Controllable Emphasis with zero data for text-to-speech

Arnaud Joly, Marco Nicolis, Ekaterina Peterova, Alessandro Lombardi, Ammar Abbas, Arent van Korlaar, Aman Hussain, Parul Sharma, Alexis Moinet, Mateusz Lajszczak, Penny Karanasou, Antonio Bonafonte, Thomas Drugman, Elena Sokolova

Keywords: text-to-speech, emphasis control

Abstract: We present a scalable method to produce high quality emphasis for text-to-speech (TTS) that does not require recordings or annotations. Many TTS models include a phoneme duration model. A simple but effective emphasis method consists in increasing the predicted duration of emphasised word. We show that this is significantly better than signal processing based techniques improving naturalness by $7.3\%$ and identifiability by $40\%$ on a reference female en-US voice, and significantly closing the gaps to methods that require explicit recordings. The method proves to be effective in 4 languages (English, Spanish, Italian, German) for different voices and multiple speaking styles.