Level-Dependent Changes in Perception of Speech Envelope Cues

Abstract

Level-dependent changes in temporal envelope fluctuations in speech and related changes in speech recognition may reveal effects of basilar-membrane nonlinearities. As a result of compression in the basilar-membrane response, the "effective" magnitude of envelope fluctuations may be reduced as speech level increases from lower level (more linear) to midlevel (more compressive) regions. With further increases to a more linear region, speech envelope fluctuations may become more pronounced. To assess these effects, recognition of consonants and key words in sentences was measured as a function of speech level for younger adults with normal hearing. Consonant-vowel syllables and sentences were spectrally degraded using "noise vocoder" processing to maximize perceptual effects of changes to the speech envelope. Broadband noise at a fixed signal-to-noise ratio maintained constant audibility as speech level increased. Results revealed significant increases in scores and envelope-dependent feature transmission from 45 to 60 dB SPL and decreasing scores and feature transmission from 60 to 85 dB SPL. This quadratic pattern, with speech recognition maximized at mid levels and poorer at lower and higher levels, is consistent with a role of cochlear nonlinearities in perception of speech envelope cues.K e y w o r d s : b a s i l a r -m e m b r a n e r e s p o n s e s , compression, human, speech envelope, vocoder

Compression and amplification algorithms in hearing aids impair the selectivity of neural responses to speech

“…The failure of both the hearing aid and linear amplification to restore consonant identification in noise could reflect a supra-threshold auditory processing deficit with hearing loss that is only manifest in difficult listening conditions, but this is not necessarily the case. Even with normal hearing, the intelligibility of speech in noise decreases as overall intensity increases (an effect known as ‘rollover’ with a complex physiological basis 13 , 15 , 36 ). When the background noise is dominated by low frequencies (as is the case for multi-talker noise), speech intelligibility decreases by approximately 5% for every 10 dB increase in overall intensity above moderate levels, even when the speech-to-noise ratio remains constant 14 , 37 .…”

“…One likely contributor is the broadening of cochlear frequency tuning with increasing sound level, which decreases the frequency selectivity of individual auditory nerve fibres and increases the spread of masking from one frequency to another 48 . But rollover is also apparent when speech is processed to contain primarily temporal cues, suggesting that there are contributions from additional factors such as increased cochlear compression at high intensities that distorts the speech envelope or reduced differential sensitivity of auditory nerve fibres at intensities that exceed their dynamic range 36 . The simplest way to avoid rollover is, of course, to decrease the intensity of incoming sounds.…”

“…The crossover frequencies between channels were 200, 500, 1000, 1750, 2750, 4000, 5500, 7000, and 8500 Hz. The intensity thresholds below which amplification was linear for each channel were 45,43,40,38,35,33,28,30,36, and 44 dB SPL. The attack and release times (the time constants of the changes in gain following an increase or decrease in the intensity of the incoming sound, respectively) for all channels were 5 and 40 ms, respectively.…”

Adaptation to Noise in Human Speech Recognition Depends on Noise-Level Statistics and Fast Dynamic-Range Compression

“…Therefore, it is tempting to hypothesize that auditory compression facilitates noise adaptation. The data from Dubno et al (2012) appear to be consistent with this idea. They did not measure noise adaptation but showed that the recognition of vocoded speech-in-noise at a constant SNR was better at midlevels (60 dB SPL), where basilar membrane responses are more compressive, than at lower or higher speech levels, where basilar membrane responses are more linear.…”

The hearing aid dilemma: amplification, compression, and distortion of the neural code

“…The failure of both the hearing aid and linear amplification to restore consonant identification in noise could reflect a supra-threshold auditory processing deficit with hearing loss that is only manifest in difficult listening conditions, but this is not necessarily the case. Even with normal hearing, the intelligibility of speech in noise decreases as overall intensity increases (an effect known as 'rollover' with a complex physiological basis (Dubno et al, 2012;Horvath and Lesica, 2011;Wong et al, 1998)). When the background noise is dominated by low frequencies (as is the case for multi-talker noise), speech intelligibility decreases by approximately 5% for every 10 dB increase in overall intensity above moderate levels, even when the speech-to-noise ratio remains constant (Hornsby et al, 2005;Studebaker et al, 1999).…”