SpeedySpeech: Efficient Neural Speech Synthesis
This work addresses the need for efficient and high-quality speech synthesis systems, particularly for applications requiring low computational resources, and is incremental as it builds on existing neural sequence-to-sequence models.
The paper tackles the problem of achieving fast training, fast inference, and high-quality audio synthesis simultaneously in neural speech synthesis, proposing a student-teacher network that uses simple convolutional blocks and a single attention layer, resulting in significantly higher voice quality ratings than Tacotron 2 and real-time performance on a CPU.
While recent neural sequence-to-sequence models have greatly improved the quality of speech synthesis, there has not been a system capable of fast training, fast inference and high-quality audio synthesis at the same time. We propose a student-teacher network capable of high-quality faster-than-real-time spectrogram synthesis, with low requirements on computational resources and fast training time. We show that self-attention layers are not necessary for generation of high quality audio. We utilize simple convolutional blocks with residual connections in both student and teacher networks and use only a single attention layer in the teacher model. Coupled with a MelGAN vocoder, our model's voice quality was rated significantly higher than Tacotron 2. Our model can be efficiently trained on a single GPU and can run in real time even on a CPU. We provide both our source code and audio samples in our GitHub repository.