A Revised Speech Intelligibility Rating (RSIR) test: listeners with normal hearing

Area of Science:

• Audiology
• Speech Science
• Psychoacoustics

Background:

• The Speech Intelligibility Rating (SIR) test by Cox and McDaniel is a key tool for assessing speech intelligibility.
• Developing a large set of equivalent passages is crucial for reliable and valid intelligibility measures.
• Previous methods using overall signal-to-babble (S/B) ratios showed variability in passage equivalence.

Purpose of the Study:

• To develop a Revised Speech Intelligibility Rating (RSIR) test with a large number of equivalent passages.
• To enhance the reliability and accuracy of speech intelligibility measurements.
• To evaluate two methods for calculating the signal-to-babble ratio (dB S/B) for passage equivalence.

Main Methods:

• Two experiments utilized 72 connected discourse passages and a method-of-adjustment (MOA) procedure.
• Experiment 1 calculated S/B ratio based on overall root-mean-square (rms) speech and babble levels.
• Experiment 2 refined S/B ratio using long-term rms levels across 16 one-third-octave bands, adjusting for passage-specific differences.

Main Results:

• Experiment 1: Mean intelligibility was -1.43 dB S/B, with 42 passages within +/- 0.5 dB of the mean.
• Experiment 2: Mean intelligibility was -8.06 dB S/B, with 64 passages within +/- 0.5 dB of the mean.
• Experiment 2 demonstrated improved passage equivalence, with a 95% critical difference of 0.72 dB for five MOAs (10.8% estimated percentage critical difference).

Conclusions:

• The refined S/B ratio calculation in Experiment 2 significantly improved passage equivalence for the RSIR test.
• The developed RSIR test provides a more reliable measure of speech intelligibility.
• This research contributes to more accurate and dependable speech intelligibility assessments in audiology and speech science.