I am programming an automatic dubbing of children's videos into multiple languages. The missing piece is a text-to-speech with lifelike children's voices.
Amazon Polly has children's neural voices in US English only. Google Cloud has more language but does not have children's voices and raising the voice's pitch by 10 semitones does not produce enjoyable results.
I am looking for a text-to-speech system with the following characteristics:
- multiple languages, like Google Cloud
- children's voices, like Amazon Polly
- lifelike speech synthesis with neural voices, like Google Cloud or Amazon Polly
- support for SSML tags like
prosody(rate, pitch, volume) and
Could you recommend some systems?