Code cleaned and MASA added. (67389472) · Commits · IVAS Codec Public Collaboration / IVAS Codec

tests/binaural/init.py

0 → 100644

+31 −0

Original line number	Diff line number	Diff line
		#!/usr/bin/env python3

		"""
		(C) 2022-2023 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.
		"""

tests/binaural/constants.py

+22 −13

Original line number	Diff line number	Diff line
		@@ -52,23 +52,32 @@ DECODER_CMD = [
		HRTF_BINARY_FILE = "default_rom_{}kHz.bin"
		SAMPLE_RATE = ["16", "32", "48"]

		INPUT_FORMATS_MC = ["MONO", "STEREO", "5_1", "5_1_2", "5_1_4", "7_1", "7_1_4"]
		INPUT_FORMATS_SBA = {
		"FOA":"1",
		"HOA2":"2",
		"HOA3":"3"
		INPUT_FORMATS_MC = ["5_1", "5_1_2", "5_1_4", "7_1", "7_1_4"] # "MONO", "STEREO",
		INPUT_FORMATS_SBA = ["1", "2", "3"]
		INPUT_FORMATS_ISM = ["1", "2", "3", "4"]
		INPUT_FORMATS_MASA = {
		"tc" : ["1", "2"],
		"dir" : ["1", "2"]
		}

		OUTPUT_FORMATS_BINAURAL = ["BINAURAL", "BINAURAL_ROOM_IR"] # "BINAURAL_ROOM_REVERB"
		HR_TRAJECTORIES_TO_TEST = ["headrot_case00_3000_q"]

		FORMAT_TO_FILE_WOEXT = {
		FORMAT_TO_FILE_MC_WOEXT = {
		"5_1": "stv51MC{}c",
		"7_1": "stv71MC{}c",
		"5_1_2": "stv512MC{}c",
		"5_1_4": "stv514MC{}c",
		"7_1_4": "stv714MC{}c",
		"FOA": "stvFOA{}c",
		"HOA2": "stv2OA{}c",
		"HOA3": "stv3OA{}c",
		}
		FORMAT_TO_FILE_SBA_WOEXT = {
		"1": "stvFOA{}c",
		"2": "stv2OA{}c",
		"3": "stv3OA{}c",
		}
		FORMAT_TO_FILE_ISM_WOEXT = "stv{}ISM{}s"
		FORMAT_TO_FILE_MASA_WOEXT = "stv{}MASA{}TC{}c"

		FORMAT_TO_METADATA_FILES = {
		"MASA": "stv{}MASA{}TC{}c.met",
		}

		OUTPUT_FORMATS_BINAURAL = ["BINAURAL", "BINAURAL_ROOM_IR"] # "BINAURAL_ROOM_REVERB"
		HR_TRAJECTORIES_TO_TEST = ["headrot_case00_3000_q", "headrot"]
		No newline at end of file

tests/binaural/test_binaural.py

+113 −17

Original line number	Diff line number	Diff line
		@@ -40,28 +40,50 @@ from tests.binaural.utils import *
		@pytest.mark.parametrize("in_fmt", INPUT_FORMATS_MC)
		def test_multichannel_binaural_static(test_info, in_fmt, out_fmt, out_fs):

		if in_fmt in ["MONO", "STEREO"]:
		pytest.skip("MONO or STEREO to Binaural rendering unsupported")
		#if in_fmt in ["MONO", "STEREO"]:
		# pytest.skip("MONO or STEREO to Binaural rendering unsupported")

		compare_multichannel_rom_vs_binary(
		# -mc InputConf : Multi-channel format
		# where InputConf specifies the channel configuration (5_1, 7_1, 5_1_2, 5_1_4, 7_1_4)

		bitrate = 512000
		in_fs = 48
		option_list = ["-mc", in_fmt]
		in_file = FORMAT_TO_FILE_MC_WOEXT[in_fmt].format(in_fs)

		compare_rom_vs_binary(
		test_info,
		in_fmt,
		option_list,
		bitrate,
		in_file,
		in_fs,
		out_fmt,
		out_fs
		out_fs,
		)

		@pytest.mark.parametrize("trj_file", HR_TRAJECTORIES_TO_TEST)
		@pytest.mark.parametrize("trj_file", [HR_TRAJECTORIES_TO_TEST[0]])
		@pytest.mark.parametrize("out_fmt", OUTPUT_FORMATS_BINAURAL)
		@pytest.mark.parametrize("out_fs", SAMPLE_RATE)
		@pytest.mark.parametrize("in_fmt", INPUT_FORMATS_MC)
		def test_multichannel_binaural_headrotation(test_info, in_fmt, out_fmt, out_fs, trj_file):

		if in_fmt in ["MONO", "STEREO"]:
		pytest.skip("MONO or STEREO to Binaural rendering unsupported")
		#if in_fmt in ["MONO", "STEREO"]:
		# pytest.skip("MONO or STEREO to Binaural rendering unsupported")

		# -mc InputConf : Multi-channel format
		# where InputConf specifies the channel configuration (5_1, 7_1, 5_1_2, 5_1_4, 7_1_4)

		compare_multichannel_rom_vs_binary(
		bitrate = 512000
		in_fs = 48
		option_list = ["-mc", in_fmt]
		in_file = FORMAT_TO_FILE_MC_WOEXT[in_fmt].format(in_fs)

		compare_rom_vs_binary(
		test_info,
		in_fmt,
		option_list,
		bitrate,
		in_file,
		in_fs,
		out_fmt,
		out_fs,
		trj_file=str(TESTV_DIR.joinpath(f"{trj_file}.csv"))
		@@ -74,23 +96,97 @@ def test_multichannel_binaural_headrotation(test_info, in_fmt, out_fmt, out_fs,
		@pytest.mark.parametrize("in_fmt", INPUT_FORMATS_SBA)
		def test_sba_binaural_static(test_info, in_fmt, fs, out_fmt):

		compare_ambisonics_rom_vs_binary(
		# -sba +/-Order : Scene Based Audio input format (Ambisonics ACN/SN3D)
		# where Order specifies the Ambisionics order (1-3)

		bitrate = 256000
		option_list = ["sba", in_fmt]
		in_file = FORMAT_TO_FILE_SBA_WOEXT[in_fmt].format(fs)

		compare_rom_vs_binary(
		test_info,
		in_fmt,
		option_list,
		bitrate,
		in_file,
		fs,
		out_fmt,
		fs,
		)

		@pytest.mark.parametrize("trj_file", HR_TRAJECTORIES_TO_TEST)
		@pytest.mark.parametrize("trj_file", [HR_TRAJECTORIES_TO_TEST[0]])
		@pytest.mark.parametrize("out_fmt", OUTPUT_FORMATS_BINAURAL)
		@pytest.mark.parametrize("fs", SAMPLE_RATE[1:])
		@pytest.mark.parametrize("in_fmt", INPUT_FORMATS_SBA)
		def test_sba_binaural_headrotation(test_info, in_fmt, fs, out_fmt, trj_file):

		compare_ambisonics_rom_vs_binary(
		# -sba +/-Order : Scene Based Audio input format (Ambisonics ACN/SN3D)
		# where Order specifies the Ambisionics order (1-3)

		bitrate = 256000
		option_list = ["-sba", in_fmt]
		in_file = FORMAT_TO_FILE_SBA_WOEXT[in_fmt].format(fs)

		compare_rom_vs_binary(
		test_info,
		option_list,
		bitrate,
		in_file,
		fs,
		out_fmt,
		fs,
		trj_file=str(TESTV_DIR.joinpath(f"{trj_file}.csv"))
		)

		""" MASA """

		@pytest.mark.parametrize("out_fmt", OUTPUT_FORMATS_BINAURAL)
		@pytest.mark.parametrize("fs", SAMPLE_RATE[-1:])
		@pytest.mark.parametrize("in_dir", INPUT_FORMATS_MASA["dir"])
		@pytest.mark.parametrize("in_tc", INPUT_FORMATS_MASA["tc"])
		def test_masa_binaural_static(test_info, in_tc, in_dir, fs, out_fmt):

		# -masa Ch File : MASA format
		# where Ch specifies the number of MASA input/transport channels (1 or 2)
		# and File specifies input file containing parametric MASA metadata

		bitrate = 256000
		test = FORMAT_TO_METADATA_FILES["MASA"].format(in_dir, in_tc, fs)
		metadata_file=str(TESTV_DIR.joinpath(test))
		option_list = ["-masa", in_tc, metadata_file]
		in_file = FORMAT_TO_FILE_MASA_WOEXT.format(in_dir, in_tc, fs)

		compare_rom_vs_binary(
		test_info,
		option_list,
		bitrate,
		in_file,
		fs,
		out_fmt,
		fs,
		)

		@pytest.mark.parametrize("trj_file", [HR_TRAJECTORIES_TO_TEST[1]])
		@pytest.mark.parametrize("out_fmt", OUTPUT_FORMATS_BINAURAL)
		@pytest.mark.parametrize("fs", SAMPLE_RATE[-1:])
		@pytest.mark.parametrize("in_dir", INPUT_FORMATS_MASA["dir"])
		@pytest.mark.parametrize("in_tc", INPUT_FORMATS_MASA["tc"])
		def test_masa_binaural_headrotation(test_info, in_tc, in_dir, fs, out_fmt, trj_file):

		# -masa Ch File : MASA format
		# where Ch specifies the number of MASA input/transport channels (1 or 2)
		# and File specifies input file containing parametric MASA metadata

		bitrate = 256000
		test = FORMAT_TO_METADATA_FILES["MASA"].format(in_dir, in_tc, fs)
		metadata_file=str(TESTV_DIR.joinpath(test))
		option_list = ["-masa", in_tc, metadata_file]
		in_file = FORMAT_TO_FILE_MASA_WOEXT.format(in_dir, in_tc, fs)

		compare_rom_vs_binary(
		test_info,
		in_fmt,
		option_list,
		bitrate,
		in_file,
		fs,
		out_fmt,
		fs,

tests/binaural/utils.py

+30 −55

Original line number	Diff line number	Diff line
		@@ -36,6 +36,7 @@ import subprocess as sp
		import sys
		from typing import Dict, Optional, Tuple
		import numpy as np
		import os

		from .constants import *
		from .compare_audio import compare_audio_arrays
		@@ -172,78 +173,52 @@ def run_decoder(
		run_cmd(cmd)


		def compare_multichannel_rom_vs_binary(
		test_info,
		in_fmt,
		out_fmt,
		out_fs,
		trj_file: Optional[str] = None
		):
		ivas_format = "mc"
		option_list = [f"-{ivas_format}", in_fmt]
		in_fs = 48
		bitrate = 512000
		def get_option_list_str(option_list):

		filename_ext = f"_{ivas_format}_{bitrate}_{in_fmt}_{in_fs}-{out_fmt}_{out_fs}"
		option_list_str = option_list.copy()
		for i in range(len(option_list_str)):
		if option_list_str[i][0] == '-':
		option_list_str[i] = option_list_str[i][1:]
		if os.path.isfile(option_list_str[i]):
		option_list_str[i] = Path(option_list_str[i]).stem

		input_path = TESTV_DIR.joinpath(FORMAT_TO_FILE_WOEXT[in_fmt].format(in_fs)).with_suffix(".wav")
		bitstream_path = BITSTREAM_DIR.joinpath(FORMAT_TO_FILE_WOEXT[in_fmt].format(in_fs) + filename_ext)
		run_encoder(bitrate, in_fs, input_path, bitstream_path, add_option_list=option_list)
		return option_list_str

		out_rom_path = DEC_ROM_DIR.joinpath(FORMAT_TO_FILE_WOEXT[in_fmt].format(in_fs) + filename_ext).with_suffix(".wav")
		if trj_file is not None:
		option_list = ["-t", trj_file]
		else :
		option_list = None
		run_decoder(out_fmt, out_fs, bitstream_path, out_rom_path, add_option_list=option_list)
		out_rom, out_rom_fs = pyaudio3dtools.audiofile.readfile(out_rom_path)

		out_bin_path = DEC_BINARY_DIR.joinpath(FORMAT_TO_FILE_WOEXT[in_fmt].format(in_fs) + filename_ext).with_suffix(".wav")
		hrtf_file = HRTF_BINARY_FILE.format(out_fs)
		if trj_file is not None:
		option_list.extend(["-hrtf", str(HRTF_BINARY_DIR.joinpath(hrtf_file))])
		else :
		option_list = ["-hrtf", str(HRTF_BINARY_DIR.joinpath(hrtf_file))]
		run_decoder(out_fmt, out_fs, bitstream_path, out_bin_path, add_option_list=option_list)
		out_bin, out_bin_fs = pyaudio3dtools.audiofile.readfile(out_rom_path)

		check_BE(test_info, out_rom, out_rom_fs, out_bin, out_bin_fs)


		def compare_ambisonics_rom_vs_binary(
		def compare_rom_vs_binary(
		test_info,
		in_fmt,
		option_list_enc,
		bitrate,
		in_file,
		in_fs,
		out_fmt,
		out_fs,
		trj_file: Optional[str] = None
		):
		ivas_format = "sba"
		option_list = [f"-{ivas_format}", INPUT_FORMATS_SBA[in_fmt]]
		bitrate = 256000

		filename_ext = f"_{ivas_format}_{bitrate}_{in_fmt}_{in_fs}-{out_fmt}_{out_fs}"
		option_str = "_".join(get_option_list_str(option_list_enc))
		bit_file_ext = f"_{option_str}_{bitrate}_{in_fs}"
		out_file_ext = bit_file_ext + f"-{out_fs}_{out_fmt}"

		input_path = TESTV_DIR.joinpath(FORMAT_TO_FILE_WOEXT[in_fmt].format(in_fs)).with_suffix(".wav")
		bitstream_path = BITSTREAM_DIR.joinpath(FORMAT_TO_FILE_WOEXT[in_fmt].format(in_fs) + filename_ext)
		input_path = TESTV_DIR.joinpath(in_file).with_suffix(".wav")
		bitstream_path = BITSTREAM_DIR.joinpath(in_file + bit_file_ext)
		run_encoder(bitrate, in_fs, input_path, bitstream_path, add_option_list=option_list_enc)

		run_encoder(bitrate, in_fs, input_path, bitstream_path, add_option_list=option_list)

		out_rom_path = DEC_ROM_DIR.joinpath(FORMAT_TO_FILE_WOEXT[in_fmt].format(in_fs) + filename_ext).with_suffix(".wav")
		if trj_file is not None:
		option_list = ["-t", trj_file]
		option_list_dec = ["-t", trj_file]
		out_file_ext += "_headrot"
		else :
		option_list = None
		run_decoder(out_fmt, out_fs, bitstream_path, out_rom_path, add_option_list=option_list)
		option_list_dec = None
		out_rom_path = DEC_ROM_DIR.joinpath(in_file + out_file_ext).with_suffix(".wav")
		run_decoder(out_fmt, out_fs, bitstream_path, out_rom_path, add_option_list=option_list_dec)
		out_rom, out_rom_fs = pyaudio3dtools.audiofile.readfile(out_rom_path)

		out_bin_path = DEC_BINARY_DIR.joinpath(FORMAT_TO_FILE_WOEXT[in_fmt].format(in_fs) + filename_ext).with_suffix(".wav")
		hrtf_file = HRTF_BINARY_FILE.format(out_fs)
		if trj_file is not None:
		option_list.extend(["-hrtf", str(HRTF_BINARY_DIR.joinpath(hrtf_file))])
		option_list_dec.extend(["-hrtf", str(HRTF_BINARY_DIR.joinpath(hrtf_file))])
		else :
		option_list = ["-hrtf", str(HRTF_BINARY_DIR.joinpath(hrtf_file))]
		run_decoder(out_fmt, out_fs, bitstream_path, out_bin_path, add_option_list=option_list)
		option_list_dec = ["-hrtf", str(HRTF_BINARY_DIR.joinpath(hrtf_file))]
		out_bin_path = DEC_BINARY_DIR.joinpath(in_file + out_file_ext).with_suffix(".wav")
		run_decoder(out_fmt, out_fs, bitstream_path, out_bin_path, add_option_list=option_list_dec)
		out_bin, out_bin_fs = pyaudio3dtools.audiofile.readfile(out_rom_path)

		check_BE(test_info, out_rom, out_rom_fs, out_rom, out_rom_fs)
		No newline at end of file
		check_BE(test_info, out_rom, out_rom_fs, out_bin, out_bin_fs)
		No newline at end of file