I've implemented basic crosstalk cancellation in both the time and frequency domain and I'm noticing that the quality of sound localization of the binaural signal processed in the frequency domain seems "more realistic" than the one in the time domain.
Is their any known studies of this sort of potential perceptual difference when filtering binaural audio in the time domain vs the frequency domain?
I guess , Bose (is a headset system) echo cancellation (beat based) . Beat => frequency
Practically , our ears are sensitive to frequency hence inevitable frequency domain (decimation of the spectrum would be improve clarity (i hope))
The questions you are asking fall within the domain of auditory subjective preference theory, which generally attempts first to find what the subjective preferences of human (and other) populations are (I may like the sound of concert hall A better than concert hall B, but you may like B better than A - how prevalent is either preference among Germans? Americans? Vietnamese? Infants? Aged dolphins?, etc.). If there are significant differences among populations (i.e., are these differences in preference reliable, or are they common misconceptions?), then what can be identified (and tested experimentally) that accounts for this difference?
In audio, such differences have been studied most extensively in concert hall design, on which subject there is a vast literature that establish both what accounts for such preferences, and how to achieve "better-sounding" designs (quadratic residue diffusers and "shoebox" hall geometries come to mind).
In principle, since many time- and frequency-domain techniques can be shown mathematically to be both identity systems and/or equivalent to each other, there probably is no fundamental difference in sound quality associated with either approach. However, the auditory system can can be remarkably sensitive to tiny differences in implementations and/or applications of any particular processing method (cf.. the old "controversy" regarding audio "quality" between digital versus analog methods). The differences that have been identified are mostly due to detailed differences in processing methods and implementations.
Good places to start looking into such questions include Rossing (ed., 2007), Springer Handbook of Acoustics, eReference ISBN 978-0-387-30425-0, and Yoichi Ando (1985), Concert Hall Acoustics, eBook ISBN 978-3-642-69810-1.
- Badges
- Science topic
- Similar topics
- Engineering
- Electrical Engineering
- Signal Process