Complex Cepstrum Based Voice Conversion Using Radial Basis Function

Table 2

Prediction performance of MCEP based for formant frequencies.

Transformation model	Formant frequencies	% Predicted frame within deviation
		2%	5%	10%	15%	20%	25%	50%	𝜇RMSE	Υ𝑋,𝑦
M1-F1	F1	51	74	80	81	83	85	90	4.45	0.7235
	F2	45	63	68	78	82	87	89	3.73	0.8182
	F3	57	62	79	86	87	89	92	3.34	0.8703
	F4	69	79	84	89	88	90	100	2.39	0.8629
F2-M2	F1	36	58	67	74	82	86	90	4.28	0.7190
	F2	57	82	86	87	87	89	91	6.30	0.7238
	F3	72	77	89	91	92	94	95	5.23	0.7474
	F4	66	74	89	90	93	95	100	4.91	0.7957