Merge branch 'main' into basop-ci/add-peaq-scores-comparison (11ac0cdf) · Commits · IVAS Codec Public Collaboration / IVAS Codec

.gitlab-ci.yml

+17 −15

Original line number	Diff line number	Diff line
		@@ -1012,11 +1012,13 @@ voip-be-on-merge-request:
		- .test-job-linux-needs-testv-dir
		- .rules-merge-request
		stage: test
		needs: ["build-codec-linux-make", "codec-smoke-test"]
		needs: ["build-codec-linux-make"]
		timeout: "10 minutes"
		script:
		- *print-common-info
		- bash ci/ivas_voip_be_test.sh
		- make clean
		- make -j
		- python3 -m pytest tests/test_be_for_jbm_neutral_dly_profile.py

		clang-format-check:
		extends:
		@@ -2205,7 +2207,7 @@ coverage-test-on-main-scheduled:
		- python3 -m pytest $TESTS_DIR_CODEC_BE_ON_MR/test_param_file.py -v -n auto --update_ref 1 -m create_ref --param_file scripts/config/self_test_evs.prm --ref_encoder_path ./IVAS_cod --ref_decoder_path ./IVAS_dec
		- bash ci/smoke_test.sh coverage
		- python3 -m pytest $TESTS_DIR_CODEC_BE_ON_MR/test_param_file.py -v -n auto --update_ref 1 -m create_ref --param_file scripts/config/self_test_ltv.prm --ref_encoder_path ./IVAS_cod --ref_decoder_path ./IVAS_dec
		- bash ci/ivas_voip_be_test.sh coverage
		- python3 -m pytest tests/test_be_for_jbm_neutral_dly_profile.py
		- lcov -c -d obj -o coverage.info
		# remove apps and lib_util files from coverage
		- lcov -r coverage.info "apps" -o coverage.info
		@@ -2289,8 +2291,8 @@ complexity-stereo-in-stereo-out:
		- *complexity-measurements-setup
		- in_format=stereo
		- out_format=stereo
		- bash ci/complexity_measurements/getWmops.sh "$in_format" "$out_format"
		- ret_val=$?
		- ret_val=0
		- bash ci/complexity_measurements/getWmops.sh "$in_format" "$out_format" \|\| ret_val=$?
		- *complexity-measurements-prepare-artifacts
		- exit $ret_val

		@@ -2307,8 +2309,8 @@ complexity-ism-in-binaural-out:
		- *complexity-measurements-setup
		- in_format=ISM
		- out_format=BINAURAL
		- bash ci/complexity_measurements/getWmops.sh "ISM+1 ISM+2 ISM+3 ISM+4" "$out_format"
		- ret_val=$?
		- ret_val=0
		- bash ci/complexity_measurements/getWmops.sh "ISM+1 ISM+2 ISM+3 ISM+4" "$out_format" \|\| ret_val=$?
		- *complexity-measurements-prepare-artifacts
		- exit $ret_val

		@@ -2325,8 +2327,8 @@ complexity-sba-hoa3-in-hoa3-out:
		- *complexity-measurements-setup
		- in_format=HOA3
		- out_format=HOA3
		- bash ci/complexity_measurements/getWmops.sh "$in_format" "$out_format"
		- ret_val=$?
		- ret_val=0
		- bash ci/complexity_measurements/getWmops.sh "$in_format" "$out_format" \|\| ret_val=$?
		- *complexity-measurements-prepare-artifacts
		- exit $ret_val

		@@ -2343,8 +2345,8 @@ complexity-mc-in-7_1_4-out:
		- *complexity-measurements-setup
		- in_format=MC
		- out_format=7_1_4
		- bash ci/complexity_measurements/getWmops.sh "$in_format" "$out_format"
		- ret_val=$?
		- ret_val=0
		- bash ci/complexity_measurements/getWmops.sh "$in_format" "$out_format" \|\| ret_val=$?
		- *complexity-measurements-prepare-artifacts
		- exit $ret_val

		@@ -2361,8 +2363,8 @@ complexity-masa-in-7_1_4-out:
		- *complexity-measurements-setup
		- in_format=MASA
		- out_format=7_1_4
		- bash ci/complexity_measurements/getWmops.sh "$in_format" "$out_format"
		- ret_val=$?
		- ret_val=0
		- bash ci/complexity_measurements/getWmops.sh "$in_format" "$out_format" \|\| ret_val=$?
		- *complexity-measurements-prepare-artifacts
		- exit $ret_val

		@@ -2379,8 +2381,8 @@ complexity-StereoDmxEVS-stereo-in-mono-out:
		- *complexity-measurements-setup
		- in_format=StereoDmxEVS
		- out_format=mono
		- bash ci/complexity_measurements/getWmops.sh "$in_format" "$out_format"
		- ret_val=$?
		- ret_val=0
		- bash ci/complexity_measurements/getWmops.sh "$in_format" "$out_format" \|\| ret_val=$?
		- *complexity-measurements-prepare-artifacts
		- exit $ret_val

apps/decoder.c

+48 −0

Original line number	Diff line number	Diff line
		@@ -128,7 +128,11 @@ typedef struct
		bool customLsOutputEnabled;
		char *customLsSetupFilename;
		int16_t orientation_tracking;
		#ifdef FIX_745_FIX_DATA_TYPE_CONVERSION
		bool non_diegetic_pan_enabled;
		#else
		int16_t Opt_non_diegetic_pan;
		#endif
		float non_diegetic_pan_gain;
		bool renderConfigEnabled;
		char *renderConfigFilename;
		@@ -148,7 +152,11 @@ typedef struct
		#endif
		#endif
		AcousticEnvironmentSequence aeSequence;
		#ifdef FIX_745_FIX_DATA_TYPE_CONVERSION
		bool dpidEnabled;
		#else
		int16_t Opt_dpid_on;
		#endif
		uint16_t directivityPatternId[IVAS_MAX_NUM_OBJECTS];

		} DecArguments;
		@@ -407,8 +415,12 @@ int main(
		if ( arg.renderConfigEnabled )
		{
		/* sanity check */
		#ifdef FIX_745_FIX_DATA_TYPE_CONVERSION
		if ( arg.outputConfig != IVAS_AUDIO_CONFIG_BINAURAL && arg.outputConfig != IVAS_AUDIO_CONFIG_BINAURAL_ROOM_IR && arg.outputConfig != IVAS_AUDIO_CONFIG_BINAURAL_ROOM_REVERB && arg.outputConfig != IVAS_AUDIO_CONFIG_BINAURAL_SPLIT_CODED && arg.outputConfig != IVAS_AUDIO_CONFIG_BINAURAL_SPLIT_PCM && arg.non_diegetic_pan_enabled == false )
		#else
		if ( arg.outputConfig != IVAS_AUDIO_CONFIG_BINAURAL && arg.outputConfig != IVAS_AUDIO_CONFIG_BINAURAL_ROOM_IR && arg.outputConfig != IVAS_AUDIO_CONFIG_BINAURAL_ROOM_REVERB && arg.outputConfig != IVAS_AUDIO_CONFIG_BINAURAL_SPLIT_CODED && arg.outputConfig != IVAS_AUDIO_CONFIG_BINAURAL_SPLIT_PCM &&
		arg.Opt_non_diegetic_pan == 0 )
		#endif
		{
		fprintf( stderr, "\nError: Renderer configuration file cannot be used in this output configuration.\n\n" );
		goto cleanup;
		@@ -427,8 +439,14 @@ int main(

		asked_frame_size = arg.renderFramesize;
		uint16_t aeID = arg.aeSequence.count > 0 ? arg.aeSequence.pID[0] : 65535;

		#ifdef FIX_745_FIX_DATA_TYPE_CONVERSION
		if ( ( error = IVAS_DEC_Configure( hIvasDec, arg.output_Fs, arg.outputConfig, arg.tsmEnabled, arg.renderFramesize, arg.customLsOutputEnabled, arg.hrtfReaderEnabled, arg.enableHeadRotation, arg.enableExternalOrientation, arg.orientation_tracking, arg.renderConfigEnabled, arg.non_diegetic_pan_enabled, arg.non_diegetic_pan_gain,
		arg.dpidEnabled, aeID, arg.delayCompensationEnabled ) ) != IVAS_ERR_OK )
		#else
		if ( ( error = IVAS_DEC_Configure( hIvasDec, arg.output_Fs, arg.outputConfig, arg.tsmEnabled, arg.renderFramesize, 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, aeID, arg.delayCompensationEnabled ) ) != IVAS_ERR_OK )
		#endif
		{
		fprintf( stderr, "\nConfigure failed: %s\n\n", IVAS_DEC_GetErrorMessage( error ) );
		goto cleanup;
		@@ -594,8 +612,12 @@ int main(

		/* sanity check */
		if ( arg.outputConfig != IVAS_AUDIO_CONFIG_BINAURAL && arg.outputConfig != IVAS_AUDIO_CONFIG_BINAURAL_ROOM_IR && arg.outputConfig != IVAS_AUDIO_CONFIG_BINAURAL_ROOM_REVERB &&
		#ifdef FIX_745_FIX_DATA_TYPE_CONVERSION
		arg.outputConfig != IVAS_AUDIO_CONFIG_BINAURAL_SPLIT_CODED && arg.outputConfig != IVAS_AUDIO_CONFIG_BINAURAL_SPLIT_PCM && arg.non_diegetic_pan_enabled == false )
		#else
		arg.outputConfig != IVAS_AUDIO_CONFIG_BINAURAL_SPLIT_CODED && arg.outputConfig != IVAS_AUDIO_CONFIG_BINAURAL_SPLIT_PCM &&
		arg.Opt_non_diegetic_pan == 0 )
		#endif
		{
		fprintf( stderr, "\nExternal Renderer Config is supported only when binaural output configurations is used as output OR when Split rendering mode is enabled. Exiting. \n" );
		goto cleanup;
		@@ -1066,12 +1088,20 @@ static bool parseCmdlIVAS_dec(

		arg->renderConfigEnabled = false;
		arg->renderConfigFilename = NULL;
		#ifdef FIX_745_FIX_DATA_TYPE_CONVERSION
		arg->dpidEnabled = false;
		#else
		arg->Opt_dpid_on = 0;
		#endif

		arg->outputMdFilename = NULL;

		arg->inputFormat = IVAS_DEC_INPUT_FORMAT_G192;
		#ifdef FIX_745_FIX_DATA_TYPE_CONVERSION
		arg->non_diegetic_pan_enabled = false;
		#else
		arg->Opt_non_diegetic_pan = 0;
		#endif
		arg->non_diegetic_pan_gain = 0.f;
		arg->tsmEnabled = false;
		arg->renderFramesize = IVAS_RENDER_FRAMESIZE_20MS;
		@@ -1410,7 +1440,11 @@ static bool parseCmdlIVAS_dec(
		else if ( strcmp( argv_to_upper, "-NON_DIEGETIC_PAN" ) == 0 )
		{
		i++;
		#ifdef FIX_745_FIX_DATA_TYPE_CONVERSION
		arg->non_diegetic_pan_enabled = true;
		#else
		arg->Opt_non_diegetic_pan = 1;
		#endif
		strncpy( argv_to_upper, argv[i], sizeof( argv_to_upper ) - 1 );
		argv_to_upper[sizeof( argv_to_upper ) - 1] = '\0';
		to_upper( argv_to_upper );
		@@ -1507,7 +1541,11 @@ static bool parseCmdlIVAS_dec(
		{
		int16_t id, tmp;

		#ifdef FIX_745_FIX_DATA_TYPE_CONVERSION
		arg->dpidEnabled = true;
		#else
		arg->Opt_dpid_on = 1;
		#endif
		++i;
		tmp = 0;
		while ( is_number( argv[i + tmp] ) && tmp < IVAS_MAX_NUM_OBJECTS )
		@@ -1587,7 +1625,12 @@ static bool parseCmdlIVAS_dec(
		arg->customLsSetupFilename = argv[i];
		}
		i++;

		#ifdef FIX_745_FIX_DATA_TYPE_CONVERSION
		if ( arg->non_diegetic_pan_enabled && arg->outputConfig != IVAS_AUDIO_CONFIG_STEREO )
		#else
		if ( ( arg->Opt_non_diegetic_pan ) && ( arg->outputConfig != IVAS_AUDIO_CONFIG_STEREO ) )
		#endif
		{
		fprintf( stderr, "Error: non-diegetic panning is supported in stereo only\n\n" );
		usage_dec();
		@@ -1605,7 +1648,12 @@ static bool parseCmdlIVAS_dec(
		{
		arg->outputConfig = IVAS_AUDIO_CONFIG_MONO;
		arg->decMode = IVAS_DEC_MODE_EVS;

		#ifdef FIX_745_FIX_DATA_TYPE_CONVERSION
		if ( arg->non_diegetic_pan_enabled )
		#else
		if ( ( arg->Opt_non_diegetic_pan ) )
		#endif
		{
		arg->outputConfig = IVAS_AUDIO_CONFIG_STEREO;
		}

ci/basop-pages/create_report_pages.py

+19 −6

Original line number	Diff line number	Diff line
		@@ -92,11 +92,12 @@ ARROW_DOWN = '<span class="arrowdown">⬂</span>'

		# expected columns. actual columns are filtered from the incoming data later, this
		# is mainly for controlling the order in the output table
		COLUMNS = ["testcase", "Result", "MLD", "MAXIMUM ABS DIFF"]
		COLUMNS = ["testcase", "Result", "MLD", "MAXIMUM ABS DIFF", "MIN_SSNR"]
		COLUMNS_GLOBAL = COLUMNS[:1]
		COLUMNS_DIFFERENTIAL = COLUMNS[1:]
		COLUMNS_DIFFERENTIAL_NOT_MLD = COLUMNS_DIFFERENTIAL[2:]


		def create_subpage(
		html_out,
		csv_out,
		@@ -111,11 +112,18 @@ def create_subpage(
		)
		write_out_csv(merged_reports, merged_reports[0].keys(), csv_out)

		table_header_a = "".join([TH_TMPL_GLOBAL.format(c) for c in COLUMNS_GLOBAL] + [TH_TMPL_DIFFERENTIAL.format(c) for c in COLUMNS_DIFFERENTIAL])
		table_header_a = "".join(
		[TH_TMPL_GLOBAL.format(c) for c in COLUMNS_GLOBAL]
		+ [TH_TMPL_DIFFERENTIAL.format(c) for c in COLUMNS_DIFFERENTIAL]
		)
		table_header_b = list()
		for c in COLUMNS_DIFFERENTIAL:
		table_header_b.append(TH_TMPL_SECOND_ROW.format(f"Previous Run<br>ID: {id_previous}"))
		table_header_b.append(TH_TMPL_SECOND_ROW.format(f"Current Run<br>ID: {id_current}"))
		table_header_b.append(
		TH_TMPL_SECOND_ROW.format(f"Previous Run<br>ID: {id_previous}")
		)
		table_header_b.append(
		TH_TMPL_SECOND_ROW.format(f"Current Run<br>ID: {id_current}")
		)
		table_header_b = "".join(table_header_b)
		table_body = "\n".join(
		tr_from_row(row, id_current, id_previous) for row in merged_reports
		@@ -241,8 +249,13 @@ def merge_and_cleanup_mld_reports(

		return diff

		other_col_pairs = [(f"{col}-{id_previous}", f"{col}-{id_current}") for col in COLUMNS_DIFFERENTIAL_NOT_MLD]
		merged = sorted(merged, key=partial(sort_func, other_col_pairs=other_col_pairs), reverse=True)
		other_col_pairs = [
		(f"{col}-{id_previous}", f"{col}-{id_current}")
		for col in COLUMNS_DIFFERENTIAL_NOT_MLD
		]
		merged = sorted(
		merged, key=partial(sort_func, other_col_pairs=other_col_pairs), reverse=True
		)

		# remove the unecessary whole path from the testcase names
		for row in merged:

ci/complexity_measurements/check_for_changes.py

+4 −1

Original line number	Diff line number	Diff line
		import argparse
		import csv
		import sys
		import numpy as np


		THRESH = 0.01
		@@ -32,6 +31,10 @@ def check_linewise_logfile(filepath, cols):
		change_ratios = [abs(float(curr[i]) / float(prev[i]) - 1) > THRESH for i in cols]
		changes_found = any(change_ratios)

		if changes_found:
		print("Previous log line:", prev)
		print("Current log line:", curr)

		return changes_found

ci/ivas_voip_be_test.sh

deleted100755 → 0

+0 −145

Original line number	Diff line number	Diff line
		#! /usr/bin/bash

		# (C) 2022-2024 IVAS codec Public Collaboration with portions copyright Dolby International AB, Ericsson AB,
		# Fraunhofer-Gesellschaft zur Foerderung der angewandten Forschung e.V., Huawei Technologies Co. LTD.,
		# Koninklijke Philips N.V., Nippon Telegraph and Telephone Corporation, Nokia Technologies Oy, Orange,
		# Panasonic Holdings Corporation, Qualcomm Technologies, Inc., VoiceAge Corporation, and other
		# contributors to this repository. All Rights Reserved.

		# This software is protected by copyright law and by international treaties.
		# The IVAS codec Public Collaboration consisting of Dolby International AB, Ericsson AB,
		# Fraunhofer-Gesellschaft zur Foerderung der angewandten Forschung e.V., Huawei Technologies Co. LTD.,
		# Koninklijke Philips N.V., Nippon Telegraph and Telephone Corporation, Nokia Technologies Oy, Orange,
		# Panasonic Holdings Corporation, Qualcomm Technologies, Inc., VoiceAge Corporation, and other
		# contributors to this repository retain full ownership rights in their respective contributions in
		# the software. This notice grants no license of any kind, including but not limited to patent
		# license, nor is any license granted by implication, estoppel or otherwise.

		# Contributors are required to enter into the IVAS codec Public Collaboration agreement before making
		# contributions.

		# This software is provided "AS IS", without any express or implied warranties. The software is in the
		# development stage. It is intended exclusively for experts who have experience with such software and
		# solely for the purpose of inspection. All implied warranties of non-infringement, merchantability
		# and fitness for a particular purpose are hereby disclaimed and excluded.

		# Any dispute, controversy or claim arising under or in relation to providing this software shall be
		# submitted to and settled by the final, binding jurisdiction of the courts of Munich, Germany in
		# accordance with the laws of the Federal Republic of Germany excluding its conflict of law rules and
		# the United Nations Convention on Contracts on the International Sales of Goods.

		function usage {
		echo
		echo "Usage:"
		echo " ivas_voip_be_test.sh [MODE]"
		echo
		echo " MODE - test (default) or coverage"
		exit
		}

		if [ ! -d "lib_com" ]; then
		echo "not in root directory! - please run in IVAS root"
		exit 1
		fi

		if [ -z "$1" ] \|\| [ "$1" == "test" ]; then
		WORKERS=""
		BUILD=1
		COVERAGE=0
		elif [ "$1" == "coverage" ]; then
		WORKERS="-t 1"
		BUILD=0
		COVERAGE=1
		else
		usage
		fi

		if [ $BUILD -eq 1 ];then
		make clean
		make all -j
		fi

		# Configuration
		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"

		cfg=./scripts/config/ci_linux.json
		dly_profile=./scripts/dly_error_profiles/dly_error_profile_0.dat

		output_dir_default="out"
		output_dir_voip="out_voip"

		# Run the same modes in VoIP and non-VoIP mode with a neutral delay profile
		./scripts/runIvasCodec.py $verbosity_cmd -p $cfg $WORKERS -U $limit_input_to_x_seconds -m "${modes[@]}" --oc "${output_formats[@]}" -o $output_dir_default \| tee voip_be_test_output.txt
		./scripts/runIvasCodec.py $verbosity_cmd -p $cfg $WORKERS -U $limit_input_to_x_seconds -m "${modes[@]}" --oc "${output_formats[@]}" -o $output_dir_voip -J "$dly_profile" \| tee -a voip_be_test_output.txt

		# Check if Python scripts above failed. They return status 0 even when running a mode fails, so we have to parse log file
		if grep -iq failed voip_be_test_output.txt ; then
		echo "Run errors in runIvasCodec.py"
		exit 1
		fi

		if [ $COVERAGE -eq 1 ];then
		# Coverage analysis requires only running the codec and may exit before the comparison part
		exit 0
		fi

		# Set up Python path
		python_audio_module_path=$(pwd)/scripts
		export PYTHONPATH=$python_audio_module_path:$PYTHONPATH
		python_audiofile_script_path=$python_audio_module_path/pyaudio3dtools/audiofile.py

		# Trim JBM delay from VoIP output files
		output_dir_voip_dec="$output_dir_voip"/dec
		output_dir_voip_dec_trimmed="$output_dir_voip"/dec_trimmed

		if [[ ! -d $output_dir_voip_dec_trimmed ]]; then
		mkdir $output_dir_voip_dec_trimmed
		fi

		for cut in "$output_dir_voip_dec"/*.wav; do
		output_path=${cut/$output_dir_voip_dec/$output_dir_voip_dec_trimmed}
		output_path=${output_path/".wav"/".raw"}
		python3 "$python_audiofile_script_path" pre-trim 60 "$cut" "$output_path" \| tee -a voip_be_test_output.txt
		done

		# Convert non-VoIP output from wav to pcm (comparison script doesn't support wav)
		output_dir_default_dec="$output_dir_default"/dec
		output_dir_default_dec_pcm="$output_dir_default"/dec_pcm

		if [[ ! -d $output_dir_default_dec_pcm ]]; then
		mkdir $output_dir_default_dec_pcm
		fi

		for ref in "$output_dir_default_dec"/*.wav; do
		output_path=${ref/$output_dir_default_dec/$output_dir_default_dec_pcm}
		output_path=${output_path/".wav"/".raw"}
		python3 "$python_audiofile_script_path" convert "$ref" "$output_path" \| tee -a voip_be_test_output.txt
		done

		# Assert BE between non-VoIP and VoIP modes
		all_be=1

		cmp_tool_path=$(pwd)/tests/cmp_pcm.py

		for ref in "$output_dir_default_dec_pcm"/*; do
		cut=${ref/$output_dir_default_dec_pcm/$output_dir_voip_dec_trimmed}
		cut=${cut/".dec."/"_jbm_dly_error_profile_0_dat.dec."}

		# Print paths of compared files, since the script doesn't do it
		printf "\nComparing %s and %s\n" "$ref" "$cut" \| tee -a voip_be_test_output.txt
		printout=$($cmp_tool_path "$ref" "$cut")
		if [ $? -ne 0 ]; then
		all_be=0
		fi
		printf "%s\n" "$printout" \| tee -a voip_be_test_output.txt
		done

		if [ $all_be -eq 1 ]; then
		printf "\n\nAll tested conditions are bit-exact\n" \| tee -a voip_be_test_output.txt
		else
		printf "\n\nBitexactness problems found!\n" \| tee -a voip_be_test_output.txt
		exit 1;
		fi