Applies to version: 0.9.8

View the code

Go to function

baumgartner2016_calibration - Calibration of listener-specific sensitivity thresholds to experimental performance


scalib = baumgartner2016_calibration(s)

Input parameter:
s : strucure containing subject's data. It must include the
fields Obj, baseline.pe_exp, and baseline.qe_exp, representing the listener's HRTF as SOFA object, the baseline local polar RMS error, and the baseline quadrant error rate, respectively.
Output parameter:
scalib : strucure containing subject's data with calibrated u

baumgartner2016_calibration returns listener data with listener-specific sensitivity thresholds calibrated by joint optimization of PE and QE to minimize mismatch between experimental and predicted results.

baumgartner2016_calibration accepts the following optional parameters:

'Srange',Sr Define the sensitivity range. Default is [eps,10].
'prange',pr Define the prior range. Default is [0,1].
'latseg',ls Define lateral segment(s) of data used for calibration. Default value is 0 deg.
'c',c Structure for optional definition of listener-specific settings like playback level or stimulus.
'TolX',tx Minimum tolerance of optimization argument (see help optimset for details).
'MaxIter',mi Maximum number of optimization iterations (see help optimset for details).

baumgartner2016_calibration accepts the following flags:

'calibprior' Calibrate also expectation prior.
'fminbnd' Use fminbnd routine for calibration. This is the default.
'fminsearch' Use fminsearch routine for calibration.
'search' Try all possibilities.