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

VOWEL DURATION MEASUREMENT USING DEEP NEURAL NETWORKS.

Yossi Adi1, Joseph Keshet1, Matthew Goldrick2

  • 1Dept. of Computer Science, Bar-Ilan University, Ramat-Gan, Israel.

IEEE International Workshop on Machine Learning for Signal Processing : [Proceedings]. IEEE International Workshop on Machine Learning for Signal Processing
|October 17, 2017
PubMed
Summary

This study developed an automatic algorithm for measuring vowel duration using deep neural networks. The convolutional neural network (CNN) approach showed promise, offering an efficient alternative to manual phonetic annotation.

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

Unique Challenges of Multidisciplinary Clinical Science: Perspectives from a Multidisciplinary Team.

Clinical psychological science : a journal of the Association for Psychological Science·2026
Same author

Speech as an objective measure of psychomotor dysfunction in major depressive disorder: validation from non-speech motor measures.

Psychiatry research·2026
Same author

Open-vocabulary Keyword Spotting with Hyper-Matched Filters for Small Footprint Devices.

Computer speech & language·2026
Same author

Automatic Measurement of Voice Onset Time and Prevoicing using Recurrent Neural Networks.

Interspeech·2026
Same author

How does a deep neural network look at lexical stress in English words?

The Journal of the Acoustical Society of America·2026
Same author

Predicting relative intelligibility from inter-talker distances in a perceptual similarity space for speech.

Psychonomic bulletin & review·2025

Area of Science:

  • Phonetics
  • Computational Linguistics
  • Speech Processing

Background:

  • Vowel duration is crucial in phonetic studies but traditionally relies on time-consuming manual annotation.
  • Existing methods for measuring vowel duration are often subjective and labor-intensive, hindering large-scale research.

Purpose of the Study:

  • To develop an accurate, automated algorithm for measuring vowel duration in consonant-vowel-consonant (CVC) speech segments.
  • To compare the performance of deep neural network architectures against traditional methods.

Main Methods:

  • Implemented and trained two deep neural network architectures: Convolutional Neural Network (CNN) and Deep Belief Network (DBN).
  • Utilized frame-level analysis on manually annotated phonetic data.
  • Compared the accuracy of CNN and DBN against a Hidden Markov Model (HMM)-based forced aligner.
Keywords:
convolution neural networksdeep belief networksforced alignmenthidden Markov modelsvowel duration measurement

Related Experiment Videos

Main Results:

  • The Convolutional Neural Network (CNN) architecture outperformed the Deep Belief Network (DBN).
  • Both CNN and the HMM-based forced aligner achieved comparable accuracy.
  • Neither automated method fully replicated the precision of manual annotation models.

Conclusions:

  • Deep neural networks, particularly CNNs, offer a viable and efficient alternative for automatic vowel duration measurement.
  • Further refinement is needed to match the accuracy of manual phonetic annotation.
  • The developed algorithm can significantly streamline phonetic research by reducing manual annotation efforts.