Zhaoheng Ni

Yen-Ju Lu, Xuankai Chang, Chenda Li et al. · cmu

This paper presents recent progress on integrating speech separation and enhancement (SSE) into the ESPnet toolkit. Compared with the previous ESPnet-SE work, numerous features have been added, including recent state-of-the-art speech enhancement models with their respective training and evaluation recipes. Importantly, a new interface has been designed to flexibly combine speech enhancement front-ends with other tasks, including automatic speech recognition (ASR), speech translation (ST), and spoken language understanding (SLU). To showcase such integration, we performed experiments on carefully designed synthetic datasets for noisy-reverberant multi-channel ST and SLU tasks, which can be used as benchmark corpora for future research. In addition to these new tasks, we also use CHiME-4 and WSJ0-2Mix to benchmark multi- and single-channel SE approaches. Results show that the integration of SE front-ends with back-end tasks is a promising research direction even for tasks besides ASR, especially in the multi-channel scenario. The code is available online at https://github.com/ESPnet/ESPnet. The multi-channel ST and SLU datasets, which are another contribution of this work, are released on HuggingFace.

Brian Yan, Jiatong Shi, Yun Tang et al. · cmu, nvidia

ESPnet-ST-v2 is a revamp of the open-source ESPnet-ST toolkit necessitated by the broadening interests of the spoken language translation community. ESPnet-ST-v2 supports 1) offline speech-to-text translation (ST), 2) simultaneous speech-to-text translation (SST), and 3) offline speech-to-speech translation (S2ST) -- each task is supported with a wide variety of approaches, differentiating ESPnet-ST-v2 from other open source spoken language translation toolkits. This toolkit offers state-of-the-art architectures such as transducers, hybrid CTC/attention, multi-decoders with searchable intermediates, time-synchronous blockwise CTC/attention, Translatotron models, and direct discrete unit models. In this paper, we describe the overall design, example models for each task, and performance benchmarking behind ESPnet-ST-v2, which is publicly available at https://github.com/espnet/espnet.

William Chen, Xuankai Chang, Yifan Peng et al. · nvidia

Self-supervised learning (SSL) has led to great strides in speech processing. However, the resources needed to train these models has become prohibitively large as they continue to scale. Currently, only a few groups with substantial resources are capable of creating SSL models, which harms reproducibility. In this work, we optimize HuBERT SSL to fit in academic constraints. We reproduce HuBERT independently from the original implementation, with no performance loss. Our code and training optimizations make SSL feasible with only 8 GPUs, instead of the 32 used in the original work. We also explore a semi-supervised route, using an ASR model to skip the first pre-training iteration. Within one iteration of pre-training, our models improve over HuBERT on several tasks. Furthermore, our HuBERT Large variant requires only 8 GPUs, achieving similar performance to the original trained on 128. As our contribution to the community, all models, configurations, and code are made open-source in ESPnet.

Yang Li, Liangzhen Lai, Yuan Shangguan et al.

Transformer-based models excel in speech recognition. Existing efforts to optimize Transformer inference, typically for long-context applications, center on simplifying attention score calculations. However, streaming speech recognition models usually process a limited number of tokens each time, making attention score calculation less of a bottleneck. Instead, the bottleneck lies in the linear projection layers of multi-head attention and feedforward networks, constituting a substantial portion of the model size and contributing significantly to computation, memory, and power usage. To address this bottleneck, we propose folding attention, a technique targeting these linear layers, significantly reducing model size and improving memory and power efficiency. Experiments on on-device Transformer-based streaming speech recognition models show that folding attention reduces model size (and corresponding memory consumption) by up to 24% and power consumption by up to 23%, all without compromising model accuracy or computation overhead.

Yangyang Shi, Gael Le Lan, Varun Nagaraja et al.

This paper presents an innovative approach to enhance control over audio generation by emphasizing the alignment between audio and text representations during model training. In the context of language model-based audio generation, the model leverages input from both textual and audio token representations to predict subsequent audio tokens. However, the current configuration lacks explicit regularization to ensure the alignment between the chosen text representation and the language model's predictions. Our proposal involves the incorporation of audio and text representation regularization, particularly during the classifier-free guidance (CFG) phase, where the text condition is excluded from cross attention during language model training. The aim of this proposed representation regularization is to minimize discrepancies in audio and text similarity compared to other samples within the same training batch. Experimental results on both music and audio generation tasks demonstrate that our proposed methods lead to improvements in objective metrics for both audio and music generation, as well as an enhancement in the human perception for audio generation.

Ernie Chang, Sidd Srinivasan, Mahi Luthra et al.

Text-to-audio generation (TTA) produces audio from a text description, learning from pairs of audio samples and hand-annotated text. However, commercializing audio generation is challenging as user-input prompts are often under-specified when compared to text descriptions used to train TTA models. In this work, we treat TTA models as a ``blackbox'' and address the user prompt challenge with two key insights: (1) User prompts are generally under-specified, leading to a large alignment gap between user prompts and training prompts. (2) There is a distribution of audio descriptions for which TTA models are better at generating higher quality audio, which we refer to as ``audionese''. To this end, we rewrite prompts with instruction-tuned models and propose utilizing text-audio alignment as feedback signals via margin ranking learning for audio improvements. On both objective and subjective human evaluations, we observed marked improvements in both text-audio alignment and music audio quality.

Hao Shi, Yuan Gao, Zhaoheng Ni et al.

Serialized output training (SOT) attracts increasing attention due to its convenience and flexibility for multi-speaker automatic speech recognition (ASR). However, it is not easy to train with attention loss only. In this paper, we propose the overlapped encoding separation (EncSep) to fully utilize the benefits of the connectionist temporal classification (CTC) and attention hybrid loss. This additional separator is inserted after the encoder to extract the multi-speaker information with CTC losses. Furthermore, we propose the serialized speech information guidance SOT (GEncSep) to further utilize the separated encodings. The separated streams are concatenated to provide single-speaker information to guide attention during decoding. The experimental results on LibriMix show that the single-speaker encoding can be separated from the overlapped encoding. The CTC loss helps to improve the encoder representation under complex scenarios. GEncSep further improved performance.

Yao-Yuan Yang, Moto Hira, Zhaoheng Ni et al.

This document describes version 0.10 of TorchAudio: building blocks for machine learning applications in the audio and speech processing domain. The objective of TorchAudio is to accelerate the development and deployment of machine learning applications for researchers and engineers by providing off-the-shelf building blocks. The building blocks are designed to be GPU-compatible, automatically differentiable, and production-ready. TorchAudio can be easily installed from Python Package Index repository and the source code is publicly available under a BSD-2-Clause License (as of September 2021) at https://github.com/pytorch/audio. In this document, we provide an overview of the design principles, functionalities, and benchmarks of TorchAudio. We also benchmark our implementation of several audio and speech operations and models. We verify through the benchmarks that our implementations of various operations and models are valid and perform similarly to other publicly available implementations.

Shinji Watanabe, Michael Mandel, Jon Barker et al.

Following the success of the 1st, 2nd, 3rd, 4th and 5th CHiME challenges we organize the 6th CHiME Speech Separation and Recognition Challenge (CHiME-6). The new challenge revisits the previous CHiME-5 challenge and further considers the problem of distant multi-microphone conversational speech diarization and recognition in everyday home environments. Speech material is the same as the previous CHiME-5 recordings except for accurate array synchronization. The material was elicited using a dinner party scenario with efforts taken to capture data that is representative of natural conversational speech. This paper provides a baseline description of the CHiME-6 challenge for both segmented multispeaker speech recognition (Track 1) and unsegmented multispeaker speech recognition (Track 2). Of note, Track 2 is the first challenge activity in the community to tackle an unsegmented multispeaker speech recognition scenario with a complete set of reproducible open source baselines providing speech enhancement, speaker diarization, and speech recognition modules.

Zhaoheng Ni, Michael I Mandel

Speech separation is an essential task for multi-talker speech recognition. Recently many deep learning approaches are proposed and have been constantly refreshing the state-of-the-art performances. The lack of algorithm implementations limits researchers to use the same dataset for comparison. Building a generic platform can benefit researchers by easily implementing novel separation algorithms and comparing them with the existing ones on customized datasets. We introduce "onssen": an open-source speech separation and enhancement library. onssen is a library mainly for deep learning separation and enhancement algorithms. It uses LibRosa and NumPy libraries for the feature extraction and PyTorch as the back-end for model training. onssen supports most of the Time-Frequency mask-based separation algorithms (e.g. deep clustering, chimera net, chimera++, and so on) and also supports customized datasets. In this paper, we describe the functionality of modules in onssen and show the algorithms implemented by onssen achieve the same performances as reported in the original papers.

Ruizhe Huang, Xiaohui Zhang, Zhaoheng Ni et al.

Connectionist temporal classification (CTC) models are known to have peaky output distributions. Such behavior is not a problem for automatic speech recognition (ASR), but it can cause inaccurate forced alignments (FA), especially at finer granularity, e.g., phoneme level. This paper aims at alleviating the peaky behavior for CTC and improve its suitability for forced alignment generation, by leveraging label priors, so that the scores of alignment paths containing fewer blanks are boosted and maximized during training. As a result, our CTC model produces less peaky posteriors and is able to more accurately predict the offset of the tokens besides their onset. It outperforms the standard CTC model and a heuristics-based approach for obtaining CTC's token offset timestamps by 12-40% in phoneme and word boundary errors (PBE and WBE) measured on the Buckeye and TIMIT data. Compared with the most widely used FA toolkit Montreal Forced Aligner (MFA), our method performs similarly on PBE/WBE on Buckeye, yet falls behind MFA on TIMIT. Nevertheless, our method has a much simpler training pipeline and better runtime efficiency. Our training recipe and pretrained model are released in TorchAudio.

Jiahui Zhao, Hao Shi, Chenrui Cui et al.

Code-switching (CS) automatic speech recognition (ASR) faces challenges due to the language confusion resulting from accents, auditory similarity, and seamless language switches. Adaptation on the pre-trained multi-lingual model has shown promising performance for CS-ASR. In this paper, we adapt Whisper, which is a large-scale multilingual pre-trained speech recognition model, to CS from both encoder and decoder parts. First, we propose an encoder refiner to enhance the encoder's capacity of intra-sentence swithching. Second, we propose using two sets of language-aware adapters with different language prompt embeddings to achieve language-specific decoding information in each decoder layer. Then, a fusion module is added to fuse the language-aware decoding. The experimental results using the SEAME dataset show that, compared with the baseline model, the proposed approach achieves a relative MER reduction of 4.1% and 7.2% on the dev_man and dev_sge test sets, respectively, surpassing state-of-the-art methods. Through experiments, we found that the proposed method significantly improves the performance on non-native language in CS speech, indicating that our approach enables Whisper to better distinguish between the two languages.

Vineel Pratap, Andros Tjandra, Bowen Shi et al.

Expanding the language coverage of speech technology has the potential to improve access to information for many more people. However, current speech technology is restricted to about one hundred languages which is a small fraction of the over 7,000 languages spoken around the world. The Massively Multilingual Speech (MMS) project increases the number of supported languages by 10-40x, depending on the task. The main ingredients are a new dataset based on readings of publicly available religious texts and effectively leveraging self-supervised learning. We built pre-trained wav2vec 2.0 models covering 1,406 languages, a single multilingual automatic speech recognition model for 1,107 languages, speech synthesis models for the same number of languages, as well as a language identification model for 4,017 languages. Experiments show that our multilingual speech recognition model more than halves the word error rate of Whisper on 54 languages of the FLEURS benchmark while being trained on a small fraction of the labeled data.

Yen-Ju Lu, Samuele Cornell, Xuankai Chang et al.

This paper describes our submission to the L3DAS22 Challenge Task 1, which consists of speech enhancement with 3D Ambisonic microphones. The core of our approach combines Deep Neural Network (DNN) driven complex spectral mapping with linear beamformers such as the multi-frame multi-channel Wiener filter. Our proposed system has two DNNs and a linear beamformer in between. Both DNNs are trained to perform complex spectral mapping, using a combination of waveform and magnitude spectrum losses. The estimated signal from the first DNN is used to drive a linear beamformer, and the beamforming result, together with this enhanced signal, are used as extra inputs for the second DNN which refines the estimation. Then, from this new estimated signal, the linear beamformer and second DNN are run iteratively. The proposed method was ranked first in the challenge, achieving, on the evaluation set, a ranking metric of 0.984, versus 0.833 of the challenge baseline.

Qiquan Zhang, Qi Song, Zhaoheng Ni et al.

Most studies on speech enhancement generally don't consider the energy distribution of speech in time-frequency (T-F) representation, which is important for accurate prediction of mask or spectra. In this paper, we present a simple yet effective T-F attention (TFA) module, where a 2-D attention map is produced to provide differentiated weights to the spectral components of T-F representation. To validate the effectiveness of our proposed TFA module, we use the residual temporal convolution network (ResTCN) as the backbone network and conduct extensive experiments on two commonly used training targets. Our experiments demonstrate that applying our TFA module significantly improves the performance in terms of five objective evaluation metrics with negligible parameter overhead. The evaluation results show that the proposed ResTCN with the TFA module (ResTCN+TFA) consistently outperforms other baselines by a large margin.

Felix Grezes, Zhaoheng Ni, Viet Anh Trinh et al.

Recurrent neural networks using the LSTM architecture can achieve significant single-channel noise reduction. It is not obvious, however, how to apply them to multi-channel inputs in a way that can generalize to new microphone configurations. In contrast, spatial clustering techniques can achieve such generalization, but lack a strong signal model. This paper combines the two approaches to attain both the spatial separation performance and generality of multichannel spatial clustering and the signal modeling performance of multiple parallel single-channel LSTM speech enhancers. The system is compared to several baselines on the CHiME3 dataset in terms of speech quality predicted by the PESQ algorithm and word error rate of a recognizer trained on mis-matched conditions, in order to focus on generalization. Our experiments show that by combining the LSTM models with the spatial clustering, we reduce word error rate by 4.6\% absolute (17.2\% relative) on the development set and 11.2\% absolute (25.5\% relative) on test set compared with spatial clustering system, and reduce by 10.75\% (32.72\% relative) on development set and 6.12\% absolute (15.76\% relative) on test data compared with LSTM model.

Zhaoheng Ni, Felix Grezes, Viet Anh Trinh et al.

Spatial clustering techniques can achieve significant multi-channel noise reduction across relatively arbitrary microphone configurations, but have difficulty incorporating a detailed speech/noise model. In contrast, LSTM neural networks have successfully been trained to recognize speech from noise on single-channel inputs, but have difficulty taking full advantage of the information in multi-channel recordings. This paper integrates these two approaches, training LSTM speech models to clean the masks generated by the Model-based EM Source Separation and Localization (MESSL) spatial clustering method. By doing so, it attains both the spatial separation performance and generality of multi-channel spatial clustering and the signal modeling performance of multiple parallel single-channel LSTM speech enhancers. Our experiments show that when our system is applied to the CHiME-3 dataset of noisy tablet recordings, it increases speech quality as measured by the Perceptual Evaluation of Speech Quality (PESQ) algorithm and reduces the word error rate of the baseline CHiME-3 speech recognizer, as compared to the default BeamformIt beamformer.

Felix Grezes, Zhaoheng Ni, Viet Anh Trinh et al.

Recent works have shown that Deep Recurrent Neural Networks using the LSTM architecture can achieve strong single-channel speech enhancement by estimating time-frequency masks. However, these models do not naturally generalize to multi-channel inputs from varying microphone configurations. In contrast, spatial clustering techniques can achieve such generalization but lack a strong signal model. Our work proposes a combination of the two approaches. By using LSTMs to enhance spatial clustering based time-frequency masks, we achieve both the signal modeling performance of multiple single-channel LSTM-DNN speech enhancers and the signal separation performance and generality of multi-channel spatial clustering. We compare our proposed system to several baselines on the CHiME-3 dataset. We evaluate the quality of the audio from each system using SDR from the BSS\_eval toolkit and PESQ. We evaluate the intelligibility of the output of each system using word error rate from a Kaldi automatic speech recognizer.

Zhaoheng Ni, Yong Xu, Meng Yu et al.

This paper aims at eliminating the interfering speakers' speech, additive noise, and reverberation from the noisy multi-talker speech mixture that benefits automatic speech recognition (ASR) backend. While the recently proposed Weighted Power minimization Distortionless response (WPD) beamformer can perform separation and dereverberation simultaneously, the noise cancellation component still has the potential to progress. We propose an improved neural WPD beamformer called "WPD++" by an enhanced beamforming module in the conventional WPD and a multi-objective loss function for the joint training. The beamforming module is improved by utilizing the spatio-temporal correlation. A multi-objective loss, including the complex spectra domain scale-invariant signal-to-noise ratio (C-Si-SNR) and the magnitude domain mean square error (Mag-MSE), is properly designed to make multiple constraints on the enhanced speech and the desired power of the dry clean signal. Joint training is conducted to optimize the complex-valued mask estimator and the WPD++ beamformer in an end-to-end way. The results show that the proposed WPD++ outperforms several state-of-the-art beamformers on the enhanced speech quality and word error rate (WER) of ASR.