DurIAN: Duration Informed Attention Network For Multimodal Synthesis
This addresses the need for robust and efficient multimodal synthesis systems for applications like virtual assistants or entertainment, though it is incremental as it builds on existing autoregressive and WaveRNN models.
The paper tackles the problem of generating natural speech and synchronized facial expressions in multimodal synthesis by introducing the Duration Informed Attention Network (DurIAN), which avoids artifacts like word skipping/repeating errors common in end-to-end systems, and achieves speech quality on par with state-of-the-art while reducing computational complexity from 9.8 to 5.5 GFLOPS with 6x real-time audio generation on a CPU.
In this paper, we present a generic and robust multimodal synthesis system that produces highly natural speech and facial expression simultaneously. The key component of this system is the Duration Informed Attention Network (DurIAN), an autoregressive model in which the alignments between the input text and the output acoustic features are inferred from a duration model. This is different from the end-to-end attention mechanism used, and accounts for various unavoidable artifacts, in existing end-to-end speech synthesis systems such as Tacotron. Furthermore, DurIAN can be used to generate high quality facial expression which can be synchronized with generated speech with/without parallel speech and face data. To improve the efficiency of speech generation, we also propose a multi-band parallel generation strategy on top of the WaveRNN model. The proposed Multi-band WaveRNN effectively reduces the total computational complexity from 9.8 to 5.5 GFLOPS, and is able to generate audio that is 6 times faster than real time on a single CPU core. We show that DurIAN could generate highly natural speech that is on par with current state of the art end-to-end systems, while at the same time avoid word skipping/repeating errors in those systems. Finally, a simple yet effective approach for fine-grained control of expressiveness of speech and facial expression is introduced.