Super-human multi-talker speech recognition: A graphical modeling approach

A Facial Feature and Lip Movement Enhanced Audio-Visual Speech Separation Model

The cocktail party problem can be more effectively addressed by leveraging the speaker's visual and audio information. This paper proposes a method to improve the audio's separation using two visual cues: facial features and lip movement. Firstly, residual connections are introduced in the audio separation module to extract detailed features. Secondly, considering the video stream contains information other than the face, which has a minimal correlation with the audio, an attention mechanism is employed in the face module to focus on crucial information. Then, the loss function considers the audio-visual similarity to take advantage of the relationship between audio and visual completely. Experimental results on the public VoxCeleb2 dataset show that the proposed model significantly enhanced SDR, PSEQ, and STOI, especially 4 dB improvements in SDR.

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

Single-channel, speaker-independent speech separation methods have recently seen great progress. However, the accuracy, latency, and computational cost of such methods remain insufficient. The majority of the previous methods have formulated the separation problem through the time-frequency representation of the mixed signal, which has several drawbacks, including the decoupling of the phase and magnitude of the signal, the suboptimality of time-frequency representation for speech separation, and the long latency of the entire system. To address these shortcomings, we propose a fully-convolutional time-domain audio separation network (Conv-TasNet), a deep learning framework for end-to-end time-domain speech separation. Conv-TasNet uses a linear encoder to generate a representation of the speech waveform optimized for separating individual speakers. Speaker separation is achieved by applying a set of weighting functions (masks) to the encoder output. The modified encoder representations are then inverted back to the waveforms using a linear decoder. The masks are found using a temporal convolutional network (TCN) consisting of stacked 1-D dilated convolutional blocks, which allows the network to model the long-term dependencies of the speech signal while maintaining a small model size. The proposed Conv-TasNet system significantly outperforms previous time-frequency masking methods in separating two- and three-speaker mixtures. Additionally, Conv-TasNet surpasses several ideal time-frequency magnitude masks in two-speaker speech separation as evaluated by both objective distortion measures and subjective quality assessment by human listeners. Finally, Conv-TasNet has a significantly smaller model size and a much shorter minimum latency, making it a suitable solution for both offline and real-time speech separation applications. This study therefore represents a major step toward the realization of speech separation systems for real-world speech processing technologies.

Impact of phase estimation on single-channel speech separation based on time-frequency masking

Time-frequency masking is a common solution for the single-channel source separation (SCSS) problem where the goal is to find a time-frequency mask that separates the underlying sources from an observed mixture. An estimated mask is then applied to the mixed signal to extract the desired signal. During signal reconstruction, the time-frequency-masked spectral amplitude is combined with the mixture phase. This article considers the impact of replacing the mixture spectral phase with an estimated clean spectral phase combined with the estimated magnitude spectrum using a conventional model-based approach. As the proposed phase estimator requires estimated fundamental frequency of the underlying signal from the mixture, a robust pitch estimator is proposed. The upper-bound clean phase results show the potential of phase-aware processing in single-channel source separation. Also, the experiments demonstrate that replacing the mixture phase with the estimated clean spectral phase consistently improves perceptual speech quality, predicted speech intelligibility, and source separation performance across all signal-to-noise ratio and noise scenarios.

On Training Targets for Supervised Speech Separation

Formulation of speech separation as a supervised learning problem has shown considerable promise. In its simplest form, a supervised learning algorithm, typically a deep neural network, is trained to learn a mapping from noisy features to a time-frequency representation of the target of interest. Traditionally, the ideal binary mask (IBM) is used as the target because of its simplicity and large speech intelligibility gains. The supervised learning framework, however, is not restricted to the use of binary targets. In this study, we evaluate and compare separation results by using different training targets, including the IBM, the target binary mask, the ideal ratio mask (IRM), the short-time Fourier transform spectral magnitude and its corresponding mask (FFT-MASK), and the Gammatone frequency power spectrum. Our results in various test conditions reveal that the two ratio mask targets, the IRM and the FFT-MASK, outperform the other targets in terms of objective intelligibility and quality metrics. In addition, we find that masking based targets, in general, are significantly better than spectral envelope based targets. We also present comparisons with recent methods in non-negative matrix factorization and speech enhancement, which show clear performance advantages of supervised speech separation.