A mapping model of spectral tilt in normal-to-Lombard speech conversion for intelligibility enhancement

Li, Gang; Hu, Ruimin; Zhang, Rui; Wang, Xiaochen

doi:10.1007/s11042-020-08838-1

Cited by 6 publications

(4 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Shreyas Seshadri [20] used parallel DNN model instead of BGMM to map normal speech to lombard speech. In 2020, a mapping model based on linear prediction and tilt correction is proposed [16] . Compared with previous studies, the deep neural network is used to map high dimensions instead of Gaussian model, and tilt correction module is creatively added to further reduce the mapping error of formant amplitude.…”

Section: Transformation Methods Based On Feature Mappingmentioning

confidence: 99%

Compensation of Lombard effect in speech enhancement

2023

Third International Conference on Digital Signal and Computer Communications (DSCC 2023)

View full text Add to dashboard Cite

Influenced by Lombard effect, the synthetic noisy speech training data commonly used in speech enhancement algorithm based on deep learning is different from the pronunciation under actual noise, which leads to the limited application performance of the algorithm in the actual environment. This paper summarizes the data compensation strategy proposed by researchers for speech enhancement algorithm under Lombard effect. However, these compensation strategies mainly focus on the front-end data processing part of speech enhancement, which improves the complexity of speech enhancement system.

show abstract

Section: Transformation Methods Based On Feature Mappingmentioning

confidence: 99%

Compensation of Lombard effect in speech enhancement

2023

Third International Conference on Digital Signal and Computer Communications (DSCC 2023)

View full text Add to dashboard Cite

show abstract

“…The data-driven approach uses a large amount of speech data to build a feature mapping model from normal speech to Lombard speech by machine learning algorithms to achieve speech style conversion (SSC). In recent years, Bayesian-Gaussian mixture models (BGMM) [20,28], deep neural networks (DNNs) [29,30], recurrent neural networks (RNNs) and their variants such as long-short-term memory (LSTM) [31] networks have been widely used for mapping acoustic features. However, current parallel SSC methods require a parallel corpus of source and target speech and usually require a temporal alignment operation on the training data, an operation that may lead to some feature distortions.…”

Section: Related Workmentioning

confidence: 99%

D2StarGAN: A Near-Far End Noise Adaptive StarGAN for Speech Intelligibility Enhancement

Li,

Zhu,

Zhao

2023

Electronics

View full text Add to dashboard Cite

When using mobile communication, the voice output from the device is already relatively clear, but in a noisy environment, it is difficult for the listener to obtain the information expressed by the speaker with clarity. Consequently, speech intelligibility enhancement technology has emerged to help alleviate this problem. Speech intelligibility enhancement (IENH) is a technique that enhances speech intelligibility during the reception phase. Previous research has focused on IENH through normal versus different levels of Lombardic speech conversion, inspired by a well-known acoustic mechanism called the Lombard effect. However, these methods often lead to speech distortion and impair the overall speech quality. To address the speech quality degradation problem, we propose an improved (StarGAN)-based IENH framework by combining StarGAN networks with the dual discriminator idea to construct the conversion framework. This approach offers two main advantages: (1) Addition of a speech metric discriminator on top of StarGAN to optimize multiple intelligibility and quality-related metrics simultaneously; (2) a framework that is adaptive to different distal and proximal noise levels with different noise types. Experimental results from objective experiments and subjective preference tests show that our approach outperforms the baseline approach, and these enable IENH to be more widely used.

show abstract

“…At present, DNN speech enhancement can be roughly divided into two methods. The first is to seek the mapping between noise and pure speech spectrum [16], and the other is based on masking [17]. DNN-based speech enhancement usually produces excessively smooth speech, resulting in speech distortion and loss of intelligibility, and large DNNS take up large memory and slow training speed [18].…”

Section: Introductionmentioning

confidence: 99%

Research on Speech Enhancement Algorithm of Multiresolution Cochleagram Based on Skip Connection Deep Neural Network

Lan

Wang

Zhang³

et al. 2022

Journal of Sensors

View full text Add to dashboard Cite

The speech enhancement effect of traditional deep learning algorithms is not ideal under low signal-to-noise ratios (SNR). Skip connections-deep neural network (Skip-DNN) improves the traditional deep neural network (DNN) by adding skip connections between each layer of the neural network to solve the degradation problem of DNN. In this paper, the Multiresolution Cochleagram (MRCG) features in the gammachirp transform domain are denoised to obtain the improved MRCG (I-MRCG). The noise reduction method adopts the Minimum Mean-Square Error Short-Time Spectral Amplitude Estimator (MMSE-STSA) and takes I-MRCG as the input feature and Skip-DNN as the training network to improve the speech enhancement effect of the model. This paper also proposes an improved source-to-distortion ratio (SDR) loss function. When the loss function uses the improved SDR, it will improve the performance of Skip-DNN speech enhancement model. The experiments in this paper are performed on the Edinburgh dataset. When using I-MRCG as the input feature of Skip-DNN, the average perceptual evaluation of speech quality (PESQ) is 2.9137, and the average short-time objective intelligibility (STOI) is 0.8515. Compared with MRCG as Skip-DNN input features, the improvements are 0.91% and 0.71%, respectively. When the improved SDR is used as the loss function of the speech model, the average PESQ is 2.9699 and the average STOI is 0.8547. Compared with other loss functions, the improved SDR has a better enhancement effect when used as the loss function of the speech enhancement model.

show abstract

A mapping model of spectral tilt in normal-to-Lombard speech conversion for intelligibility enhancement

Cited by 6 publications

References 30 publications

Compensation of Lombard effect in speech enhancement

Compensation of Lombard effect in speech enhancement

D2StarGAN: A Near-Far End Noise Adaptive StarGAN for Speech Intelligibility Enhancement

Research on Speech Enhancement Algorithm of Multiresolution Cochleagram Based on Skip Connection Deep Neural Network

Contact Info

Product

Resources

About