These tools will no longer be maintained as of December 31, 2024. Archived website can be found here. PubMed4Hh GitHub repository can be found here. Contact NLM Customer Service if you have questions.


Pubmed for Handhelds

PUBMED FOR HANDHELDS

Search MEDLINE/PubMed

  • Title: Acoustic parameters for classification of breathiness in continuous speech according to the GRBAS scale.
    Author: Stráník A, Čmejla R, Vokřál J.
    Journal: J Voice; 2014 Sep; 28(5):653.e9-653.e17. PubMed ID: 24755168.
    Abstract:
    OBJECTIVES: The purpose of this study was to classify objectively breathiness in continuous speech according to a subjective evaluation of voice based on the GRBAS scale. STUDY DESIGN: A retrospective, experimental study. METHODS: A total of 593 records with read text were twice evaluated by five experts according to the GRBAS scale within two sessions with a time delay of at least 2 weeks. The records were subsequently subjected to acoustic analysis using parameters which do not rely on the accurate estimation of fundamental frequency: Glottal-to-Noise Excitation ratio, Cepstral Peak Prominence Pearson r at autocorrelation peak, Breathiness Index, and the ratio of high- to mid/low-frequency energy. These parameters were subsequently analyzed and a total of 92 features were created for each record. After feature space reduction based on Correlation Feature Selection and Information Gain, the feature space was reduced to four parameters. These four parameters were used for classification of breathiness. RESULTS: In the final set of four, the acoustic parameters have significantly different mean ranks in every grade of breathiness according to the GRBAS scale (Kruskal-Wallis test [P < 0.001]). The accuracy of classifier for objective evaluation of level of breathiness based on the discrete scale of breathiness reached 77%. Assuming continuous grades of breathiness, the classifier reached ρ = 0.92 (P < 0.001). CONCLUSIONS: The level of breathiness in continuous speech can be effectively described by automatic system-based analysis of acoustic measures. The proposed automatic system is able to determine the level of breathiness in continuous speech with sufficient precision.
    [Abstract] [Full Text] [Related] [New Search]