Test on Simulated Data

Tests on simulated data

(a) High signal-to-noise ratio case : SNR=0 dB, New algorithm β= 1e-5

(b) High signal-to-noise ratio case : SNR=0dB, Old algorithm,β’= 1e-5

(c) Low signal-to-noise ratio case : SNR = -20 dB , New algorithm β= 1e-5

(d) Low signal-to-noise ratio case : SNR= -20 dB, Old algorithm β’= 1e-5

(e) Low signal-to-noise ratio case : SNR = -20 dB , Old algorithm β’= 1e-7

(f) Low signal-to-noise ratio case : SNR = -20 dB , Old algorithm β’=1e-9 β= 1e-5

Fig.6 The tests on the simulated data showed that both the new algorithm and the old Kalman Filter based algorithm work well at high SNR (a), (b); but at low SNR, the old algorithm does not filtered out the noise sufficiently (d), to reduce the noise, it needs narrow down the filter bandwidth ( which is controlled by parameter β’, the smaller β’ the narrower band). However, narrowing the band causes distortions in DP amplitude tracking (e) (f). The new algorithm allows wide bandwidth and has better tracking performance (c) .
Back to main page of A New Real-time Method for Direct Measurement of DPOAE Contra-lateral Suppression Dynamics document.