72
Participants
Start Date
August 1, 2024
Primary Completion Date
August 31, 2024
Study Completion Date
August 31, 2024
Solo: Unmasked Speech Stimuli
Speech stimuli recorded using non-moving speakers and mics. No masking sources present. No BSS applied to multi-channel recordings. Very high output QoS values.
Raw: Fully masked speech--no motion stimuli
Speech stimuli recorded using non-moving speakers and mics. All masking sources present. No speech separation or extraction methods applied to multi-channel recordings. Very low output QoS values.
StatScrub: Extracted Speech--no motion stimuli
Speech stimuli recorded using non-moving speakers and mics. All masking sources present. Joint ACES scrubbing of both noise sources applied to multi-channel recordings. Very high output QoS values.
SlideSpch: Scrubbed Speech emitted from linearly moving speaker stimuli
Speech stimuli recorded using linearly moving speech source and stationary masking sources and mics. All masking sources present. Joint ACES scrubbing of both noise sources applied to multi-channel recordings. Moderately high output QoS values.
SlideNoise: Speech Scrubbed from linearly moving and stationary noise stimuli
Mixed speech and noise sources recorded using a stationary speech source, a stationary noise source, and a linearly moving noise source. A valid source hypothesis of the speech source is used to extract the speech source. High output QoS values.
SlideMic: Stationary sources scrubbed from a linearly moving mic stimuli
Mixed speech and noise sources recorded using all stationary sources, and a linearly moving microphone (mic 1). Joint ACES scrubbing of both noise sources is used to reduce the response of Mic 1 to a residue of speech. Low output QoS values.
University of Cincinnati, Cincinnati
National Institute on Deafness and Other Communication Disorders (NIDCD)
NIH
University of Cincinnati
OTHER
Speech Technology and Applied Research Corp.
INDUSTRY