Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

A statistical model based fundamental frequency synthesizer for Mandarin speech.

S H Chen1, S Chang, S M Lee

  • 1Department of Communication Engineering, Tung University, Hsinchu, Taiwan, Republic of China.

The Journal of the Acoustical Society of America
|July 1, 1992
PubMed
Summary

A new statistical model synthesizes Mandarin speech's fundamental frequency (F0) by learning linguistic context. This approach accurately reconstructs F0 contours, improving naturalness in text-to-speech systems.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Transverse-momentum and pseudorapidity distributions of charged hadrons in pp collisions at square root of s = 7 TeV.

Physical review letters·2010
Same author

Determination of the absolute order of shadow moiré fringes by using two differently colored light sources.

Applied optics·2010
Same author

Genetics and vaccine efficacy: host genetic variation affecting Marek's disease vaccine efficacy in White Leghorn chickens.

Poultry science·2010
Same author

Sonographic evaluation of renal transplant complications.

Journal of medical imaging and radiation oncology·2010
Same author

Observation of the 3-D structure of a small object using a phase conjugate fiberscope.

Applied optics·2010
Same author

Dynamic interferometry using two phase conjugate waves.

Applied optics·2010

Area of Science:

  • Speech Technology
  • Computational Linguistics
  • Acoustic Phonetics

Background:

  • Fundamental-frequency (F0) contour is crucial for prosody in Mandarin text-to-speech (TTS).
  • Accurate F0 synthesis remains a challenge, particularly in capturing linguistic influences on pitch patterns.
  • Existing methods often struggle to implicitly learn and apply phonological rules.

Purpose of the Study:

  • To propose a novel statistical model for Mandarin F0 synthesis in TTS.
  • To automatically deduce and implicitly memorize phonological rules within the model.
  • To enhance the naturalness and accuracy of synthesized speech.

Main Methods:

  • Utilizing a statistical model to correlate syllable F0 contour patterns with contextual linguistic features.

Related Experiment Videos

  • Empirically estimating model parameters from a large corpus of sentential utterances.
  • Employing the Viterbi algorithm with extracted contextual features for F0 contour synthesis.

    • Main Results:

    • The model implicitly learned declination effects and sandhi rules.
    • 77.56% of synthesized F0 contours matched the VQ-quantized original speech.
    • Informal listening tests indicated naturalness in the synthesized speech.

    Conclusions:

    • The proposed statistical model effectively synthesizes Mandarin F0 contours.
    • The method implicitly captures complex phonological phenomena.
    • This approach offers a promising direction for improving Mandarin TTS systems.