Publication | Closed Access
Differentiated perceptual evaluation of pathological voice quality: reliability and correlations with acoustic measurements.
275
Citations
0
References
1996
Year
Voice DisordersPathological SpeechPathological Voice QualityPerceptual EvaluationSpeech ScienceVoice AnalysisVoice EvaluationSpeech RecognitionDeviant Voice QualityVoice QualityAcoustic MeasurementsAcoustic AnalysisHealth SciencesSpeech PerceptionAudiologyRehabilitationSpeech AcousticSpeech CommunicationSpeech AnalysisVoiceSpeech AcousticsSpeech ProcessingClinical PracticeArts
The perceptual GRBAS scale for deviant voice quality (completed with a "I" parameter: Instability = fluctuation of voice quality over time) was tested in 5 different institutes on 943 voice patients. Each voice was evaluated separately by 2 professionals. The interrater correlation reaches 0.87 for G (grade), 0.70 for R (roughness), 0.69 for B (breathiness) and 0.65 for A/S (asthenicity/strain). Experience with the scale significantly improves the interrater agreement. The intrarater correlation (80 voices; time interval 2-6 months) is systematically slightly stronger than the interrater correlation. Further, GIRBAS/acoustical correlations were investigated on 80 pathological voices, by using the Multidimensional Voice Program (MDVP) of Kay. Principal component analysis allows reduction of 21 MDVP parameters to 5 clinically relevant aspects, with a typical acoustical measurement for each of them. The strongest correlations are found between G and Shimmer/Noise to harmonics ratio, R and Jitter, and B and Shimmer. As tremor and diplophonia are easily identified perceptually and acoustically, it seems useful to add this information respectively to AS and R. The GIRBAS scale seems to be a valuable instrument for clinical practice.