Add VFR support to ratecontrol and slicetype, fix HRD timing.

doc/reST/cli.rst

@@ -2536,14 +2536,26 @@ VUI fields must be manually specified.
 
 	Set the picture structure and emits it in the picture timing SEI message.
 	Values in the range 0..12. See D.3.3 of the HEVC spec. for a detailed explanation.
-	Required for HLG (Hybrid Log Gamma) signaling. Not signaled by default.
+
+.. option:: --psfile <filename>
+
+	Specify a text file which contains the picture structure for some or all frames.
+	Allows for dynamic pulldown (VFR in CFR containers). The format of each line is
+
+	framenumber framefieldcoding picstruct
+
+	Framefieldcoding shall be 0, 1, 2 (progressive, bottom-first or top-first, resp.),
+	It shall not change for an encoded sequence and match the encoder configuration.
+
+	Picstruct is the structure to specify in the framenumber's Picture Timing SEI.
+	It shall be compatible with the encoder configuration for the given sequence.
 
 .. option:: --video-signal-type-preset <string>
 
 	Specify combinations of color primaries, transfer characteristics, color matrix,
 	range of luma and chroma signals, and chroma sample location.
 	String format: <system-id>[:<color-volume>]
-	
+
 	This has higher precedence than individual VUI parameters. If any individual VUI option
 	is specified together with this, which changes the values set corresponding to the system-id
 	or color-volume, it will be discarded.
@@ -2657,7 +2669,7 @@ Bitstream options
 	Picture Timing SEI messages providing timing information to the
 	decoder. Default disabled
 
-    	
+
 .. option:: --hrd-concat, --no-hrd-concat
 
     Set concatenation flag for the first keyframe in the HRD buffering period SEI. This

source/CMakeLists.txt

@@ -32,7 +32,7 @@ option(NATIVE_BUILD "Target the build CPU" OFF)
 option(STATIC_LINK_CRT "Statically link C and C++ runtimes 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 217)
+set(X265_BUILD 218)
 configure_file("${PROJECT_SOURCE_DIR}/x265.def.in"
                "${PROJECT_BINARY_DIR}/x265.def")
 configure_file("${PROJECT_SOURCE_DIR}/x265_config.h.in"

source/abrEncApp.cpp

@@ -763,6 +763,16 @@ namespace X265_NS {
                         }
                     }
 
+                    if (m_cliopt.psfile)
+                    {
+                        if (!m_cliopt.parsePSFile(pic_orig[view], m_param->interlaceMode, m_param->bField))
+                        {
+                            x265_log(NULL, X265_LOG_ERROR, "can't parse psfile for frame %d in %s\n",
+                                pic_in[view]->poc, profileName);
+                            fclose(m_cliopt.psfile);
+                            m_cliopt.psfile = NULL;
+                        }
+                    }
                     if (m_cliopt.framesToBeEncoded && inFrameCount >= m_cliopt.framesToBeEncoded)
                         pic_in[view] = NULL;
                     else if (readPicture(pic_in[view], view)){

source/abrEncApp.h

@@ -91,6 +91,7 @@ namespace X265_NS {
         CLIOptions m_cliopt;
         InputFile* m_input[MAX_VIEWS];
         const char* m_reconPlayCmd;
+        FILE*    m_psfile;
         FILE*    m_qpfile;
         FILE*    m_zoneFile;
         FILE*    m_dolbyVisionRpu;/* File containing Dolby Vision BL RPU metadata */

source/common/frame.cpp

@@ -58,7 +58,7 @@ Frame::Frame()
     m_addOnPrevChange = NULL;
     m_classifyFrame = false;
     m_fieldNum = 0;
-    m_picStruct = 0;
+    m_picStruct = PIC_STRUCT_AUTO;
     m_edgePic = NULL;
     m_gaussianPic = NULL;
     m_thetaPic = NULL;
@@ -85,6 +85,12 @@ Frame::Frame()
     m_targetBitrate = 0;
     m_targetCrf = 0;
     m_targetQp = 0;
+
+    /* HRD */
+    m_duration = m_cpbDuration = 0;
+    m_cpbDelay = m_dpbDelay = 0;
+    m_displayPicCount = m_codedPicCount = 0;
+    m_timebase = 0.;
 }
 
 bool Frame::create(x265_param *param, float* quantOffsets)

source/common/frame.h

@@ -182,6 +182,15 @@ class Frame
     /* target QP for this picture.*/
     int                    m_targetQp;
 
+    /* HRD timing for this frame */
+    uint32_t               m_duration;           /* display duration [clock ticks] */
+    uint32_t               m_cpbDuration;        /* lifetime in the CPB [clock ticks] */
+    uint32_t               m_cpbDelay;           /* CPB removal delay [clock ticks] */
+    uint32_t               m_dpbDelay;           /* DPB output delay [clock ticks] */
+    int64_t                m_displayPicCount;    /* PTS in clock ticks */
+    int64_t                m_codedPicCount;      /* DTS in clock ticks */
+    double                 m_timebase;           /* timebase for this picture [seconds] */
+
     Frame();
 
     bool create(x265_param *param, float* quantOffsets);

source/common/lowres.cpp

@@ -353,6 +353,11 @@ void Lowres::init(PicYuv* origPic, int poc, bool bEnableTemporalFilter)
     frameNum = poc;
     leadingBframes = 0;
     indB = 0;
+
+    cpbDurationSecs = 0.;
+    dispDurationSecs = 0.;
+    dispPicCount = durationPicCount = 0;
+
     memset(costEst, -1, sizeof(costEst));
     memset(weightedCostDelta, 0, sizeof(weightedCostDelta));
 
@@ -372,8 +377,13 @@ void Lowres::init(PicYuv* origPic, int poc, bool bEnableTemporalFilter)
     for (int i = 0; i < bframes + 2; i++)
         intraMbs[i] = 0;
     if (origPic->m_param->rc.vbvBufferSize)
+    {
         for (int i = 0; i < X265_LOOKAHEAD_MAX + 1; i++)
+        {
             plannedType[i] = X265_TYPE_AUTO;
+            plannedCpbDuration[i] = origPic->m_param->fpsDenom / (double)origPic->m_param->fpsNum;
+        }
+    }
 
     /* downscale and generate 4 hpel planes for lookahead */
     if (bEnableTemporalFilter)
@@ -386,7 +396,7 @@ void Lowres::init(PicYuv* origPic, int poc, bool bEnableTemporalFilter)
     extendPicBorder(lowresPlane[1], lumaStride, width, lines, origPic->m_lumaMarginX, origPic->m_lumaMarginY);
     extendPicBorder(lowresPlane[2], lumaStride, width, lines, origPic->m_lumaMarginX, origPic->m_lumaMarginY);
     extendPicBorder(lowresPlane[3], lumaStride, width, lines, origPic->m_lumaMarginX, origPic->m_lumaMarginY);
-    
+
     if (origPic->m_param->bEnableHME || origPic->m_param->bEnableTemporalFilter)
     {
         if (bEnableTemporalFilter)

source/common/lowres.h

@@ -207,6 +207,7 @@ struct Lowres : public ReferencePlanes
     /* used for vbvLookahead */
     int       plannedType[X265_LOOKAHEAD_MAX + 1];
     int64_t   plannedSatd[X265_LOOKAHEAD_MAX + 1];
+    double    plannedCpbDuration[X265_LOOKAHEAD_MAX + 1];
     int       indB;
     int       bframes;
 
@@ -222,6 +223,11 @@ struct Lowres : public ReferencePlanes
     double    frameVariance;
     int*      edgeInclined;
 
+    /* Vbv Lookahead & cuTree durations */
+    double   cpbDurationSecs;
+    double   dispDurationSecs;
+    int64_t  dispPicCount;
+    uint32_t durationPicCount;
 
     /* cutree intermediate data */
     PicQPAdaptationLayer* pAQLayer;

source/common/param.cpp

@@ -1020,7 +1020,7 @@ int x265_param_parse(x265_param* p, const char* name, const char* value)
         else
         {
             float fps = (float)atof(value);
-            if (fps > 0 && fps <= INT_MAX / 1000)
+            if (fps > 0 && fps <= INT_MAX / 1000 && ceilf(fps) != floorf(fps))
             {
                 p->fpsNum = (int)(fps * 1000 + .5);
                 p->fpsDenom = 1000;

source/encoder/api.cpp

@@ -1014,7 +1014,7 @@ void x265_picture_init(x265_param *param, x265_picture *pic)
     pic->userSEI.numPayloads = 0;
     pic->rpu.payloadSize = 0;
     pic->rpu.payload = NULL;
-    pic->picStruct = 0;
+    pic->picStruct = PIC_STRUCT_AUTO;
     pic->vbvEndFlag = 0;
 
     if ((strlen(param->analysisSave) || strlen(param->analysisLoad)) || (param->bAnalysisType == AVC_INFO))