THE AUDITORY MODELING TOOLBOX

Applies to version: 1.1.0

View the code

Go to function

DEMO_BARUMERLI2021 - Demo for full sphere localization model from Barumerli et al. (2021)

Description

demo_barumerli2021(flag) demonstrates how to compute and visualize the baseline prediction (localizing broadband sounds with own ears) on the full sphere using the localization model from Barumerli et al. (2021).

demo_barumerli2021_1.png

Baseline prediction

This demo computes the baseline prediction (localizing broadband sounds with own ears) for an exemplary listener (NH12).

Averaged polar and lateral accuracy

This code produces the following output:

Loading real data from Majdak et al. 2010
Metrics from actual data
lateral_bias []:    -1.54
lateral_rms_error []:       8.35
elevation_bias []:  6.55
local_rms_polar []: 27.29
quadrant_err [%]:     2.92
Model simulation starting...

Generating feature space over the uniformily sampled sphere
Simulating subject answers
Metrics from simulated data
lateral_bias []:    0.42
lateral_rms_error []:       11.25
elevation_bias []:  4.85
local_rms_polar []: 29.21
quadrant_err [%]:     6.82

Simulating single estimation
target [az,el]: [0.00, 45.00]
      response [az,el]: [163.91, 12.15]