Merge pull request #100 from adamjw24/develop_1.2

Changes for 1.2.0 (5% BD-rate improvement for faster)

.github/workflows/Build.yml

@@ -13,7 +13,6 @@ jobs:
     strategy:
       matrix:
         include:
-          - os: ubuntu-16.04
           - os: ubuntu-18.04
           - os: ubuntu-20.04
           - os: macos-10.15

CMakeLists.txt

@@ -10,7 +10,7 @@ if( NOT CMAKE_VERSION VERSION_LESS 3.13.0 )
 endif()
 
 # project name
-project( vvenc VERSION 1.1.0 )
+project( vvenc VERSION 1.2.0 )
 
 set( VVENC_ENABLE_X86_SIMD TRUE )
 set( VVENC_ENABLE_ARM_SIMD FALSE )
@@ -54,10 +54,10 @@ if( NOT CMAKE_SOURCE_DIR STREQUAL CMAKE_CURRENT_SOURCE_DIR )
   if( MSVC )
     add_compile_options( "/EHsc" )
   endif()
-  
+
   list( FIND ${PROJECT_NAME}_ADD_SUBDIRECTORIES "source/App/vvencFFapp" vvEncFfAppFound )
   list( FIND ${PROJECT_NAME}_ADD_SUBDIRECTORIES "source/App/vvencapp" vvEncAppFound )
-  
+
   if( vvEncFfAppFound GREATER_EQUAL 0 OR vvEncAppFound GREATER_EQUAL 0 )
     list( PREPEND ${PROJECT_NAME}_ADD_SUBDIRECTORIES "source/Lib/apputils" )
   endif()
@@ -75,19 +75,20 @@ endif()
 
 # enable install target
 set( VVENC_ENABLE_INSTALL                   ON  CACHE BOOL   "Enable or disable install target" )
-                                                             
+
 # enable postfix                                             
 set( VVENC_ENABLE_BUILD_TYPE_POSTFIX        OFF CACHE BOOL   "Enable or disable build type postfix for apps and libs" )
-                                                             
+
 set( VVENC_ENABLE_LINK_TIME_OPT             ON  CACHE BOOL   "Enable link time optimization for release and profile builds" )
 
+set( VVENC_ENABLE_THIRDPARTY_JSON           ON  CACHE BOOL   "Enable use of thirdparty json library" )
+
 if( CMAKE_CXX_COMPILER_ID STREQUAL "GNU" OR
     CMAKE_CXX_COMPILER_ID STREQUAL "Clang" )
-    
+
   # enable address sanitizer
   set( VVENC_USE_ADDRESS_SANITIZER          OFF CACHE BOOL   "Enable or disable address sanitizer" )
-
-  # add -march=native
+
   set( VVENC_OPT_TARGET_ARCH                ""  CACHE STRING "Enable or disable building with architecture specific optimization" )
 endif()
 
@@ -146,9 +147,6 @@ if( NOT CMAKE_BUILD_TYPE )
   set( CMAKE_BUILD_TYPE "Release" CACHE STRING "Choose the type of build, options are: Debug Release RelWithDebInfo MinSizeRel." FORCE )
 endif()
 
-#set( ENABLE_LINK_TIME_OPT ON       CACHE BOOL   "Enable link time optimization for release and profile builds" )
-#set( OPT_TARGET_ARCH      "native" CACHE STRING "Create code and optimize for this architecture (default native)" )
-
 # set c++11
 set( CMAKE_CXX_STANDARD 14 )
 set( CMAKE_CXX_STANDARD_REQUIRED ON )
@@ -343,8 +341,26 @@ add_test( NAME Test_compare_output-medium_noqpa_0thr COMMAND ${CMAKE_COMMAND} -E
 add_test( NAME Test_vvencFFapp-lowdelay_medium_enc COMMAND vvencFFapp -c ../../cfg/experimental/lowdelay_medium.cfg -c ../../test/data/RTn23.cfg --IntraPeriod=-1 -dph 1 -f 8 -b outf.vvc )
 add_test( NAME Test_vvencFFapp-lowdelay_medium_dec COMMAND vvencFFapp --decode -b outf.vvc )
 
-add_test( NAME Test_remove_temp_files COMMAND ${CMAKE_COMMAND} -E remove out.vvc tout.vvc rec.yuv outf.vvc )
-
+add_test( NAME Test_vvencapp-medium_rc2p COMMAND vvencapp --preset medium -s 80x44 -r 15 -i ../../test/data/RTn23_80x44p15_f15.yuv -f 5 --Bitrate=10000 --Passes=2 -o out.vvc )
+set_tests_properties( Test_vvencapp-medium_rc2p PROPERTIES TIMEOUT 30 )
+add_test( NAME Test_vvencFFapp-medium_rc2p COMMAND vvencFFapp -c ../../cfg/randomaccess_medium.cfg -c ../../test/data/RTn23.cfg -f 5 --TargetBitrate=10000 --Passes=2 -b outf.vvc )
+set_tests_properties( Test_vvencFFapp-medium_rc2p PROPERTIES TIMEOUT 30 )
+add_test( NAME Test_compare_output-medium_rc2p COMMAND ${CMAKE_COMMAND} -E compare_files out.vvc outf.vvc )
+
+add_test( NAME Test_vvencFFapp-medium_rc2p_statsFile1_exp COMMAND vvencFFapp -c ../../cfg/randomaccess_medium.cfg -c ../../test/data/RTn23.cfg -f 5 --TargetBitrate=10000 --Pass=1 --RCStatsFile=stats.json -b outf.vvc )
+set_tests_properties( Test_vvencFFapp-medium_rc2p_statsFile1_exp PROPERTIES TIMEOUT 30 )
+add_test( NAME Test_vvencFFapp-medium_rc2p_statsFile2_exp COMMAND vvencFFapp -c ../../cfg/randomaccess_medium.cfg -c ../../test/data/RTn23.cfg -f 5 --TargetBitrate=10000 --Pass=2 --RCStatsFile=stats.json -b outf.vvc )
+set_tests_properties( Test_vvencFFapp-medium_rc2p_statsFile2_exp PROPERTIES TIMEOUT 30 )
+add_test( NAME Test_compare_output-medium_rc2p_statsFile_exp COMMAND ${CMAKE_COMMAND} -E compare_files out.vvc outf.vvc )
+
+add_test( NAME Test_vvencapp-medium_rc2p_statsFile1_easy COMMAND vvencapp --preset medium -s 80x44 -r 15 -i ../../test/data/RTn23_80x44p15_f15.yuv -f 5 --Bitrate=10000 --Pass=1 --RCStatsFile=stats.json -o out.vvc )
+set_tests_properties( Test_vvencapp-medium_rc2p_statsFile1_easy PROPERTIES TIMEOUT 30 )
+add_test( NAME Test_vvencapp-medium_rc2p_statsFile2_easy COMMAND vvencapp --preset medium -s 80x44 -r 15 -i ../../test/data/RTn23_80x44p15_f15.yuv -f 5 --Bitrate=10000 --Pass=2 --RCStatsFile=stats.json -o out.vvc )
+set_tests_properties( Test_vvencapp-medium_rc2p_statsFile2_easy PROPERTIES TIMEOUT 30 )
+add_test( NAME Test_compare_output-medium_rc2p_statsFile_easy COMMAND ${CMAKE_COMMAND} -E compare_files out.vvc outf.vvc )
+
+add_test( NAME Test_remove_temp_files COMMAND ${CMAKE_COMMAND} -E remove out.vvc tout.vvc rec.yuv outf.vvc stats.json )
+
 
 if( VVENC_ENABLE_INSTALL )
   # include installer

Makefile

@@ -46,6 +46,10 @@ ifneq ($(disable-lto),)
 CONFIG_OPTIONS += -DVVENC_ENABLE_LINK_TIME_OPT=OFF
 endif
 
+ifneq ($(disable-json),)
+CONFIG_OPTIONS += -DVVENC_ENABLE_THIRDPARTY_JSON=OFF
+endif
+
 ifneq ($(enable-build-type-postfix),)
 CONFIG_OPTIONS += -DVVENC_ENABLE_BUILD_TYPE_POSTFIX=ON
 endif

README.md

@@ -49,6 +49,8 @@ The standard encoder (**vvencapp**) can be used in one of five predefined preset
 | --qp,-q <int>          | 32                               | Quantization parameter (0..63)                                                                       |
 | --bitrate,-b <int>     | 0                                | Bitrate for rate control (0: constant-QP encoding without rate control, otherwise bits per second). <br> Rate control requires correct `--framerate` (see option above). |
 | --passes,-p <int>      | -1                               | Number of rate control passes (1: single-pass rate control, 2: two-pass rate control)                |
+| --pass <int>           | not set                          | Set current rate control pass. If not set, encoder will run both passes. If set to [1,2], <br> encoder will execute first or second pass only. Requires `--rcstatsfile` to be set (see below). |
+| --rcstatsfile <str>    | not set                          | Rate control statistics file, to store or load first pass rate control statistics data.              |
 | --qpa <int>            | 1                                | Perceptual QP adaptation (QPA) to improve subjective video quality (0: off, 1: on)                   |
 | --refreshsec,-rs <int> | 1                                | Intra period/refresh in seconds                                                                      |
 | --threads,-t <int>     | size >= 1280x720: <br> 8, else: 4  | Number of threads (1-N)                                                                              |
@@ -72,7 +74,7 @@ The expert mode encoder (**vvencFFapp**) is based on the [VVC test model (VTM)](
 **Example usage:** In order to start your first experiments with the expert mode encoder, adapt the sequence.cfg configuration file to your input YUV source file and use the following command:
 
     vvencFFapp -c randomaccess_medium.cfg -c sequence.cfg
-    
+
 ## How to map command line parameters of the standard encoder into the full featured expert mode encoder?
 The export mode encoder (**vvencFFapp**) can be used in the same manner as the standard encoder (**vvencapp**) by using the adapted expert option names.
 Be aware of that some options have different default values. The following table shows only expert options that differs from the standard encoder:
@@ -91,7 +93,9 @@ Be aware of that some options have different default values. The following table
 | --output,-o <str>      | not set                          | --BitstreamFile,-b <str> | not set                        |
 | --qp,-q <int>          | 32                               | --QP <int>             | 32                               |
 | --bitrate,-b <int>     | 0                                | --TargetBitrate <int>  | 0                                |
-| --passes,-p <int>      | -1                                | --NumPasses <int>     | -1                               |
+| --passes,-p <int>      | -1                               | --NumPasses <int>      | -1                               |
+| --pass <int>           | -1                               | --Pass <int>           | -1                               |
+| --rcstatsfile <str>    | -                                | --RCStatsFile <str>    | -                                |
 | --qpa <int>            | 1                                | --PerceptQPA,-qpa <int> | 0                               |
 | --refreshtype,-rt <str> | cra                             | --DecodingRefreshType <str> | cra                         |
 | --refreshsec,rs <int>  | 1                                | --RefreshSec <int>       | 1                              |
@@ -104,20 +108,20 @@ Be aware of that some options have different default values. The following table
 | --accessunitdelimiter,-aud <int> | auto                   | --AccessUnitDelimiter,-aud <int> | auto                   |
 | --vuiparameterspresent,-vui <int> | auto                  | --VuiParametersPresent,-vui <int> | auto                  |
 | --hrdparameterspresent,-hrd <int> | auto                  | --HrdParametersPresent,-hrd <int> | auto                  |
-| --decodedpicturehash,-dph <int> | off                     | --SEIDecodedPictureHash,-dph <int> | auto                 |
+| --decodedpicturehash,-dph <int> | off                     | --SEIDecodedPictureHash,-dph <int> | off                  |
 
 **Example usage:** Given a YUV 4:2:0 input file with a bit-depth of 8bit and a resolution of 176x144 pixels,
 the following calls will encode the input file with the medium speedup preset with 1Mbit/s by using Two pass rate control. 
 Both calls will produce the same output.
 
-**standard encoder:** 
+**standard encoder:**
 
     vvencapp --preset medium -i BUS_176x144_75@15.yuv -s 176x144 -r 15 -b 1000000 -p 2 -o str.266
 
-**full featured expert mode encoder:** 
+**full featured expert mode encoder:**
 
     vvencFFapp --preset medium --InputFile BUS_176x144_75@15.yuv -s 176x144 -fr 15 -TargetBitrate 1000000 --NumPasses 2 -qpa 1 -t -1 -b str.266
-    
+
 # Contributing
 
 Feel free to contribute. To do so:
@@ -126,6 +130,20 @@ Feel free to contribute. To do so:
 * Apply the desired changes
 * Create a pull-request to the upstream repository
 
+# Third party tools
+
+## nlohmann/json: JSON for Modern C++
+
+JSON serialization is used for writing / reading rate control statistics. The external library is available under
+MIT license at https://github.com/nlohmann/json. Please see [LICENSE.MIT](./thirdparty/nlohmann_json/LICENSE.MIT) for
+the terms of use of the contents of this library.
+
+In case you don't want to include the external nlohmann/json library, please compile with
+
+    make install-release disable-json=1
+
+In this case writing / reading rate control statistic files is not supported by the VVenC encoders.
+
 # License
 
 Please see [LICENSE.txt](./LICENSE.txt) file for the terms of use of the contents of this repository.

cfg/experimental/lowdelay_fast.cfg

@@ -74,8 +74,8 @@ MaxNumMergeCand               : 6      # Maximum number of merge candidates
 # Tools configuration
 Affine                        : 2      # Affine prediction: 0: disabled, 1: vtm, 2: fast
 ALF                           : 1      # Adpative Loop Filter: 0: disabled, 1: enabled
-ALFSpeed                      : 0      # ALF speed (skip filtering of non-referenced frames) [0-1]
-AllowDisFracMMVD              : 0      # Disable fractional MVD in MMVD mode adaptively
+ALFSpeed                      : 1      # ALF speed (skip filtering of non-referenced frames) [0-1]
+AllowDisFracMMVD              : 1      # Disable fractional MVD in MMVD mode adaptively
 BCW                           : 0      # Enable Generalized Bi-prediction(Bcw) 0: disabled, 1: enabled, 2: fast
 BIO                           : 1      # Bi-directional optical flow: 0: disabled, 1: enabled
 CCALF                         : 1      # Cross-component Adaptive Loop Filter: 0: disabled, 1: enabled
@@ -86,7 +86,7 @@ EncDbOpt                      : 0      # Encoder optimization with deblocking fi
 Geo                           : 0      # Geometric partitioning mode: 0: disabled, 1: vtm, 2: good quality, 3: fast
 IMV                           : 5      # Adaptive MV precision Mode (IMV): 0: disabled, 1:vtm, 2-7: fast modes
 ISP                           : 0      # Intra subpartitions: 0: disabled, 1: enabled, 2: fast, 3: faster
-JointCbCr                     : 0      # Joint coding of chroma residuals: 0: disabled, 1: enabled
+JointCbCr                     : 1      # Joint coding of chroma residuals: 0: disabled, 1: enabled
 LFNST                         : 1      # LFNST: 0: disabled, 1: enabled
 LMChroma                      : 1      # LMChroma prediction: 0: disabled, 1: enabled
 LMCSEnable                    : 2      # LMCS: 0: disable, 1:enable, 2: use SCC detection to disable for screen coded content
@@ -121,7 +121,7 @@ ChromaTS                      : 0      # Chroma transform skip: 0: disabled, 1:
 TransformSkipLog2MaxSize      : 4
 
 # Fast options                
-QtbttExtraFast                : 2      # Non-VTM compatible QTBTT speed-ups: 0: disabled, 1: enabled
+QtbttExtraFast                : 3      # Non-VTM compatible QTBTT speed-ups: 0: disabled, 1: enabled
 ContentBasedFastQtbt          : 1      # Signal based QTBT speed-up: 0: disabled, 1: enabled
 PBIntraFast                   : 1      # Fast assertion if the intra mode is probable: 0: disabled, 1: enabled
 FastMrg                       : 2      # Fast methods for inter merge: 0: disabled, 1: vtm, 2: fast
@@ -131,6 +131,8 @@ FastIntraTools                : 0      # Speedup intra tools: LFNST, ISP, MTS
 FastInferMerge                : 2      # Fast method to skip Inter/Intra modes. 0: off, [1..4] speedups
 ECU                           : 1      # Early CU setting
 IntegerET                     : 0      # Enable early termination for integer motion search
-IntraEstDecBit                : 3      # Test every eight intra directional mode in the first pass
+IntraEstDecBit                : 2      # Test every eight intra directional mode in the first pass
+NumIntraModesFullRD           : -1     # Number modes for full RD intra search
+ReduceIntraChromaModesFullRD  : 1      # Reduce modes for chroma full RD intra search
 
 

cfg/experimental/lowdelay_faster.cfg

@@ -73,12 +73,12 @@ MaxNumMergeCand               : 6      # Maximum number of merge candidates
 
 # Tools configuration
 Affine                        : 0      # Affine prediction: 0: disabled, 1: vtm, 2: fast
-ALF                           : 0      # Adpative Loop Filter: 0: disabled, 1: enabled
-ALFSpeed                      : 0      # ALF speed (skip filtering of non-referenced frames) [0-1]
+ALF                           : 1      # Adpative Loop Filter: 0: disabled, 1: enabled
+ALFSpeed                      : 1      # ALF speed (skip filtering of non-referenced frames) [0-1]
 AllowDisFracMMVD              : 0      # Disable fractional MVD in MMVD mode adaptively
 BCW                           : 0      # Enable Generalized Bi-prediction(Bcw) 0: disabled, 1: enabled, 2: fast
 BIO                           : 0      # Bi-directional optical flow: 0: disabled, 1: enabled
-CCALF                         : 0      # Cross-component Adaptive Loop Filter: 0: disabled, 1: enabled
+CCALF                         : 1      # Cross-component Adaptive Loop Filter: 0: disabled, 1: enabled
 CIIP                          : 0      # CIIP mode: 0: disable, 1: vtm, 2: fast, 3: faster
 DepQuant                      : 0      # Dependent quantization: 0: disabled, 1: enabled
 DMVR                          : 1      # Decoder-side Motion Vector Refinement: 0: disabled, 1: enabled
@@ -121,7 +121,7 @@ ChromaTS                      : 0      # Chroma transform skip: 0: disabled, 1:
 TransformSkipLog2MaxSize      : 3
 
 # Fast options                
-QtbttExtraFast                : 3      # Non-VTM compatible QTBTT speed-ups: 0: disabled, 1: enabled
+QtbttExtraFast                : 7      # Non-VTM compatible QTBTT speed-ups: 0: disabled, 1: enabled
 ContentBasedFastQtbt          : 1      # Signal based QTBT speed-up: 0: disabled, 1: enabled
 PBIntraFast                   : 1      # Fast assertion if the intra mode is probable: 0: disabled, 1: enabled
 FastMrg                       : 2      # Fast methods for inter merge: 0: disabled, 1: vtm, 2: fast
@@ -132,5 +132,7 @@ FastInferMerge                : 4      # Fast method to skip Inter/Intra modes.
 ECU                           : 1      # Early CU setting
 IntegerET                     : 1      # Enable early termination for integer motion search
 IntraEstDecBit                : 3      # Test every eight intra directional mode in the first pass
+NumIntraModesFullRD           : 1      # Number modes for full RD intra search
+ReduceIntraChromaModesFullRD  : 1      # Reduce modes for chroma full RD intra search
 
 

cfg/experimental/lowdelay_medium.cfg

@@ -121,8 +121,8 @@ ChromaTS                      : 0      # Chroma transform skip: 0: disabled, 1:
 TransformSkipLog2MaxSize      : 4
 
 # Fast options                
-QtbttExtraFast                : 2      # Non-VTM compatible QTBTT speed-ups: 0: disabled, 1: enabled
-ContentBasedFastQtbt          : 1      # Signal based QTBT speed-up: 0: disabled, 1: enabled
+QtbttExtraFast                : 3      # Non-VTM compatible QTBTT speed-ups: 0: disabled, 1: enabled
+ContentBasedFastQtbt          : 0      # Signal based QTBT speed-up: 0: disabled, 1: enabled
 PBIntraFast                   : 1      # Fast assertion if the intra mode is probable: 0: disabled, 1: enabled
 FastMrg                       : 2      # Fast methods for inter merge: 0: disabled, 1: vtm, 2: fast
 FastLocalDualTreeMode         : 2      # Fast intra pass coding for local dual-tree in intra coding region: 0: disable, 1: use threshold, 2: one intra mode only
@@ -132,5 +132,7 @@ FastInferMerge                : 0      # Fast method to skip Inter/Intra modes.
 ECU                           : 0      # Early CU setting
 IntegerET                     : 0      # Enable early termination for integer motion search
 IntraEstDecBit                : 2      # Test every fourth intra directional mode in the first pass
+NumIntraModesFullRD           : -1     # Number modes for full RD intra search
+ReduceIntraChromaModesFullRD  : 0      # Reduce modes for chroma full RD intra search