Speech Source Separation Using Variational Autoencoder and Bandpass Filter

Duc, Hao; Son, Tran Hung; Chau, Duc Thanh

doi:10.1109/access.2020.3019495

Cited by 18 publications

(6 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In [ 116 ], the researchers used two data sets, TIMIT and VIVOS. TIMIT is a speech corpus for American English; it has 8 dialects and 630 speakers.…”

Section: Applicationsmentioning

confidence: 99%

An Overview of Variational Autoencoders for Source Separation, Finance, and Bio-Signal Applications

Singh

Ogunfunmi

2021

Entropy

View full text Add to dashboard Cite

Autoencoders are a self-supervised learning system where, during training, the output is an approximation of the input. Typically, autoencoders have three parts: Encoder (which produces a compressed latent space representation of the input data), the Latent Space (which retains the knowledge in the input data with reduced dimensionality but preserves maximum information) and the Decoder (which reconstructs the input data from the compressed latent space). Autoencoders have found wide applications in dimensionality reduction, object detection, image classification, and image denoising applications. Variational Autoencoders (VAEs) can be regarded as enhanced Autoencoders where a Bayesian approach is used to learn the probability distribution of the input data. VAEs have found wide applications in generating data for speech, images, and text. In this paper, we present a general comprehensive overview of variational autoencoders. We discuss problems with the VAEs and present several variants of the VAEs that attempt to provide solutions to the problems. We present applications of variational autoencoders for finance (a new and emerging field of application), speech/audio source separation, and biosignal applications. Experimental results are presented for an example of speech source separation to illustrate the powerful application of variants of VAE: VAE, β-VAE, and ITL-AE. We conclude the paper with a summary, and we identify possible areas of research in improving performance of VAEs in particular and deep generative models in general, of which VAEs and generative adversarial networks (GANs) are examples.

show abstract

“…In [ 116 ], the researchers used two data sets, TIMIT and VIVOS. TIMIT is a speech corpus for American English; it has 8 dialects and 630 speakers.…”

Section: Applicationsmentioning

confidence: 99%

An Overview of Variational Autoencoders for Source Separation, Finance, and Bio-Signal Applications

Singh

Ogunfunmi

2021

Entropy

View full text Add to dashboard Cite

show abstract

“…In signal and speech processing, monaural speech source separation is challenging because it separates the target speaker from the mixture of speakers and the background noises and interferences in a single microphone recording. Speaker separation [10]- [13], speech enhancement [14], [15], and speech de-reverberation and denoising [16] come under single-channel source separation categories, as in Fig. 1.…”

Section: Introductionmentioning

confidence: 99%

“…Speaker separation allows extracting more than one speaker from the mixture of two and more than two speakers [10]- [13]. Speech enhancement improves noisy speech signals' intelligibility and perceptual quality [14], [15] and attempts to separate speech from noisy mixture signals.…”

Section: Introductionmentioning

confidence: 99%

State-of-the-Art Analysis of Deep Learning-Based Monaural Speech Source Separation Techniques

2023

View full text Add to dashboard Cite

The monaural speech source separation problem is an important application in the signal processing field. But recent interaction of deep learning algorithms with signal processing achieves remarkable performance improvement for speech source separation problems. This paper explores the numerous state-of-the-art deep learning-based monaural speech source separation algorithms in the timefrequency (T-F), time, and hybrid domains. The motivation, algorithm, and framework of different deep learning models for monaural speech source separation are analyzed. The benchmarked algorithms in the T-F domain can be categorized as deep neural networks (DNN), clustering, permutation, multi-task learning, computational auditory sense analysis (CASA), and phase reconstruction-based techniques, whereas the state-of-the-art time-domain approaches can be categorized as CNN, RNN, multi-scale fusion (MSF), and transformer-based techniques. The end-to-end post filter (E2EPF) is a hybrid algorithm combining T-F and time-domain works to achieve enhanced results. Time-domain models have shown improvement in separation performance compared to the T-F and hybrid domain models with small model sizes. Methods in T-F, time, and hybrid domains are compared using 𝑆𝐷𝑅, 𝑆𝐼 − 𝑆𝐷𝑅, 𝑆𝐼 − 𝑆𝑁𝑅, PESQ, and 𝑆𝑇𝑂𝐼 as quality assessment metrics on some benchmark datasets.

show abstract

“…B LIND source separation (BSS) refers to the process of separating and recovering signals by using only the observed signals when the source signals and the transmission channel are unknown [1]. After years of research and development, BSS has been successfully applied in many fields, including military communications [2], image processing [3], speech signal processing [4], and other fields. BSS is defined as underdetermined blind source separation (UBSS) when the number of observed signals is less than the number of source signals [5].…”

Section: Introductionmentioning

confidence: 99%

Mixing Matrix Estimation Algorithm for Underdetermined Blind Source Separation

Bai

Shi³

et al. 2021

IEEE Access

View full text Add to dashboard Cite

An original algorithm for underdetermined mixing matrix estimation is proposed in this paper, which can estimate the mixing matrix effectively when the number of source signals is unknown. Firstly, a new single source point (SSP) detection algorithm based on transformation matrix is proposed, which can effectively detect single source time-frequency (TF) points by using the characteristics of complex ratio and improve the sparsity of source signals. In view of the fact that the number of source signals is unknown, a novel estimation algorithm based on element sorting is proposed, which can significantly improve the estimation accuracy of the number of source signals. Finally, the mixing matrix is estimated by using the cluster center obtained by agglomerative hierarchical clustering (AHC) algorithm. The simulation results show that the proposed algorithm can improve the estimation accuracy of the number of source signals and underdetermined mixing matrix obviously, and the algorithm has higher robustness compared with other algorithms.

show abstract

Speech Source Separation Using Variational Autoencoder and Bandpass Filter

Cited by 18 publications

References 31 publications

An Overview of Variational Autoencoders for Source Separation, Finance, and Bio-Signal Applications

An Overview of Variational Autoencoders for Source Separation, Finance, and Bio-Signal Applications

State-of-the-Art Analysis of Deep Learning-Based Monaural Speech Source Separation Techniques

Mixing Matrix Estimation Algorithm for Underdetermined Blind Source Separation

Contact Info

Product

Resources

About