Merge branch 'main' into add-plc-conditions-for-OSBA-in-self-test (90e32cf7) · Commits · IVAS Codec Public Collaboration / IVAS Codec

.gitlab-ci.yml

+13 −11

Original line number	Diff line number	Diff line
		@@ -909,6 +909,7 @@ clang-format-check:
		- run: git apply $PATCH_FILE_NAME\n
		- commit new changes"

		- format_problems=0
		- scripts/check-format.sh -af -p 8 \|\| format_problems=$?
		- if [ $format_problems == 0 ] ; then exit 0; fi

		@@ -926,7 +927,7 @@ clang-format-check:
		expire_in: 1 day
		paths:
		- tmp-formatting-fix/
		when: always
		when: on_failure
		name: "$ARTIFACT_BASE_NAME"
		expose_as: "formatting patch"

		@@ -948,18 +949,19 @@ check-first-frame-is-sid:
		- cmake .
		- make -j

		# TODO: for some MASA modes, we currently do not have testvectors that actually trigger DTX
		# SBA modes are run separately to use shorter part of file
		- exit_code_no_sba=0
		- exit_code_sba=0
		- modes=$(scripts/runIvasCodec.py -l \| grep dtx \| grep -v MASA \| grep -vE "FOA\|HOA" )
		- scripts/runIvasCodec.py -z console -p scripts/config/ci_linux_sidstart_test.json -m $modes -s --bs_length 200 -U 0:20 \|\| exit_code_no_sba=$?
		- exit_code_hoa=0
		- exit_code_foa=0

		# run all modes and cut bitstream to start with an SID. Use stereo output to limit runtime, test is only about decoding
		- modes=$(scripts/runIvasCodec.py -l \| grep dtx \| grep -vE "FOA\|HOA" )
		- scripts/runIvasCodec.py -z console -p scripts/config/ci_linux_sidstart_test.json -m $modes -s --bs_length 200 -U 0:20 --oc stereo \|\| exit_code_no_sba=$?
		- modes=$(scripts/runIvasCodec.py -l \| grep dtx \| grep -E "HOA")
		- scripts/runIvasCodec.py -z console -p scripts/config/ci_linux_sidstart_test.json -m $modes -s --bs_length 100 -U 70:80 \|\| exit_code_sba=$?
		# need to do FOA separately as VAD apparently behaves differently there
		- scripts/runIvasCodec.py -z console -p scripts/config/ci_linux_sidstart_test.json -m $modes -s --bs_length 100 -U 70:80 --oc stereo \|\| exit_code_hoa=$?
		- modes=$(scripts/runIvasCodec.py -l \| grep dtx \| grep "FOA")
		- scripts/runIvasCodec.py -z console -p scripts/config/ci_linux_sidstart_test.json -m $modes -s --bs_length 100 -U 75:110 \|\| exit_code_sba=$?
		- if [ $exit_code_no_sba -ne 0 ] \|\| [ $exit_code_sba -ne 0 ]; then exit 1; fi
		- scripts/runIvasCodec.py -z console -p scripts/config/ci_linux_sidstart_test.json -m $modes -s --bs_length 100 -U 75:110 --oc stereo \|\| exit_code_foa=$?

		- if [ $exit_code_no_sba -ne 0 ] \|\| [ $exit_code_hoa -ne 0 ] \|\| [ $exit_code_hoa -ne 0 ]; then exit 1; fi
		artifacts:
		paths:
		- out/logs
		@@ -1332,7 +1334,7 @@ sanitizer-test-masa:
		- if: $SANITIZER_SCHEDULE_A
		when: delayed
		start_in: 21 hours
		timeout: 3 hours
		timeout: 5 hours
		script:
		- *update-ltv-repo
		- python3 ci/run_scheduled_sanitizer_test.py MASA $OUT_FORMATS_CHANNEL_BASED $OUT_FORMATS_SCENE_BASED $OUT_FORMATS_BINAURAL EXT --tests $SANITIZER_TESTS

apps/decoder.c

+28 −58

Original line number	Diff line number	Diff line
		@@ -145,10 +145,9 @@ typedef struct
		#endif
		#endif
		uint16_t acousticEnvironmentId;
		#ifdef FIX_708_DPID_COMMAND_LINE
		int16_t Opt_dpid_on;
		#endif
		uint16_t directivityPatternId[IVAS_MAX_NUM_OBJECTS];

		} DecArguments;


		@@ -197,7 +196,11 @@ int main(
		RotFileReader *refRotReader = NULL;
		Vector3PairFileReader *referenceVectorReader = NULL;
		ivas_error error = IVAS_ERR_UNKNOWN;
		#ifdef FIX_847_OUTPUT_PCM_BUFFER
		int16_t *pcmBuf = NULL;
		#else
		int16_t pcmBuf[MAX_OUTPUT_PCM_BUFFER_SIZE];
		#endif
		#ifdef SPLIT_REND_WITH_HEAD_ROT
		uint8_t splitRendBitsBuf[IVAS_MAX_SPLIT_REND_BITS_BUFFER_SIZE_IN_BYTES];
		#endif
		@@ -434,15 +437,8 @@ int main(
		* Configure the decoder
		------------------------------------------------------------------------------------------/

		if ( ( error = IVAS_DEC_Configure( hIvasDec, arg.output_Fs, arg.outputConfig, arg.tsmEnabled, arg.enable5ms,
		arg.customLsOutputEnabled, arg.hrtfReaderEnabled, arg.enableHeadRotation, arg.enableExternalOrientation, arg.orientation_tracking, arg.renderConfigEnabled, arg.Opt_non_diegetic_pan, arg.non_diegetic_pan_gain,
		#ifdef FIX_708_DPID_COMMAND_LINE
		arg.Opt_dpid_on,
		#endif
		#ifdef FIX_708_AEID_COMMAND_LINE
		arg.acousticEnvironmentId,
		#endif
		arg.delayCompensationEnabled ) ) != IVAS_ERR_OK )
		if ( ( error = IVAS_DEC_Configure( hIvasDec, arg.output_Fs, arg.outputConfig, arg.tsmEnabled, arg.enable5ms, arg.customLsOutputEnabled, arg.hrtfReaderEnabled, arg.enableHeadRotation, arg.enableExternalOrientation,
		arg.orientation_tracking, arg.renderConfigEnabled, arg.Opt_non_diegetic_pan, arg.non_diegetic_pan_gain, arg.Opt_dpid_on, arg.acousticEnvironmentId, arg.delayCompensationEnabled ) ) != IVAS_ERR_OK )
		{
		fprintf( stderr, "\nConfigure failed: %s\n\n", IVAS_DEC_GetErrorMessage( error ) );
		goto cleanup;
		@@ -640,11 +636,7 @@ int main(

		if ( ( error = RenderConfigReader_getDirectivity( renderConfigReader, arg.directivityPatternId, renderConfig.directivity ) ) != IVAS_ERR_OK )
		{
		#ifdef FIX_730_DPID_NOT_SET_CORRECTLY
		fprintf( stderr, "Failed to get directivity patterns for one or more of IDs: %d %d %d %d\n\n", arg.directivityPatternId[0], arg.directivityPatternId[1], arg.directivityPatternId[2], arg.directivityPatternId[3] );
		#else
		fprintf( stderr, "Failed to get directivity for objects: %d %d %d %d\n\n", arg.directivityPatternId[0], arg.directivityPatternId[1], arg.directivityPatternId[2], arg.directivityPatternId[3] );
		#endif
		goto cleanup;
		}

		@@ -658,11 +650,7 @@ int main(
		goto cleanup;
		}
		}
		#ifdef FIX_708_AEID_COMMAND_LINE
		else
		#else
		else if ( error != IVAS_ERR_ACOUSTIC_ENVIRONMENT_MISSING )
		#endif
		{
		fprintf( stderr, "Failed to get acoustic environment with ID: %d\n\n", arg.acousticEnvironmentId );
		goto cleanup;
		@@ -759,6 +747,21 @@ int main(
		}
		}

		#ifdef FIX_847_OUTPUT_PCM_BUFFER
		/------------------------------------------------------------------------------------------
		* Allocate output data buffer
		------------------------------------------------------------------------------------------/

		int16_t pcmBufSize;
		if ( ( error = IVAS_DEC_GetOutputBufferSize( hIvasDec, &pcmBufSize ) ) != IVAS_ERR_OK )
		{
		fprintf( stderr, "\nGetOutputBufferSize failed: %s\n\n", IVAS_DEC_GetErrorMessage( error ) );
		goto cleanup;
		}

		pcmBuf = malloc( pcmBufSize * sizeof( int16_t ) );

		#endif

		/-----------------------------------------------------------------
		* Decoding
		@@ -813,10 +816,14 @@ int main(

		cleanup:

		#ifdef FIX_847_OUTPUT_PCM_BUFFER
		free( pcmBuf );
		#endif

		#ifdef DEBUG_SBA_AUDIO_DUMP
		IVAS_DEC_GetSbaDebugParams( hIvasDec, &numOutChannels, &numTransportChannels, &pca_ingest_channels );
		#endif

		#endif
		if ( arg.hrtfReaderEnabled )
		{
		IVAS_DEC_HRTF_HANDLE hHrtfTD = NULL;
		@@ -1008,9 +1015,7 @@ static bool parseCmdlIVAS_dec(

		arg->renderConfigEnabled = false;
		arg->renderConfigFilename = NULL;
		#ifdef FIX_708_DPID_COMMAND_LINE
		arg->Opt_dpid_on = 0;
		#endif

		#ifdef SPLIT_REND_WITH_HEAD_ROT
		arg->outputMdFilename = NULL;
		@@ -1034,18 +1039,10 @@ static bool parseCmdlIVAS_dec(
		arg->noBadFrameDelay = false;
		#endif
		#endif
		#ifdef FIX_708_AEID_COMMAND_LINE
		arg->acousticEnvironmentId = 65535;
		#else
		arg->acousticEnvironmentId = 0;
		#endif
		for ( i = 0; i < IVAS_MAX_NUM_OBJECTS; ++i )
		{
		#ifdef FIX_730_DPID_NOT_SET_CORRECTLY
		arg->directivityPatternId[i] = 65535;
		#else
		arg->directivityPatternId[i] = 0;
		#endif
		}

		/-----------------------------------------------------------------
		@@ -1417,7 +1414,6 @@ static bool parseCmdlIVAS_dec(
		else if ( strcmp( argv_to_upper, "-AEID" ) == 0 )
		{
		++i;
		#ifdef FIX_708_AEID_COMMAND_LINE
		if ( argc - i <= 4 )
		{
		fprintf( stderr, "Error: Acoustic environment ID not specified!\n\n" );
		@@ -1431,22 +1427,17 @@ static bool parseCmdlIVAS_dec(
		usage_dec();
		return false;
		}
		#endif
		arg->acousticEnvironmentId = (int16_t) atoi( argv[i++] );
		}
		else if ( strcmp( argv_to_upper, "-DPID" ) == 0 )
		{
		#ifdef FIX_708_DPID_COMMAND_LINE
		int16_t id;
		int16_t id, tmp;

		arg->Opt_dpid_on = 1;
		#endif
		++i;
		int16_t tmp;
		tmp = 0;
		while ( is_number( argv[i + tmp] ) && tmp < IVAS_MAX_NUM_OBJECTS )
		{
		#ifdef FIX_708_DPID_COMMAND_LINE
		id = (int16_t) atoi( argv[i + tmp] );

		if ( !is_digits_only( argv[i + tmp] ) \|\| id < 0 )
		@@ -1457,13 +1448,9 @@ static bool parseCmdlIVAS_dec(
		}

		arg->directivityPatternId[tmp] = id;
		#else
		arg->directivityPatternId[tmp] = (int16_t) atoi( argv[i + tmp] );
		#endif
		++tmp;
		}

		#ifdef FIX_708_DPID_COMMAND_LINE
		if ( tmp == 0 )
		{
		if ( argc - i <= 4 \|\| argv[i][0] == '-' )
		@@ -1480,7 +1467,6 @@ static bool parseCmdlIVAS_dec(
		return false;
		}
		}
		#endif

		i += tmp;
		}
		@@ -1576,12 +1562,7 @@ static bool parseCmdlIVAS_dec(
		return false;
		}

		#ifdef FIX_807_VARIABLE_SPEED_DECODING
		if ( ( !arg->enableHeadRotation ) && ( !arg->enableExternalOrientation ) && ( !arg->tsmEnabled ) )
		#else
		if ( ( !arg->enableHeadRotation ) && ( !arg->enableExternalOrientation ) )
		#endif

		{
		arg->enable5ms = false;
		}
		@@ -1675,15 +1656,11 @@ static void usage_dec( void )
		fprintf( stdout, "-force R : Force specific binaural rendering mode, R = (TDREND, CLDFBREND),\n" );
		#endif
		fprintf( stdout, "-exof File : External orientation File for external orientation trajectory\n" );
		#ifdef FIX_708_DPID_COMMAND_LINE
		fprintf( stdout, "-dpid ID : Directivity pattern ID(s) = [ID1, ID2, ID3, ID4]. Space-separated list of up\n" );
		fprintf( stdout, " to 4 numbers (unsigned integers) can be specified for BINAURAL and BINAURAL_ROOM_REVERB\n" );
		fprintf( stdout, " output configuration. ID1, ID2, ID3, ID4 specify the directivity pattern IDs used for\n" );
		fprintf( stdout, " ISMs 1,2,3 and 4 respectively. This options needs to be accompanied by a render_config file,\n" );
		fprintf( stdout, " otherwise a default directivity pattern is used.\n" );
		#else
		fprintf( stdout, "-dpid ID : Directivity pattern ID(s) (space-separated list of up to 4 numbers can be specified) for binaural output configuration\n" );
		#endif
		fprintf( stdout, "-aeid ID : Acoustic environment ID (number >= 0) for BINAURAL_ROOM_REVERB output configuration\n" );
		fprintf( stdout, "-level level : Complexity level, level = (1, 2, 3), will be defined after characterisation. \n" );
		fprintf( stdout, " Currently, all values default to level 3 (full functionality).\n" );
		@@ -2481,18 +2458,11 @@ static ivas_error decodeG192(
		{
		// update_mem(); TODO: verify this
		update_wmops();
		#ifdef FIX_RAM_COUNTING_5MS_RENDERING
		update_mem();
		#ifdef MEM_COUNT_DETAILS
		export_mem( "mem_analysis.csv" );
		#endif
		#endif
		}
		#ifndef FIX_RAM_COUNTING_5MS_RENDERING
		#ifdef MEM_COUNT_DETAILS
		export_mem( "mem_analysis.csv" );
		#endif
		#endif
		#endif
		}

apps/renderer.c

+8 −22

Original line number	Diff line number	Diff line
		@@ -100,7 +100,7 @@ typedef struct
		uint16_t numObjects;
		IsmFileReader *ismReaders[RENDERER_MAX_ISM_INPUTS];
		uint32_t numPositions[RENDERER_MAX_ISM_INPUTS];
		IVAS_REND_AudioObjectPosition positions[RENDERER_MAX_ISM_INPUTS]; / size: [RENDERER_MAX_ISM_INPUTS][numPositions[object_index]] */
		IVAS_ISM_METADATA positions[RENDERER_MAX_ISM_INPUTS]; / size: [RENDERER_MAX_ISM_INPUTS][numPositions[object_index]] */
		uint16_t positionDurations[RENDERER_MAX_ISM_INPUTS]; / size: [RENDERER_MAX_ISM_INPUTS][numPositions[object_index]] */
		uint32_t currentPositionIdxs[RENDERER_MAX_ISM_INPUTS]; /* Index of current position as listed in the metadata file */
		uint16_t durationCounters[RENDERER_MAX_ISM_INPUTS]; /* Number of frames spent at current position */
		@@ -123,7 +123,7 @@ typedef struct

		typedef struct
		{
		IVAS_REND_AudioObjectPosition positions[RENDERER_MAX_ISM_INPUTS];
		IVAS_ISM_METADATA positions[RENDERER_MAX_ISM_INPUTS];
		int16_t numObjects;
		} ObjectPositionBuffer;

		@@ -384,11 +384,7 @@ static const CmdLnParser_Option cliOptions[] = {
		.id = CmdLnOptionId_directivityPatternId,
		.match = "ism_directivity_pattern_id",
		.matchShort = "dpid",
		#ifdef FIX_708_DPID_COMMAND_LINE
		.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.",
		#else
		.description = "Directivity pattern ID(s) (space-separated list of up to 4 numbers can be specified) for binaural output configuration",
		#endif
		},
		{
		.id = CmdLnOptionId_acousticEnvironmentId,
		@@ -437,7 +433,7 @@ static int8_t parseUint32( const char line, uint32_t ret );

		static int8_t parseInt32( const char line, int32_t ret );

		static void parseObjectPosition( char line, IVAS_REND_AudioObjectPosition position, uint16_t *positionDuration );
		static void parseObjectPosition( char line, IVAS_ISM_METADATA position, uint16_t *positionDuration );

		static void parseMetadata( char metadataString, char inDir, InputConfig inConfig, IsmPositionProvider positionProvider, MasaFileReader masaReaders, LfeRoutingConfig lfeRoutingConfigs );

		@@ -2640,18 +2636,10 @@ static CmdlnArgs defaultArgs(

		for ( int32_t i = 0; i < RENDERER_MAX_ISM_INPUTS; ++i )
		{
		#ifdef FIX_730_DPID_NOT_SET_CORRECTLY
		args.directivityPatternId[i] = 65535;
		#else
		args.directivityPatternId[i] = 0;
		#endif
		}

		#ifdef FIX_708_AEID_COMMAND_LINE
		args.acousticEnvironmentId = 65535;
		#else
		args.acousticEnvironmentId = 0;
		#endif

		return args;
		}
		@@ -2819,13 +2807,11 @@ static void parseOption(
		break;
		case CmdLnOptionId_acousticEnvironmentId:
		assert( numOptionValues == 1 );
		#ifdef FIX_708_AEID_COMMAND_LINE
		if ( !is_digits_only( optionValues[0] ) )
		{
		fprintf( stderr, "Invalid acousting environment ID specified: %s\n", optionValues[0] );
		exit( -1 );
		}
		#endif
		args->acousticEnvironmentId = (int16_t) strtol( optionValues[0], NULL, 10 );
		break;
		case CmdLnOptionId_syncMdDelay:
		@@ -3297,7 +3283,7 @@ static void parseOptionalInputValues(

		static void parseObjectPosition(
		char *line,
		IVAS_REND_AudioObjectPosition *position,
		IVAS_ISM_METADATA *position,
		uint16_t *positionDuration )
		{
		char *endptr;
		@@ -3346,7 +3332,7 @@ static void parseIsm(
		if ( parseUint32( line, &numberOfObjectPositionsToRead ) == 0 )
		{
		positionProvider->numPositions[idx] = numberOfObjectPositionsToRead;
		positionProvider->positions[idx] = malloc( numberOfObjectPositionsToRead * sizeof( IVAS_REND_AudioObjectPosition ) );
		positionProvider->positions[idx] = malloc( numberOfObjectPositionsToRead * sizeof( IVAS_ISM_METADATA ) );
		positionProvider->positionDurations[idx] = malloc( numberOfObjectPositionsToRead * sizeof( uint16_t ) );

		for ( i = 0; i < numberOfObjectPositionsToRead; ++i )

ci/ivas_voip_be_test.sh

+1 −1

Original line number	Diff line number	Diff line
		@@ -60,7 +60,7 @@ if [ $BUILD -eq 1 ];then
		fi

		# Configuration
		modes=('SBA_b128_wb_cbr' 'MC_7_1_b96_fb_cbr' 'ISM2_b48_fb_cbr')
		modes=('HOA3_b128_wb_cbr' 'MC_7_1_b96_fb_cbr' 'ISM2_b48_fb_cbr')
		output_formats=('STEREO' 'FOA' '7_1' 'HOA3')
		limit_input_to_x_seconds=30
		verbosity_cmd="-z console"

lib_com/bits_alloc.c

+0 −4

Original line number	Diff line number	Diff line
		@@ -476,10 +476,8 @@ static ivas_error acelp_FCB_allocator(
		}

		/* try to increase the FCB of the first subframe in cases when the next step is lower than the current step */
		#ifdef NONBE_FIX_694_OMASA_ACELP
		if ( fixed_cdk_index[sfr] < ACELP_FIXED_CDK_NB - 1 )
		{
		#endif
		step = fcb_table( fixed_cdk_index[sfr] + 1, L_subfr ) - fcb_table( fixed_cdk_index[sfr], L_subfr );
		if ( *nBits >= step && cdbk >= 0 )
		{
		@@ -493,9 +491,7 @@ static ivas_error acelp_FCB_allocator(
		*nBits -= step;
		}
		}
		#ifdef NONBE_FIX_694_OMASA_ACELP
		}
		#endif
		}
		/* TRANSITION coding: allocate highest FCBQ bit-budget to the subframe with the glottal-shape codebook */
		if ( tc_subfr >= L_SUBFR )