Adaptation of Tacotron2-based Text-To-Speech for Articulatory-to-Acoustic Mapping using Ultrasound Tongue Imaging

TL;DR

This paper adapts a pre-trained Tacotron2 TTS model for articulatory-to-acoustic mapping using ultrasound tongue imaging, employing transfer learning to improve speech synthesis quality with limited data.

Contribution

It introduces a transfer learning approach that adapts Tacotron2 for ultrasound-based articulatory-to-acoustic conversion, combining CNN, Tacotron2, and WaveGlow for improved naturalness.

Findings

Enhanced speech naturalness over previous models

Effective transfer learning with limited ultrasound data

Successful integration of CNN, Tacotron2, and WaveGlow

Abstract

For articulatory-to-acoustic mapping, typically only limited parallel training data is available, making it impossible to apply fully end-to-end solutions like Tacotron2. In this paper, we experimented with transfer learning and adaptation of a Tacotron2 text-to-speech model to improve the final synthesis quality of ultrasound-based articulatory-to-acoustic mapping with a limited database. We use a multi-speaker pre-trained Tacotron2 TTS model and a pre-trained WaveGlow neural vocoder. The articulatory-to-acoustic conversion contains three steps: 1) from a sequence of ultrasound tongue image recordings, a 3D convolutional neural network predicts the inputs of the pre-trained Tacotron2 model, 2) the Tacotron2 model converts this intermediate representation to an 80-dimensional mel-spectrogram, and 3) the WaveGlow model is applied for final inference. This generated speech contains the timing of the original articulatory data from the ultrasound recording, but the F0 contour and the spectral information is predicted by the Tacotron2 model. The F0 values are independent of the original ultrasound images, but represent the target speaker, as they are inferred from the pre-trained Tacotron2 model. In our experiments, we demonstrated that the synthesized speech quality is more natural with the proposed solutions than with our earlier model.