Effects of high-frequency suppression for speech recognition in noise in Spanish normal-hearing subjects.

Hypothesis: Subjects with normal hearing (NH) experience lower performance in speech understanding in noise when frequency components of speech above 8 kHz are removed.

Background: Previous studies have explored speech perception in noise under various circumstances, but none have been specifically designed to explore the relevance of frequencies above 8 kHz and, more specifically, for the Spanish language.

Methods: Twenty-nine subjects with NH, aged 19 to 55 years, and native speakers of the Spanish language listened to two sets of three disyllabic word lists with background noise. One set of words was unfiltered and included frequency components up to 22 kHz, and another set was band-pass filtered to include frequency components between 70 Hz and 8 kHz. Words were presented at 65 dB sound pressure level. Each set of words was presented with background noise with the same bandwidth limit as the accompanying list and adjusted to construct signal-to-noise ratio (SNR) conditions of +5, 0, and -5 dB.

Results: Results demonstrate a higher performance for unfiltered words at +5 and -5 dB SNR, although not at 0 dB SNR. For SNR +5 dB and -5 dB, the average success rate in word recognition was 17% higher when words were not filtered. For the case SNR = 0, however, both conditions yield statistically similar results.

Conclusion: This study suggests that high-frequency components above 8 kHz contribute to speech understanding in noise for subjects with NH of the Spanish language. Given these findings, it would be interesting to determine if bandwidth limitations in current hearing prostheses may contribute to known difficulties with perception of speech in noise in hearing-impaired subjects.