Target identification using relative level in multi-talker listening

Abstract

Previous studies have suggested that listeners can identify words spoken by a target talker amidst competing talkers if they are distinguished by their spatial location or vocal characteristics. This "direct" identification of individual words is distinct from an "indirect" identification based on an association with other words (call-signs) that uniquely label the target. The present study assessed listeners' ability to use differences in presentation level between a target and overlapping maskers to identify target words. A new sentence was spoken every 800 ms by an unpredictable talker from an unpredictable location. Listeners reported color and number words in a target sentence distinguished by a unique call-sign. When masker levels were fixed, target words could be identified directly based on their relative level. Speech-reception thresholds (SRTs) were low (-12.9 dB) and were raised by 5 dB when direct identification was disrupted by randomizing masker levels. Thus, direct identification is possible using relative level. The underlying psychometric functions were monotonic even when relative level was a reliable cue. In a further experiment, indirect identification was prevented by removing the unique call-sign cue. SRTs did not change provided that other cues were available to identify target words directly. Thus, direct identification is possible without indirect identification.

Predicting the effect of hearing loss and audibility on amplified speech reception in a multi-talker listening scenario

“…1), but the results reported here do not include these effects as they made minor changes to the predictions, which is consistent with the background noise being essentially diffuse. 1 Kitterick et al (2013) have changed the naming of these methods from "orienting" and "cueing" (Kitterick et al, 2010) to indirect and direct, respectively. 2 The gains at the octave frequencies from 0.25 to 8 kHz were 7.9, 14, 29, 31, 25, and 2.2 dB, respectively.…”

Cortical and Sensory Causes of Individual Differences in Selective Attention Ability Among Listeners With Normal Hearing Thresholds

“…The features used to direct attention can be based on spatial attributes of the sound (e.g., Ahveninen et al, 2006;Banerjee, Snyder, Molholm, & Foxe, 2011;Ihlefeld & Shinn-Cunningham, 2008;Kidd, Arbogast, Mason, & Gallun, 2005;Ruggles & Shinn-Cunningham, 2011) or on spectrotemporal source attributes, such as timbre, pitch, or level (e.g., Brungart, Simpson, Ericson, & Scott, 2001;Culling, Hodder, & Toh, 2003;Darwin, Brungart, & Simpson, 2003;Devergie, Grimault, Tillmann, & Berthommier, 2010;Greenberg & Larkin, 1968;Jones, Kidd, & Wetzel, 1981;Kitterick, Clarke, O'Shea, Seymour, & Summerfield, 2013;Maddox & Shinn-Cunningham, 2012;Scharf, Quigley, Aoki, Peachey, & Reeves, 1987;Varghese, Ozmeral, Best, & Shinn-Cunningham, 2012;Wright & Fitzgerald, 2004). Depending on exactly how a listener focuses attention, different cortical control networks are engaged.…”