Learning and Modeling Unit Embeddings Using Deep Neural Networks for Unit-Selection-Based Mandarin Speech Synthesis

2020 ◽  
Vol 19 (3) ◽  
pp. 1-14
Author(s):  
Xiao Zhou ◽  
Zhen-Hua Ling ◽  
Li-Rong Dai
Keyword(s):  
Neural Networks ◽  
Speech Synthesis ◽  
Deep Neural Networks ◽  
Unit Selection

LSTM Deep Neural Networks Postfiltering for Enhancing Synthetic Voices

2017 ◽  
Vol 32 (01) ◽  
pp. 1860008 ◽  
Author(s):  
Marvin Coto-Jiménez ◽  
John Goddard-Close
Keyword(s):  
Neural Networks ◽  
Speech Synthesis ◽  
Deep Neural Networks ◽  
Short Term Memory ◽  
Markov Models ◽  
Natural Speech ◽  
Objective Measures ◽  
Recent Developments ◽  
Small Footprint ◽  
Synthetic Voices

Recent developments in speech synthesis have produced systems capable of producing speech which closely resembles natural speech, and researchers now strive to create models that more accurately mimic human voices. One such development is the incorporation of multiple linguistic styles in various languages and accents. Speech synthesis based on Hidden Markov Models (HMM) is of great interest to researchers, due to its ability to produce sophisticated features with a small footprint. Despite some progress, its quality has not yet reached the level of the current predominant unit-selection approaches, which select and concatenate recordings of real speech, and work has been conducted to try to improve HMM-based systems. In this paper, we present an application of long short-term memory (LSTM) deep neural networks as a postfiltering step in HMM-based speech synthesis. Our motivation stems from a similar desire to obtain characteristics which are closer to those of natural speech. The paper analyzes four types of postfilters obtained using five voices, which range from a single postfilter to enhance all the parameters, to a multi-stream proposal which separately enhances groups of parameters. The different proposals are evaluated using three objective measures and are statistically compared to determine any significance between them. The results described in the paper indicate that HMM-based voices can be enhanced using this approach, specially for the multi-stream postfilters on the considered objective measures.

scholarly journals Expiry date recognition using deep neural networks

2020 ◽  
Vol 13 (1) ◽  
pp. 1-17
Author(s):  
Traian Rebedea ◽  
Vlad Florea
Keyword(s):  
Neural Networks ◽  
Character Recognition ◽  
Optical Character Recognition ◽  
Speech Synthesis ◽  
Efficient Solution ◽  
Deep Neural Networks ◽  
Accuracy Improvement ◽  
Expiry Date ◽  
Food Items ◽  
Optical Character

This paper proposes a deep learning solution for optical character recognition, specifically tuned to detect expiration dates that are printed on the packaging of food items. This method can be used to reduce food waste, having a significant impact on the design of smart refrigerators and can prove especially useful for persons with vision difficulties, by combining it with a speech synthesis engine. The main problem in designing an efficient solution for expiry date recognition is the lack of a large enough dataset to train deep neural networks. To tackle this issue, we propose to use an additional dataset composed of synthetically generated images. Both the synthetic and real image datasets are detailed in the paper and we show that the proposed method offers a 9.4% accuracy improvement over using real images alone.

scholarly journals Development of Vietnamese Speech Synthesis System using Deep Neural Networks

2019 ◽  
Vol 34 (4) ◽  
pp. 349-363 ◽  
Author(s):  
Thinh Van Nguyen ◽  
Bao Quoc Nguyen ◽  
Kinh Huy Phan ◽  
Hai Van Do
Keyword(s):  
Neural Networks ◽  
Markov Model ◽  
Hidden Markov Model ◽  
Speech Synthesis ◽  
Deep Neural Networks ◽  
Hidden Markov ◽  
Experimental Results ◽  
Training Data ◽  
The Internet ◽  
Synthesis System

In this paper, we present our first Vietnamese speech synthesis system based on deep neural networks. To improve the training data collected from the Internet, a cleaning method is proposed. The experimental results indicate that by using deeper architectures we can achieve better performance for the TTS than using shallow architectures such as hidden Markov model. We also present the effect of using different amounts of data to train the TTS systems. In the VLSP TTS challenge 2018, our proposed DNN-based speech synthesis system won the first place in all three subjects including naturalness, intelligibility, and MOS.

scholarly journals A review of Serbian parametric speech synthesis based on deep neural networks

2017 ◽  
Vol 9 (1) ◽  
pp. 32-37 ◽  
Author(s):  
Tijana Delic ◽  
Milan Secujski ◽  
Sinisa Suzic
Keyword(s):  
Neural Networks ◽  
Speech Synthesis ◽  
Deep Neural Networks ◽  
Parametric Speech Synthesis
Sign in / Sign up

Export Citation Format

Share Document