Skip to main content
EURASIP Journal on Audio, Speech, and Music Processing
Fig. 8 | EURASIP Journal on Audio, Speech, and Music Processing

Fig. 8

From: Emotional voice conversion using neural networks with arbitrary scales F0 based on wavelet transform

Fig. 8

Examples of original F0 signal and F0 signal reconstructed with different numbers of scales. The top graph shows the original F0 signal (Os) and the F0 signal reconstructed with eight scales in each temporal level (R8 has 8×5 scales in total); the second graph shows Os and the reconstructed signal with five scales in each temporal level (R5 has 5×5 scales in total); the bottom pan shows Os and reconstructed signal with three scales in each temporal level (R3 has 3×5 scales in total)

Back to article page