Automated Child Voice Generation: Methodology and Implementation
Authors: Shaimaa Alwaisi, Mohammed Salah Al-Radhi, Géza Németh
An overview of proposed child speech synthesis architectures.
you can find synthesized samples here
The image below demonstrates The comparison of Mel spectrograms and F0 values between the ground truth and three models: (b) AutoVocoder, (c) ParallelWavGan, and (d) WORLD Vocoder for both Boy and Girl.
