arXiv subjects

Kohei Saijo

Publications and source records attributed to Kohei Saijo.

30 records · Page 2Linked to original sources

arXiv2024

TF-Locoformer: Transformer with Local Modeling by Convolution for Speech Separation and Enhancement

Time-frequency (TF) domain dual-path models achieve high-fidelity speech separation. While some previous state-of-the-art (SoTA) models rely on RNNs, this reliance means they lack the parallelizability, scalability, and versatility of Transformer blocks. Given the wide-ranging success of pure Transformer-based architectures in other fields, in this work we focus on removing the RNN from TF-domain dual-path models, while maintaining SoTA performance. This work presents TF-Locoformer, a Transformer-based model with LOcal-modeling by COnvolution. The model uses feed-forward networks (FFNs) with convolution layers, instead of linear layers, to capture local information, letting the self-attention focus on capturing global patterns. We place two such FFNs before and after self-attention to enhance the local-modeling capability. We also introduce a novel normalization for TF-domain dual-path models. Experiments on separation and enhancement datasets show that the proposed model meets or exceeds SoTA in multiple benchmarks with an RNN-free architecture.

eess.AS↗

arXiv2024

URGENT Challenge: Universality, Robustness, and Generalizability For Speech Enhancement

The last decade has witnessed significant advancements in deep learning-based speech enhancement (SE). However, most existing SE research has limitations on the coverage of SE sub-tasks, data diversity and amount, and evaluation metrics. To fill this gap and promote research toward universal SE, we establish a new SE challenge, named URGENT, to focus on the universality, robustness, and generalizability of SE. We aim to extend the SE definition to cover different sub-tasks to explore the limits of SE models, starting from denoising, dereverberation, bandwidth extension, and declipping. A novel framework is proposed to unify all these sub-tasks in a single model, allowing the use of all existing SE approaches. We collected public speech and noise data from different domains to construct diverse evaluation data. Finally, we discuss the insights gained from our preliminary baseline experiments based on both generative and discriminative SE methods with 12 curated metrics.

eess.AS↗

arXiv2024

Beyond Performance Plateaus: A Comprehensive Study on Scalability in Speech Enhancement

Deep learning-based speech enhancement (SE) models have achieved impressive performance in the past decade. Numerous advanced architectures have been designed to deliver state-of-the-art performance; however, their scalability potential remains unrevealed. Meanwhile, the majority of research focuses on small-sized datasets with restricted diversity, leading to a plateau in performance improvement. In this paper, we aim to provide new insights for addressing the above issues by exploring the scalability of SE models in terms of architectures, model sizes, compute budgets, and dataset sizes. Our investigation involves several popular SE architectures and speech data from different domains. Experiments reveal both similarities and distinctions between the scaling effects in SE and other tasks such as speech recognition. These findings further provide insights into the under-explored SE directions, e.g., larger-scale multi-domain corpora and efficiently scalable architectures.

eess.AS↗

arXiv2024

Toward Universal Speech Enhancement for Diverse Input Conditions

The past decade has witnessed substantial growth of data-driven speech enhancement (SE) techniques thanks to deep learning. While existing approaches have shown impressive performance in some common datasets, most of them are designed only for a single condition (e.g., single-channel, multi-channel, or a fixed sampling frequency) or only consider a single task (e.g., denoising or dereverberation). Currently, there is no universal SE approach that can effectively handle diverse input conditions with a single model. In this paper, we make the first attempt to investigate this line of research. First, we devise a single SE model that is independent of microphone channels, signal lengths, and sampling frequencies. Second, we design a universal SE benchmark by combining existing public corpora with multiple conditions. Our experiments on a wide range of datasets show that the proposed single model can successfully handle diverse conditions with strong performance.

eess.AS↗

arXiv2023

A Single Speech Enhancement Model Unifying Dereverberation, Denoising, Speaker Counting, Separation, and Extraction

We propose a multi-task universal speech enhancement (MUSE) model that can perform five speech enhancement (SE) tasks: dereverberation, denoising, speech separation (SS), target speaker extraction (TSE), and speaker counting. This is achieved by integrating two modules into an SE model: 1) an internal separation module that does both speaker counting and separation; and 2) a TSE module that extracts the target speech from the internal separation outputs using target speaker cues. The model is trained to perform TSE if the target speaker cue is given and SS otherwise. By training the model to remove noise and reverberation, we allow the model to tackle the five tasks mentioned above with a single model, which has not been accomplished yet. Evaluation results demonstrate that the proposed MUSE model can successfully handle multiple tasks with a single model.

eess.AS↗

arXiv2023

Exploring Speech Recognition, Translation, and Understanding with Discrete Speech Units: A Comparative Study

Speech signals, typically sampled at rates in the tens of thousands per second, contain redundancies, evoking inefficiencies in sequence modeling. High-dimensional speech features such as spectrograms are often used as the input for the subsequent model. However, they can still be redundant. Recent investigations proposed the use of discrete speech units derived from self-supervised learning representations, which significantly compresses the size of speech data. Applying various methods, such as de-duplication and subword modeling, can further compress the speech sequence length. Hence, training time is significantly reduced while retaining notable performance. In this study, we undertake a comprehensive and systematic exploration into the application of discrete units within end-to-end speech processing models. Experiments on 12 automatic speech recognition, 3 speech translation, and 1 spoken language understanding corpora demonstrate that discrete units achieve reasonably good results in almost all the settings. We intend to release our configurations and trained models to foster future research efforts.

cs.CL↗

arXiv2023

Exploring Speech Enhancement for Low-resource Speech Synthesis

High-quality and intelligible speech is essential to text-to-speech (TTS) model training, however, obtaining high-quality data for low-resource languages is challenging and expensive. Applying speech enhancement on Automatic Speech Recognition (ASR) corpus mitigates the issue by augmenting the training data, while how the nonlinear speech distortion brought by speech enhancement models affects TTS training still needs to be investigated. In this paper, we train a TF-GridNet speech enhancement model and apply it to low-resource datasets that were collected for the ASR task, then train a discrete unit based TTS model on the enhanced speech. We use Arabic datasets as an example and show that the proposed pipeline significantly improves the low-resource TTS system compared with other baseline methods in terms of ASR WER metric. We also run empirical analysis on the correlation between speech enhancement and TTS performances.

eess.AS↗

arXiv2023

Self-Remixing: Unsupervised Speech Separation via Separation and Remixing

We present Self-Remixing, a novel self-supervised speech separation method, which refines a pre-trained separation model in an unsupervised manner. The proposed method consists of a shuffler module and a solver module, and they grow together through separation and remixing processes. Specifically, the shuffler first separates observed mixtures and makes pseudo-mixtures by shuffling and remixing the separated signals. The solver then separates the pseudo-mixtures and remixes the separated signals back to the observed mixtures. The solver is trained using the observed mixtures as supervision, while the shuffler's weights are updated by taking the moving average with the solver's, generating the pseudo-mixtures with fewer distortions. Our experiments demonstrate that Self-Remixing gives better performance over existing remixing-based self-supervised methods with the same or less training costs under unsupervised setup. Self-Remixing also outperforms baselines in semi-supervised domain adaptation, showing effectiveness in multiple setups.

eess.AS↗

arXiv2023

Remixing-based Unsupervised Source Separation from Scratch

We propose an unsupervised approach for training separation models from scratch using RemixIT and Self-Remixing, which are recently proposed self-supervised learning methods for refining pre-trained models. They first separate mixtures with a teacher model and create pseudo-mixtures by shuffling and remixing the separated signals. A student model is then trained to separate the pseudo-mixtures using either the teacher's outputs or the initial mixtures as supervision. To refine the teacher's outputs, the teacher's weights are updated with the student's weights. While these methods originally assumed that the teacher is pre-trained, we show that they are capable of training models from scratch. We also introduce a simple remixing method to stabilize training. Experimental results demonstrate that the proposed approach outperforms mixture invariant training, which is currently the only available approach for training a monaural separation model from scratch.

eess.AS↗

arXiv2022

Independence-based Joint Dereverberation and Separation with Neural Source Model

We propose an independence-based joint dereverberation and separation method with a neural source model. We introduce a neural network in the framework of time-decorrelation iterative source steering, which is an extension of independent vector analysis to joint dereverberation and separation. The network is trained in an end-to-end manner with a permutation invariant loss on the time-domain separation output signals. Our proposed method can be applied in any situation with at least as many microphones as sources, regardless of their number. In experiments, we demonstrate that our method results in high performance in terms of both speech quality metrics and word error rate (WER), even for mixtures with a different number of speakers than training. Furthermore, the model, trained on synthetic mixtures, without any modifications, greatly reduces the WER on the recorded dataset LibriCSS.

eess.AS↗

arXiv2022

Spatial Loss for Unsupervised Multi-channel Source Separation

We propose a spatial loss for unsupervised multi-channel source separation. The proposed loss exploits the duality of direction of arrival (DOA) and beamforming: the steering and beamforming vectors should be aligned for the target source, but orthogonal for interfering ones. The spatial loss encourages consistency between the mixing and demixing systems from a classic DOA estimator and a neural separator, respectively. With the proposed loss, we train the neural separators based on minimum variance distortionless response (MVDR) beamforming and independent vector analysis (IVA). We also investigate the effectiveness of combining our spatial loss and a signal loss, which uses the outputs of blind source separation as the reference. We evaluate our proposed method on synthetic and recorded (LibriCSS) mixtures. We find that the spatial loss is most effective to train IVA-based separators. For the neural MVDR beamformer, it performs best when combined with a signal loss. On synthetic mixtures, the proposed unsupervised loss leads to the same performance as a supervised loss in terms of word error rate. On LibriCSS, we obtain close to state-of-the-art performance without any labeled training data.

eess.AS↗

arXiv2022

Remix-cycle-consistent Learning on Adversarially Learned Separator for Accurate and Stable Unsupervised Speech Separation

A new learning algorithm for speech separation networks is designed to explicitly reduce residual noise and artifacts in the separated signal in an unsupervised manner. Generative adversarial networks are known to be effective in constructing separation networks when the ground truth for the observed signal is inaccessible. Still, weak objectives aimed at distribution-to-distribution mapping make the learning unstable and limit their performance. This study introduces the remix-cycle-consistency loss as a more appropriate objective function and uses it to fine-tune adversarially learned source separation models. The remix-cycle-consistency loss is defined as the difference between the mixed speech observed at microphones and the pseudo-mixed speech obtained by alternating the process of separating the mixed sound and remixing its outputs with another combination. The minimization of this loss leads to an explicit reduction in the distortions in the output of the separation network. Experimental comparisons with multichannel speech separation demonstrated that the proposed method achieved high separation accuracy and learning stability comparable to supervised learning.

eess.AS↗