Related Experiment Video
Updated: Aug 14, 2026
A Two-interval Forced-choice Task for Multisensory Comparisons
Published on: November 9, 2018
Modified Multiple Stimulus With Hidden Reference and Anchors-Gabrielsson Total Impression Sound Quality Rating
Mohamed Rahme1,2, Paula Folkeard1, Steve Beaulac1
1National Centre for Audiology, Western University, London, Ontario, Canada.
Purpose:
The purpose of the study was to obtain, analyze, and compare subjective sound quality data for the same test stimuli using modified multistimulus MUSHRA (Multiple Stimulus with Hidden Reference and Anchors) based procedures (viz., MUSHRA with custom anchors and MUSHRA without anchor) and the single-stimulus Gabrielsson's total impression rating procedure.
Method:
Twenty normally hearing young adults were recruited in this study. Participants completed sound quality ratings on two different hearing aid recording data sets-Data Set A contained speech recordings from four different hearing aids under a variety of noisy and processing conditions, and Data Set B contained speech recordings from a single hearing aid under a combination of different noisy, reverberant, and signal processing conditions. Recordings in both data sets were rated for their quality using the total impression rating procedure. In addition, quality ratings of Data Set A recordings were obtained using a MUSHRA with custom anchors, while the ratings of Data Set B recordings were collected using a MUSHRA without anchor.
Results:
Statistical analyses revealed a high test-retest reliability of quality ratings for the same stimuli that were rated multiple times. In addition, high-interrater reliability was observed with all three rating procedures. Further analyses indicated (a) a high correlation between the total impression rating and the two modified MUSHRA ratings and (b) a similar relationship between the average and standard deviation of the subjective rating data obtained by the total impression rating and MUSHRA with custom anchors on Data Set A, and the total impression rating and the MUSHRA without anchor on Data Set B.
Conclusion:
Both sound quality procedures, namely, the MUSHRA-based procedures and the total impression rating scale, obtained similar quality ratings of varied hearing aid speech recordings with high reliability.
