A$^3$T: Alignment-Aware Acoustic and Text Pretraining for Speech Synthesis and Editing
This work addresses speech synthesis and editing for applications requiring natural speech generation, representing an incremental advance in leveraging representation learning for synthesis tasks.
The paper tackles the challenge of generating high-quality speech in synthesis tasks by proposing A^3T, a framework that reconstructs masked acoustic signals using text and alignment during pretraining, resulting in improved performance on speech editing and multi-speaker synthesis without external models.
Recently, speech representation learning has improved many speech-related tasks such as speech recognition, speech classification, and speech-to-text translation. However, all the above tasks are in the direction of speech understanding, but for the inverse direction, speech synthesis, the potential of representation learning is yet to be realized, due to the challenging nature of generating high-quality speech. To address this problem, we propose our framework, Alignment-Aware Acoustic-Text Pretraining (A$^3$T), which reconstructs masked acoustic signals with text input and acoustic-text alignment during training. In this way, the pretrained model can generate high quality reconstructed spectrogram, which can be applied to the speech editing and unseen speaker TTS directly. Experiments show A$^3$T outperforms SOTA models on speech editing, and improves multi-speaker speech synthesis without the external speaker verification model.