Despite the tremendous success of automatic speech recognition (ASR) with the introduction of deep learning, its performance is still unsatisfactory in many real-world multi-talker scenarios. Speaker separation excels in separating individual talkers but, as a frontend, it introduces processing artifacts that degrade the ASR backend trained on clean speech. As a result, mainstream robust ASR systems train the backend on noisy speech to mitigate processing artifacts. In this work, we propose to decouple the training of the speaker separation frontend and the ASR backend, and evaluate the proposed system with the backend trained on clean speech only. Our decoupled system achieves 5.1% word error rates (WER) on the Libri2Mix dev/test sets, significantly outperforming other multi-talker ASR baselines. Its effectiveness is also demonstrated with the state-of-the-art 7.60%/5.74% WERs on 1-ch and 6-ch SMS-WSJ. Furthermore, on recorded LibriCSS, we achieve the speaker-attributed WER of 2.92%. These state-of-the-art results suggest that decoupling speaker separation and recognition is an effective approach to elevate robust multi-talker ASR performance. Finally, we provide insights into the acoustic conditions where the decoupled approach is expected to outperform the mainstream approach of training on noisy speech.

This paper investigates robust speaker localization at the frame level on the basis of complex spectral mapping, which is capable of learning both the magnitude and phase of the target signal. Unlike prevailing deep learning methods for speaker localization, we perform MIMO (multi-input multi-output) based multi-channel speech enhancement first and then localize the enhanced speaker using weighted generalized cross correlation. In addition, we propose new multi-channel loss functions that incorporate phase differences in order to preserve inter-channel phase relations, which is key to accurate sound localization. Systematic evaluations using simulated and recorded room impulse responses demonstrate that the proposed model yields excellent frame-level speaker localization results in reverberant and noisy environments and outperforms related methods by a large margin, even surpassing their utterance-level results.

Despite the tremendous success of automatic speech recognition (ASR) with the introduction of deep learning, its performance is still unsatisfactory in many real-world multi-talker scenarios. Speaker separation excels in separating individual talkers but, as a frontend, it introduces processing artifacts that degrade the ASR backend trained on clean speech. As a result, mainstream robust ASR systems train on noisy speech to avoid processing artifacts. In this work, we propose to decouple the training of the multi-channel speaker separation frontend and the ASR backend, with the latter trained only on clean speech. On SMS-WSJ, the proposed approach achieves a word error rate (WER) of 5.74%, outperforming the previous best by 14.3%. Furthermore, on recorded LibriCSS, we achieve the speaker-attributed WER of 3.86%, outperforming the previous best system trained on the same data by 24.8%. These state-of-the-art results suggest that decoupling speech separation and recognition is a potentially effective approach to robust ASR.

Continuous speaker separation aims to separate overlapping speakers in real-world environments like meetings, but it often falls short in isolating speech segments of a single speaker. This leads to split signals that adversely affect downstream applications such as automatic speech recognition and speaker diarization. Existing solutions like speaker counting have limitations. This paper presents a novel multi-channel approach for continuous speaker separation based on multi-input multi-output (MIMO) complex spectral mapping. This MIMO approach enables robust speaker localization by preserving inter-channel phase relations. Speaker localization as a byproduct of the MIMO separation model is then used to identify single-talker frames and reduce speaker splitting. We demonstrate that this approach achieves superior frame-level sound localization. Systematic experiments on the LibriCSS dataset further show that the proposed approach outperforms other methods, advancing state-of-the-art speaker separation performance.

When dealing with overlapped speech, the performance of automatic speech recognition (ASR) systems substantially degrades as they are designed for single-talker speech. To enhance ASR performance in conversational or meeting environments, continuous speaker separation (CSS) is commonly employed. However, CSS requires a short separation window to avoid many speakers inside the window and sequential grouping of discontinuous speech segments. To address these limitations, we introduce a new multi-channel framework called “speaker separation via neural diarization” (SSND) for meeting environments. Our approach utilizes an end-to-end diarization system to identify the speech activity of each individual speaker. By leveraging estimated speaker boundaries, we generate a sequence of embeddings, which in turn facilitate the assignment of speakers to the outputs of a multi-talker separation model. SSND addresses the permutation ambiguity issue of talker-independent speaker separation during the diarization phase through location-based training, rather than during the separation process. This unique approach allows multiple non-overlapped speakers to be assigned to the same output stream, making it possible to efficiently process long segments-a task impossible with CSS. Additionally, SSND is naturally suitable for speaker-attributed ASR. We evaluate our proposed diarization and separation methods on the open LibriCSS dataset, advancing state-of-the-art diarization and ASR results by a large margin.

Personalized speech enhancement (PSE) models achieve promising results compared with unconditional speech enhancement models due to their ability to remove interfering speech in addition to background noise. Unlike unconditional speech enhancement, causal PSE models may occasionally remove the target speech by mistake. The PSE models also tend to leak interfering speech when the target speaker is silent for an extended period. We show that existing PSE methods suffer from a trade-off between speech over-suppression and interference leakage by addressing one problem at the expense of the other. We propose a new PSE model training framework using cross-task knowledge distillation to mitigate this trade-off. Specifically, we utilize a personalized voice activity detector (pVAD) during training to exclude the non-target speech frames that are wrongly identified as containing the target speaker with hard or soft classification. This prevents the PSE model from being too aggressive while still allowing the model to learn to suppress the input speech when it is likely to be spoken by interfering speakers. Comprehensive evaluation results are presented, covering various PSE usage scenarios.

The performance of automatic speech recognition (ASR) systems severely degrades when multi-talker speech overlap occurs. In meeting environments, speech separation is typically performed to improve the robustness of ASR systems. Recently, location-based training (LBT) was proposed as a new training criterion for multi-channel talker-independent speaker separation. Assuming fixed array geometry, LBT outperforms widely-used permutation-invariant training in fully overlapped utterances and matched reverberant conditions. This paper extends LBT to conversational multi-channel speaker separation. We introduce multi-resolution LBT to estimate the complex spectrograms from low to high time and frequency resolutions. With multi-resolution LBT, convolutional kernels are assigned consistently based on speaker locations in physical space. Evaluation results show that multi-resolution LBT consistently outperforms other competitive methods on the recorded LibriCSS corpus.

With the recent surge of video conferencing tools usage, providing high-quality speech signals and accurate captions have become essential to conduct day-to-day business or connect with friends and families. Single-channel personalized speech enhancement (PSE) methods show promising results compared with the unconditional speech enhancement (SE) methods in these scenarios due to their ability to remove interfering speech in addition to the environmental noise. In this work, we leverage spatial information afforded by microphone arrays to improve such systems’ performance further. We investigate the relative importance of speaker embeddings and spatial features. Moreover, we propose a new causal array-geometry-agnostic multi-channel PSE model, which can generate a high-quality enhanced signal from arbitrary microphone geometry. Experimental results show that the proposed geometry agnostic model outperforms the model trained on a specific microphone array geometry in both speech quality and automatic speech recognition accuracy. We also demonstrate the effectiveness of the proposed approach for unseen array geometries.

Permutation-invariant training (PIT) is a dominant approach for addressing the permutation ambiguity problem in talker-independent speaker separation. Leveraging spatial information afforded by microphone arrays, we propose a new training approach to resolving permutation ambiguities for multi-channel speaker separation. The proposed approach, named location-based training (LBT), assigns speakers on the basis of their spatial locations. This training strategy is easy to apply, and organizes speakers according to their positions in physical space. Specifically, this study investigates azimuth angles and source distances for location-based training. Evaluation results on separating two- and three-speaker mixtures show that azimuth-based training consistently outperforms PIT, and distance-based training further improves the separation performance when speaker azimuths are close. Furthermore, we dynamically select azimuth-based or distance-based training by estimating the azimuths of separated speakers, which further improves separation performance. LBT has a linear training complexity with respect to the number of speakers, as opposed to the factorial complexity of PIT. We further demonstrate the effectiveness of LBT for the separation of four and five concurrent speakers.

Permutation ambiguity is a crucial issue for deep learning based talker-independent speaker separation. Deep clustering and permutation invariant training (PIT) have been widely used to address the permutation ambiguity problem in monaural scenarios. Although both approaches have been extended to multi-microphone scenarios, we believe that the permutation ambiguity problem can be naturally avoided by leveraging the spatial relations of multiple speakers. In this study, we present location-based training (LBT), a new approach to achieve talker independency in multi-channel speaker separation. Unlike PIT that examines all possible permutations, LBT assigns speakers according to their positions in physical space. With a linear training complexity to the number of concurrent speakers, LBT is computationally much more efficient than PIT with a factorial complexity, particularly when a large number of overlapping speakers needs to be separated. Specifically, we propose two training criteria: azimuth-based and distance-based training, using speaker azimuths and distances relative to a microphone array, respectively. Evaluation results show that LBT significantly outperforms PIT on two-speaker and three-speaker mixtures with different array geometries and in various acoustic conditions. In addition, we propose a joint training strategy to integrate azimuth-based and distance-based training, which further improves separation performance.

Existing speaker separation methods deliver excellent performance on fully overlapped signal mixtures. To apply these methods in daily conversations that include occasional concurrent speakers, recent studies incorporate both overlapped and non-overlapped segments in the training data. However, such training data can degrade the separation performance due to triviality of non-overlapped segments where the model reflects the input to the output. We propose a new loss function for speaker separation based on permutation invariant training that dynamically reweighs losses using the segment overlap ratio. The new loss function emphasizes overlapped regions while deemphasizing the segments with single speakers. We demonstrate the effectiveness of the proposed loss function on an automatic speech recognition (ASR) task. Experiments on the recently introduced LibriCSS corpus show that our proposed single-channel method produces consistent improvements compared to baseline methods.

Deep neural network (DNN) embeddings for speaker recognition have recently attracted much attention. Compared to i-vectors, they are more robust to noise and room reverberation as DNNs leverage large-scale training. This article addresses the question of whether speech enhancement approaches are still useful when DNN embeddings are used for speaker recognition. We investigate single- and multi-channel speech enhancement for text-independent speaker verification based on x-vectors in conditions where strong diffuse noise and reverberation are both present. Single-channel (monaural) speech enhancement is based on complex spectral mapping and is applied to individual microphones. We use masking-based minimum variance distortion-less response (MVDR) beamformer and its rank-1 approximation for multi-channel speech enhancement. We propose a novel method of deriving time-frequency masks from the estimated complex spectrogram. In addition, we investigate gammatone frequency cepstral coefficients (GFCCs) as robust speaker features. Systematic evaluations and comparisons on the NIST SRE 2010 retransmitted corpus show that both monaural and multi-channel speech enhancement significantly outperform x-vector's performance, and our covariance matrix estimate is effective for the MVDR beamformer.