Browsing by Author "Hsu, Chin-Cheng"

Now showing 1 - 3 of 3

Locally Linear Embedding Based Post-Filtering for Speech Enhancement
(2018-11-01)
Hwang, Hsin-Te
;
Wu, Yi-Chiao
;
Wang, Syu-Siang
;
Hsu, Chin-Cheng
;
Tsao, Yu
We present a novel speech enhancement method based on locally linear embedding (LLE). The proposed method works as a post-filter to further suppress the residual noises in the enhanced speech signals obtained by a speech enhancement system to attain improved speech quality and intelligibility. We design two types of LLE-based post-filters: the direct LLE-based post-filter (called the DL post-filter) and the LLE-based difference compensation post-filter (called the LDC post-filter). The key technique of the proposed post-filters is to apply the LLE-based feature prediction method, which integrates the LLE algorithm, a classical manifold learning method, with the exemplar-based feature prediction method, to predict either the spectral features of the clean speech from those of the enhanced speech (for DL) or the spectral difference of {clean speech; noisy speech} from that of {enhanced speech; noisy speech} (for LDC). As a result, for DL, the predicted clean speech signals can be directly reconstructed from the predicted clean spectral features. On the other hand, for LDC, the predicted clean spectral features are obtained by compensating the spectral features of the noisy speech with the predicted clean-noisy spectral difference, and then the predicted clean speech signals can be reconstructed accordingly. Experimental results demonstrate the effectiveness of the proposed post-filters for two representative speech enhancement methods, namely the deep denoising autoencoder (DDAE) and the minimum mean-square-error (MMSE) spectral estimation methods.
Robust emotion recognition by spectro-temporal modulation statistic features
(2012-03-01)
Chi, Tai-Shih
;
Yeh, Lan-Ying
;
Hsu, Chin-Cheng
;
電機工程學系
;
Department of Electrical and Computer Engineering
Most speech emotion recognition studies consider clean speech. In this study, statistics of joint spectro-temporal modulation features are extracted from an auditory perceptual model and are used to detect the emotion status of speech under noisy conditions. Speech samples were extracted from the Berlin Emotional Speech database and corrupted with white and babble noise under various SNR levels. This study investigates a clean train/noisy test scenario to simulate practical conditions with unknown noisy sources. Simulations demonstrate the redundancy of the proposed spectro-temporal modulation features and further consider the dimensionality reduction. The proposed modulation features achieve higher recognition rates of speech emotions under noisy conditions than (1) conventional mel-frequency cepstral coefficients combined with prosodic features; (2) official acoustic features adopted in the INTERSPEECH 2009 Emotion Challenge. Adding modulation features increased the recognition rates of INTERSPEECH proposed features by approximately 7% for all tested SNR conditions (20-0 dB).
Voice Conversion Based on Locally Linear Embedding
(2018-11-01)
Hwang, Hsin-Te
;
Wu, Yi-Chiao
;
Peng, Yu-Huai
;
Hsu, Chin-Cheng
;
Tsao, Yu
This paper presents a novel locally linear embedding (LLE)-based framework for exemplar -based spectral conversion (SC). The key feature of the proposed SC framework is that it integrates the LLE algorithm, a manifold learning method, with the conventional exemplar -based SC method. One important advantage of the LLE-based SC framework is that it can be applied to either one-to-one SC or many-to-one SC. For one-to-one SC, a parallel speech corpus consisting of the pre-specified source and target speakers' speeches is used to construct the paired source and target dictionaries in advance. During online conversion, the LLE-based SC method converts the source spectral features to the target like spectral features based on the paired dictionaries. On the other hand, when applied to many-to-one SC, our system is capable of converting the voice of any unseen source speaker to that of a desired target speaker, without the requirement of collecting parallel training speech utterances from them beforehand. To further improve the quality of the converted speech, the maximum likelihood parameter generation (MLPG) and global variance (GV) methods are adopted in the proposed SC systems. Experimental results demonstrate that the proposed one-to-one SC system is comparable with the state-of-the-art Gaussian mixture model (GMM)-based one-to-one SC system in terms of speech quality and speaker similarity, and the many-to-one SC system can approximate the performance of the one-to-one SC system.