THE AUDITORY MODELING TOOLBOX

Applies to version: 1.1.0

View the code

Go to function

MCLACHLAN2021_METRICS - - extract localization metrics

Usage

metric = mclachlan2021_metrics(doa)

Input parameters

Description

doa : Struct returned from the mclachlan2021 model with
estimated and real directions of arrival

Output parameters

'',metric metrics for evaluation of localisation performance
.reversal_fb
: Percentage of front-back reversals, omitting
small errors occurring within 10 degrees of the plane dividing the hemifields
.reversal_ud : Percentage of up-down reversals, omitting
small errors occurring within 10 degrees of the plane dividing the hemifields

System Message: WARNING/2 (<string>, line 36)

Definition list ends without a blank line; unexpected unindent.

.rmsL : Lateral root mean squared error .rmsP : Polar root mean squared error, omitting reversals

System Message: ERROR/3 (<string>, line 38)

Unexpected indentation.
and restricted to directions within 35 degrees of the vertical midline

mclachlan2021_metrics(...) returns psychoacoustic performance parameters for experimental response patterns. doa is a struct where actual and estimated directions of arrival must be provided.

References:

D. Schonstein, L. Ferre, and B. F. G. Katz. Comparison of headphones and equalization for virtual auditory source localization. Proc. Euronoise, 2008.

J. Reijniers, D. Vanderleist, C. Jin, C. S., and H. Peremans. An ideal-observer model of human sound localization. Biological Cybernetics, 108:169--181, 2014.