Unit Selection Synthesis

We now turn to unit-selection synthesis which is the dominant synthesis technique in text-to-speech today. Unit selection is the natural extension of second-generation concatenative systems, and deals with the issues of how to manage large numbers of units, how to extend prosody beyond just FO and timing control, and how to alleviate the distortions caused by signal processing. 

---

Faced with the choice between fully natural but inflexible canned speech, and somewhat unnatural but fully flexible TTS, some researchers have proposed limited domain synthesis systems which aim to combine the benefits of both. There are as many different approaches to this. Some systems attempt to mix canned speech and TTS. Black and Lenzo [52] proposed a system for cleverly joining words and carrier phrases, for use in applications such as a talking clock. The phrase splicing approach of Donovan et al [139] used recorded carrier phrases and cleverly spliced with unit selection synthesis from recordings of the same speaker. A somewhat different approach is to use what is basically the same system as for normal unit selection synthesis but to load the database of recordings with words and phrases fi-om the required domain [440], [436], [394], [9]. 

The unit selection synthesis technique described in Chapter 16 uses an entirely data driven approach, whereby recorded speech waveforms are cut up, rearranged and concatenated to say new sentences. Given the success of this approach in normal synthesis, a number of researchers have applied these algorithms to FO synthesis [296], [310], [311]. 

While many possible approaches to statistical synthesis are possible, most work has focused on using hidden Markov models (HMMs). This along with the unit selection techniques of the next chapter are termed third generation techniques. This chapter gives a full introduction to these and explains how they can be used in synthesis. In addition we also show how these can be used to automatically label speech databases, which finds use in many areas of speech technology, including unit selection synthesis. Finally, we introduce some other statistical synthesis techniques. 

Unit selection synthesis operates by selecting units from a large speech database according to how well they match a specification and how well they join together. 

Amdal, 1., AND SvENDSEN, T. Unit selection synthesis database development using utterance verification. In Proceedings of Eurospeech 2005 (2005). 

Rutten, P., Aylett, M., Fackrell, J., and Taylor, P. A Statistically Motivated Database Pruning Technique For Unit Selection Synthesis. In Proceedings of the ICSLP (2002), pp. 125-128. 

---