Fig. 1From: Text-to-speech system for low-resource language using cross-lingual transfer learning and data augmentationOverview of the base TTS system. Speaker embedding is used to train the multi-speaker model, but is not used for training the single-speaker modelBack to article page