Diphone

A diphone describes in concatenative speech synthesis the short section ( block) speech, which begins in the middle of phon, and ends in the middle of the following phon. A diphone will then contain the damage caused by coarticulation transition between the two sounds. The concatenation of blocks, each of which only one phone include ( Allophonsynthese ) results in speech synthesis only at extremely unsatisfactory results since the coarticulation between the sounds in this case can not be considered. In contrast, diphone already leads to remarkably good results, the sound and sufficiently understandable of course. The quality can be increased by the use of longer blocks instead of diphones (for example, syllables, common words or sequences of sounds ) further, but this often is not practical for reasons of asset size.

The Diphonbausteine ​​used are for example using the PSOLA algorithm in their prosodic information ( intensity, fundamental frequency, duration) manipulation during the synthesis to produce a natural speech melody.