Effects of Expansion Algorithms on Speech Reception Thresholds

Abstract

Expansion is commonly used to reduce microphone noise and low-level environmental noises that can be annoying to hearing aid users. It may also improve or reduce the perception of low-level speech. This study assessed the impact of two expansion algorithms, single and multiple channel, on speech reception thresholds (SRT) with 10 hearing impaired listeners wearing hearing aids with ADRO processing. The single-channel algorithm suppressed sounds below 45 dB A, while the multiple-channel algorithm suppressed sounds below the long-term average spectrum of speech at either 55 or 45 dB SPL. The mean HINT SRTs in quiet were 39.4, 40.7, 40.6, and 41.8 dB A without expansion, with single-channel expansion, and with multiple-channel expansion at expansion thresholds of 45 and 55 dB SPL, respectively. The difference in mean SRT was only statistically significant between no expansion and multiple-channel expansion at a 55 dB SPL threshold. A regression analysis between the change in individual SRT for each expansion condition and pure tone average hearing loss showed no correlation. Our calculations indicate that only those with exceptionally good hearing will find microphone noise audible. The current practice of prescribing expansion algorithms based on hearing thresholds alone is questioned, and other rationales are discussed.

Hearing in Noise Test Brazil: standardization for young adults with normal hearing

“…In contrast, many international studies use HINT in free field conditions to assess the performance and specific characteristics of auxiliary hearing devices. 19 , 20 , 21 , 22 , 23 …”

Environmental noise reduction configuration: Effects on preferences, satisfaction, and speech understanding

“…The expansion ratio was set to 0.7:1 and the attack (rising level) and release (falling level) times were set to 20 and 200 ms, respectively, for all channels. This strategy of fitting multiple-channel expansion preserves the intelligibility of speech at the soft level of 55 dB SPL without inducing 'pumping' artefacts (Zakis & Wise, 2007;Wise & Zakis, 2008).…”