Application research of bel canto performance based on artificial intelligence technology
Publié en ligne: 30 nov. 2022
Pages: 133 - 144
Reçu: 09 mai 2022
Accepté: 08 août 2022
© 2023 Zhen Wu, published by Sciendo
This work is licensed under the Creative Commons Attribution 4.0 International License.
Fig. 1
Schematic diagram of a neural networkFig. 2
Autoencoder (AE)Fig. 3
Transformer model mechanismFig. 4
Overall network structure diagram. BAP, band aperiodicity; F0, xxx; FFT, yyy; MGC, mel-generalised coefficient; V/UV, voiced/unvoicedFig. 5
Input: ‘score’ encoderFig. 6
Example of music notationFig. 7
Length predictor and duration regulatorFig. 8
Decoder decoder.BAP, band aperiodicity; F0, xxx; FFT, yyy; MGC, mel-generalised coefficient; V/UV, voiced/unvoicedFig. 9
Residual connection of pitch. BAP, band aperiodicity; MGC, mel-generalised coefficient; V/UV, voiced/unvoiced