Speech-in-noise measures: variable versus fixed speech and noise levels

Area of Science:

• Audiology
• Speech-Language Pathology
• Psychoacoustics

Background:

• Assessing speech perception in noise is crucial for understanding hearing abilities.
• The revised speech perception in noise (R-SPIN) test is a standard tool for this assessment.
• Variations in testing paradigms, such as fixed vs. variable signal-to-noise ratios, may influence results.

Purpose of the Study:

• To compare speech recognition performance under two auditory conditions: fixed speech level with varying noise, and fixed noise level with varying speech.
• To determine if experimental design impacts speech perception in noise outcomes.

Main Methods:

• A quasi-experimental design using the R-SPIN test (Lists 3 & 4) with high (HP) and low (LP) predictability words.
• Modified R-SPIN into a multiple signal-to-noise ratio paradigm (23 to -1 dB).
• Calculated 50% points using the Spearman-Kärber equation for 16 young normal-hearing and 48 older hearing-impaired listeners.

Main Results:

• Normal-hearing listeners outperformed hearing-impaired listeners on both HP and LP conditions.
• HP sentences yielded better performance than LP sentences for all participants.
• Mean 50% points were marginally better (0.1–0.4 dB lower) in the speech-variable, babble-fixed condition compared to the speech-fixed, babble-variable condition.

Conclusions:

• The observed differences (≤ 0.4 dB) between the two testing paradigms are negligible for practical speech recognition assessment.
• These small variations do not significantly alter speech perception in noise outcomes.
• The R-SPIN test demonstrates robust performance measurement across these tested conditions.