we randomly choose two source speeches from four speakers(SF1, SF2, TM1, TM2) for conversion.

• SF1 : source female 1
• SF2 : source female 2
• TM1: target male 1
• TM2: target male 2

• SF1-SF2+200020.wav : source female1’s speech convert to source female2’s speech.
• SF1-TM2+200025.wav : source female1’s speech convert to target male2’s speech.