search

MTG / WGANSing

Multi-voice singing voice synthesis
235 stars 44 forks source link

language independent #15

Closed MAnal0025 closed 3 years ago

MAnal0025 commented 4 years ago

hi I want to implement your model using another language corpus so I want to ask you that is your model is language-dependent or independent? thanks in advance

ghost commented 4 years ago

All is in the paper : WGANSing: A Multi-Voice Singing VoiceSynthesizer Based on the Wasserstein-GAN and in The NUS sung and spoken lyrics corpus: A quantitative comparison of singing and speech

The process use the NUS database that use a subset of the ARPABET phonemes that is common in the CMU Pronouncing Dictionary. The CMU dictionary is US English based but you can easily adapt the system for a more generic phonetic system like IPA. ARPABET is a subset of IPA and is interchangeable.