Improving Voice Separation by Incorporating End-to-end Speech Recognition
This work addresses voice separation challenges for applications in noisy environments, though it is incremental by building on existing transfer learning approaches.
The paper tackled voice separation in noisy and data-limited scenarios by incorporating phonetic and linguistic features from an end-to-end speech recognition system, resulting in significant improvements in signal-to-distortion ratio over baselines and outperforming an audio-visual model on the AVSpeech dataset.
Despite recent advances in voice separation methods, many challenges remain in realistic scenarios such as noisy recording and the limits of available data. In this work, we propose to explicitly incorporate the phonetic and linguistic nature of speech by taking a transfer learning approach using an end-to-end automatic speech recognition (E2EASR) system. The voice separation is conditioned on deep features extracted from E2EASR to cover the long-term dependence of phonetic aspects. Experimental results on speech separation and enhancement task on the AVSpeech dataset show that the proposed method significantly improves the signal-to-distortion ratio over the baseline model and even outperforms an audio visual model, that utilizes visual information of lip movements.