Merge remote-tracking branch 'origin/main' into... (872d7806) · Commits · SA4 / Audio / IVAS BASOP

.gitlab-ci.yml

+62 −2

Original line number	Diff line number	Diff line
		variables:
		# note: GitLab cannot reference variables defined by users in the include ref:, we need to use a YAML anchor for this
		# see https://docs.gitlab.com/ci/yaml/includes/#use-variables-with-include for more information
		IVAS_CODEC_CI_REF: &IVAS_CODEC_CI_REF 41186736915f5c0a857ac92e8ab438c434e5247d
		IVAS_CODEC_CI_REF: &IVAS_CODEC_CI_REF a31272de16bd1b556269a50bc179321a60f2a500

		include:
		- local: .gitlab-ci/variables.yml
		@@ -62,6 +62,9 @@ workflow:
		- if: $CI_PIPELINE_SOURCE == 'web' && $MANUAL_PIPELINE_TYPE == 'voip-be-test'
		variables:
		IVAS_PIPELINE_NAME: 'Voip BE test on $CI_COMMIT_BRANCH'
		- if: $CI_PIPELINE_SOURCE == 'web' && $MANUAL_PIPELINE_TYPE == 'renderer-framesize-be'
		variables:
		IVAS_PIPELINE_NAME: 'Renderer framesize BE test on $CI_COMMIT_BRANCH'
		- if: $CI_PIPELINE_SOURCE == 'web' && $MANUAL_PIPELINE_TYPE == 'peaq-enc-passthrough'
		variables:
		IVAS_PIPELINE_NAME: 'PEAQ encoder pass-through test: $CI_COMMIT_BRANCH'
		@@ -189,6 +192,7 @@ workflow:
		.test-job-linux-needs-testv-dir:
		extends: .test-job-linux
		before_script:
		- !reference [ .job-linux, before_script ]
		- bash "${CI_PROJECT_DIR}"/ivas-codec-ci/snippets/basop/update-scripts-repo.sh
		- if [ ! -d "$TESTV_DIR" ]; then mkdir -p $TESTV_DIR; fi
		- cp -r scripts/testv/* $TESTV_DIR/
		@@ -1059,6 +1063,63 @@ check-regressions-short-dec--10db:
		DUT_ENCODER_PATH: "$REF_ENCODER_PATH"
		MERGE_TARGET_ENCODER_PATH: "$REF_ENCODER_PATH"

		check-be-between-renderer-framesizes:
		extends:
		- .test-job-linux
		# - .rules-pytest-to-main-short
		rules:
		- if: $CI_PIPELINE_SOURCE == 'web' && $MANUAL_PIPELINE_TYPE == "renderer-framesize-be"
		stage: test
		needs: ["build-codec-linux-make"]
		script:
		- bash "${CI_PROJECT_DIR}"/ivas-codec-ci/snippets/disable-limiter.sh
		- make clean
		- make -j
		- echo "$SHORT_TEST_SUITE"

		- bash "${CI_PROJECT_DIR}"/ivas-codec-ci/snippets/basop/build-float-ref-and-dut-binaries.sh
		- python3 ci/remove_unsupported_testcases.py $PRM_FILES

		- exit_code=0
		- python3 -m pytest tests/codec_be_on_mr_nonselection --tb=no -v --ref_encoder_path ./IVAS_cod_ref --ref_decoder_path ./IVAS_dec --update_ref 1 --html=report-20ms.html --self-contained-html --junit-xml=report-junit-20ms.xml \|\| exit_code=$?

		- exit_code5=0
		- exit_code10=0
		- python3 -m pytest tests/codec_be_on_mr_nonselection --tb=no -v --ref_encoder_path ./IVAS_cod_ref --ref_decoder_path ./IVAS_dec --html=report-5ms.html --self-contained-html --junit-xml=report-junit-5ms.xml --dut_fr 5 --decoder_only \|\| exit_code5=$?
		- python3 -m pytest tests/codec_be_on_mr_nonselection --tb=no -v --ref_encoder_path ./IVAS_cod_ref --ref_decoder_path ./IVAS_dec --html=report-10ms.html --self-contained-html --junit-xml=report-junit-10ms.xml --dut_fr 10 --decoder_only \|\| exit_code10=$?

		- zero_errors5=$(cat report-junit-5ms.xml \| grep -c 'errors="0"') \|\| true
		- zero_errors10=$(cat report-junit-10ms.xml \| grep -c 'errors="0"') \|\| true
		- zero_errors=1

		- *print-results-banner
		- echo "!! Encoder command lines are in the 20ms log files only !!!"
		- if [ $exit_code -ne 0 ]; then echo "20 ms framesize run already failed with errors."; zero_errors=0; fi
		- if [ $zero_errors5 != 1 ]; then echo "run error with 5ms rendering encountered"; zero_errors=0 ; fi
		- if [ $zero_errors10 != 1 ]; then echo "run error with 10ms rendering encountered"; zero_errors=0 ; fi
		- if [ $zero_errors != 1 ]; then exit $EXIT_CODE_FAIL; fi
		- if [ $exit_code5 -ne 0 ]; then echo "Non-bitexact cases encountered with 5ms rendering!"; exit_code=1; fi
		- if [ $exit_code10 -ne 0 ]; then echo "Non-bitexact cases encountered with 10ms rendering!"; exit_code=1; fi
		- if [ $exit_code -ne 0 ]; then exit $EXIT_CODE_FAIL; fi
		artifacts:
		name: "mr-$CI_MERGE_REQUEST_IID--sha-$CI_COMMIT_SHORT_SHA--stage-$CI_JOB_STAGE--results"
		expire_in: 1 week
		when: always
		expose_as: "pytest 5ms and 10ms vs 20ms results"
		paths:
		- report-junit-5ms.xml
		- report-5ms.html
		- report-junit-10ms.xml
		- report-10ms.html
		- report-junit-20ms.xml
		- report-20ms.html
		reports:
		junit:
		- report-junit-5ms.xml
		- report-junit-10ms.xml
		- report-junit-20ms.xml


		# ---------------------------------------------------------------
		# Short test jobs for running from web interface or schedule
		# ---------------------------------------------------------------
		@@ -2042,7 +2103,6 @@ pages:
		- if: $UPDATE_PAGES
		script:
		- !reference [ .job-linux, before_script ]

		- bash "${CI_PROJECT_DIR}"/ivas-codec-ci/snippets/print-common-info.sh
		- bash "${CI_PROJECT_DIR}"/ivas-codec-ci/snippets/basop/update-scripts-repo.sh
		- python3 ci/setup_pages.py

.gitlab-ci/variables.yml

+1 −0

Original line number	Diff line number	Diff line
		@@ -63,4 +63,5 @@ variables:
		- 'complexity'
		- 'coverage'
		- 'voip-be-test'
		- 'renderer-framesize-be'
		- 'peaq-enc-passthrough'

apps/renderer.c

+100 −49

Original line number	Diff line number	Diff line
		@@ -149,7 +149,7 @@ typedef struct
		int32_t sampleRate;
		InputConfig inConfig;
		OutputConfig outConfig;
		char inMetadataFilePaths[RENDERER_MAX_ISM_INPUTS][RENDERER_MAX_CLI_ARG_LENGTH];
		char inMetadataFilePaths[RENDERER_MAX_ISM_INPUTS + RENDERER_MAX_MASA_INPUTS][RENDERER_MAX_CLI_ARG_LENGTH];
		int16_t numInMetadataFiles;
		char headRotationFilePath[RENDERER_MAX_CLI_ARG_LENGTH];
		char referenceVectorFilePath[RENDERER_MAX_CLI_ARG_LENGTH];
		@@ -217,13 +217,13 @@ static const CmdLnParser_Option cliOptions[] = {
		.id = CmdLnOptionId_inputFormat,
		.match = "input_format",
		.matchShort = "if",
		.description = "Audio format of input file (e.g. 5_1 or HOA3 or META, use -l for a list)",
		.description = "Audio format of input file (e.g. 5_1 or HOA3 or META,\nuse -l for a list)",
		},
		{
		.id = CmdLnOptionId_inputMetadata,
		.match = "input_metadata",
		.matchShort = "im",
		.description = "Space-separated list of path to metadata files for ISM or MASA inputs",
		.description = "Space-separated list of path to metadata files for ISM/MASA/OMASA/\nOSBA inputs. \nFor OMASA, ISM files must be specified first.",
		},
		{
		.id = CmdLnOptionId_outputFile,
		@@ -241,57 +241,59 @@ static const CmdLnParser_Option cliOptions[] = {
		.id = CmdLnOptionId_sampleRate,
		.match = "sample_rate",
		.matchShort = "fs",
		.description = "Input sampling rate in kHz (16, 32, 48) - required only with raw PCM inputs",
		.description = "Input sampling rate in kHz (16, 32, 48) - required only with raw\nPCM inputs",
		},
		{
		.id = CmdLnOptionId_trajFile,
		.match = "trajectory_file",
		.matchShort = "T",
		.description = "Head rotation trajectory file for simulation of head tracking (only for binaural outputs)",
		.description = "Head rotation trajectory file for simulation of head tracking\n(only for binaural outputs)",
		},
		{
		.id = CmdLnOptionId_refRotFile,
		.match = "reference_rotation_file",
		.matchShort = "rf",
		.description = "Reference rotation trajectory file for simulation of head tracking (only for binaural outputs)",
		.description = "Reference rotation trajectory file for simulation of head tracking\n(only for binaural outputs)",
		},
		{
		.id = CmdLnOptionId_customHrtfFile,
		.match = "custom_hrtf",
		.matchShort = "hrtf",
		.description = "Custom HRTF file for binaural rendering (only for binaural outputs)",
		.description = "Custom HRTF file for binaural rendering\n(only for binaural outputs)",
		},
		{
		.id = CmdLnOptionId_renderConfigFile,
		.match = "render_config_parameters",
		.matchShort = "render_config",
		.description = "Binaural renderer configuration parameters in file (only for binaural outputs)",
		.description = "Binaural renderer configuration parameters in file\n(only for binaural outputs)",
		},
		{
		.id = CmdLnOptionId_nonDiegeticPan,
		.match = "non_diegetic_panning",
		.matchShort = "non_diegetic_pan",
		.description = "Panning mono non diegetic sound to stereo -90<= pan <= 90\nleft or l or 90->left, right or r or -90->right, center or c or 0 ->middle\n",
		.description = "Panning mono non diegetic sound to stereo -90<= pan <= 90\nleft or l or 90->left, right or r or -90->right,\ncenter or c or 0 ->middle",
		},
		{
		.id = CmdLnOptionId_orientationTracking,
		.match = "tracking_type",
		.matchShort = "otr",
		.description = "Head orientation tracking type: 'none', 'ref', 'avg' or `ref_vec` or `ref_vec_lev` (only for binaural outputs)",
		.description = "Head orientation tracking type: 'none', 'ref', 'avg' or `ref_vec`\nor `ref_vec_lev` (only for binaural outputs)",
		},
		{
		.id = CmdlnOptionId_lfePosition,
		.match = "lfe_position",
		.matchShort = "lp",
		.description = "Output LFE position. Comma-delimited triplet of [gain, azimuth, elevation] where gain is linear (like --gain, -g) and azimuth, elevation are in degrees.\nIf specified, overrides the default behavior which attempts to map input to output LFE channel(s)",
		.description = "Output LFE position. Comma-delimited triplet of [gain, azimuth,\nelevation] where gain is linear (like --gain, -g) and azimuth,\nelevation are in degrees. If specified, overrides the default\nbehavior which attempts to map input to output LFE channel(s)",
		},
		{ .id = CmdlnOptionId_lfeMatrix,
		{
		.id = CmdlnOptionId_lfeMatrix,
		.match = "lfe_matrix",
		.matchShort = "lm",
		.description = "LFE panning matrix. File (CSV table) containing a matrix of dimensions [ num_input_lfe x num_output_channels ] with elements specifying linear routing gain (like --gain, -g). \nIf specified, overrides the output LFE position option and the default behavior which attempts to map input to output LFE channel(s)" },
		.description = "LFE panning matrix. File (CSV table) containing a matrix of\ndimensions [ num_input_lfe x num_output_channels ] with elements\nspecifying linear routing gain (like --gain, -g). If specified,\noverrides the output LFE position option and the default\nbehavior which attempts to map input to output LFE channel(s)",
		},
		{
		.id = CmdLnOptionId_noDelayCmp,
		.match = "no_delay_comparison",
		.match = "no_delay_compensation",
		.matchShort = "no_delay_cmp",
		.description = "[flag] Turn off delay compensation",
		},
		@@ -299,7 +301,7 @@ static const CmdLnParser_Option cliOptions[] = {
		.id = CmdLnOptionId_complexityLevel,
		.match = "complexity_level",
		.matchShort = "level",
		.description = "Complexity level, level = (1, 2, 3), will be defined after characterisation.",
		.description = "Complexity level, level = (1, 2, 3), will be defined after\ncharacterisation.",
		},
		{
		.id = CmdLnOptionId_quietModeEnabled,
		@@ -323,13 +325,13 @@ static const CmdLnParser_Option cliOptions[] = {
		.id = CmdLnOptionId_referenceVectorFile,
		.match = "reference_vector_file",
		.matchShort = "rvf",
		.description = "Reference vector trajectory file for simulation of head tracking (only for binaural outputs)",
		.description = "Reference vector trajectory file for simulation of head tracking\n(only for binaural outputs)",
		},
		{
		.id = CmdLnOptionId_exteriorOrientationFile,
		.match = "exterior_orientation_file",
		.matchShort = "exof",
		.description = "External orientation trajectory file for simulation of external orientations",
		.description = "External orientation trajectory file for simulation of external\norientations",
		},
		{
		.id = CmdLnOptionId_framing,
		@@ -341,19 +343,19 @@ static const CmdLnParser_Option cliOptions[] = {
		.id = CmdLnOptionId_syncMdDelay,
		.match = "sync_md_delay",
		.matchShort = "smd",
		.description = "Metadata Synchronization Delay in ms, Default is 0. Quantized by 5ms subframes for TDRenderer (13ms -> 10ms -> 2subframes)",
		.description = "Metadata Synchronization Delay in ms, Default is 0. Quantized by\n5ms subframes for TDRenderer (13ms -> 10ms -> 2subframes)",
		},
		{
		.id = CmdLnOptionId_directivityPatternId,
		.match = "ism_directivity_pattern_id",
		.matchShort = "dpid",
		.description = "Directivity pattern ID(s) = [ID1, ID2, ID3, ID4]. Space-separated list of up to 4 numbers (unsigned integers) can be specified for BINAURAL and BINAURAL_ROOM_REVERB output configuration.\nID1, ID2, ID3, ID4 specify the directivity pattern IDs used for ISMs 1,2,3 and 4 respectively.\nThis options needs to be accompanied by a render_config file, otherwise a default directivity pattern is used.",
		.description = "Directivity pattern ID(s) = [ID1, ID2, ID3, ID4]. Space-separated\nlist of up to 4 numbers (unsigned integers) can be specified for\nBINAURAL and BINAURAL_ROOM_REVERB output.\nID1, ID2, ID3, ID4 specify the directivity pattern IDs used for\nISMs 1,2,3 and 4 respectively. \nThis option needs to be accompanied by a render_config file,\notherwise a default directivity pattern is used.",
		},
		{
		.id = CmdLnOptionId_acousticEnvironmentId,
		.match = "acoustic_environment_id",
		.matchShort = "aeid",
		.description = "Acoustic environment ID (number >= 0) for BINAURAL_ROOM_REVERB output configuration",
		.description = "Acoustic environment ID (number >= 0) for BINAURAL_ROOM_REVERB output.",
		},
		};

		@@ -368,6 +370,8 @@ static IVAS_AUDIO_CONFIG ambisonicsOrderToEnum( const int16_t order );

		static void parseSceneDescriptionFile( char path, char audioFilePath, InputConfig inConfig, IsmPositionProvider positionProvider, MasaFileReader masaReaders, LfeRoutingConfig lfeRoutingConfigs );

		static void parseCombinedFormatInput( InputConfig inConfig, char *configString );

		static ivas_error parseCustomLayoutFile( const char filePath, IVAS_CUSTOM_LS_DATA pLsSetupCustom );

		static CmdlnArgs parseCmdlnArgs( const int argc, char **argv );
		@@ -584,25 +588,7 @@ static void setupWithSingleFormatInput(
		strncpy( audioFilePath, args.inputFilePath, FILENAME_MAX - 1 );

		/* Depending on input format, prepare metadata reading for ISM or MASA */
		if ( args.inConfig.numMasaBuses != 0 )
		{
		if ( args.inConfig.numMasaBuses != args.numInMetadataFiles )
		{
		fprintf( stderr, "Error: all MASA inputs must have a corresponding metadata file" );
		exit( -1 );
		}

		for ( int16_t i = 0; i < args.numInMetadataFiles; ++i )
		{
		masaReaders[i] = MasaFileReader_open( args.inMetadataFilePaths[i] );
		if ( masaReaders[i] == NULL )
		{
		fprintf( stderr, "Could not open MASA metadata file %s\n", args.inMetadataFilePaths[i] );
		exit( -1 );
		}
		}
		}
		else if ( args.inConfig.numAudioObjects != 0 )
		if ( args.inConfig.numAudioObjects != 0 )
		{
		positionProvider->numObjects = args.inConfig.numAudioObjects;
		for ( int16_t i = 0; i < positionProvider->numObjects; ++i )
		@@ -625,6 +611,26 @@ static void setupWithSingleFormatInput(
		}
		}
		}
		if ( args.inConfig.numMasaBuses != 0 )
		{
		if ( args.inConfig.numMasaBuses != ( args.numInMetadataFiles - args.inConfig.numAudioObjects ) )
		{
		fprintf( stderr, "Error: all MASA inputs must have a corresponding metadata file" );
		exit( -1 );
		}

		int16_t reader_idx = 0;
		for ( int16_t i = args.inConfig.numAudioObjects; i < args.numInMetadataFiles; ++i )
		{
		reader_idx = i - args.inConfig.numAudioObjects;
		masaReaders[reader_idx] = MasaFileReader_open( args.inMetadataFilePaths[i] );
		if ( masaReaders[reader_idx] == NULL )
		{
		fprintf( stderr, "Could not open MASA metadata file %s\n", args.inMetadataFilePaths[i] );
		exit( -1 );
		}
		}
		}

		return;
		}
		@@ -1804,16 +1810,11 @@ static bool parseInConfig(
		inConfig->masaBuses[0].gain_dB = 0.0f;
		break;
		case IVAS_AUDIO_CONFIG_OBA:
		/* If input format is objects, parse the characters after "ISM" to get number of objects */
		/* If input format is objects, parse the characters after "ISM" to get
		* number of objects and check for combined formats. */
		{
		char *ptr = NULL;
		inConfig->numAudioObjects = (uint16_t) strtol( inFormatStr + 3, &ptr, 10 );
		if ( ptr == NULL \|\| *ptr != '\0' )
		{
		/* Failed to parse string as a number */
		fprintf( stderr, "Cannot parse string \"%s\" as a valid input format", inFormatStr );
		return false;
		}
		if ( inConfig->numAudioObjects > RENDERER_MAX_ISM_INPUTS )
		{
		fprintf( stderr, "Too many objects at input. Max %d supported.", RENDERER_MAX_ISM_INPUTS );
		@@ -1825,6 +1826,17 @@ static bool parseInConfig(
		inConfig->audioObjects[i].inputChannelIndex = i;
		inConfig->audioObjects[i].gain_dB = 0.0f;
		}
		if ( *ptr != '\0' )
		{
		/* Try to parse combined format */
		parseCombinedFormatInput( inConfig, &ptr );
		}
		if ( ptr == NULL \|\| *ptr != '\0' )
		{
		/* Failed to parse string as a number */
		fprintf( stderr, "Cannot parse string \"%s\" as a valid input format", inFormatStr );
		return false;
		}
		}
		break;
		case IVAS_AUDIO_CONFIG_INVALID:
		@@ -2255,7 +2267,7 @@ static void parseOption(
		}
		break;
		case CmdLnOptionId_inputMetadata:
		assert( numOptionValues <= RENDERER_MAX_ISM_INPUTS );
		assert( numOptionValues <= RENDERER_MAX_ISM_INPUTS + RENDERER_MAX_MASA_INPUTS );
		for ( int16_t i = 0; i < numOptionValues; ++i )
		{
		strncpy( args->inMetadataFilePaths[i], optionValues[i], RENDERER_MAX_CLI_ARG_LENGTH - 1 );
		@@ -3086,6 +3098,41 @@ static void parseMasa(
		return;
		}

		static void parseCombinedFormatInput(
		InputConfig *inConfig,
		char **configString )
		{
		IVAS_AUDIO_CONFIG audioConfig;
		audioConfig = parseAudioConfig( *configString );

		if ( audioConfig == IVAS_AUDIO_CONFIG_FOA \|\| audioConfig == IVAS_AUDIO_CONFIG_HOA2 \|\| audioConfig == IVAS_AUDIO_CONFIG_HOA3 )
		{
		/* OSBA */
		inConfig->numAmbisonicsBuses = 1;
		inConfig->ambisonicsBuses[0].audioConfig = audioConfig;
		inConfig->ambisonicsBuses[0].inputChannelIndex = inConfig->numAudioObjects;
		inConfig->ambisonicsBuses[0].gain_dB = -6.f;
		*configString += 4;

		/* Modify input gain for objects too */
		for ( int16_t i = 0; i < inConfig->numAudioObjects; ++i )
		{
		inConfig->audioObjects[i].gain_dB = -6.f;
		}
		}
		else if ( audioConfig == IVAS_AUDIO_CONFIG_MASA1 \|\| audioConfig == IVAS_AUDIO_CONFIG_MASA2 )
		{
		/* OMASA */
		inConfig->numMasaBuses = 1;
		inConfig->masaBuses[0].audioConfig = audioConfig;
		inConfig->masaBuses[0].inputChannelIndex = inConfig->numAudioObjects;
		inConfig->masaBuses[0].gain_dB = 0.0f;
		*configString += 5;
		}

		return;
		}

		static ivas_error parseCustomLayoutFile(
		const char *filePath,
		IVAS_CUSTOM_LS_DATA *pLsSetupCustom )
		@@ -3302,10 +3349,13 @@ static void printSupportedAudioConfigs( void )
		"HOA2",
		"HOA3",
		"ISMx (input only)",
		"ISMxSBAy (OSBA, input only)",
		"ISMxMASAy (OMASA, input only)",
		"MASAx",
		"BINAURAL (output only)",
		"BINAURAL_ROOM_IR (output only)",
		"BINAURAL_ROOM_REVERB (output only)",
		"META (Scene description, input only)",
		};

		fprintf( stdout, "Supported audio formats:\n" );
		@@ -3313,6 +3363,7 @@ static void printSupportedAudioConfigs( void )
		{
		fprintf( stdout, "%s\n", supportedFormats[i] );
		}
		fprintf( stdout, "\n" );

		return;
		}

lib_com/basop32.c

+2 −2

Original line number	Diff line number	Diff line
		@@ -3582,13 +3582,13 @@ Word16 i_mult_o( Word16 a, Word16 b, Flag *Overflow )
		#ifdef ORIGINAL_G7231
		return a * b;
		#else
		register Word32 c = a * b;
		register Word32 c = (Word32) a * b;
		return saturate_o( c, Overflow );
		#endif
		}
		Word16 i_mult( Word16 a, Word16 b )
		{
		return i_mult_o( a, b, NULL );
		return i_mult_sat( a, b );
		}
		Word16 i_mult_sat( Word16 a, Word16 b )
		{

lib_com/basop_util.c

+20 −2

Original line number	Diff line number	Diff line
		@@ -1062,7 +1062,6 @@ Word32 BASOP_Util_Divide3232_Scale_cadence( Word32 x, Word32 y, Word16 *s )
		return z;
		}

		#ifdef DIV32_OPT_NEWTON
		Word32 div_w_newton( Word32 num, Word32 den );
		/*
		Table of 256 precalculated estimates to be used by the "div_w_newton"
		@@ -1462,7 +1461,6 @@ Word32 BASOP_Util_Divide3232_Scale_newton( Word32 x, Word32 y, Word16 *s )

		return z;
		}
		#endif /* DIV32_OPT_NEWTON */

		Word16 BASOP_Util_Divide3232_Scale( Word32 x, Word32 y, Word16 *s )
		{
		@@ -1918,6 +1916,26 @@ Word16 findIndexOfMinWord32( Word32 *x, const Word16 len )
		return indx;
		}

		Word16 findIndexOfMinWord64( Word64 *x, const Word16 len )
		{
		Word16 i, indx;


		indx = 0;
		move16();
		FOR( i = 1; i < len; i++ )
		{
		if ( LT_64( x[i], x[indx] ) )
		{
		indx = i;
		move16();
		}
		}


		return indx;
		}


		Word16 imult1616( Word16 x, Word16 y )
		{