Phonetic Synthesizer

The Phonetic Synthesizer allows you to generate highly natural speech that can say anything. With the Phonetic Synthesizer, there is no need to store recorded speech samples on disk; all the speech in the game can be generated on the fly. Because the speech can be generated dynamically, you can make the characters in the game genuinely respond to the action by talking about what is actually going on rather than have to resort to a few stock phrases. Using the Phonetic Voice Building tool you can add new voices, or use your current speech samples as the basis of a new voice in the Phonetic Synthesizer. The speech produced sounds natural and full of character; in shooters you can voice gangsters, cops or soldiers each in their own style; in sports games you can create genuinely live commentary that responds to the game, mentions players by name and talks about the actual game unfolding. The Phonetic Synthesizer runs in the game itself and requires very little in the way of memory or CPU. As everything is generated on the fly, an unlimited number of speech samples can be produced with no additional memory use.

Phonetic Morpher

The Phonetic Morpher allows you to modify natural or synthetic speech to create a rich variety of effects. The speaker morpher allows you to change a speech sample by mimicking changes in the physical characteristics of the original speaker; in this way you can make one sample sound as if it was spoken by a whole range of people. Alternatively, the Prosody Morpher maintains a speaker’s characteristics, but allows you to make many subtly different versions of the same sentence, changing inflection, emphasis and style. This allows you to avoid the annoying repetition often found in games because of memory constraints or development budgets.
The Phonetic Morpher can be integrated into the game and work on the speech generated by the Synthesizer, or normal speech stored on disk. In this way, a whole range of unique sounding characters can be generated live within the game, without any additional memory overhead. The Phonetic Morpher can also of course be used offline, if a developer simply wants different versions of some speech they have already recorded.

Phonetic LipSync

The Phonetic LipSync provides accurate timing information to drive facial animation. By using sophisticated phonetic alignment technology, it can determine where each word, syllable and phoneme starts and stops in any speech sample. This can be used on natural speech samples or driven by the output of the Phonetic Synthesizer or Phonetic Morpher. Achieving realism in facial animation is one of the hardest challenges in games today, and the Phonetic LipSync helps solve many of the problems by providing extremely accurate and detailed information on the timing of speech. As with all products in the Phonetic Arts range, the Phonetic Lip Sync works within the game engine allowing lip syncing on the fly.