Limited error based event localizing temporal decomposition and its application to variable-rate speech coding

Nguyen, Phu Chien; Akagi, Masato; Nguyen, Binh P.

doi:10.1016/j.specom.2007.02.007

Cited by 16 publications

(26 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A number of modifications have been explored to overcome these drawbacks. In this study, we employ the MRTD algorithm [13]. The reasons for using the MRTD algorithm in this work are twofold: (i) the MRTD algorithm enforces a new property on event functions, named the ''well-shapedness'' property, to model the temporal structure of speech more effectively [13]; (ii) event targets can convey the speaker's identity [19].…”

Section: Temporal Decompositionmentioning

confidence: 99%

“…In this study, we employ the MRTD algorithm [13]. The reasons for using the MRTD algorithm in this work are twofold: (i) the MRTD algorithm enforces a new property on event functions, named the ''well-shapedness'' property, to model the temporal structure of speech more effectively [13]; (ii) event targets can convey the speaker's identity [19]. In the MRTD algorithm, LSF parameters are chosen for the input of TD, because of their spectral sensitivity (an adverse alteration of one coefficient results in a spectral change only around that frequency [20]) and their stability and interpolation advantages (LSFs result in low spectral distortion when being interpolated and/or quantized [21]).…”

Section: Temporal Decompositionmentioning

confidence: 99%

“…The STRAIGHT spectra are suitable for TD, because they are smooth in the time-frequency domain. More details of the extraction of LSF parameters from STRAIGHT can be found in [13]. To determine the initial event locations, the MRTD algorithm uses a spectral stability criterion based on LSF parameters.…”

Section: Temporal Decompositionmentioning

confidence: 99%

“…This paper proposes a new spectral modification method to address two issues, insufficient smoothness of modified spectra between frames and ineffective spectral modification. First, we propose a new modeling of speech spectra for spectral modification based on temporal decomposition (TD) [12,13] and Gaussian mixture model (GMM) [14][15][16]. To model the spectral evolution, we employ the modified restricted temporal decomposition (MRTD) algorithm [13].…”

Section: Introductionmentioning

confidence: 99%

“…First, we propose a new modeling of speech spectra for spectral modification based on temporal decomposition (TD) [12,13] and Gaussian mixture model (GMM) [14][15][16]. To model the spectral evolution, we employ the modified restricted temporal decomposition (MRTD) algorithm [13]. To model the speech spectrum, we use GMM parameters [14][15][16].…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

A flexible spectral modification method based on temporal decomposition and Gaussian mixture model

Nguyen

Akagi

2009

Acoust. Sci. & Tech.

Self Cite

View full text Add to dashboard Cite

Manipulating spectral structure often leads to degradation of speech quality, which is mainly due to insufficient smoothness of the modified spectra between frames, and ineffective spectral modification. This paper presents a new spectral modification method to improve the quality of modified speech. If frames are processed independently, discontinuous features may be generated. Therefore, a speech analysis technique called temporal decomposition (TD), which decomposes speech into event targets and event functions, is used to model the spectral evolution effectively. Instead of modifying the speech spectra frame by frame, we only need to modify event targets and event functions. This feature leads to easy modification of the speech spectra, and the smoothness of modified speech is ensured by the shape of event functions. To improve spectral modification, we explore Gaussian mixture model parameters (spectral-GMM parameters) to model the spectral envelope of each event target, and develop a new algorithm for modifying spectral-GMM parameters in accordance with formant scaling factors. We first evaluate the effectiveness of our proposed method in spectra modeling, and then apply it to two areas which require different amounts of spectral modification, emotional speech synthesis and voice gender conversion. Experimental results show that the effectiveness of our proposed method is verified for spectra modeling and spectral modification.

show abstract

Section: Temporal Decompositionmentioning

confidence: 99%

Section: Temporal Decompositionmentioning

confidence: 99%

Section: Temporal Decompositionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

A flexible spectral modification method based on temporal decomposition and Gaussian mixture model

Nguyen

Akagi

2009

Acoust. Sci. & Tech.

Self Cite

View full text Add to dashboard Cite

show abstract

A three-layered model for expressive speech perception

Huang

Akagi

2008

Speech Communication

Self Cite

View full text Add to dashboard Cite

English syllabic consonants vs. schwa at discourse level

Arboleda

Monroy

2015

RESLA

View full text Add to dashboard Cite

This is a perceptual study on the production of English syllabic consonants vs. schwa in word final position at discourse level. It is intended to find out whether the speaker’s gender, accent, speech rate and emphasis placed upon words have any bearing on the production of this alternation. The effect the aforementioned factors may have on the production of this apparently anarchic phonetic alternation demands further exploration (see, however, O’Shaughnessy, 1981; Byrd, 1994; Wells, 1995; Töft, 2002). The informants for this study were 80 non-rhotic native newsreaders (40 males and 40 females) taken from theBBC learning English website(2009). Three female listeners not knowledgeable about the purposes of the study had to decide whether they perceived a syllabic consonant in certain words (800 overall) or not. Results show that there is a statistically significant relationship between each of these factors and the production of English syllabic consonants vs. schwa.

show abstract

Limited error based event localizing temporal decomposition and its application to variable-rate speech coding

Cited by 16 publications

References 17 publications

A flexible spectral modification method based on temporal decomposition and Gaussian mixture model

A flexible spectral modification method based on temporal decomposition and Gaussian mixture model

A three-layered model for expressive speech perception

English syllabic consonants vs. schwa at discourse level

Contact Info

Product

Resources

About