This page contains following speech samples:

1. Source (Normal): Natural speech (source speaker, male, normal)
2. Source (Falsetto): Natural speech (source speaker, male, falsetto)
3. Target: Natural speech (target speaker, female)
4. Converted (Normal-MSE): Converted speech (Source (Normal) to Target, trained w/ MSE)
5. Converted (Normal-GAN): Converted speech (Source (Normal) to Target, trained w/ GAN)
6. Converted (Falsetto-MSE): Converted speech (Source (Falsetto) to Target, trained w/ MSE)
7. Converted (Falsetto-GAN): Converted speech (Source (Falsetto) to Target, trained w/ GAN)

(1) Speech feature conversion + WORLD vocoder

	Source (Normal)	Source (Falsetto)	Target	Converted (Normal-MSE)	Converted (Normal-GAN)	Converted (Falsetto-MSE)	Converted (Falsetto-GAN)
1
2
3
4
5

(2) Spectral differential filtering

	Source (Normal)	Source (Falsetto)	Target	Converted (Normal-MSE)	Converted (Normal-GAN)	Converted (Falsetto-MSE)	Converted (Falsetto-GAN)
1
2
3
4
5