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

Speech feature extraction method using subband-based periodicity and nonperiodicity decomposition.

Kentaro Ishizuka1, Tomohiro Nakatani, Yasuhiro Minami

  • 1NTT Communication Science Laboratories, NTT Corporation, Hikaridai 2-4, Seikacho, Sourakugun, Kyoto 619-0237, Japan. ishizuka@cslab.kecl.ntt.co.jp

The Journal of the Acoustical Society of America
|August 1, 2006
PubMed
Summary

This study introduces a novel speech feature extraction method using periodicity and nonperiodicity for improved automatic speech recognition. The technique enhances noise robustness, especially when training and testing data have different noise characteristics.

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

Deficiency of C/EBPβ in pancreatic acinar cells exacerbates inflammation in the early phase of acute pancreatitis.

Biochemical and biophysical research communications·2026
Same author

Remote Ischemic Conditioning for Acute Ischemic Stroke: The RICAIS Randomized Clinical Trial.

Journal of atherosclerosis and thrombosis·2026
Same author

High-Sensitivity C-Reactive Protein and Risk of Recurrent Vascular Events in Patients With Stroke Associated With Complex Aortic Atheroma.

Circulation reports·2026
Same author

Association Between High-Density Lipoprotein Cholesterol Levels and Prognoses in Patients With Stroke.

Journal of the American Heart Association·2026
Same author

Cumulative association of triglycerides and interleukin-6 with recurrent vascular events after ischemic stroke.

Journal of clinical lipidology·2026
Same author

Answer to the Letter to the Editor of D. Liao, et al. concerning "Do postoperative changes in physical function affect patient-reported outcomes in patients with lumbar spinal stenosis undergoing rehabilitation? A secondary analysis of a randomized controlled trial" by M. Minetama, et al. (Eur Spine J [2025]; doi: 10.1007/s00586-025-09302-0).

European spine journal : official publication of the European Spine Society, the European Spinal Deformity Society, and the European Section of the Cervical Spine Research Society·2025

Area of Science:

  • Speech Processing
  • Auditory Perception
  • Signal Analysis

Background:

  • Automatic speech recognition (ASR) systems often struggle with noise robustness.
  • Existing methods may not effectively handle variations in noise between training and testing data.
  • Speech perception research offers insights into robust auditory processing.

Purpose of the Study:

  • To develop a novel speech feature extraction method for robust automatic speech recognition.
  • To leverage the auditory comb filtering hypothesis for improved signal decomposition.
  • To enhance ASR performance in noisy conditions, particularly with differing noise characteristics.

Main Methods:

  • Decomposition of subband signals into periodic and nonperiodic components using independently designed comb filters.

Related Experiment Videos

  • Utilizing both periodic and nonperiodic features for speech representation.
  • Employing a parallel distributed processing framework inspired by speech perception.

    • Main Results:

    • The proposed method demonstrates superior performance compared to conventional methods in noisy environments.
    • Robustness against noise spectrum bias is achieved through independent subband periodicity estimation.
    • The feature representation maintains speech information content while exploiting noise resilience of periodicity measurements.

    Conclusions:

    • The proposed periodicity-nonperiodicity based feature extraction offers enhanced robustness for automatic speech recognition.
    • The method's design, motivated by speech perception, provides advantages over speech production-based approaches.
    • Effective performance is confirmed in continuous digit recognition tasks with varying noise conditions.