromaSadValue);
+ }
+
+ if (m_dupBuffer[1]->bufUpdated)
+ m_hist_of_adj_frames[0] = m_hist_of_adj_frames[1];
+ }
+ }
+
+ }
+
+/* TBD
+- to be updated for missing parameters in case of re-use else where and
improvised to copy constructor / assignment operator of x265 picture data
structure.
+- benefits avoid function and use language features appropriately.
+*/
+
void Encoder::copyPicture(x265_picture *dest, const x265_picture *src)
{
dest->poc = src->poc;
@@ -1299,6 +1332,25 @@
memcpy(dest->planes[0], src->planes[0], src->framesize * sizeof(char));
dest->planes[1] = (char*)dest->planes[0] + src->stride[0] * src->height;
dest->planes[2] = (char*)dest->planes[1] + src->stride[1] * (src->height
>> x265_cli_csps[src->colorSpace].height[1]);
+ memcpy(&dest->analysisData, &src->analysisData, sizeof(src->analysisData));
+
+}
+
+void Encoder::setPictureFlags(int idx) {
+ m_dupBuffer[idx]->bOccupied = true;
+ m_dupBuffer[idx]->bufUpdated = true;
+ m_dupBuffer[idx]->bDup = false;
+}
+
+void Encoder::unsetPictureFlags(int idx) {
+ if (idx == 1) {
+ m_dupBuffer[idx]->bOccupied = false;
+ m_dupBuffer[idx]->bufUpdated = false;
+ m_dupBuffer[idx]->bDup = false;
+ }
+ else if (idx == 0) {
+ m_dupBuffer[idx]->bufUpdated = false;
+ }
}
/**
@@ -1327,7 +1379,9 @@
const x265_picture* inputPic = NULL;
static int written = 0, read = 0;
bool dontRead = false;
-
+ bool isScenecutEnabled = m_param->bHistbasedScenecut;
+ bool dropflag = false;
+
if (m_exportedPic)
{
if (!m_param->bUseAnalysisFile && m_param->analysisSave)
@@ -1338,7 +1392,7 @@
}
if ((pic_in && (!m_param->chunkEnd || (m_encodedFrameNum <
m_param->chunkEnd))) || (m_param->bEnableFrameDuplication && !pic_in && (read <
written)))
{
- if ((m_param->bEnableFrameDuplication && !pic_in && (read < written)))
+ if ((m_param->bEnableFrameDuplication && !pic_in && (read < written))
|| (isScenecutEnabled && !pic_in && (read < written)))
dontRead = true;
else
{
@@ -1361,7 +1415,7 @@
}
}
- if (m_param->bEnableFrameDuplication)
+ if (m_param->bEnableFrameDuplication || isScenecutEnabled )
{
double psnrWeight = 0;
@@ -1372,6 +1426,11 @@
copyPicture(m_dupBuffer[0]->dupPic, pic_in);
m_dupBuffer[0]->bOccupied = true;
written++;
+ if (m_param->bHistbasedScenecut) {
+ setPictureFlags(0);
+ updateSceneCutAndFrameDuplicateFlags();
+ unsetPictureFlags(0);
+ }
return 0;
}
else if (!m_dupBuffer[1]->bOccupied)
@@ -1379,31 +1438,54 @@
copyPicture(m_dupBuffer[1]->dupPic, pic_in);
m_dupBuffer[1]->bOccupied = true;
written++;
+ if (m_param->bHistbasedScenecut) {
+ setPictureFlags(1);
+ updateSceneCutAndFrameDuplicateFlags();
+ unsetPictureFlags(1);
+ }
}
- psnrWeight = ComputePSNR(m_dupBuffer[0]->dupPic,
m_dupBuffer[1]->dupPic, m_param);
-
- if (psnrWeight >= m_param->dupThreshold)
- {
- if (m_dupBuffer[0]->bDup)
- {
- m_dupBuffer[0]->dupPic->picStruct = tripling;
- m_dupBuffer[0]->bDup = false;
- read++;
+ if (m_param->bEnableFrameDuplication &&
m_param->bHistbasedScenecut) {
+ if (m_dupBuffer[1]->bDup == false &&
m_dupBuffer[1]->dupPic->analysisData.bScenecut == false) {
+ psnrWeight = ComputePSNR(m_dupBuffer[0]->dupPic,
m_dupBuffer[1]->dupPic, m_param);
+ if (psnrWeight >= m_param->dupThreshold)
+ dropflag = true;
}
- else
- {
- m_dupBuffer[0]->dupPic->picStruct = doubling;
- m_dupBuffer[0]->bDup = true;
- m_dupBuffer[1]->bOccupied = false;
- read++;
- return 0;
+ else {
+ dropflag = true;
}
}
- else if (m_dupBuffer[0]->bDup)
+ else if (m_param->bEnableFrameDuplication) {
+ psnrWeight = ComputePSNR(m_dupBuffer[0]->dupPic,
m_dupBuffer[1]->dupPic, m_param);
+ if (psnrWeight >= m_param->dupThreshold)
+ dropflag = true;
+ }
+
+ if (m_param->bEnableFrameDuplication)
+ {
+ if (dropflag)
+ {
+ if (m_dupBuffer[0]->bDup)
+ {
+ m_dupBuffer[0]->dupPic->picStruct = tripling;
+ m_dupBuffer[0]->bDup = false;
+ read++;
+ }
+ else
+ {
+ m_dupBuffer[0]->dupPic->picStruct = doubling;
+ m_dupBuffer[0]->bDup = true;
+ m_dupBuffer[1]->bOccupied = false;
+ read++;
+ return 0;
+ }
+ }
+ else if (m_dupBuffer[0]->bDup)
m_dupBuffer[0]->bDup = false;
- else
- m_dupBuffer[0]->dupPic->picStruct = 0;
+ else
+ m_dupBuffer[0]->dupPic->picStruct = 0;
+ }
+
}
if (read < written)
@@ -1485,7 +1567,10 @@
inFrame->m_poc = ++m_pocLast;
inFrame->m_userData = inputPic->userData;
- inFrame->m_pts = inputPic->pts;
+ inFrame->m_pts = inputPic->pts;
+ if (m_param->bHistbasedScenecut) {
+ inFrame->m_lowres.bScenecut = inputPic->analysisData.bScenecut;
+ }
inFrame->m_forceqp = inputPic->forceqp;
inFrame->m_param = (m_reconfigure || m_reconfigureRc) ?
m_latestParam : m_param;
inFrame->m_picStruct = inputPic->picStruct;
@@ -1613,7 +1698,7 @@
m_param->bUseRcStats = 0;
}
- if (m_param->bEnableFrameDuplication && ((read < written) ||
(m_dupBuffer[0]->dupPic->picStruct == tripling && (read <= written))))
+ if ( (m_param->bEnableFrameDuplication || isScenecutEnabled) && ((read
< written) || (m_dupBuffer[0]->dupPic->picStruct == tripling && (read <=
written))))
{
if (m_dupBuffer[0]->dupPic->picStruct == tripling)
m_dupBuffer[0]->bOccupied = m_dupBuffer[1]->bOccupied = false;
@@ -3162,6 +3247,7 @@
* adaptive I frame placement */
p->keyframeMax = INT_MAX;
p->scenecutThreshold = 0;
+ p->bHistbasedScenecut = 0;
}
else if (p->keyframeMax <= 1)
{
@@ -3175,6 +3261,7 @@
p->lookaheadDepth = 0;
p->bframes = 0;
p->scenecutThreshold = 0;
+ p->bHistbasedScenecut = 0;
p->bFrameAdaptive = 0;
p->rc.cuTree = 0;
p->bEnableWeightedPred = 0;
@@ -3828,6 +3915,20 @@
m_param->searchMethod = m_param->hmeSearchMethod[2];
}
}
+
+ if (p->bHistbasedScenecut && p->scenecutThreshold) {
+ p->scenecutThreshold = 0;
+ p->bHistbasedScenecut = false;
+ x265_log(p, X265_LOG_WARNING, "Amibigious choice. disabling scene cut
detection \n");
+ }
+ else if (p->scenecutThreshold && p->edgeTransitionThreshold != 0.01) {
+ x265_log(p, X265_LOG_WARNING, "using scenecut-bias %d for scene cut
detection\n",p->scenecutBias);
+ }
+ else if (p->bHistbasedScenecut && p->edgeTransitionThreshold == 0.0) {
+ p->edgeTransitionThreshold = 0.01;
+ x265_log(p, X265_LOG_INFO, "using default threshold %.2lf for scene
cut detection\n", p->edgeTransitionThreshold);
+ }
+
}
void Encoder::readAnalysisFile(x265_analysis_data* analysis, int curPoc, const
x265_picture* picIn, int paramBytes)
diff -r 37648fca915b -r deaecadc4306 source/encoder/encoder.h
--- a/source/encoder/encoder.h Fri Oct 11 12:45:52 2019 +0530
+++ b/source/encoder/encoder.h Tue Oct 15 12:16:17 2019 +0530
@@ -32,6 +32,8 @@
#include "nal.h"
#include "framedata.h"
#include "svt.h"
+#include "scenecut.h"
+
#ifdef ENABLE_HDR10_PLUS
#include "dynamicHDR10/hdr10plus.h"
#endif
@@ -154,6 +156,9 @@
//Flag to check whether the picture has duplicated.
bool bDup;
+
+ bool bufUpdated;
+
};
@@ -195,6 +200,9 @@
ThreadPool* m_threadPool;
FrameEncoder* m_frameEncoder[X265_MAX_FRAME_THREADS];
+
+ YuvHistogram* m_hist_of_adj_frames;
+ sad_stats* m_sad_stats;
DPB* m_dpb;
Frame* m_exportedPic;
FILE* m_analysisFileIn;
@@ -279,6 +287,10 @@
if (m_prevTonemapPayload.payload != NULL)
X265_FREE(m_prevTonemapPayload.payload);
#endif
+ delete m_sad_stats;
+ m_sad_stats = NULL;
+ delete[] m_hist_of_adj_frames;
+ m_hist_of_adj_frames = NULL;
};
void create();
@@ -349,6 +361,12 @@
void copyPicture(x265_picture *dest, const x265_picture *src);
+ void unsetPictureFlags(int index);
+
+ void setPictureFlags(int index);
+
+ void updateSceneCutAndFrameDuplicateFlags();
+
void initRefIdx();
void analyseRefIdx(int *numRefIdx);
void updateRefIdx();
@@ -364,6 +382,7 @@
void initSPS(SPS *sps);
void initPPS(PPS *pps);
};
+
}
#endif // ifndef X265_ENCODER_H
diff -r 37648fca915b -r deaecadc4306 source/encoder/ratecontrol.cpp
--- a/source/encoder/ratecontrol.cpp Fri Oct 11 12:45:52 2019 +0530
+++ b/source/encoder/ratecontrol.cpp Tue Oct 15 12:16:17 2019 +0530
@@ -493,6 +493,7 @@
CMP_OPT_FIRST_PASS("open-gop", m_param->bOpenGOP);
CMP_OPT_FIRST_PASS(" keyint", m_param->keyframeMax);
CMP_OPT_FIRST_PASS("scenecut", m_param->scenecutThreshold);
+ CMP_OPT_FIRST_PASS("hist-threshold",
m_param->edgeTransitionThreshold);
CMP_OPT_FIRST_PASS("intra-refresh", m_param->bIntraRefresh);
if (m_param->bMultiPassOptRPS)
{
@@ -1183,6 +1184,7 @@
m_param->rc.bStatRead = 0;
m_param->bFrameAdaptive = 0;
m_param->scenecutThreshold = 0;
+ m_param->bHistbasedScenecut = false;
m_param->rc.cuTree = 0;
if (m_param->bframes > 1)
m_param->bframes = 1;
@@ -2173,7 +2175,7 @@
if (m_isVbv && m_currentSatd > 0 && curFrame)
{
if (m_param->lookaheadDepth || m_param->rc.cuTree ||
- m_param->scenecutThreshold ||
+ (m_param->scenecutThreshold || m_param->bHistbasedScenecut) ||
(m_param->bFrameAdaptive && m_param->bframes))
{
/* Lookahead VBV: If lookahead is done, raise the quantizer as
necessary
diff -r 37648fca915b -r deaecadc4306 source/encoder/slicetype.cpp
--- a/source/encoder/slicetype.cpp Fri Oct 11 12:45:52 2019 +0530
+++ b/source/encoder/slicetype.cpp Tue Oct 15 12:16:17 2019 +0530
@@ -30,6 +30,7 @@
#include "primitives.h"
#include "lowres.h"
#include "mv.h"
+#include "scenecut.h"
#include "slicetype.h"
#include "motion.h"
@@ -114,8 +115,8 @@
//Applying Gaussian filter on the picture
src = (pixel*)curFrame->m_fencPic->m_picOrg[0];
refPic = curFrame->m_gaussianPic + curFrame->m_fencPic->m_lumaMarginY *
stride + curFrame->m_fencPic->m_lumaMarginX;
+ edgePic = pic1 + curFrame->m_fencPic->m_lumaMarginY * stride +
curFrame->m_fencPic->m_lumaMarginX;
pixel pixelValue = 0;
-
for (int rowNum = 0; rowNum < height; rowNum++)
{
for (int colNum = 0; colNum < width; colNum++)
@@ -127,7 +128,8 @@
1 [4 9 12 9 4]
--- [5 12 15 12 5]
159 [4 9 12 9 4]
- [2 4 5 4 2]*/
+ [2 4 5 4 2]
+ */
const intptr_t rowOne = (rowNum - 2)*stride, colOne = colNum -
2;
const intptr_t rowTwo = (rowNum - 1)*stride, colTwo = colNum -
1;
@@ -145,52 +147,7 @@
}
}
}
-
-#if HIGH_BIT_DEPTH //10-bit build
- float threshold = 1023;
- pixel whitePixel = 1023;
-#else
- float threshold = 255;
- pixel whitePixel = 255;
-#endif
-#define PI 3.14159265
-
- float gradientH = 0, gradientV = 0, radians = 0, theta = 0;
- float gradientMagnitude = 0;
- pixel blackPixel = 0;
- edgePic = curFrame->m_edgePic + curFrame->m_fencPic->m_lumaMarginY *
stride + curFrame->m_fencPic->m_lumaMarginX;
- //Applying Sobel filter on the gaussian filtered picture
- for (int rowNum = 0; rowNum < height; rowNum++)
- {
- for (int colNum = 0; colNum < width; colNum++)
- {
- edgeTheta[(rowNum*stride) + colNum] = 0;
- if ((rowNum != 0) && (colNum != 0) && (rowNum != height - 1) &&
(colNum != width - 1)) //Ignoring the border pixels of the picture
- {
- /*Horizontal and vertical gradients
- [ -3 0 3 ] [-3 -10 -3 ]
- gH = [ -10 0 10] gV = [ 0 0 0 ]
- [ -3 0 3 ] [ 3 10 3 ]*/
-
- const intptr_t rowOne = (rowNum - 1)*stride, colOne = colNum
-1;
- const intptr_t rowTwo = rowNum * stride, colTwo = colNum;
- const intptr_t rowThree = (rowNum + 1)*stride, colThree =
colNum + 1;
- const intptr_t index = (rowNum*stride) + colNum;
-
- gradientH = (float)(-3 * refPic[rowOne + colOne] + 3 *
refPic[rowOne + colThree] - 10 * refPic[rowTwo + colOne] + 10 * refPic[rowTwo +
colThree] - 3 * refPic[rowThree + colOne] + 3 * refPic[rowThree + colThree]);
- gradientV = (float)(-3 * refPic[rowOne + colOne] - 10 *
refPic[rowOne + colTwo] - 3 * refPic[rowOne + colThree] + 3 * refPic[rowThree +
colOne] + 10 * refPic[rowThree + colTwo] + 3 * refPic[rowThree + colThree]);
-
- gradientMagnitude = sqrtf(gradientH * gradientH + gradientV *
gradientV);
- radians = atan2(gradientV, gradientH);
- theta = (float)((radians * 180) / PI);
- if (theta < 0)
- theta = 180 + theta;
- edgeTheta[(rowNum*stride) + colNum] = (pixel)theta;
-
- edgePic[index] = gradientMagnitude >= threshold ? whitePixel :
blackPixel;
- }
- }
- }
+ computeEdge(edgePic, refPic, edgeTheta, stride, height, width);
}
//Find the angle of a block by averaging the pixel angles
@@ -1471,7 +1428,7 @@
if (m_lastNonB && !m_param->rc.bStatRead &&
((m_param->bFrameAdaptive && m_param->bframes) ||
- m_param->rc.cuTree || m_param->scenecutThreshold ||
+ m_param->rc.cuTree || m_param->scenecutThreshold ||
m_param->bHistbasedScenecut ||
(m_param->lookaheadDepth && m_param->rc.vbvBufferSize)))
{
slicetypeAnalyse(frames, false);
@@ -1962,10 +1919,15 @@
int numBFrames = 0;
int numAnalyzed = numFrames;
- bool isScenecut = scenecut(frames, 0, 1, true, origNumFrames);
+ bool isScenecut = false;
/* When scenecut threshold is set, use scenecut detection for I frame
placements */
- if (m_param->scenecutThreshold && isScenecut)
+ if (m_param->scenecutThreshold)
+ isScenecut = scenecut(frames, 0, 1, true, origNumFrames);
+ else if (m_param->bHistbasedScenecut)
+ isScenecut = frames[1]->bScenecut;
+
+ if (isScenecut)
{
frames[1]->sliceType = X265_TYPE_I;
return;
@@ -1976,14 +1938,24 @@
m_extendGopBoundary = false;
for (int i = m_param->bframes + 1; i < origNumFrames; i +=
m_param->bframes + 1)
{
- scenecut(frames, i, i + 1, true, origNumFrames);
+ if (m_param->scenecutThreshold)
+ 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) )
- {
- m_extendGopBoundary = true;
- break;
- }
+ if (m_param->scenecutThreshold)
+ {
+ if (frames[j]->bScenecut && scenecutInternal(frames, j
- 1, j, true))
+ {
+ m_extendGopBoundary = true;
+ break;
+ }
+ }
+ else if(m_param->bHistbasedScenecut &&
frames[j]->bScenecut)
+ {
+ m_extendGopBoundary = true;
+ break;
+ }
}
if (m_extendGopBoundary)
break;
@@ -2088,13 +2060,23 @@
{
for (int j = 1; j < numBFrames + 1; j++)
{
- if (scenecut(frames, j, j + 1, false, origNumFrames) ||
- (bForceRADL && (frames[j]->frameNum == preRADL)))
- {
- frames[j]->sliceType = X265_TYPE_P;
- numAnalyzed = j;
- break;
+ if (m_param->bHistbasedScenecut) {
+ if (frames[j]->bScenecut || (bForceRADL &&
(frames[j]->frameNum == preRADL)))
+ {
+ frames[j]->sliceType = X265_TYPE_P;
+ numAnalyzed = j;
+ break;
+ }
}
+ else if (m_param->scenecutThreshold){
+ if ( scenecut(frames, j, j + 1, false, origNumFrames) ||
(bForceRADL && (frames[j]->frameNum == preRADL)) )
+ {
+ frames[j]->sliceType = X265_TYPE_P;
+ numAnalyzed = j;
+ break;
+ }
+ }
+
}
}
resetStart = bKeyframe ? 1 : X265_MIN(numBFrames + 2, numAnalyzed + 1);
diff -r 37648fca915b -r deaecadc4306 source/encoder/slicetype.h
--- a/source/encoder/slicetype.h Fri Oct 11 12:45:52 2019 +0530
+++ b/source/encoder/slicetype.h Tue Oct 15 12:16:17 2019 +0530
@@ -43,6 +43,14 @@
#define AQ_EDGE_BIAS 0.5
#define EDGE_INCLINATION 45
+#ifdef HIGH_BIT_DEPTH
+#define edge_threshold 1023.0
+#define whitePixel 1023.0
+#else
+#define edge_threshold 255.0
+#define pixel whitePixel 255.0
+#endif
+
/* Thread local data for lookahead tasks */
struct LookaheadTLD
{
diff -r 37648fca915b -r deaecadc4306 source/test/regression-tests.txt
--- a/source/test/regression-tests.txt Fri Oct 11 12:45:52 2019 +0530
+++ b/source/test/regression-tests.txt Tue Oct 15 12:16:17 2019 +0530
@@ -157,6 +157,9 @@
ducks_take_off_420_720p50.y4m,--preset medium --aq-mode 4 --crf 22 --no-cutree
ducks_take_off_420_1_720p50.y4m,--preset medium --selective-sao 4 --sao --crf
20
Traffic_4096x2048_30p.y4m, --preset medium --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 0.01
+Traffic_4096x2048_30p.y4m, --preset medium --frame-dup --dup-threshold 60
--hrd --bitrate 10000 --vbv-bufsize 15000 --vbv-maxrate 12000 --hist-scenecut
--hist-threshold 0.01
+sintel_trailer_2k_1920x1080_24.yuv, --preset medium --scenecut 40
--scenecut-bias 20
# Main12 intraCost overflow bug test
720p50_parkrun_ter.y4m,--preset medium
diff -r 37648fca915b -r deaecadc4306 source/x265.h
--- a/source/x265.h Fri Oct 11 12:45:52 2019 +0530
+++ b/source/x265.h Tue Oct 15 12:16:17 2019 +0530
@@ -210,7 +210,9 @@
uint32_t numCUsInFrame;
uint32_t numPartitions;
uint32_t depthBytes;
- int bScenecut;
+ bool bScenecut;
+ double edgeSadValue;
+ double chromaSadValue;
x265_weight_param* wt;
x265_analysis_inter_data* interData;
x265_analysis_intra_data* intraData;
@@ -291,6 +293,9 @@
char sliceType;
int bScenecut;
double ipCostRatio;
+ double yedgeSadValue;
+ double chromaSadValue;
+
int frameLatency;
x265_cu_stats cuStats;
x265_pu_stats puStats;
@@ -465,6 +470,9 @@
//Dolby Vision RPU metadata
x265_dolby_vision_rpu rpu;
+ //Flag to determine the latest frame in the buffer
+ bool bufUpdated;
+
int fieldNum;
//SEI picture structure message
@@ -1017,8 +1025,9 @@
* decisions. Default is 0 - disabled. 1 is the same as 0. Max 16 */
int lookaheadSlices;
- /* An arbitrary threshold which determines how aggressively the lookahead
- * should detect scene cuts. The default (40) is recommended. */
+ /* An arbitrary threshold which determines how aggressively the lookahead
+ * should detect scene cuts. The default (40) is recommended.
+ * Used for encoding cost based scenecut detection */
int scenecutThreshold;
/* Replace keyframes by using a column of intra blocks that move across
the video
@@ -1803,6 +1812,7 @@
/*Emit content light level info SEI*/
int bEmitCLL;
+
/*
* Signals picture structure SEI timing message for every frame
@@ -1819,6 +1829,17 @@
/*Input sequence bit depth. It can be either 8bit, 10bit or 12bit.*/
int sourceBitDepth;
+
+ /* A genuine threshold which determines whether a frame is a scenecut or
not
+ * when compared against edge and color sad values of a frames
histograms.Default 0.01
+ * Range:real number in range (0,2)
+ * Used for histogram based scene cut detection */
+ double edgeTransitionThreshold;
+
+ /*enables improved scenecut detection algorithm to detect scenecuts for
slice type
+ decision and rate control */
+ bool bHistbasedScenecut;
+
} x265_param;
/* x265_param_alloc:
* Allocates an x265_param instance. The returned param structure is not
diff -r 37648fca915b -r deaecadc4306 source/x265cli.h
--- a/source/x265cli.h Fri Oct 11 12:45:52 2019 +0530
+++ b/source/x265cli.h Tue Oct 15 12:16:17 2019 +0530
@@ -129,6 +129,9 @@
{ "scenecut", required_argument, NULL, 0 },
{ "no-scenecut", no_argument, NULL, 0 },
{ "scenecut-bias", required_argument, NULL, 0 },
+ { "hist-scenecut", no_argument, NULL, 0},
+ { "no-hist-scenecut", no_argument, NULL, 0},
+ { "hist-threshold", required_argument, NULL, 0},
{ "fades", no_argument, NULL, 0 },
{ "no-fades", no_argument, NULL, 0 },
{ "radl", required_argument, NULL, 0 },
@@ -485,7 +488,10 @@
H0(" --gop-lookahead <integer> Extends gop boundary if a scenecut is
found within this from keyint boundary. Default 0\n");
H0(" --no-scenecut Disable adaptive I-frame decision\n");
H0(" --scenecut <integer> How aggressively to insert extra
I-frames. Default %d\n", param->scenecutThreshold);
- H1(" --scenecut-bias <0..100.0> Bias for scenecut detection. Default
%.2f\n", param->scenecutBias);
+ H0(" --hist-scenecut ..... Enables improved scene-cut detection
using histogram based algorithm.");
+ H0(" --no-hist-scenecut Disables improved scene-cut detection
using histogram based algorithm. ");
+ H0(" --scenecut-bias <0..100.0> Bias for scenecut detection. Default
%.2f\n", param->scenecutBias);
+ H0(" --hist-threshold <0.0..2.0> Threshold for histogram based
scenecut detection Default %.2f\n", param->edgeTransitionThreshold);
H0(" --[no-]fades Enable detection and handling of
fade-in regions. Default %s\n", OPT(param->bEnableFades));
H0(" --radl <integer> Number of RADL pictures allowed in
front of IDR. Default %d\n", param->radl);
H0(" --intra-refresh Use Periodic Intra Refresh instead of
IDR frames\n");
# HG changeset patch
# User gopi Satykrishna Akisetty
# Date 1571121977 -19800
# Tue Oct 15 12:16:17 2019 +0530
# Node ID deaecadc43060ba37a85d9724a1a306a86433432
# Parent 37648fca915b389bafe923d8443818359e80ebf2
Improved scenecut detection
This patch does the following.
1. identifies scenecuts using sad of edge and chroma histogram based thresholding.
2. Add option "--hist-scenecut" and "--hist-threshold' to enable improved scenecut method for slice type decisions,rate control and a threshold for determining scene-cuts.
3. Removed duplicate edgefilter code and created a global function for use in scene cut detection and aq in Lookahead.
diff -r 37648fca915b -r deaecadc4306 doc/reST/cli.rst
--- a/doc/reST/cli.rst Fri Oct 11 12:45:52 2019 +0530
+++ b/doc/reST/cli.rst Tue Oct 15 12:16:17 2019 +0530
@@ -1427,6 +1427,18 @@
intra cost of a frame used in scenecut detection. For example, a value of 5 indicates,
if the inter cost of a frame is greater than or equal to 95 percent of the intra cost of the frame,
then detect this frame as scenecut. Values between 5 and 15 are recommended. Default 5.
+
+.. option:: --hist-scenecut, --no-hist-scenecut
+
+ indicates that I-frames need to be inserted using edge and color histogram based scenecut algorithm.
+ option: `--hist-scencut` 1 enables adaptive I frame placement using this method and disables the default scene cut algorithm.
+ option:`--no-hist-scenecut` adaptive I frame placement.
+
+.. option:: --hist-threshold <0.0..2.0>
+
+ This value represents the threshold for SAD for edge histograms used in scenecut detection. This requires hist-scenecut to be enabled.
+ For example, a value of 0.2 indicates that a frame with SAD value greater than 0.2 against the previous frame as scenecut.
+ Values between 0.0 and 2.0 are recommended. Default 0.1.
.. option:: --radl <integer>
diff -r 37648fca915b -r deaecadc4306 source/CMakeLists.txt
--- a/source/CMakeLists.txt Fri Oct 11 12:45:52 2019 +0530
+++ b/source/CMakeLists.txt Tue Oct 15 12:16:17 2019 +0530
@@ -29,7 +29,7 @@
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 180)
+set(X265_BUILD 181)
configure_file("${PROJECT_SOURCE_DIR}/x265.def.in"
"${PROJECT_BINARY_DIR}/x265.def")
configure_file("${PROJECT_SOURCE_DIR}/x265_config.h.in"
diff -r 37648fca915b -r deaecadc4306 source/common/CMakeLists.txt
--- a/source/common/CMakeLists.txt Fri Oct 11 12:45:52 2019 +0530
+++ b/source/common/CMakeLists.txt Tue Oct 15 12:16:17 2019 +0530
@@ -151,4 +151,5 @@
predict.cpp predict.h
scalinglist.cpp scalinglist.h
quant.cpp quant.h contexts.h
- deblock.cpp deblock.h)
+ deblock.cpp deblock.h
+ scenecut.h scenecut.cpp)
diff -r 37648fca915b -r deaecadc4306 source/common/common.h
--- a/source/common/common.h Fri Oct 11 12:45:52 2019 +0530
+++ b/source/common/common.h Tue Oct 15 12:16:17 2019 +0530
@@ -129,14 +129,20 @@
typedef uint64_t sum2_t;
typedef uint64_t pixel4;
typedef int64_t ssum2_t;
+#define HISTOGRAM_SIZE 1024
+#define SHIFT 1
#else
typedef uint8_t pixel;
typedef uint16_t sum_t;
typedef uint32_t sum2_t;
typedef uint32_t pixel4;
typedef int32_t ssum2_t; // Signed sum
+#define HISTOGRAM_SIZE 256
+#define SHIFT 0
#endif // if HIGH_BIT_DEPTH
+#define PI 3.14159265
+
#if X265_DEPTH < 10
typedef uint32_t sse_t;
#else
diff -r 37648fca915b -r deaecadc4306 source/common/param.cpp
--- a/source/common/param.cpp Fri Oct 11 12:45:52 2019 +0530
+++ b/source/common/param.cpp Tue Oct 15 12:16:17 2019 +0530
@@ -167,6 +167,8 @@
param->bFrameAdaptive = X265_B_ADAPT_TRELLIS;
param->bBPyramid = 1;
param->scenecutThreshold = 40; /* Magic number pulled in from x264 */
+ param->edgeTransitionThreshold = 0.01;
+ param->bHistbasedScenecut = false;
param->lookaheadSlices = 8;
param->lookaheadThreads = 0;
param->scenecutBias = 5.0;
@@ -567,6 +569,7 @@
param->bframes = 0;
param->lookaheadDepth = 0;
param->scenecutThreshold = 0;
+ param->bHistbasedScenecut = false;
param->rc.cuTree = 0;
param->frameNumThreads = 1;
}
@@ -609,7 +612,7 @@
return 0;
}
-static int x265_atobool(const char* str, bool& bError)
+static bool x265_atobool(const char* str, bool& bError)
{
if (!strcmp(str, "1") ||
!strcmp(str, "true") ||
@@ -920,6 +923,7 @@
{
bError = false;
p->scenecutThreshold = atoi(value);
+ p->bHistbasedScenecut = false;
}
}
OPT("temporal-layers") p->bEnableTemporalSubLayers = atobool(value);
@@ -1186,6 +1190,31 @@
OPT("opt-ref-list-length-pps") p->bOptRefListLengthPPS = atobool(value);
OPT("multi-pass-opt-rps") p->bMultiPassOptRPS = atobool(value);
OPT("scenecut-bias") p->scenecutBias = atof(value);
+ OPT("hist-scenecut")
+ {
+ p->bHistbasedScenecut = atobool(value);
+
+ if (bError)
+ {
+ bError = false;
+ p->bHistbasedScenecut = false;
+ }
+
+ if (p->bHistbasedScenecut) {
+ bError = false;
+ p->scenecutThreshold = 0;
+ }
+
+ }
+ OPT("hist-threshold") {
+ p->edgeTransitionThreshold = atof(value);
+ if (bError)
+ {
+ bError = false;
+ p->edgeTransitionThreshold = 0.01;
+ x265_log(p, X265_LOG_INFO, "using default threshold %.2lf for scene cut detection\n", p->edgeTransitionThreshold);
+ }
+ }
OPT("lookahead-threads") p->lookaheadThreads = atoi(value);
OPT("opt-cu-delta-qp") p->bOptCUDeltaQP = atobool(value);
OPT("multi-pass-opt-analysis") p->analysisMultiPassRefine = atobool(value);
@@ -1623,8 +1652,14 @@
"Valid Logging level -1:none 0:error 1:warning 2:info 3:debug 4:full");
CHECK(param->scenecutThreshold < 0,
"scenecutThreshold must be greater than 0");
- CHECK(param->scenecutBias < 0 || 100 < param->scenecutBias,
- "scenecut-bias must be between 0 and 100");
+ if (param->scenecutThreshold) {
+ CHECK(param->scenecutBias < 0 || 100 < param->scenecutBias,
+ "scenecut-bias must be between 0 and 100");
+ }
+ else if (param->bHistbasedScenecut) {
+ CHECK(param->edgeTransitionThreshold < 0.0 || 2.0 < param->edgeTransitionThreshold,
+ "hist-threshold must be between 0.0 and 2.0");
+ }
CHECK(param->radl < 0 || param->radl > param->bframes,
"radl must be between 0 and bframes");
CHECK(param->rdPenalty < 0 || param->rdPenalty > 2,
@@ -1780,10 +1815,20 @@
x265_log(param, X265_LOG_INFO, "ME / range / subpel / merge : %s / %d / %d / %d\n",
x265_motion_est_names[param->searchMethod], param->searchRange, param->subpelRefine, param->maxNumMergeCand);
- if (param->keyframeMax != INT_MAX || param->scenecutThreshold)
- x265_log(param, X265_LOG_INFO, "Keyframe min / max / scenecut / bias: %d / %d / %d / %.2lf\n", param->keyframeMin, param->keyframeMax, param->scenecutThreshold, param->scenecutBias * 100);
+ if (param->scenecutThreshold && param->keyframeMax != INT_MAX)
+ param->edgeTransitionThreshold = 0.0;
+ else if (param->bHistbasedScenecut && param->keyframeMax != INT_MAX)
+ param->scenecutBias = 0.0;
+ else if (param->keyframeMax != INT_MAX) {
+ param->edgeTransitionThreshold = 0.0;
+ param->scenecutBias = 0.0;
+ }
+
+ if (param->keyframeMax == INT_MAX)
+ x265_log(param, X265_LOG_INFO, "Keyframe min / max / scenecut : disabled\n");
else
- x265_log(param, X265_LOG_INFO, "Keyframe min / max / scenecut : disabled\n");
+ x265_log(param, X265_LOG_INFO, "Keyframe min / max / scenecut / bias / threshold : %d / %d / %d / %.2lf / %.2lf\n", param->keyframeMin, param->keyframeMax, ( param->bHistbasedScenecut || param->scenecutThreshold ), param->scenecutBias * 100, param->edgeTransitionThreshold);
+
if (param->cbQpOffset || param->crQpOffset)
x265_log(param, X265_LOG_INFO, "Cb/Cr QP Offset : %d / %d\n", param->cbQpOffset, param->crQpOffset);
@@ -1949,6 +1994,8 @@
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);
+ s += sprintf(s, " hist-threshold=%.2f", p->edgeTransitionThreshold);
s += sprintf(s, " radl=%d", p->radl);
BOOL(p->bEnableHRDConcatFlag, "splice");
BOOL(p->bIntraRefresh, "intra-refresh");
@@ -2096,6 +2143,8 @@
BOOL(p->bOptRefListLengthPPS, "opt-ref-list-length-pps");
BOOL(p->bMultiPassOptRPS, "multi-pass-opt-rps");
s += sprintf(s, " scenecut-bias=%.2f", p->scenecutBias);
+ s += sprintf(s, " hist-threshold=%.2f", p->edgeTransitionThreshold);
+
BOOL(p->bOptCUDeltaQP, "opt-cu-delta-qp");
BOOL(p->bAQMotion, "aq-motion");
BOOL(p->bEmitHDRSEI, "hdr");
@@ -2246,6 +2295,7 @@
dst->lookaheadSlices = src->lookaheadSlices;
dst->lookaheadThreads = src->lookaheadThreads;
dst->scenecutThreshold = src->scenecutThreshold;
+ dst->bHistbasedScenecut = src->bHistbasedScenecut;
dst->bIntraRefresh = src->bIntraRefresh;
dst->maxCUSize = src->maxCUSize;
dst->minCUSize = src->minCUSize;
@@ -2403,6 +2453,7 @@
dst->bOptRefListLengthPPS = src->bOptRefListLengthPPS;
dst->bMultiPassOptRPS = src->bMultiPassOptRPS;
dst->scenecutBias = src->scenecutBias;
+ dst->edgeTransitionThreshold = src->edgeTransitionThreshold;
dst->gopLookahead = src->lookaheadDepth;
dst->bOptCUDeltaQP = src->bOptCUDeltaQP;
dst->analysisMultiPassDistortion = src->analysisMultiPassDistortion;
diff -r 37648fca915b -r deaecadc4306 source/common/scenecut.cpp
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/source/common/scenecut.cpp Tue Oct 15 12:16:17 2019 +0530
@@ -0,0 +1,674 @@
+#include <iostream>
+#include <sstream>
+#include <vector>
+#include <algorithm>
+#include <math.h>
+#include "encoder.h"
+#include "scenecut.h"
+#include "slicetype.h"
+
+using namespace std;
+using namespace X265_NS;
+
+#define EDGE_PLANE_COUNT 1
+
+namespace X265_NS {
+
+void computeEdge(pixel * edgePic, pixel *refPic, pixel * edgeTheta, intptr_t stride, int height, int width)
+ {
+ float gradientH = 0, gradientV = 0, radians = 0, theta = 0;
+ float gradientMagnitude = 0;
+ pixel blackPixel = 0;
+
+ //Applying Sobel filter
+ for (int rowNum = 0; rowNum < height; rowNum++)
+ {
+ for (int colNum = 0; colNum < width; colNum++)
+ {
+ edgeTheta[(rowNum*stride) + colNum] = 0;
+ if ((rowNum != 0) && (colNum != 0) && (rowNum != height - 1) && (colNum != width - 1)) //Ignoring the border pixels of the picture
+ {
+ /*Horizontal and vertical gradients
+ [ -3 0 3 ] [-3 -10 -3 ]
+ gH = [ -10 0 10] gV = [ 0 0 0 ]
+ [ -3 0 3 ] [ 3 10 3 ]*/
+
+ const intptr_t rowOne = (rowNum - 1)*stride, colOne = colNum - 1;
+ const intptr_t rowTwo = rowNum * stride, colTwo = colNum;
+ const intptr_t rowThree = (rowNum + 1)*stride, colThree = colNum + 1;
+ const intptr_t index = (rowNum*stride) + colNum;
+
+ gradientH = (float)(-3 * refPic[rowOne + colOne] + 3 * refPic[rowOne + colThree] - 10 * refPic[rowTwo + colOne] + 10 * refPic[rowTwo + colThree] - 3 * refPic[rowThree + colOne] + 3 * refPic[rowThree + colThree]);
+ gradientV = (float)(-3 * refPic[rowOne + colOne] - 10 * refPic[rowOne + colTwo] - 3 * refPic[rowOne + colThree] + 3 * refPic[rowThree + colOne] + 10 * refPic[rowThree + colTwo] + 3 * refPic[rowThree + colThree]);
+
+ gradientMagnitude = sqrtf(gradientH * gradientH + gradientV * gradientV);
+ radians = atan2(gradientV, gradientH);
+ theta = (float)((radians * 180) / PI);
+ if (theta < 0)
+ theta = 180 + theta;
+ edgeTheta[(rowNum*stride) + colNum] = (pixel)theta;
+
+ edgePic[index] = (pixel)(gradientMagnitude >= edge_threshold ? whitePixel : blackPixel);
+ }
+ }
+ }
+ }
+
+ Histogram::Histogram()
+ {
+ memset(frequency_distribution, 0, HISTOGRAM_SIZE * sizeof(int32_t));
+ }
+
+ Histogram::Histogram(Histogram const& hist) {
+ memset(frequency_distribution, 0, HISTOGRAM_SIZE * sizeof(int32_t));
+ for (int i = 0; i < HISTOGRAM_SIZE; i++) {
+ frequency_distribution[i] = hist.frequency_distribution[i];
+ }
+ }
+
+ Histogram & Histogram::operator=(Histogram const& hist)
+ {
+ memset(frequency_distribution, 0, HISTOGRAM_SIZE * sizeof(int32_t));
+ memcpy(frequency_distribution, hist.frequency_distribution, sizeof(int32_t)*HISTOGRAM_SIZE);
+ return *this;
+ }
+
+ Histogram::~Histogram() {}
+
+ YuvHistogram::YuvHistogram() {}
+
+ void YuvHistogram::initHistograms(int32_t planecount) {
+ this->plane_count = planecount;
+ m_isalloc = false;
+ bisUpdated = false;
+ param = NULL;
+
+ yuv_hist = edge_hist = NULL;
+ plane_sizes = plane_heights = plane_widths = NULL;
+ edgePic = edgeThetaPic = NULL;
+
+ plane_sizes = X265_MALLOC(int32_t, plane_count);
+ plane_heights = X265_MALLOC(int32_t, plane_count);
+ plane_widths = X265_MALLOC(int32_t, plane_count);
+
+ if (!plane_sizes || !plane_heights || !plane_widths) {
+ x265_log(param, X265_LOG_ERROR, "unable to allocate memory for plane dimensions\n");
+ m_isalloc &= false;
+ }
+ else {
+ memset(plane_sizes, 0, plane_count * sizeof(int32_t));
+ memset(plane_heights, 0, plane_count * sizeof(int32_t));
+ memset(plane_widths, 0, plane_count * sizeof(int32_t));
+ m_isalloc &= true;
+ }
+
+ yuv_hist = X265_MALLOC(Histogram, plane_count);
+ edge_hist = X265_MALLOC(Histogram, plane_count);
+
+ if (!yuv_hist || !edge_hist) {
+ m_isalloc &= false;
+ x265_log(param, X265_LOG_ERROR, "unable to allocate memory for histograms\n");
+ }
+
+ }
+
+ void YuvHistogram::initHistograms(x265_param *p) {
+ param = p;
+ plane_count = x265_cli_csps[param->internalCsp].planes;
+ bisUpdated = false;
+ m_isalloc = false;
+
+ yuv_hist = edge_hist = NULL;
+ plane_sizes = plane_heights = plane_widths = NULL;
+ edgePic = edgeThetaPic = NULL;
+
+ plane_sizes = X265_MALLOC(int32_t, plane_count);
+ plane_heights = X265_MALLOC(int32_t, plane_count);
+ plane_widths = X265_MALLOC(int32_t, plane_count);
+
+ if (!plane_sizes || !plane_heights || !plane_widths) {
+ x265_log(param, X265_LOG_ERROR, "unable to allocate memory for plane dimensions\n");
+ m_isalloc &= false;
+ }
+ else {
+ memset(plane_sizes, 0, plane_count * sizeof(int32_t));
+ memset(plane_heights, 0, plane_count * sizeof(int32_t));
+ memset(plane_widths, 0, plane_count * sizeof(int32_t));
+ m_isalloc &= true;
+ }
+
+ yuv_hist = X265_MALLOC(Histogram, plane_count);
+ edge_hist = X265_MALLOC(Histogram, plane_count);
+
+ if (!yuv_hist || !edge_hist) {
+ m_isalloc &= false;
+ x265_log(param, X265_LOG_ERROR, "unable to allocate memory for histograms\n");
+ }
+
+ }
+
+ bool YuvHistogram::allocHistogramBuffers() {
+ //allocate memory for edge filter output and histograms
+ bool isalloc = true;
+
+ edgePic = X265_MALLOC(pixel*, plane_count);
+ edgeThetaPic = X265_MALLOC(pixel*, plane_count);
+
+ if (!edgePic || !edgeThetaPic) {
+ isalloc &= false;
+ x265_log(param, X265_LOG_ERROR, "unable to allocate memory for edge buffers\n");
+ return isalloc;
+ }
+
+ for (int i = 0; i < plane_count; i++) {
+
+ edgePic[i] = edgeThetaPic[i] = NULL;
+ edgePic[i] = X265_MALLOC(pixel, plane_sizes[i]);
+ edgeThetaPic[i] = X265_MALLOC(pixel, plane_sizes[i]);
+
+ if (edgePic[i] && edgeThetaPic[i]) {
+ memset(edgePic[i], 0, plane_sizes[i] * sizeof(pixel));
+ memset(edgeThetaPic[i], 0, plane_sizes[i] * sizeof(pixel));
+ isalloc &= true;
+ }
+ else
+ isalloc &= false;
+ }
+ return isalloc;
+ }
+
+ YuvHistogram::~YuvHistogram()
+ {
+ freeHistogramBuffers(); //change implementation based on allocation changes
+ }
+
+ YuvHistogram::YuvHistogram(YuvHistogram const& hist) {
+
+ maxuv_hist = hist.maxuv_hist;
+ plane_count = hist.plane_count;
+ bisUpdated = hist.bisUpdated;
+ param = hist.param;
+ memcpy(plane_sizes, hist.plane_sizes, plane_count * sizeof(int32_t));
+ memcpy(plane_heights, hist.plane_heights, plane_count * sizeof(int32_t));
+ memcpy(plane_widths, hist.plane_widths, plane_count * sizeof(int32_t));
+ memcpy(yuv_hist, hist.yuv_hist, plane_count * sizeof(Histogram));
+ memcpy(edge_hist, hist.edge_hist, plane_count * sizeof(Histogram));
+
+ if (!m_isalloc) {
+ m_isalloc = false;
+ m_isalloc = allocHistogramBuffers();
+ }
+
+ if (m_isalloc) {
+ for (int i = 0; i < plane_count; i++) {
+ if (edgePic[i] && edgeThetaPic[i]) {
+ memcpy(edgePic[i], hist.edgePic[i], plane_sizes[i] * sizeof(pixel));
+ memcpy(edgeThetaPic[i], hist.edgeThetaPic[i], plane_sizes[i] * sizeof(pixel));
+ }
+ }
+ }
+
+ }
+
+ YuvHistogram & YuvHistogram ::operator=(const YuvHistogram & copy_hist)
+ {
+ maxuv_hist = copy_hist.maxuv_hist;
+ plane_count = copy_hist.plane_count;
+ bisUpdated = copy_hist.bisUpdated;
+ param = copy_hist.param;
+ memcpy(plane_sizes, copy_hist.plane_sizes, plane_count * sizeof(int32_t));
+ memcpy(plane_heights, copy_hist.plane_heights, plane_count * sizeof(int32_t));
+ memcpy(plane_widths, copy_hist.plane_widths, plane_count * sizeof(int32_t));
+ memcpy(yuv_hist, copy_hist.yuv_hist, plane_count * sizeof(Histogram));
+ memcpy(edge_hist, copy_hist.edge_hist, plane_count * sizeof(Histogram));
+
+ if (!m_isalloc) {
+ m_isalloc = false;
+ m_isalloc = allocHistogramBuffers();
+ }
+
+ if (m_isalloc) {
+ for (int i = 0; i < plane_count; i++) {
+ if (edgePic[i] && edgeThetaPic[i]) {
+ memcpy(edgePic[i], copy_hist.edgePic[i], plane_sizes[i] * sizeof(pixel));
+ memcpy(edgeThetaPic[i], copy_hist.edgeThetaPic[i], plane_sizes[i] * sizeof(pixel));
+ }
+ }
+ }
+
+ return *this;
+ }
+
+ void YuvHistogram::initFrameDimensions(x265_picture & pic) {
+
+ for (int i = 0; i < plane_count; i++) {
+ plane_widths[i] = pic.width;
+ plane_heights[i] = pic.height >> x265_cli_csps[pic.colorSpace].height[i];
+ plane_sizes[i] = plane_widths[i] * plane_heights[i];
+ }
+ }
+
+ void YuvHistogram::freeHistogramBuffers() {
+ //de allocate memory for histograms and edge filtered output
+ if (edgePic && edgeThetaPic) {
+ for (int i = 0; i < plane_count; i++) {
+ if (edgePic[i] && edgeThetaPic[i]) {
+ X265_FREE_ZERO(edgePic[i]);
+ X265_FREE_ZERO(edgeThetaPic[i]);
+ }
+ }
+ X265_FREE_ZERO(edgePic);
+ X265_FREE_ZERO(edgeThetaPic);
+ }
+
+ if (plane_sizes && plane_heights && plane_widths) {
+ X265_FREE_ZERO(plane_sizes);
+ X265_FREE_ZERO(plane_heights);
+ X265_FREE_ZERO(plane_widths);
+ }
+
+ if (yuv_hist && edge_hist) {
+ X265_FREE_ZERO(yuv_hist);
+ X265_FREE_ZERO(edge_hist);
+ }
+ }
+
+ bool YuvHistogram::edgeFilter(x265_picture *frame) {
+
+ if (!m_isalloc) {
+ initFrameDimensions(*frame);
+ m_isalloc = allocHistogramBuffers();
+ }
+
+ if (m_isalloc) {
+ for (int idx = 0; idx < EDGE_PLANE_COUNT; idx++) {
+
+ memset(edgePic[idx], 0, sizeof(pixel) * plane_sizes[idx]);
+ memset(edgeThetaPic[idx], 0, sizeof(pixel) * plane_sizes[idx]);
+
+ pixel *src = (pixel*)frame->planes[idx];
+ pixel *edge_pic = edgePic[idx];
+ pixel *ref_pic = src;
+ pixel *edge_theta = edgeThetaPic[idx];
+
+ assert(edge_pic != NULL);
+ assert(ref_pic != NULL);
+ memcpy(edge_pic, src, plane_sizes[idx] * sizeof(pixel));
+ memcpy(ref_pic, src, plane_sizes[idx] * sizeof(pixel));
+
+ computeEdge(edge_pic, ref_pic, edge_theta,plane_widths[idx], plane_heights[idx], plane_widths[idx]);
+ }
+ return true;
+ }
+ else {
+ return false;
+ }
+ }
+
+ bool YuvHistogram::computeHistograms(x265_picture &cur_frame) {
+
+ bool bsuccess = false;
+ bsuccess = computeLumaEdgeHistogram(cur_frame);
+ if (bsuccess) {
+ if (plane_count > 1) {
+ bsuccess &= computeChromaHistogram(cur_frame);
+ }
+ return bsuccess;
+ }
+ else {
+ return bsuccess;
+ }
+
+ }
+
+ bool YuvHistogram::computeLumaEdgeHistogram(x265_picture &frame) {
+
+ pixel pixel_val = 0;
+
+ memset(edge_hist[0].frequency_distribution, 0, HISTOGRAM_SIZE * sizeof(int32_t));
+
+ int size = frame.height*(frame.stride[0] >> SHIFT);
+
+ for (int i = 0; i < size; i++) {
+ pixel_val = edgePic[0][i];
+ edge_hist[0].frequency_distribution[pixel_val]++;
+ }
+ return true;
+ }
+
+ bool YuvHistogram::computeChromaHistogram(x265_picture &frame) {
+ /*u hist calculation*/
+ pixel pixel_val = 0;
+ int32_t pixel_ucount = 0, pixel_vcount = 0;
+
+ int u_height = (frame.height >> x265_cli_csps[frame.colorSpace].height[1]);
+ int size = u_height * (frame.stride[1] >> SHIFT);
+ memset(yuv_hist[1].frequency_distribution, 0, HISTOGRAM_SIZE * sizeof(int32_t));
+
+ for (int i = 0; i < size; i++) {
+ pixel_val = *((pixel *)frame.planes[1] + i);
+ yuv_hist[1].frequency_distribution[pixel_val]++;
+ pixel_ucount++;
+ }
+
+ /*v hist calculation for independent uv planes */
+
+ if (plane_count == 3) {
+ pixel_val = 0;
+ int v_height = (frame.height >> x265_cli_csps[frame.colorSpace].height[2]);
+ size = v_height * (frame.stride[2] >> SHIFT);
+ memset(yuv_hist[2].frequency_distribution, 0, HISTOGRAM_SIZE * sizeof(int32_t));
+
+ for (int i = 0; i < size; i++) {
+ pixel_val = *((pixel *)frame.planes[2] + i);
+ yuv_hist[2].frequency_distribution[pixel_val]++;
+ pixel_vcount++;
+ }
+
+ for (int i = 0; i < HISTOGRAM_SIZE; i++) {
+ maxuv_hist.frequency_distribution[i] = max(yuv_hist[1].frequency_distribution[i], yuv_hist[2].frequency_distribution[i]);
+ }
+
+ }
+ else {
+ maxuv_hist = yuv_hist[1]; //for two planes scenario
+ }
+
+ return true;
+ }
+
+ bool YuvHistogram::isUpdated() {
+ return bisUpdated;
+ }
+
+ void YuvHistogram::setUpdateFlag(bool flag) {
+ bisUpdated = flag;
+ }
+
+ bool YuvHistogram::getUpdateFlag() {
+ return bisUpdated;
+ }
+
+ SadYuv::SadYuv() { }
+
+ void SadYuv::initSadYuv(int planecount) {
+ this->plane_count = planecount;
+ sad_yuv = NULL;
+ psad_yuv = NULL;
+ sad_yuv = X265_MALLOC(int32_t, plane_count);
+ psad_yuv = X265_MALLOC(double, plane_count);
+ if (sad_yuv && psad_yuv) {
+ memset(sad_yuv, 0, plane_count * sizeof(int32_t));
+ memset(psad_yuv, 0, sizeof(double) * plane_count);
+ }
+ }
+
+ SadYuv & SadYuv::operator=(SadYuv const& sad_val) {
+ this->plane_count = sad_val.plane_count;
+ if (!sad_yuv && !psad_yuv) {
+ sad_yuv = NULL;
+ psad_yuv = NULL;
+ sad_yuv = X265_MALLOC(int32_t, plane_count);
+ psad_yuv = X265_MALLOC(double, plane_count);
+ if (sad_yuv && psad_yuv) {
+ memcpy(sad_yuv, sad_val.sad_yuv, plane_count * sizeof(int32_t));
+ memcpy(psad_yuv, sad_val.psad_yuv, sizeof(double) * plane_count);
+ }
+ }
+ else {
+ if (sad_yuv)
+ memcpy(sad_yuv, sad_val.sad_yuv, plane_count * sizeof(int32_t));
+ if (psad_yuv)
+ memcpy(psad_yuv, sad_val.psad_yuv, sizeof(double) * plane_count);
+ }
+ return *this;
+ }
+
+ SadYuv::~SadYuv() {
+ if (sad_yuv && psad_yuv) {
+ X265_FREE(sad_yuv);
+ X265_FREE(psad_yuv);
+ }
+ }
+
+ int sad_stats::frames_scanned=0;
+ int sad_stats::line_number=0;
+
+ sad_stats::sad_stats(int planecount, double threshold) {
+ this->plane_count = planecount;
+ calculateThresholds(threshold);
+ allocateBuffers();
+ }
+
+ void sad_stats::calculateThresholds(double threshold) {
+ edge_hist_threshold = threshold;
+ strength_factor = 2.0;
+ chroma_hist_threshold = threshold * 10.0;
+ scaled_edge_threshold = edge_hist_threshold * strength_factor;
+ scaled_chroma_threshold = chroma_hist_threshold * strength_factor;
+ }
+
+ void sad_stats::init() {
+ bscene_cut = NULL;
+ bdrop_frame = NULL;
+ sad_vals = NULL;
+ maxuv_sad_vals = NULL;
+ edge_sad_vals = NULL;
+ prev_hist = NULL;
+ }
+
+ sad_stats::~sad_stats() {
+ releaseBuffers();
+ }
+
+ void sad_stats::allocateBuffers() {
+
+ init();
+ sad_vals = new SadYuv[DUP_BUFFER]();
+ maxuv_sad_vals = new SadYuv[DUP_BUFFER]();
+ edge_sad_vals = new SadYuv[DUP_BUFFER]();
+ prev_hist = new YuvHistogram();
+ prev_hist->initHistograms(plane_count);
+
+ for (int i = 0; i < DUP_BUFFER; i++) {
+ sad_vals[i].initSadYuv(plane_count);
+ maxuv_sad_vals[i].initSadYuv(plane_count);
+ edge_sad_vals[i].initSadYuv(plane_count);
+ }
+
+ bscene_cut = new bool[DUP_BUFFER];
+ bdrop_frame = new bool[DUP_BUFFER];
+
+ if (!sad_vals || !maxuv_sad_vals || !edge_sad_vals || !bscene_cut || !bdrop_frame) {
+ x265_log(NULL, X265_LOG_ERROR, "Heap Error !");
+ exit(101);
+ }
+ else {
+ memset(bscene_cut, false, 2 * sizeof(bool));
+ memset(bdrop_frame, false, 2 * sizeof(bool));
+ }
+
+ }
+
+ void sad_stats::releaseBuffers() {
+ if (sad_vals && maxuv_sad_vals && edge_sad_vals && bscene_cut && bdrop_frame && prev_hist) {
+ delete[] sad_vals;
+ delete[] maxuv_sad_vals;
+ delete[] edge_sad_vals;
+ delete[] bscene_cut;
+ delete[] bdrop_frame;
+ delete prev_hist;
+ }
+ }
+
+ bool sad_stats::computeSadValue(YuvHistogram *input_frames, int32_t* plane_sizes) {
+
+ int32_t *yuv_sad_val = NULL, *edge_sad_val = NULL, *maxuv_sad_val = NULL;
+
+ double *maxuv_normalized_sad = NULL, *yuv_norm_sad = NULL, *edge_normalized_sads = NULL;
+
+ YuvHistogram * ref_hist = NULL, *cur_hist = NULL;
+
+ /*inorder to process frames as per poc's updated by frame duplication */
+ if (frames_scanned > 0) {
+
+ if (!input_frames[0].isUpdated() && input_frames[1].isUpdated()) {
+ ref_hist = prev_hist;
+ cur_hist = input_frames + 1;
+
+ yuv_sad_val = sad_vals[1].sad_yuv,
+ edge_sad_val = edge_sad_vals[1].sad_yuv,
+ maxuv_sad_val = maxuv_sad_vals[1].sad_yuv;
+ maxuv_normalized_sad = maxuv_sad_vals[1].psad_yuv,
+ yuv_norm_sad = sad_vals[1].psad_yuv,
+ edge_normalized_sads = edge_sad_vals[1].psad_yuv;
+ input_frames[1].setUpdateFlag(false);
+
+ }
+ else if (input_frames[0].isUpdated() && input_frames[1].isUpdated()) {
+ ref_hist = prev_hist;
+ cur_hist = input_frames;
+
+ yuv_sad_val = sad_vals[0].sad_yuv,
+ edge_sad_val = edge_sad_vals[0].sad_yuv,
+ maxuv_sad_val = maxuv_sad_vals[0].sad_yuv;
+ maxuv_normalized_sad = maxuv_sad_vals[0].psad_yuv,
+ yuv_norm_sad = sad_vals[0].psad_yuv,
+ edge_normalized_sads = edge_sad_vals[0].psad_yuv;
+ input_frames[0].setUpdateFlag(false);
+
+ }
+ else if (input_frames[0].isUpdated() && !input_frames[1].isUpdated()) {
+ ref_hist = prev_hist;
+ cur_hist = input_frames;
+
+ yuv_sad_val = sad_vals[0].sad_yuv,
+ edge_sad_val = edge_sad_vals[0].sad_yuv,
+ maxuv_sad_val = maxuv_sad_vals[0].sad_yuv;
+ maxuv_normalized_sad = maxuv_sad_vals[0].psad_yuv,
+ yuv_norm_sad = sad_vals[0].psad_yuv,
+ edge_normalized_sads = edge_sad_vals[0].psad_yuv;
+ input_frames[0].setUpdateFlag(false);
+ }
+ else {
+ return true;
+ }
+ }
+ else {
+ cur_hist = input_frames;
+
+ yuv_sad_val = sad_vals[0].sad_yuv,
+ edge_sad_val = edge_sad_vals[0].sad_yuv,
+ maxuv_sad_val = maxuv_sad_vals[0].sad_yuv;
+
+ maxuv_normalized_sad = maxuv_sad_vals[0].psad_yuv,
+ yuv_norm_sad = sad_vals[0].psad_yuv,
+ edge_normalized_sads = edge_sad_vals[0].psad_yuv;
+ input_frames[0].setUpdateFlag(false);
+ }
+
+ if (frames_scanned == 0) { //first frame is scenecut by default no sad computation for the same.
+
+ maxuv_sad_val[0] = 0;
+ maxuv_normalized_sad[0] = 0.0;
+ memset(yuv_sad_val, 0 , plane_count * sizeof(int32_t));
+ memset(edge_sad_val, 0, plane_count * sizeof(int32_t));
+ memset(edge_normalized_sads, 0, plane_count * sizeof(double));
+ memset(yuv_norm_sad, 0, plane_count * sizeof(double));
+
+ }
+ else {
+ int32_t freq_diff[3];
+ int32_t maxuv_freq_diff[1];
+ int32_t edge_freq_diff[3];
+ double color_probability_diff[3], edge_probability_diff[3];
+
+ memset(yuv_sad_val, 0, plane_count*sizeof(int32_t));
+ memset(edge_sad_val, 0, plane_count*sizeof(int32_t));
+
+ memset(yuv_norm_sad, 0, plane_count * sizeof(double));
+ memset(edge_normalized_sads, 0, plane_count * sizeof(double));
+ memset(color_probability_diff, 0, plane_count * sizeof(double));
+ memset(edge_probability_diff, 0, plane_count * sizeof(double));
+
+ maxuv_normalized_sad[0] = 0.0;
+ maxuv_sad_val[0] = 0;
+
+ memset(freq_diff, 0, 3 * sizeof(int32_t));
+ memset(maxuv_freq_diff, 0, sizeof(int32_t));
+ memset(edge_freq_diff, 0, 3 * sizeof(int32_t));
+
+ for (int i = 0; i < plane_count; i++) {
+ {
+ for (int j = 0; j < HISTOGRAM_SIZE; j++) {
+
+ if (i == 0 && plane_count >= 1) {
+ maxuv_freq_diff[i] = (abs(cur_hist->maxuv_hist.frequency_distribution[j] - ref_hist->maxuv_hist.frequency_distribution[j]));
+ maxuv_sad_val[i] += maxuv_freq_diff[i];
+ maxuv_normalized_sad[i] += (double)maxuv_freq_diff[i] / plane_sizes[i];
+ edge_freq_diff[i] = abs(cur_hist->edge_hist[i].frequency_distribution[j] - ref_hist->edge_hist[i].frequency_distribution[j]);
+ edge_probability_diff[i] = double(edge_freq_diff[i]) / plane_sizes[i];
+ edge_sad_val[i] += edge_freq_diff[i];
+ edge_normalized_sads[i] += edge_probability_diff[i];
+ }
+ else {
+ freq_diff[i] = abs(cur_hist->yuv_hist[i].frequency_distribution[j] - ref_hist->yuv_hist[i].frequency_distribution[j]);
+ color_probability_diff[i] = (double)freq_diff[i] / plane_sizes[i];
+ yuv_sad_val[i] += freq_diff[i];
+ yuv_norm_sad[i] += color_probability_diff[i];
+ }
+
+ }
+
+ }
+ }
+
+ }
+
+ *prev_hist = *cur_hist;
+
+ frames_scanned++;
+
+ return true;
+ }
+
+ void sad_stats::findSceneCuts(x265_picture * picList, bool& bdup) {
+
+ if (frames_scanned == 1) {
+ //for first frame
+ bscene_cut[0] = true;
+ bdrop_frame[0] = false;
+ picList->analysisData.bScenecut = (int)getSceneCutflag(0);
+ bdup = getDropflag(0);
+ picList->analysisData.edgeSadValue = edge_sad_vals[0].psad_yuv[0];
+ picList->analysisData.chromaSadValue = maxuv_sad_vals[0].psad_yuv[0];
+ }
+ else {
+ bscene_cut[1] = bdrop_frame[1] = false;
+ if (edge_sad_vals[1].psad_yuv[0] == 0) {
+ bdrop_frame[1] = true;
+ }
+ else if (edge_sad_vals[1].psad_yuv[0] > edge_hist_threshold || maxuv_sad_vals[1].psad_yuv[0] >= chroma_hist_threshold) {
+ bscene_cut[1] = true;
+ bdrop_frame[1] = false;
+ }
+ else if (edge_sad_vals[1].psad_yuv[0] > scaled_edge_threshold || maxuv_sad_vals[1].psad_yuv[0] >= scaled_chroma_threshold) {
+ bscene_cut[1] = true;
+ bdrop_frame[1] = false;
+ }
+ picList->analysisData.bScenecut = (int)getSceneCutflag(1);
+ bdup = getDropflag(1);
+ picList->analysisData.edgeSadValue = edge_sad_vals[1].psad_yuv[0];
+ picList->analysisData.chromaSadValue = maxuv_sad_vals[1].psad_yuv[0];
+ }
+ }
+
+ bool sad_stats::getDropflag(int i) {
+ return bdrop_frame[i];
+ }
+
+ bool sad_stats::getSceneCutflag(int i) {
+ return bscene_cut[i];
+ }
+
+}
\ No newline at end of file
diff -r 37648fca915b -r deaecadc4306 source/common/scenecut.h
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/source/common/scenecut.h Tue Oct 15 12:16:17 2019 +0530
@@ -0,0 +1,147 @@
+#ifndef SCENECUT_H
+#define SCENECUT_H
+
+#include <string>
+#include <iostream>
+#include <sstream>
+#include <vector>
+#include <algorithm>
+#include <math.h>
+
+#include "yuv.h"
+#include "common.h"
+
+#ifdef HIGH_BIT_DEPTH
+#define edge_threshold 1023.0
+#define whitePixel 1023.0
+#else
+#define edge_threshold 255.0
+#define pixel whitePixel 255.0
+#endif
+
+using namespace std;
+
+namespace X265_NS {
+
+ class Histogram {
+
+ public:
+ int32_t frequency_distribution[HISTOGRAM_SIZE];
+
+ Histogram();
+
+ Histogram(Histogram const& hist);
+
+ Histogram & operator=(Histogram const& hist);
+
+ ~Histogram();
+
+ };
+
+ class YuvHistogram {
+ public:
+ Histogram *yuv_hist;
+ Histogram *edge_hist;
+ int32_t *plane_sizes;
+ int32_t *plane_heights;
+ int32_t *plane_widths;
+
+ Histogram maxuv_hist;
+ int32_t plane_count;
+ bool bisUpdated;
+
+ pixel** edgePic;
+ pixel** edgeThetaPic;
+
+ x265_param * param; /*for handling various color spaces*/
+ bool m_isalloc;
+
+ YuvHistogram();
+
+ void initHistograms(int32_t plane_count);
+
+ void initHistograms(x265_param *p);
+
+ bool allocHistogramBuffers();
+
+ YuvHistogram(YuvHistogram const& hist);
+
+ YuvHistogram & operator=(const YuvHistogram & copy_hist);
+
+ ~YuvHistogram();
+
+ void initFrameDimensions(x265_picture & pic);
+
+ void freeHistogramBuffers();
+
+ bool edgeFilter(x265_picture *frame);
+
+ bool computeHistograms(x265_picture &cur_frame);
+
+ bool computeLumaEdgeHistogram(x265_picture &frame);
+
+ bool computeChromaHistogram(x265_picture &frame);
+
+ bool isUpdated();
+
+ void setUpdateFlag(bool flag);
+
+ bool getUpdateFlag();
+
+ };
+
+ struct SadYuv {
+ int32_t *sad_yuv;
+ double *psad_yuv;
+ int plane_count;
+ ~SadYuv();
+ SadYuv();
+ void initSadYuv(int plane_count);
+ SadYuv & operator=(SadYuv const& sad_val);
+};
+
+ class sad_stats {
+
+ bool *bscene_cut;
+ bool *bdrop_frame;
+ SadYuv * sad_vals;
+ SadYuv * maxuv_sad_vals;
+ SadYuv * edge_sad_vals;
+ int plane_count;
+ static int line_number;
+ static int frames_scanned;
+ YuvHistogram *prev_hist;
+ double edge_hist_threshold;
+ double chroma_hist_threshold;
+ double scaled_chroma_threshold;
+ double scaled_edge_threshold;
+ double strength_factor;
+
+ public:
+ sad_stats(int plane_count, double threshold);
+
+ ~sad_stats();
+
+ void init();
+
+ void allocateBuffers();
+
+ void releaseBuffers();
+
+ void calculateThresholds(double threshold);
+
+ bool computeSadValue(YuvHistogram *frames, int32_t* plane_sizes);
+
+ void findSceneCuts(x265_picture * piclist,bool & bdup);
+
+ bool getDropflag(int i);
+
+ bool getSceneCutflag(int i);
+
+ };
+
+void computeEdge(pixel * edgePic, pixel *refPic, pixel * edgeTheta, intptr_t stride, int height, int width);
+
+}
+
+#endif
diff -r 37648fca915b -r deaecadc4306 source/encoder/api.cpp
--- a/source/encoder/api.cpp Fri Oct 11 12:45:52 2019 +0530
+++ b/source/encoder/api.cpp Tue Oct 15 12:16:17 2019 +0530
@@ -31,6 +31,7 @@
#include "nal.h"
#include "bitcost.h"
#include "svt.h"
+#include "scenecut.h"
#if ENABLE_LIBVMAF
#include "libvmaf.h"
@@ -117,7 +118,10 @@
x265_log(param, X265_LOG_INFO, "build info %s\n", PFX(build_info_str));
encoder = new Encoder;
-
+ encoder->m_sad_stats = new sad_stats(x265_cli_csps[p->internalCsp].planes,param->edgeTransitionThreshold);
+ encoder->m_hist_of_adj_frames = new YuvHistogram[2];
+ encoder->m_hist_of_adj_frames[0].initHistograms(p);
+ encoder->m_hist_of_adj_frames[1].initHistograms(p);
#ifdef SVT_HEVC
if (param->bEnableSvtHevc)
@@ -809,6 +813,7 @@
CHECKED_MALLOC_ZERO(interData->ref, int32_t, analysis->numCUsInFrame * X265_MAX_PRED_MODE_PER_CTU * numDir);
}
analysis->interData = interData;
+ analysis->bScenecut = false;
return;
@@ -924,6 +929,7 @@
pic->rpu.payloadSize = 0;
pic->rpu.payload = NULL;
pic->picStruct = 0;
+ pic->bufUpdated = false;
if ((param->analysisSave || param->analysisLoad) || (param->bAnalysisType == AVC_INFO))
{
@@ -933,7 +939,9 @@
uint32_t numCUsInFrame = widthInCU * heightInCU;
pic->analysisData.numCUsInFrame = numCUsInFrame;
pic->analysisData.numPartitions = param->num4x4Partitions;
+ pic->analysisData.bScenecut = false;
}
+
}
void x265_picture_free(x265_picture *p)
@@ -955,7 +963,8 @@
{
if (param && param->rc.zonefileCount) {
for (int i = 0; i < param->rc.zonefileCount; i++)
- x265_free(param->rc.zones[i].zoneParam);
+ if(param->rc.zones[i].zoneParam)
+ x265_free(param->rc.zones[i].zoneParam);
}
if (param && (param->rc.zoneCount || param->rc.zonefileCount))
x265_free(param->rc.zones);
diff -r 37648fca915b -r deaecadc4306 source/encoder/encoder.cpp
--- a/source/encoder/encoder.cpp Fri Oct 11 12:45:52 2019 +0530
+++ b/source/encoder/encoder.cpp Tue Oct 15 12:16:17 2019 +0530
@@ -119,6 +119,9 @@
m_frameEncoder[i] = NULL;
for (uint32_t i = 0; i < DUP_BUFFER; i++)
m_dupBuffer[i] = NULL;
+
+ m_hist_of_adj_frames = NULL;
+ m_sad_stats = NULL;
MotionEstimate::initScales();
#if ENABLE_HDR10_PLUS
@@ -162,7 +165,9 @@
int rows = (p->sourceHeight + p->maxCUSize - 1) >> g_log2Size[p->maxCUSize];
int cols = (p->sourceWidth + p->maxCUSize - 1) >> g_log2Size[p->maxCUSize];
- if (m_param->bEnableFrameDuplication)
+
+
+ if (m_param->bEnableFrameDuplication || m_param->bHistbasedScenecut)
{
size_t framesize = 0;
int pixelbytes = p->sourceBitDepth > 8 ? 2 : 1;
@@ -184,6 +189,7 @@
m_dupBuffer[i]->dupPlane = NULL;
m_dupBuffer[i]->dupPlane = X265_MALLOC(char, framesize);
m_dupBuffer[i]->dupPic->planes[0] = m_dupBuffer[i]->dupPlane;
+ m_dupBuffer[i]->bufUpdated = false;
m_dupBuffer[i]->bOccupied = false;
m_dupBuffer[i]->bDup = false;
}
@@ -820,7 +826,7 @@
m_exportedPic = NULL;
}
- if (m_param->bEnableFrameDuplication)
+ if (m_param->bEnableFrameDuplication || m_param->bHistbasedScenecut)
{
for (uint32_t i = 0; i < DUP_BUFFER; i++)
{
@@ -1280,6 +1286,33 @@
return psnrWeight = (psnrY * 6 + psnrU + psnrV) / 8;
}
+void Encoder::updateSceneCutAndFrameDuplicateFlags() {
+ /* SCD computation and drop flag*/
+ for (int i = 0; i < DUP_BUFFER; i++) {
+ if (m_dupBuffer[i]->bufUpdated) {
+ m_hist_of_adj_frames[i].setUpdateFlag(true);
+ m_hist_of_adj_frames[i].edgeFilter(m_dupBuffer[i]->dupPic);
+ m_hist_of_adj_frames[i].computeHistograms(*m_dupBuffer[i]->dupPic);
+ m_sad_stats->computeSadValue(m_hist_of_adj_frames, m_hist_of_adj_frames->plane_sizes);
+ m_sad_stats->findSceneCuts(m_dupBuffer[i]->dupPic, m_dupBuffer[i]->bDup);
+
+ if (m_dupBuffer[i]->dupPic->analysisData.bScenecut) {
+ x265_log(m_param, X265_LOG_DEBUG, "scene cut at %d edge hist sad: %0.4lf maxuv hist sad: %0.4lf\n",
+ m_dupBuffer[i]->dupPic->poc,m_dupBuffer[i]->dupPic->analysisData.edgeSadValue,m_dupBuffer[i]->dupPic->analysisData.chromaSadValue);
+ }
+
+ if (m_dupBuffer[1]->bufUpdated)
+ m_hist_of_adj_frames[0] = m_hist_of_adj_frames[1];
+ }
+ }
+
+ }
+
+/* TBD
+- to be updated for missing parameters in case of re-use else where and improvised to copy constructor / assignment operator of x265 picture data structure.
+- benefits avoid function and use language features appropriately.
+*/
+
void Encoder::copyPicture(x265_picture *dest, const x265_picture *src)
{
dest->poc = src->poc;
@@ -1299,6 +1332,25 @@
memcpy(dest->planes[0], src->planes[0], src->framesize * sizeof(char));
dest->planes[1] = (char*)dest->planes[0] + src->stride[0] * src->height;
dest->planes[2] = (char*)dest->planes[1] + src->stride[1] * (src->height >> x265_cli_csps[src->colorSpace].height[1]);
+ memcpy(&dest->analysisData, &src->analysisData, sizeof(src->analysisData));
+
+}
+
+void Encoder::setPictureFlags(int idx) {
+ m_dupBuffer[idx]->bOccupied = true;
+ m_dupBuffer[idx]->bufUpdated = true;
+ m_dupBuffer[idx]->bDup = false;
+}
+
+void Encoder::unsetPictureFlags(int idx) {
+ if (idx == 1) {
+ m_dupBuffer[idx]->bOccupied = false;
+ m_dupBuffer[idx]->bufUpdated = false;
+ m_dupBuffer[idx]->bDup = false;
+ }
+ else if (idx == 0) {
+ m_dupBuffer[idx]->bufUpdated = false;
+ }
}
/**
@@ -1327,7 +1379,9 @@
const x265_picture* inputPic = NULL;
static int written = 0, read = 0;
bool dontRead = false;
-
+ bool isScenecutEnabled = m_param->bHistbasedScenecut;
+ bool dropflag = false;
+
if (m_exportedPic)
{
if (!m_param->bUseAnalysisFile && m_param->analysisSave)
@@ -1338,7 +1392,7 @@
}
if ((pic_in && (!m_param->chunkEnd || (m_encodedFrameNum < m_param->chunkEnd))) || (m_param->bEnableFrameDuplication && !pic_in && (read < written)))
{
- if ((m_param->bEnableFrameDuplication && !pic_in && (read < written)))
+ if ((m_param->bEnableFrameDuplication && !pic_in && (read < written)) || (isScenecutEnabled && !pic_in && (read < written)))
dontRead = true;
else
{
@@ -1361,7 +1415,7 @@
}
}
- if (m_param->bEnableFrameDuplication)
+ if (m_param->bEnableFrameDuplication || isScenecutEnabled )
{
double psnrWeight = 0;
@@ -1372,6 +1426,11 @@
copyPicture(m_dupBuffer[0]->dupPic, pic_in);
m_dupBuffer[0]->bOccupied = true;
written++;
+ if (m_param->bHistbasedScenecut) {
+ setPictureFlags(0);
+ updateSceneCutAndFrameDuplicateFlags();
+ unsetPictureFlags(0);
+ }
return 0;
}
else if (!m_dupBuffer[1]->bOccupied)
@@ -1379,31 +1438,54 @@
copyPicture(m_dupBuffer[1]->dupPic, pic_in);
m_dupBuffer[1]->bOccupied = true;
written++;
+ if (m_param->bHistbasedScenecut) {
+ setPictureFlags(1);
+ updateSceneCutAndFrameDuplicateFlags();
+ unsetPictureFlags(1);
+ }
}
- psnrWeight = ComputePSNR(m_dupBuffer[0]->dupPic, m_dupBuffer[1]->dupPic, m_param);
-
- if (psnrWeight >= m_param->dupThreshold)
- {
- if (m_dupBuffer[0]->bDup)
- {
- m_dupBuffer[0]->dupPic->picStruct = tripling;
- m_dupBuffer[0]->bDup = false;
- read++;
+ if (m_param->bEnableFrameDuplication && m_param->bHistbasedScenecut) {
+ if (m_dupBuffer[1]->bDup == false && m_dupBuffer[1]->dupPic->analysisData.bScenecut == false) {
+ psnrWeight = ComputePSNR(m_dupBuffer[0]->dupPic, m_dupBuffer[1]->dupPic, m_param);
+ if (psnrWeight >= m_param->dupThreshold)
+ dropflag = true;
}
- else
- {
- m_dupBuffer[0]->dupPic->picStruct = doubling;
- m_dupBuffer[0]->bDup = true;
- m_dupBuffer[1]->bOccupied = false;
- read++;
- return 0;
+ else {
+ dropflag = true;
}
}
- else if (m_dupBuffer[0]->bDup)
+ else if (m_param->bEnableFrameDuplication) {
+ psnrWeight = ComputePSNR(m_dupBuffer[0]->dupPic, m_dupBuffer[1]->dupPic, m_param);
+ if (psnrWeight >= m_param->dupThreshold)
+ dropflag = true;
+ }
+
+ if (m_param->bEnableFrameDuplication)
+ {
+ if (dropflag)
+ {
+ if (m_dupBuffer[0]->bDup)
+ {
+ m_dupBuffer[0]->dupPic->picStruct = tripling;
+ m_dupBuffer[0]->bDup = false;
+ read++;
+ }
+ else
+ {
+ m_dupBuffer[0]->dupPic->picStruct = doubling;
+ m_dupBuffer[0]->bDup = true;
+ m_dupBuffer[1]->bOccupied = false;
+ read++;
+ return 0;
+ }
+ }
+ else if (m_dupBuffer[0]->bDup)
m_dupBuffer[0]->bDup = false;
- else
- m_dupBuffer[0]->dupPic->picStruct = 0;
+ else
+ m_dupBuffer[0]->dupPic->picStruct = 0;
+ }
+
}
if (read < written)
@@ -1485,7 +1567,10 @@
inFrame->m_poc = ++m_pocLast;
inFrame->m_userData = inputPic->userData;
- inFrame->m_pts = inputPic->pts;
+ inFrame->m_pts = inputPic->pts;
+ if (m_param->bHistbasedScenecut) {
+ inFrame->m_lowres.bScenecut = inputPic->analysisData.bScenecut;
+ }
inFrame->m_forceqp = inputPic->forceqp;
inFrame->m_param = (m_reconfigure || m_reconfigureRc) ? m_latestParam : m_param;
inFrame->m_picStruct = inputPic->picStruct;
@@ -1613,7 +1698,7 @@
m_param->bUseRcStats = 0;
}
- if (m_param->bEnableFrameDuplication && ((read < written) || (m_dupBuffer[0]->dupPic->picStruct == tripling && (read <= written))))
+ if ( (m_param->bEnableFrameDuplication || isScenecutEnabled) && ((read < written) || (m_dupBuffer[0]->dupPic->picStruct == tripling && (read <= written))))
{
if (m_dupBuffer[0]->dupPic->picStruct == tripling)
m_dupBuffer[0]->bOccupied = m_dupBuffer[1]->bOccupied = false;
@@ -3162,6 +3247,7 @@
* adaptive I frame placement */
p->keyframeMax = INT_MAX;
p->scenecutThreshold = 0;
+ p->bHistbasedScenecut = 0;
}
else if (p->keyframeMax <= 1)
{
@@ -3175,6 +3261,7 @@
p->lookaheadDepth = 0;
p->bframes = 0;
p->scenecutThreshold = 0;
+ p->bHistbasedScenecut = 0;
p->bFrameAdaptive = 0;
p->rc.cuTree = 0;
p->bEnableWeightedPred = 0;
@@ -3828,6 +3915,20 @@
m_param->searchMethod = m_param->hmeSearchMethod[2];
}
}
+
+ if (p->bHistbasedScenecut && p->scenecutThreshold) {
+ p->scenecutThreshold = 0;
+ p->bHistbasedScenecut = false;
+ x265_log(p, X265_LOG_WARNING, "Amibigious choice. disabling scene cut detection \n");
+ }
+ else if (p->scenecutThreshold && p->edgeTransitionThreshold != 0.01) {
+ x265_log(p, X265_LOG_WARNING, "using scenecut-bias %d for scene cut detection\n",p->scenecutBias);
+ }
+ else if (p->bHistbasedScenecut && p->edgeTransitionThreshold == 0.0) {
+ p->edgeTransitionThreshold = 0.01;
+ x265_log(p, X265_LOG_INFO, "using default threshold %.2lf for scene cut detection\n", p->edgeTransitionThreshold);
+ }
+
}
void Encoder::readAnalysisFile(x265_analysis_data* analysis, int curPoc, const x265_picture* picIn, int paramBytes)
diff -r 37648fca915b -r deaecadc4306 source/encoder/encoder.h
--- a/source/encoder/encoder.h Fri Oct 11 12:45:52 2019 +0530
+++ b/source/encoder/encoder.h Tue Oct 15 12:16:17 2019 +0530
@@ -32,6 +32,8 @@
#include "nal.h"
#include "framedata.h"
#include "svt.h"
+#include "scenecut.h"
+
#ifdef ENABLE_HDR10_PLUS
#include "dynamicHDR10/hdr10plus.h"
#endif
@@ -154,6 +156,9 @@
//Flag to check whether the picture has duplicated.
bool bDup;
+
+ bool bufUpdated;
+
};
@@ -195,6 +200,9 @@
ThreadPool* m_threadPool;
FrameEncoder* m_frameEncoder[X265_MAX_FRAME_THREADS];
+
+ YuvHistogram* m_hist_of_adj_frames;
+ sad_stats* m_sad_stats;
DPB* m_dpb;
Frame* m_exportedPic;
FILE* m_analysisFileIn;
@@ -279,6 +287,10 @@
if (m_prevTonemapPayload.payload != NULL)
X265_FREE(m_prevTonemapPayload.payload);
#endif
+ delete m_sad_stats;
+ m_sad_stats = NULL;
+ delete[] m_hist_of_adj_frames;
+ m_hist_of_adj_frames = NULL;
};
void create();
@@ -349,6 +361,12 @@
void copyPicture(x265_picture *dest, const x265_picture *src);
+ void unsetPictureFlags(int index);
+
+ void setPictureFlags(int index);
+
+ void updateSceneCutAndFrameDuplicateFlags();
+
void initRefIdx();
void analyseRefIdx(int *numRefIdx);
void updateRefIdx();
@@ -364,6 +382,7 @@
void initSPS(SPS *sps);
void initPPS(PPS *pps);
};
+
}
#endif // ifndef X265_ENCODER_H
diff -r 37648fca915b -r deaecadc4306 source/encoder/ratecontrol.cpp
--- a/source/encoder/ratecontrol.cpp Fri Oct 11 12:45:52 2019 +0530
+++ b/source/encoder/ratecontrol.cpp Tue Oct 15 12:16:17 2019 +0530
@@ -493,6 +493,7 @@
CMP_OPT_FIRST_PASS("open-gop", m_param->bOpenGOP);
CMP_OPT_FIRST_PASS(" keyint", m_param->keyframeMax);
CMP_OPT_FIRST_PASS("scenecut", m_param->scenecutThreshold);
+ CMP_OPT_FIRST_PASS("hist-threshold", m_param->edgeTransitionThreshold);
CMP_OPT_FIRST_PASS("intra-refresh", m_param->bIntraRefresh);
if (m_param->bMultiPassOptRPS)
{
@@ -1183,6 +1184,7 @@
m_param->rc.bStatRead = 0;
m_param->bFrameAdaptive = 0;
m_param->scenecutThreshold = 0;
+ m_param->bHistbasedScenecut = false;
m_param->rc.cuTree = 0;
if (m_param->bframes > 1)
m_param->bframes = 1;
@@ -2173,7 +2175,7 @@
if (m_isVbv && m_currentSatd > 0 && curFrame)
{
if (m_param->lookaheadDepth || m_param->rc.cuTree ||
- m_param->scenecutThreshold ||
+ (m_param->scenecutThreshold || m_param->bHistbasedScenecut) ||
(m_param->bFrameAdaptive && m_param->bframes))
{
/* Lookahead VBV: If lookahead is done, raise the quantizer as necessary
diff -r 37648fca915b -r deaecadc4306 source/encoder/slicetype.cpp
--- a/source/encoder/slicetype.cpp Fri Oct 11 12:45:52 2019 +0530
+++ b/source/encoder/slicetype.cpp Tue Oct 15 12:16:17 2019 +0530
@@ -30,6 +30,7 @@
#include "primitives.h"
#include "lowres.h"
#include "mv.h"
+#include "scenecut.h"
#include "slicetype.h"
#include "motion.h"
@@ -114,8 +115,8 @@
//Applying Gaussian filter on the picture
src = (pixel*)curFrame->m_fencPic->m_picOrg[0];
refPic = curFrame->m_gaussianPic + curFrame->m_fencPic->m_lumaMarginY * stride + curFrame->m_fencPic->m_lumaMarginX;
+ edgePic = pic1 + curFrame->m_fencPic->m_lumaMarginY * stride + curFrame->m_fencPic->m_lumaMarginX;
pixel pixelValue = 0;
-
for (int rowNum = 0; rowNum < height; rowNum++)
{
for (int colNum = 0; colNum < width; colNum++)
@@ -127,7 +128,8 @@
1 [4 9 12 9 4]
--- [5 12 15 12 5]
159 [4 9 12 9 4]
- [2 4 5 4 2]*/
+ [2 4 5 4 2]
+ */
const intptr_t rowOne = (rowNum - 2)*stride, colOne = colNum - 2;
const intptr_t rowTwo = (rowNum - 1)*stride, colTwo = colNum - 1;
@@ -145,52 +147,7 @@
}
}
}
-
-#if HIGH_BIT_DEPTH //10-bit build
- float threshold = 1023;
- pixel whitePixel = 1023;
-#else
- float threshold = 255;
- pixel whitePixel = 255;
-#endif
-#define PI 3.14159265
-
- float gradientH = 0, gradientV = 0, radians = 0, theta = 0;
- float gradientMagnitude = 0;
- pixel blackPixel = 0;
- edgePic = curFrame->m_edgePic + curFrame->m_fencPic->m_lumaMarginY * stride + curFrame->m_fencPic->m_lumaMarginX;
- //Applying Sobel filter on the gaussian filtered picture
- for (int rowNum = 0; rowNum < height; rowNum++)
- {
- for (int colNum = 0; colNum < width; colNum++)
- {
- edgeTheta[(rowNum*stride) + colNum] = 0;
- if ((rowNum != 0) && (colNum != 0) && (rowNum != height - 1) && (colNum != width - 1)) //Ignoring the border pixels of the picture
- {
- /*Horizontal and vertical gradients
- [ -3 0 3 ] [-3 -10 -3 ]
- gH = [ -10 0 10] gV = [ 0 0 0 ]
- [ -3 0 3 ] [ 3 10 3 ]*/
-
- const intptr_t rowOne = (rowNum - 1)*stride, colOne = colNum -1;
- const intptr_t rowTwo = rowNum * stride, colTwo = colNum;
- const intptr_t rowThree = (rowNum + 1)*stride, colThree = colNum + 1;
- const intptr_t index = (rowNum*stride) + colNum;
-
- gradientH = (float)(-3 * refPic[rowOne + colOne] + 3 * refPic[rowOne + colThree] - 10 * refPic[rowTwo + colOne] + 10 * refPic[rowTwo + colThree] - 3 * refPic[rowThree + colOne] + 3 * refPic[rowThree + colThree]);
- gradientV = (float)(-3 * refPic[rowOne + colOne] - 10 * refPic[rowOne + colTwo] - 3 * refPic[rowOne + colThree] + 3 * refPic[rowThree + colOne] + 10 * refPic[rowThree + colTwo] + 3 * refPic[rowThree + colThree]);
-
- gradientMagnitude = sqrtf(gradientH * gradientH + gradientV * gradientV);
- radians = atan2(gradientV, gradientH);
- theta = (float)((radians * 180) / PI);
- if (theta < 0)
- theta = 180 + theta;
- edgeTheta[(rowNum*stride) + colNum] = (pixel)theta;
-
- edgePic[index] = gradientMagnitude >= threshold ? whitePixel : blackPixel;
- }
- }
- }
+ computeEdge(edgePic, refPic, edgeTheta, stride, height, width);
}
//Find the angle of a block by averaging the pixel angles
@@ -1471,7 +1428,7 @@
if (m_lastNonB && !m_param->rc.bStatRead &&
((m_param->bFrameAdaptive && m_param->bframes) ||
- m_param->rc.cuTree || m_param->scenecutThreshold ||
+ m_param->rc.cuTree || m_param->scenecutThreshold || m_param->bHistbasedScenecut ||
(m_param->lookaheadDepth && m_param->rc.vbvBufferSize)))
{
slicetypeAnalyse(frames, false);
@@ -1962,10 +1919,15 @@
int numBFrames = 0;
int numAnalyzed = numFrames;
- bool isScenecut = scenecut(frames, 0, 1, true, origNumFrames);
+ bool isScenecut = false;
/* When scenecut threshold is set, use scenecut detection for I frame placements */
- if (m_param->scenecutThreshold && isScenecut)
+ if (m_param->scenecutThreshold)
+ isScenecut = scenecut(frames, 0, 1, true, origNumFrames);
+ else if (m_param->bHistbasedScenecut)
+ isScenecut = frames[1]->bScenecut;
+
+ if (isScenecut)
{
frames[1]->sliceType = X265_TYPE_I;
return;
@@ -1976,14 +1938,24 @@
m_extendGopBoundary = false;
for (int i = m_param->bframes + 1; i < origNumFrames; i += m_param->bframes + 1)
{
- scenecut(frames, i, i + 1, true, origNumFrames);
+ if (m_param->scenecutThreshold)
+ 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) )
- {
- m_extendGopBoundary = true;
- break;
- }
+ if (m_param->scenecutThreshold)
+ {
+ if (frames[j]->bScenecut && scenecutInternal(frames, j - 1, j, true))
+ {
+ m_extendGopBoundary = true;
+ break;
+ }
+ }
+ else if(m_param->bHistbasedScenecut && frames[j]->bScenecut)
+ {
+ m_extendGopBoundary = true;
+ break;
+ }
}
if (m_extendGopBoundary)
break;
@@ -2088,13 +2060,23 @@
{
for (int j = 1; j < numBFrames + 1; j++)
{
- if (scenecut(frames, j, j + 1, false, origNumFrames) ||
- (bForceRADL && (frames[j]->frameNum == preRADL)))
- {
- frames[j]->sliceType = X265_TYPE_P;
- numAnalyzed = j;
- break;
+ if (m_param->bHistbasedScenecut) {
+ if (frames[j]->bScenecut || (bForceRADL && (frames[j]->frameNum == preRADL)))
+ {
+ frames[j]->sliceType = X265_TYPE_P;
+ numAnalyzed = j;
+ break;
+ }
}
+ else if (m_param->scenecutThreshold){
+ if ( scenecut(frames, j, j + 1, false, origNumFrames) || (bForceRADL && (frames[j]->frameNum == preRADL)) )
+ {
+ frames[j]->sliceType = X265_TYPE_P;
+ numAnalyzed = j;
+ break;
+ }
+ }
+
}
}
resetStart = bKeyframe ? 1 : X265_MIN(numBFrames + 2, numAnalyzed + 1);
diff -r 37648fca915b -r deaecadc4306 source/encoder/slicetype.h
--- a/source/encoder/slicetype.h Fri Oct 11 12:45:52 2019 +0530
+++ b/source/encoder/slicetype.h Tue Oct 15 12:16:17 2019 +0530
@@ -43,6 +43,14 @@
#define AQ_EDGE_BIAS 0.5
#define EDGE_INCLINATION 45
+#ifdef HIGH_BIT_DEPTH
+#define edge_threshold 1023.0
+#define whitePixel 1023.0
+#else
+#define edge_threshold 255.0
+#define pixel whitePixel 255.0
+#endif
+
/* Thread local data for lookahead tasks */
struct LookaheadTLD
{
diff -r 37648fca915b -r deaecadc4306 source/test/regression-tests.txt
--- a/source/test/regression-tests.txt Fri Oct 11 12:45:52 2019 +0530
+++ b/source/test/regression-tests.txt Tue Oct 15 12:16:17 2019 +0530
@@ -157,6 +157,9 @@
ducks_take_off_420_720p50.y4m,--preset medium --aq-mode 4 --crf 22 --no-cutree
ducks_take_off_420_1_720p50.y4m,--preset medium --selective-sao 4 --sao --crf 20
Traffic_4096x2048_30p.y4m, --preset medium --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 0.01
+Traffic_4096x2048_30p.y4m, --preset medium --frame-dup --dup-threshold 60 --hrd --bitrate 10000 --vbv-bufsize 15000 --vbv-maxrate 12000 --hist-scenecut --hist-threshold 0.01
+sintel_trailer_2k_1920x1080_24.yuv, --preset medium --scenecut 40 --scenecut-bias 20
# Main12 intraCost overflow bug test
720p50_parkrun_ter.y4m,--preset medium
diff -r 37648fca915b -r deaecadc4306 source/x265.h
--- a/source/x265.h Fri Oct 11 12:45:52 2019 +0530
+++ b/source/x265.h Tue Oct 15 12:16:17 2019 +0530
@@ -210,7 +210,9 @@
uint32_t numCUsInFrame;
uint32_t numPartitions;
uint32_t depthBytes;
- int bScenecut;
+ bool bScenecut;
+ double edgeSadValue;
+ double chromaSadValue;
x265_weight_param* wt;
x265_analysis_inter_data* interData;
x265_analysis_intra_data* intraData;
@@ -291,6 +293,9 @@
char sliceType;
int bScenecut;
double ipCostRatio;
+ double yedgeSadValue;
+ double chromaSadValue;
+
int frameLatency;
x265_cu_stats cuStats;
x265_pu_stats puStats;
@@ -465,6 +470,9 @@
//Dolby Vision RPU metadata
x265_dolby_vision_rpu rpu;
+ //Flag to determine the latest frame in the buffer
+ bool bufUpdated;
+
int fieldNum;
//SEI picture structure message
@@ -1017,8 +1025,9 @@
* decisions. Default is 0 - disabled. 1 is the same as 0. Max 16 */
int lookaheadSlices;
- /* An arbitrary threshold which determines how aggressively the lookahead
- * should detect scene cuts. The default (40) is recommended. */
+ /* An arbitrary threshold which determines how aggressively the lookahead
+ * should detect scene cuts. The default (40) is recommended.
+ * Used for encoding cost based scenecut detection */
int scenecutThreshold;
/* Replace keyframes by using a column of intra blocks that move across the video
@@ -1803,6 +1812,7 @@
/*Emit content light level info SEI*/
int bEmitCLL;
+
/*
* Signals picture structure SEI timing message for every frame
@@ -1819,6 +1829,17 @@
/*Input sequence bit depth. It can be either 8bit, 10bit or 12bit.*/
int sourceBitDepth;
+
+ /* A genuine threshold which determines whether a frame is a scenecut or not
+ * when compared against edge and color sad values of a frames histograms.Default 0.01
+ * Range:real number in range (0,2)
+ * Used for histogram based scene cut detection */
+ double edgeTransitionThreshold;
+
+ /*enables improved scenecut detection algorithm to detect scenecuts for slice type
+ decision and rate control */
+ bool bHistbasedScenecut;
+
} x265_param;
/* x265_param_alloc:
* Allocates an x265_param instance. The returned param structure is not
diff -r 37648fca915b -r deaecadc4306 source/x265cli.h
--- a/source/x265cli.h Fri Oct 11 12:45:52 2019 +0530
+++ b/source/x265cli.h Tue Oct 15 12:16:17 2019 +0530
@@ -129,6 +129,9 @@
{ "scenecut", required_argument, NULL, 0 },
{ "no-scenecut", no_argument, NULL, 0 },
{ "scenecut-bias", required_argument, NULL, 0 },
+ { "hist-scenecut", no_argument, NULL, 0},
+ { "no-hist-scenecut", no_argument, NULL, 0},
+ { "hist-threshold", required_argument, NULL, 0},
{ "fades", no_argument, NULL, 0 },
{ "no-fades", no_argument, NULL, 0 },
{ "radl", required_argument, NULL, 0 },
@@ -485,7 +488,10 @@
H0(" --gop-lookahead <integer> Extends gop boundary if a scenecut is found within this from keyint boundary. Default 0\n");
H0(" --no-scenecut Disable adaptive I-frame decision\n");
H0(" --scenecut <integer> How aggressively to insert extra I-frames. Default %d\n", param->scenecutThreshold);
- H1(" --scenecut-bias <0..100.0> Bias for scenecut detection. Default %.2f\n", param->scenecutBias);
+ H0(" --hist-scenecut ..... Enables improved scene-cut detection using histogram based algorithm.");
+ H0(" --no-hist-scenecut Disables improved scene-cut detection using histogram based algorithm. ");
+ H0(" --scenecut-bias <0..100.0> Bias for scenecut detection. Default %.2f\n", param->scenecutBias);
+ H0(" --hist-threshold <0.0..2.0> Threshold for histogram based scenecut detection Default %.2f\n", param->edgeTransitionThreshold);
H0(" --[no-]fades Enable detection and handling of fade-in regions. Default %s\n", OPT(param->bEnableFades));
H0(" --radl <integer> Number of RADL pictures allowed in front of IDR. Default %d\n", param->radl);
H0(" --intra-refresh Use Periodic Intra Refresh instead of IDR frames\n");
_______________________________________________
x265-devel mailing list
x265-devel@videolan.org
https://mailman.videolan.org/listinfo/x265-devel
