Fetching the paper…

Conv-TasNet: Surpassing Ideal Time-Frequency Magnitude Masking for Speech Separation · Around