THE AUDITORY MODELING TOOLBOX
Go to function
MOORE2016_SPECTRUM
calculate the spectrum for an audio segment of 2048x2 samples
Usage:
[fLeftRelevant, LLeftRelevant, fRightRelevant, LRightRelevant] = moore2016_spectrum(s, Fs, dBMax, wHann, vLimitingIndizes)
Input parameters:
| s | input signal |
| Fs | sampling frequency [Hz] |
| dBMax | maximum dB |
| wHann | matrix with window coefficients in columns |
| vLimitingIndizes | indices for fft assembly |
Output parameters:
| fLeftRelevant | frequency of relevant components left ear |
| LLeftRelevant | level of relevant components left ear |
| fRightRelevant | frequency of relevant components right ear |
| LRightRelevant | level of relevant components right ear |
Description:
This code calculates the spectrum of the input signal as required by the loudness model moore2016 in the version for TVL 2016 based on ANSI S3.4-2007 and Moore & Glasberg (2007).
It returns only the relevant components, i.e. components that have at least -30 dB SPL and at least 60 dB less than the maximum component. Overall, 4 vectors are returned: frequency and level for the left ear, and same for the right ear. dBMax is the rms level of a full scale sinusoid hann windows and limiting indizes for the 6 FFTs are passed so they are calculated only once. Much is done with intensity rather than level so that nonzeros() works correctly without the need of excemptions.
Build with Bootstrap