Release from speech-on-speech masking in a front-and-back geometry

Abstract

Informational masking of a target female talker by female distracters was measured with target and distracters presented from directly in front of the listener as a baseline condition. Next, it was found that if the distracters were also presented from directly in back of the listener, advanced or delayed by a few milliseconds with respect to the distracters in front, release from informational masking occurred. Release from informational masking was found for all delays within the Haas region of +/-50 ms, with peak release of about 3.5 dB. This peak occurred for a delay of +/-2 ms and it was shown to be the result of delay-and-add filtering. Release from energetic masking was also found, but only for delays of +/-0.5 ms or less.

Spatial release from speech-on-speech masking in the median sagittal plane

“…Increasing target and masker source separation to 100 , however, did not result in significantly more masking release (À50 : F(1,7) ¼ 0.14, P ¼ 0.718; þ50 : F(1,7) ¼ 0.01, P ¼ 0.915). These results confirm earlier observations of spatial release from speech-onspeech masking in the nominal, median sagittal plane (Plomp, 1976;Bolia et al, 1999;McAnally et al, 2002;Worley and Darwin, 2002;Freyman et al, 2005;Aaronson et al, 2009). The lack of evidence of spatial release from masking for targets at 0 elevation is puzzling.…”

“…HRIRs modified in that way were used by McAnally et al (2002) to examine release from speech-on-speech masking in the nominal, median sagittal plane. McAnally et al observed modest but significant masking release when target and masker sources were separated by 50 or 100 , a finding consistent with the conclusions of Aaronson et al (2009) that ITDs were not responsible for the bulk of the masking release observed in their study.…”

“…This finding is interesting because it suggests that spatial manipulations that do not cause target and masker interaural disparities to differ can nevertheless produce release from masking. Supporting evidence for this proposition is provided by studies that show that substantially more speech-on-speech masking occurs when both target and masker sources are directly in front of a listener than when the masker source is moved directly behind the listener (Plomp, 1976;Freyman et al, 2005;Aaronson et al, 2009). Given the inherent difficulty of aligning sound-source locations precisely with a listener's median sagittal plane, it is likely that targets and maskers in the aforementioned studies were associated with slightly different interaural time disparities (ITDs).…”

“…Small differences between target and masker ITDs could have supported release from masking through binaural analysis or by giving rise to the perception of azimuthal separation of target and masker sources. This issue was explicitly addressed by Aaronson et al (2009), who estimated that the maximum angular misalignment of their front and rear sound sources with respect to the bite bar they provided to help participants maintain a constant head alignment was less than 0.5 ; this corresponds to a maximum difference between target and masker ITDs of less than 7 ls. Somewhat greater misalignment with respect to the interaural axes of the participants of Aaronson et al is probable, given that most individuals are not completely anatomically symmetrical.…”

Speech-on-speech masking with variable access to the linguistic content of the masker speech

“…Electronic mail: lauren.calandruccio@qc.cuny.edu Informational masker signals can be manipulated in various ways, resulting in interesting and informative patterns of release in masking. For example, switching the gender of the talker of the competing signals to oppose that of the target talker ͑e.g., Helfer and Freyman, 2008͒, time reversal of the competing speech ͑Freyman et al, 2001͒, spatial separation of the competing speech ͑Freyman et al., 1999;Aaronson et al, 2008͒, and interaural time delays between the target and competing speech signals ͑Carhart et al, 1967͒ all can improve listeners' recognition. Recently, several researchers have documented a release in masking on an English speech recognition task when the masker is changed from competing English speech to a language unfamiliar to the listener.…”