Mircea Giurgiu * Results on Automatic Speech Recognition in Romanian

Fig. 6. - Number of vectors from each cluster; M=64, SET1, respectively SET2, databases.

Fig. 7. - SNR in the coding process for M=64; SET1, respectively SET2, VQ databases.

Fig. 8 - SNR in the training process for M=128; SET1, respectively SET2, VQ databases.

Fig. 9. - SNR in the VQ coding process, for M=128, SET1, respectively SET2, VQ databases.

181