Exploring the effects of masker spectro-temporal coherence on the informational masking of speech

Abstract

The effect of an extraneous formant on intelligibility is influenced by the depth of variation in its frequency contour. This study explored whether masker impact also depends on spectro-temporal coherence, using a method ensuring that interference occurs only through informational masking. Three-formant analogs of natural sentences were generated using a monotonous periodic source. Target formants were presented monaurally; the target ear was assigned randomly on each trial. A competitor for F2 (F2C) was presented contralaterally; listeners must reject F2C to optimize recognition. In the reference condition, F2C was created by inverting the F2 frequency contour and using a constant RMS-matched amplitude contour. In the coherent conditions, the F2C frequency contour was preserved but the amplitude contour was divided into abutting 100- or 200-ms segments using a raised-cosine envelope (10-ms rise/fall); these values were informed by typical syllable durations. Segment order was randomized in the incoherent conditions, introducing abrupt discontinuities into the F2C frequency contour. Adding F2C lowered keyword scores, but to the same extent for the reference, coherent, and incoherent conditions. This suggests that the impact on intelligibility depends critically on the overall extent of frequency variation in the interferer, but not on its spectro-temporal coherence. [Work supported by ESRC.]