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 model of auditory perception as front end for automatic speech recognition.

J Tchorz1, B Kollmeier

  • 1Carl von Ossietzky Universität Oldenburg, AG Medizinische Physik, Germany.

The Journal of the Acoustical Society of America
|October 26, 1999
PubMed
Summary

This study introduces an auditory-based front end for speech recognition, demonstrating superior noise robustness compared to traditional methods. Key findings highlight adaptive compression and low-pass filtering as crucial for effective noise reduction in speech signals.

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

Automatic Detection of Gait Perturbations With Everyday Wearable Technology.

IEEE open journal of engineering in medicine and biology·2025
Same author

[Application of the matrix sentence test Russian version in children].

Vestnik otorinolaringologii·2020
Same author

Speech Intelligibility Enhancement by Interaural Magnification.

Acta oto-laryngologica·2020
Same author

Speech Enhancement by Filtering in the Loudness Domain.

Acta oto-laryngologica·2020
Same author

[The application of the screening questionnaires for the diagnostics of the central auditory processing disorders in the pediatric practice].

Vestnik otorinolaringologii·2018
Same author

Decision criteria for replacement of fillings: a retrospective study.

Clinical and experimental dental research·2018

Area of Science:

  • Auditory Processing
  • Speech Recognition
  • Signal Processing

Background:

  • Current automatic speech recognition (ASR) systems often struggle with noise.
  • Mel-scale cepstral features are common but can be sensitive to additive noise.
  • Peripheral auditory processing offers a biologically inspired approach to robust feature extraction.

Purpose of the Study:

  • To propose and evaluate an auditory-based front end for ASR.
  • To enhance the robustness of speech representations in noisy conditions.
  • To identify key processing stages contributing to noise resilience.

Main Methods:

  • Developed a quantitative model of peripheral auditory processing, simulating spectral and temporal auditory system properties.
  • Evaluated the model's speech representation robustness in speaker-independent, isolated word recognition tasks with additive noise.

Related Experiment Videos

  • Investigated the impact of modifying different processing stages, including adaptive compression and low-pass filtering.

    • Main Results:

    • The proposed auditory front end demonstrated higher robustness in noise compared to mel-scale cepstral features.
    • Adaptive compression was identified as the most critical stage for robust speech representation in noise.
    • Low-pass filtering of the frequency band envelope further improved noise reduction.

    Conclusions:

    • An auditory-based front end offers a significant advantage in speech recognition robustness against noise.
    • Adaptive compression and specific filtering techniques are vital for enhancing noise resilience in speech signal processing.
    • This biologically inspired approach holds promise for improving ASR performance in real-world noisy environments.