Research Article
Complex Cepstrum Based Voice Conversion Using Radial Basis Function
Table 3
Prediction performance of complex cepstrum based for formant frequencies.
| Transformation model | Formant frequencies | % Predicted frame within deviation | 2% | 5% | 10% | 15% | 20% | 25% | 50% | | |
| M1-F1 | F1 | 59 | 80 | 89 | 91 | 91 | 93 | 95 | 4.45 | 0.7197 | F2 | 52 | 72 | 85 | 88 | 91 | 92 | 95 | 3.55 | 0.8149 | F3 | 63 | 83 | 90 | 92 | 93 | 95 | 99 | 2.65 | 0.8837 | F4 | 72 | 86 | 91 | 93 | 95 | 97 | 100 | 2.049 | 0.8909 |
| F2-M2 | F1 | 38 | 60 | 71 | 78 | 80 | 86 | 90 | 8.56 | 0.757 | F2 | 60 | 82 | 89 | 92 | 92 | 93 | 98 | 3.62 | 0.790 | F3 | 72 | 87 | 92 | 95 | 95 | 95 | 100 | 2.93 | 0.778 | F4 | 70 | 86 | 91 | 96 | 97 | 99 | 100 | 2.41 | 0.756 |
|
|