Across-formant integration and speech intelligibility: Effects of acoustic source properties in the presence and absence of a contralateral interferer

Abstract

The role of source properties in across-formant integration was explored using three-formant (F1+F2+F3) analogues of natural sentences (targets). In experiment 1, F1+F3 were harmonic analogues (H1+H3) generated using a monotonous buzz source and second-order resonators; in experiment 2, F1+F3 were tonal analogues (T1+T3). F2 could take either form (H2 or T2). Target formants were always presented monaurally; the receiving ear was assigned randomly on each trial. In some conditions, only the target was present; in others, a competitor for F2 (F2C) was presented contralaterally. Buzz-excited or tonal competitors were created using the time-reversed frequency and amplitude contours of F2. Listeners must reject F2C to optimize keyword recognition. Whether or not a competitor was present, there was no effect of source mismatch between F1+F3 and F2. The impact of adding F2C was modest when it was tonal but large when it was harmonic, irrespective of whether F2C matched F1+F3. This pattern was maintained when harmonic and tonal counterparts were loudness-matched (experiment 3). Source type and competition, rather than acoustic similarity, governed the phonetic contribution of a formant. Contrary to earlier research using dichotic targets, requiring across-ear integration to optimize intelligibility, H2C was an equally effective informational masker for H2 as for T2.

Dichotic integration of acoustic-phonetic information: Competition from extraneous formants increases the effect of second-formant attenuation on intelligibility

“…The experiments reported here were not intended to resolve these issues but the results help to constrain what is needed from a full account of the perceptual organization of speech. In particular, there is an interesting parallel between the results of the current study and some of our previous work exploring how listeners integrate and segregate formants in stimulus ensembles with mixed source properties Summers et al, 2016). Most notably, our ability to use the acoustic-phonetic information carried by a sinewave analogue of F2 may be impaired greatly by the presence of an RMS-matched buzz-excited F2C in the other ear, but not when the source properties of F2 and F2C are reversed.…”

“…Even when a formant ensemble is configured such that the physical exclusion of one formant changes one intelligible stimulus (/ru/) into another (/li/), introducing a difference in F0 of nine semitones between that formant and the others is not sufficient to eliminate the /ru/ percept (Darwin, 1981;Gardner et al, 1989). Furthermore, under competitive conditions, there are circumstances in which listeners fail to combine formants with shared acoustic source properties and to exclude extraneous formants with radically mismatched source propertiesnotably, when all the target formants are sine-wave analogues and the extraneous formant is rendered as a buzz-excited resonance Summers et al, 2016;see below). Some researchers have appealed to a speech-specific notion of phonetic coherence based on the plausibility of the articulatory gestures implied by the time-varying properties of formants in an ensemble rather than on general-purpose grouping cues (see, e.g., Liberman, 1982;Mann and Liberman, 1983;Remez et al, 1994;Remez, 2001Remez, , 2003Remez, , 2005, but to our knowledge this concept has never been clearly defined acoustically.…”

“…Note that these mixtures introduced more than one candidate for each of the lowest three formants, which could then compete with one another for inclusion in a particular perceptual organization. Furthermore, some of our recent studies of perceptual organization and informational masking using formant ensembles have shown that the presence or absence of a competitor formant can profoundly affect the extent to which the acoustic-phonetic information carried by a given formant is integrated into a coherent percept Summers et al, 2016). These findingsand their implications for speech perception under adverse listening conditions-are considered further in Sec.…”

“…For example, listeners are usually able to understand sine-wave analogues of speech, a highly unnatural stimulus in which the lowest few formants of the speech signal are each replaced by a timevarying sinusoid tracking the frequency and amplitude contour of that formant (Bailey et al, 1977;Remez et al, 1981). Listeners are also capable of combining the acoustic-phonetic information carried by formants with different excitation source properties, such as differences in fundamental frequency (F0; e.g., Cutting, 1976) or stimuli for which some formants are rendered as buzz-excited resonances and others as sine-wave analogues Summers et al, 2016).…”

Informational masking of speech depends on masker spectro-temporal variation but not on its coherence

“…To the extent that a constant-amplitude extraneous formant can have a greater impact than one with a time-varying amplitude contour , this is probably a secondary effect arising from greater illumination of the underlying formant-frequency contour owing to the absence of low-amplitude intervals. To date, the only other factors shown to modulate the impact of an interferer on intelligibility to an extent comparable with that of formant-frequency change are radical differences between formants in acoustic source properties, such as harmonic vs sine-wave analogues Summers et al, 2016) and, to a lesser extent, differences in fundamental frequency (DF0) (Summers et al, 2010(Summers et al, , 2017. Nonetheless, IM can remain considerable even when there is a clearly discernible DF0 between the target and interfering formants (4 semitones) (Summers and Roberts, 2020).…”