Reverberant overlap- and self-masking in consonant identification.

Two effects of reverberation on the identification of consonants were evaluated for ten normal-hearing subjects: (1) the overlap of energy of a preceding consonant on the following consonant, called ''overlap-masking''; and (2) the internal temporal smearing of energy within each consonant, called ''self-masking.'' The stimuli were eight consonants /p,t,k,f,m,n,l,w/. The consonants were spoken in /s-at/ context (experiment 1) and generated by a speech synthesizer in /s-at/ and /-at/ contexts (experiment 2). In both experiments, identification of consonants was tested in four conditions: (1) quiet, without degradations; (2) with a babble of voices; (3) with noise that was shaped like either natural or synthetic /s/ for the two experiments, respectively; and (4) with room reverberation. The results for the natural and synthetic syllables indicated that the effect of reverberation on identification of consonants following /s/ was not comparable to masking by either the /s/-spectrum-shaped noise or the babble. In addition, the results for the synthetic syllables indicated that most of the errors in reverberation for the /s-at/ context were similar to a sum of errors in two conditions: (1) with /s/-shaped noise causing overlap masking; and (2) with reverberation causing self-masking within each consonant. [ABSTRACT FROM AUTHOR]