diff options
Diffstat (limited to 'scripts/py/gen_audio.py')
-rw-r--r-- | scripts/py/gen_audio.py | 107 |
1 files changed, 86 insertions, 21 deletions
diff --git a/scripts/py/gen_audio.py b/scripts/py/gen_audio.py index ff33bfb..4d7318c 100644 --- a/scripts/py/gen_audio.py +++ b/scripts/py/gen_audio.py @@ -1,6 +1,6 @@ #!env/bin/python3 -# SPDX-FileCopyrightText: Copyright 2021 Arm Limited and/or its affiliates <open-source-office@arm.com> +# SPDX-FileCopyrightText: Copyright 2021, 2023 Arm Limited and/or its affiliates <open-source-office@arm.com> # SPDX-License-Identifier: Apache-2.0 # # Licensed under the Apache License, Version 2.0 (the "License"); @@ -17,34 +17,99 @@ """ Utility script to convert an audio clip into eval platform desired spec. """ -import soundfile as sf - from argparse import ArgumentParser from os import path -from gen_utils import AudioUtils +import soundfile as sf + +from gen_utils import GenUtils parser = ArgumentParser() -parser.add_argument("--audio_path", help="Audio file path", required=True) -parser.add_argument("--output_dir", help="Output directory", required=True) -parser.add_argument("--sampling_rate", type=int, help="target sampling rate.", default=16000) -parser.add_argument("--mono", type=bool, help="convert signal to mono.", default=True) -parser.add_argument("--offset", type=float, help="start reading after this time (in seconds).", default=0) -parser.add_argument("--duration", type=float, help="only load up to this much audio (in seconds).", default=0) -parser.add_argument("--res_type", type=AudioUtils.res_data_type, help=f"Resample type: {AudioUtils.res_type_list()}.", default='kaiser_best') -parser.add_argument("--min_samples", type=int, help="Minimum sample number.", default=16000) -parser.add_argument("-v", "--verbosity", action="store_true") -args = parser.parse_args() + +# pylint: disable=duplicate-code +parser.add_argument( + "--audio_path", + help="Audio file path", + required=True +) + +parser.add_argument( + "--output_dir", + help="Output directory", + required=True +) + +parser.add_argument( + "--sampling_rate", + type=int, + help="target sampling rate.", + default=16000 +) + +parser.add_argument( + "--mono", + type=bool, + help="convert signal to mono.", + default=True +) + +parser.add_argument( + "--offset", + type=float, + help="start reading after this time (in seconds).", + default=0 +) + +parser.add_argument( + "--duration", + type=float, + help="only load up to this much audio (in seconds).", + default=0 +) + +parser.add_argument( + "--res_type", + type=GenUtils.res_data_type, + help=f"Resample type: {GenUtils.res_type_list()}.", + default='kaiser_best' +) + +parser.add_argument( + "--min_samples", + type=int, + help="Minimum sample number.", + default=16000 +) + +parser.add_argument( + "-v", + "--verbosity", + action="store_true" +) +# pylint: enable=duplicate-code + +parsed_args = parser.parse_args() def main(args): - audio_data, samplerate = AudioUtils.load_resample_audio_clip(args.audio_path, - args.sampling_rate, - args.mono, args.offset, - args.duration, args.res_type, - args.min_samples) - sf.write(path.join(args.output_dir, path.basename(args.audio_path)), audio_data, samplerate) + """ + Generate the new audio file + @param args: Parsed args + """ + audio_sample = GenUtils.read_audio_file( + args.audio_path, args.offset, args.duration + ) + + resampled_audio = GenUtils.resample_audio_clip( + audio_sample, args.sampling_rate, args.mono, args.res_type, args.min_samples + ) + + sf.write( + path.join(args.output_dir, path.basename(args.audio_path)), + resampled_audio.data, + resampled_audio.sample_rate + ) if __name__ == '__main__': - main(args) + main(parsed_args) |