[2Win5-61] Control of Speech Synthesis Using Music-oriented Constraints
Keywords:Speech Synthesis
This study aims to generate natural speech by incorporating music-oriented constraints into speech synthesis to enhance emotional and expressive qualities.
We propose a method using Style-Bert-VITS2 to integrate pitch-related musical elements into transcriptions.
The model is trained using the PJS corpus with pitch constraints and additional speech-text pairs from CSJ.
Experimental results demonstrate that the generated speech effectively reflects the imposed constraints.
We propose a method using Style-Bert-VITS2 to integrate pitch-related musical elements into transcriptions.
The model is trained using the PJS corpus with pitch constraints and additional speech-text pairs from CSJ.
Experimental results demonstrate that the generated speech effectively reflects the imposed constraints.
Authentication for paper PDF access
A password is required to view paper PDFs. If you are a registered participant, please log on the site from Participant Log In.
You could view the PDF with entering the PDF viewing password bellow.