Joint training framework for text-to-speech and voice conversion using multi-source Tacotron and WaveNet
This work addresses the need for unified models in speech synthesis and conversion, though it appears incremental as it builds on existing Tacotron and WaveNet architectures.
The authors tackled the problem of training a single model for both text-to-speech and voice conversion tasks by proposing a multi-source Tacotron with dual attention, achieving efficient performance in both tasks as shown in listening experiments.
We investigated the training of a shared model for both text-to-speech (TTS) and voice conversion (VC) tasks. We propose using an extended model architecture of Tacotron, that is a multi-source sequence-to-sequence model with a dual attention mechanism as the shared model for both the TTS and VC tasks. This model can accomplish these two different tasks respectively according to the type of input. An end-to-end speech synthesis task is conducted when the model is given text as the input while a sequence-to-sequence voice conversion task is conducted when it is given the speech of a source speaker as the input. Waveform signals are generated by using WaveNet, which is conditioned by using a predicted mel-spectrogram. We propose jointly training a shared model as a decoder for a target speaker that supports multiple sources. Listening experiments show that our proposed multi-source encoder-decoder model can efficiently achieve both the TTS and VC tasks.