Level-dependent changes in perception of temporal envelope cues.

Abstract

Level-dependent changes in speech recognition may reveal effects of basilar-membrane nonlinearities on temporal envelope fluctuations. It is hypothesized that, as a result of the compressive effects of the active cochlear mechanism, the “effective” magnitude of speech envelope fluctuations will be reduced as speech level increases from lower (more linear) to conversational (more compressive) regions. With further increases from conversational levels (to a more linear region), temporal envelope fluctuations will become more pronounced. Accordingly, speech recognition will be maximized at conversational levels due to the optimal “flattened” envelope and then decline at lower or higher levels. To test these assumptions, speech recognition scores were measured as a function of level for adults with normal hearing. Speech stimuli were spectrally degraded using “noise vocoder” processing so that perceptual effects of modifications to the speech temporal envelope can be revealed. As vocoded speech level increased, background noise level also increased, maintaining a fixed signal-to-noise ratio to minimize sensation-level effects on speech recognition scores. Discussion will focus on level-dependent effects for different speech stimuli and the role of nonlinearities on perception of temporal envelope cues. [Work supported by NIH/NIDCD]