EFFECTS OF SIMULATED REVERBERATION ON THE USE OF BINAURAL CUES AND FUNDAMENTAL-FREQUENCY DIFFERENCES FOR SEPARATING CONCURRENT VOWELS

Citation
Jf. Culling et al., EFFECTS OF SIMULATED REVERBERATION ON THE USE OF BINAURAL CUES AND FUNDAMENTAL-FREQUENCY DIFFERENCES FOR SEPARATING CONCURRENT VOWELS, Speech communication, 14(1), 1994, pp. 71-95
Citations number
72
Categorie Soggetti
Communication,"Language & Linguistics
Journal title
ISSN journal
01676393
Volume
14
Issue
1
Year of publication
1994
Pages
71 - 95
Database
ISI
SICI code
0167-6393(1994)14:1<71:EOSROT>2.0.ZU;2-O
Abstract
A computational simulation was used to generate impulse responses betw een points in a rectangular room and two points on opposite sides of a spherical ''head''. Sounds were convolved with the impulse responses to generate stimuli with which to study the effects of reverberation o n the ability of listeners to use differences in fundamental frequency (DELTAFos) to separate concurrent vowels. Experiment 1 verified the s uitability of the simulation by showing that it produced (i) appropria te percepts of lateralization, (ii) a larger contribution to lateraliz ation from interaural differences in timing than level, and (iii) no e ffects of reverberation on lateralization. Experiments 2-5 measured ma sked identification thresholds for synthetic harmonic ''target'' vowel s in the presence of masking sounds. In Experiment 2, listeners identi fied targets against pink-noise maskers. The experiment established a spatial geometry and a degree of reverberation for which listeners did not benefit from binaural cues arising from the spatial geometry of t he sources. Experiment 3 demonstrated that the same arrangement did no t undermine the ability to use DELTAFos to separate targets from vowel -like maskers when both had static Fo contours, but did prevent listen ers from using DELTAFos carried on coherently changing Fo contours. Ex periment 4 showed that a modulation width of +/- 1.45% was sufficient to reduce the benefits of DELTAFos, but that the benefits were not eli minated until the width of modulation exceeded the DELTAFo. It is argu ed that these results are compatible with existing models of the abili ty to use DELTAFos to separate concurrent vowels and that reverberatio n undermines the ability when the Fos are changing by diffusing the pe riodicities of the competing sources. Finally, Experiment 5 demonstrat ed that reverberation had no effect on the ability to separate a modul ated vowel from pink noise. Thus, reverberation may have its detriment al effects in these experiments by diffusing the periodicity of the ma sking sounds rather than the targets. Overall, the experiments demonst rate that DELTAFos can be more robust cues for separating concurrent s ounds than binaural cues. The relevance of these results to the percep tion of natural continuous speech is discussed.