Date of this Version
Reverberation time (RT) is a metric commonly used to describe room acoustic conditions, but different rooms which have the same reverberation time can have different reflection densities. Much less is known about how humans perceive different reflection densities and how sensitive humans are to changes in reflection density. Previous investigations in the existing literature have studied the upper limit of distinguishable reflection density using artificial impulse responses, but not with more realistic impulse responses simulated in room acoustic software or measured from real rooms. The aim of this dissertation is to investigate methods for quantifying reflection density from measured impulse responses, and to understand human perception of reflection density more completely by determining the upper limit of distinguishable reflection density and just noticeable difference of reflection density.
This dissertation presents three studies on the perception of reflection densities. What is the upper limit of distinguishable reflection density when using artificial impulse responses convolved with a clapping signal, and how does this limit change with different reverberation times (Study 1)? What if the impulse responses are simulated from room acoustic software instead? Does the upper limit change if the source signal changes from clapping to speech (Study 2)? And finally, how sensitive are humans to the change of reflection density (Study 3)? In each study, a number of listeners completed threealternative forced-choice subjective tests using the one-up two-down adaptive testing method, comprised of different RTs, reflection densities, and source signals (clapping or speech).
The results showed relatively large variation among testing subjects, possibly due to other perceptual cues rather than reflection density. After grouping the results based on how well subjects demonstrated convergence, the upper limit of distinguishable reflection density and just noticeable difference of reflection density have been identified. These results are useful for understanding how reflection density may be applied as an additional room acoustic parameter.