Loading scripts/reverb/generate_acoustic_environments_metadata.py +9 −4 Original line number Diff line number Diff line Loading @@ -31,13 +31,13 @@ """ # # Generate binary render configuration output files for testing purposes # The binary code generation is based on the MPEG-I audio standard # which defines functions to decode raw bitstream into internal parameters # Generate binary render configuration output files for testing purposes. # The binary code generation is based on the MPEG-I audio standard, # which defines functions to decode raw bitstream into internal parameters. # from bitarray import bitarray, test as bitarray_test from bitarray import bitarray import math from enum import Enum import numpy as np Loading Loading @@ -238,6 +238,11 @@ class fgdMethod(Enum): Default_Banding = '10' def get_default_grid_nr_bands(code): assert 0 <= code <= 8 return [10, 10, 31, 5, 6, 3, 41, 21, 25][code] def get_distance_code(distance, isSmallScene = True): # 0, 1, ... 99 metersCode = [ Loading scripts/reverb/text_to_binary_payload.py 0 → 100644 +151 −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. """ # # Read a text-based configuration file and generate the equivalent binary payload. # Makes use of the pyhton configuration file parser configparser, # and of the payload configuration functions in generate_acoustic_environments_metadata. # The configuration file format is as follows: # # [<section>:<index>] # <option>=<value> # # a comment # # <section> is one of the supported section names (frequencyGrid, acousticEnvironment). # <index> is an integer used as acoustic environment ID (revAcEnvID), and to refer to (revFreqGridIdx). # There is no leading white space for sections and options, and no trailing ';'. from bitarray import bitarray from configparser import ConfigParser import ast from generate_acoustic_environments_metadata import * # convert text containing an option value to a Python value, if possible def eval_option(text): try: value = ast.literal_eval(text) except ValueError: value = text return value def get_bool_code(b): assert(isinstance(b, bool)) return format(b, '01b') def parse_reverb_text_configuration_and_generate_binary_payload(file): # parse file config = ConfigParser() files_parsed = config.read(file) assert len(files_parsed) == 1, 'file {} not successfully parsed'.format(file) # collect dicts of frequency grid and acoustic environment sections sections = { key : {} for key in ['frequencyGrid', 'acousticEnvironment' ] } for section_name in config.sections(): section, index = section_name.split(':') assert section in sections, 'unknown section name' sections[section][index] = config[section_name] # parse frequency grids nr_bands = [] data = bitarray( '1' # hasAcEnv + get_count_or_index_code(len(sections['frequencyGrid'])), # fgdNrGrids endian='big') for _, fg in sections['frequencyGrid'].items(): if fg['method'] == 'individualFrequencies': nr_bands.append(len(eval_option(fg['frequencies']))) data += bitarray( fgdMethod.Individual_Frequencies.value # fgdMethod + get_count_or_index_code(len(eval_option(fg['frequencies']))) # fgdNrBands + concatenate(get_frequency_code, eval_option(fg['frequencies']))) # fgdCenterFreq elif fg['method'] == 'startHopAmount': nr_bands.append(eval_option(fg['nrBands'])) data += bitarray( fgdMethod.Start_Hop_Amount.value # fgdMethod + get_count_or_index_code(eval_option(fg['nrBands'])) # fgdNrBands + get_frequency_code(eval_option(fg['centerFrequency'])) # fgdCenterFreq + get_frequency_hop_code(eval_option(fg['hop']))) # frequencyHop elif fg['method'] == 'defaultBanding': nr_bands.append(get_default_grid_nr_bands(eval_option(fg['defaultGrid']))) data += bitarray( fgdMethod.Default_Banding.value # fgdMethod + format(eval_option(fg['defaultGrid']), '04b') # fgdDefaultGrid + get_bool_code('defaultGridOffset' in fg)) # fgdIsSubGrid if 'defaultGridOffset' in fg: data += bitarray( format(eval_option(fg['defaultGridOffset']), '03b') # fgdDefaultGridOffset ## only 3 bits used iso 6 (to offset into max 41 bands), correct? + format(eval_option(fg['defaultGridNrBands']), '06b')) # fgdDefaultGridNrBands else: assert False, 'unknow frequency grid method' # parse acoustic environments for index, ae in sections['acousticEnvironment'].items(): data += bitarray( get_count_or_index_code(len(sections['acousticEnvironment'])) # revNrElements + get_id_code(eval_option(index)) # revAcEnvID + get_count_or_index_code(eval_option(ae['frequencyGridIndex'])) # revFreqGridIdx + get_duration_code(eval_option(ae['predelay'])) # revPredelay + concatenate(get_duration_code, eval_option(ae['rt60'])) # revRT60 + concatenate(get_dsr_code, eval_option(ae['dsr'])) # revDSR + get_bool_code('earlyReflectionsSize' in ae)) # hasEarlyReflections if 'earlyReflectionsSize' in ae: assert len(eval_option(ae['absorptionCoefficients'])) == nr_bands[eval_option(ae['frequencyGridIndex'])] * 6, 'wrong number of absorption coefficients' data += bitarray( concatenate(lambda d : get_distance_code(d, True), eval_option(ae['earlyReflectionsSize'])) # erSize + get_count_or_index_code(eval_option(ae['earlyReflectionsfrequencyGridIndex'])) # erFreqGridIdx + concatenate(get_absorption_code, eval_option(ae['absorptionCoefficients'])) # erAbsCoeff + get_bool_code('listenerOrigin' in ae)) # hasListenerOrigin if 'listenerOrigin' in ae: xyz = eval_option(ae['listenerOrigin']) assert len(xyz) == 3, 'wrong number of listener origin coordinates' data += bitarray( '1' if xyz[0] >= 0 else '0' # isPositiveX + '1' if xyz[1] >= 0 else '0' # isPositiveY + concatenate(get_distance_code, xyz)) # erListenerOrigin # generate binary file data.tofile(open(file.split('.')[0] + '.dat', 'wb')) if __name__ == "__main__": import argparse parser = argparse.ArgumentParser(description='Reads a text-based configuration file and generates the equivalent binary payload file (<file>.dat)') parser.add_argument("configuration_file") args = parser.parse_args() parse_reverb_text_configuration_and_generate_binary_payload(args.configuration_file) print("\nNote: the conversion algorithm uses quantization, which may lead to quantization errors.") Loading
scripts/reverb/generate_acoustic_environments_metadata.py +9 −4 Original line number Diff line number Diff line Loading @@ -31,13 +31,13 @@ """ # # Generate binary render configuration output files for testing purposes # The binary code generation is based on the MPEG-I audio standard # which defines functions to decode raw bitstream into internal parameters # Generate binary render configuration output files for testing purposes. # The binary code generation is based on the MPEG-I audio standard, # which defines functions to decode raw bitstream into internal parameters. # from bitarray import bitarray, test as bitarray_test from bitarray import bitarray import math from enum import Enum import numpy as np Loading Loading @@ -238,6 +238,11 @@ class fgdMethod(Enum): Default_Banding = '10' def get_default_grid_nr_bands(code): assert 0 <= code <= 8 return [10, 10, 31, 5, 6, 3, 41, 21, 25][code] def get_distance_code(distance, isSmallScene = True): # 0, 1, ... 99 metersCode = [ Loading
scripts/reverb/text_to_binary_payload.py 0 → 100644 +151 −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. """ # # Read a text-based configuration file and generate the equivalent binary payload. # Makes use of the pyhton configuration file parser configparser, # and of the payload configuration functions in generate_acoustic_environments_metadata. # The configuration file format is as follows: # # [<section>:<index>] # <option>=<value> # # a comment # # <section> is one of the supported section names (frequencyGrid, acousticEnvironment). # <index> is an integer used as acoustic environment ID (revAcEnvID), and to refer to (revFreqGridIdx). # There is no leading white space for sections and options, and no trailing ';'. from bitarray import bitarray from configparser import ConfigParser import ast from generate_acoustic_environments_metadata import * # convert text containing an option value to a Python value, if possible def eval_option(text): try: value = ast.literal_eval(text) except ValueError: value = text return value def get_bool_code(b): assert(isinstance(b, bool)) return format(b, '01b') def parse_reverb_text_configuration_and_generate_binary_payload(file): # parse file config = ConfigParser() files_parsed = config.read(file) assert len(files_parsed) == 1, 'file {} not successfully parsed'.format(file) # collect dicts of frequency grid and acoustic environment sections sections = { key : {} for key in ['frequencyGrid', 'acousticEnvironment' ] } for section_name in config.sections(): section, index = section_name.split(':') assert section in sections, 'unknown section name' sections[section][index] = config[section_name] # parse frequency grids nr_bands = [] data = bitarray( '1' # hasAcEnv + get_count_or_index_code(len(sections['frequencyGrid'])), # fgdNrGrids endian='big') for _, fg in sections['frequencyGrid'].items(): if fg['method'] == 'individualFrequencies': nr_bands.append(len(eval_option(fg['frequencies']))) data += bitarray( fgdMethod.Individual_Frequencies.value # fgdMethod + get_count_or_index_code(len(eval_option(fg['frequencies']))) # fgdNrBands + concatenate(get_frequency_code, eval_option(fg['frequencies']))) # fgdCenterFreq elif fg['method'] == 'startHopAmount': nr_bands.append(eval_option(fg['nrBands'])) data += bitarray( fgdMethod.Start_Hop_Amount.value # fgdMethod + get_count_or_index_code(eval_option(fg['nrBands'])) # fgdNrBands + get_frequency_code(eval_option(fg['centerFrequency'])) # fgdCenterFreq + get_frequency_hop_code(eval_option(fg['hop']))) # frequencyHop elif fg['method'] == 'defaultBanding': nr_bands.append(get_default_grid_nr_bands(eval_option(fg['defaultGrid']))) data += bitarray( fgdMethod.Default_Banding.value # fgdMethod + format(eval_option(fg['defaultGrid']), '04b') # fgdDefaultGrid + get_bool_code('defaultGridOffset' in fg)) # fgdIsSubGrid if 'defaultGridOffset' in fg: data += bitarray( format(eval_option(fg['defaultGridOffset']), '03b') # fgdDefaultGridOffset ## only 3 bits used iso 6 (to offset into max 41 bands), correct? + format(eval_option(fg['defaultGridNrBands']), '06b')) # fgdDefaultGridNrBands else: assert False, 'unknow frequency grid method' # parse acoustic environments for index, ae in sections['acousticEnvironment'].items(): data += bitarray( get_count_or_index_code(len(sections['acousticEnvironment'])) # revNrElements + get_id_code(eval_option(index)) # revAcEnvID + get_count_or_index_code(eval_option(ae['frequencyGridIndex'])) # revFreqGridIdx + get_duration_code(eval_option(ae['predelay'])) # revPredelay + concatenate(get_duration_code, eval_option(ae['rt60'])) # revRT60 + concatenate(get_dsr_code, eval_option(ae['dsr'])) # revDSR + get_bool_code('earlyReflectionsSize' in ae)) # hasEarlyReflections if 'earlyReflectionsSize' in ae: assert len(eval_option(ae['absorptionCoefficients'])) == nr_bands[eval_option(ae['frequencyGridIndex'])] * 6, 'wrong number of absorption coefficients' data += bitarray( concatenate(lambda d : get_distance_code(d, True), eval_option(ae['earlyReflectionsSize'])) # erSize + get_count_or_index_code(eval_option(ae['earlyReflectionsfrequencyGridIndex'])) # erFreqGridIdx + concatenate(get_absorption_code, eval_option(ae['absorptionCoefficients'])) # erAbsCoeff + get_bool_code('listenerOrigin' in ae)) # hasListenerOrigin if 'listenerOrigin' in ae: xyz = eval_option(ae['listenerOrigin']) assert len(xyz) == 3, 'wrong number of listener origin coordinates' data += bitarray( '1' if xyz[0] >= 0 else '0' # isPositiveX + '1' if xyz[1] >= 0 else '0' # isPositiveY + concatenate(get_distance_code, xyz)) # erListenerOrigin # generate binary file data.tofile(open(file.split('.')[0] + '.dat', 'wb')) if __name__ == "__main__": import argparse parser = argparse.ArgumentParser(description='Reads a text-based configuration file and generates the equivalent binary payload file (<file>.dat)') parser.add_argument("configuration_file") args = parser.parse_args() parse_reverb_text_configuration_and_generate_binary_payload(args.configuration_file) print("\nNote: the conversion algorithm uses quantization, which may lead to quantization errors.")