次へ: Experimental Results
上へ: raut04ASJ03
戻る: Polynomial Approximation
To analyze the accuracy of the approximation, and to compare it with other
methods, a set of onedimensional vectors was generated for speech signal
by MonteCarlo simulation. These speech vectors were corrupted by
adding noise at different SNR. Noise vectors were also generated by
MonteCarlo simulation.
The mean of corrupted speech estimated by different methods has been
plotted in Figure 2. The result shows that Lagrange
polynomial approximation gives almost the same result as given by
MonteCarlo simulation, and thus outperforms VTS1 and logmax approximations.
図 2:
Estimated mean of corrupted speech: by MonteCarlo simulation, Logmax approximation, Vector Taylor Series1, and
Lagrange Polynomial Approximation (LPA) for ,
, and varying from 0 to
20.

平成16年4月23日