Generating the synth examples. Step 3 not clear. about conferencingspeech2021 HOT 4 CLOSED

@popcornell These config files are generated from mix_wav.py.

from conferencingspeech2021.

Aren't the ones provided here ?
https://github.com/ConferencingSpeech/ConferencingSpeech2021/tree/master/simulation/data
I have them already without running mix_wav.py.

Anyway everything is fine if i run "new parameter configuration". I think the two steps before are confusing. are they necessary ?

from conferencingspeech2021.

Thank you for your questioning, we may not write the readme clearly. For generating the data, you can choose to run the first command “python mix_wav.py --mix_config_path=./data/dev_linear_simu_mix.config --save_dir=./data/wavs/dev/simu_linear/ --chunk_len=6 --generate_config=False ”if you have generated config file or want to reproduce our baseline experiments.
You can also choose to run the second command "python mix_wav.py --clean_wav_list=./data/dev_clean.lst --noise_wav_list=./data/dev_noise.lst --rir_wav_list=./data/dev_linear_rir.lst --mix_config_path=./data/dev_linear_simu_mix.config --save_dir=./data/wavs/dev/simu_linear --chunk_len=6 --generate_config=True
”if you want to generate your own config file to expand training data.

from conferencingspeech2021.

Thank you now i understand !!

from conferencingspeech2021.