TABLE OF CONTENTS In text-to-speech synthesis the form of the sign carrier is an important criterion for speech quality, not only with respect to the system input (as discussed above) but also to the system output. The requirement that a text should be unobtrusive when it is put into the system can be easily met by only choosing ideal texts for synthesis. In contrast it is much more difficult when applied to the system output, i.e. when the synthesized speech signal itself is analyzed from the point of view of it being a sign carrier as well. As the acoustic text is a medium, a sign carrier itself, it should, of course, be unobtrusive and not form a distraction either. When listening to speech the auditory perceived signal should hint at the form-content-function relationship, not at the form itself. Certain forms hint at certain contents and functions, and in turn, certain contents and functions require specific forms. The materialism of the sign carrier synthetic speech can only be located, or remain, in the background when it does not catch listener's attention with regard to a violation of this form-content-function relationship.
For many text-to-speech syntheses there is a marked discrepancy between form criteria of this sign carrier and the contents and functions correlated with the signs, in other words, between the acoustic-synthetic form and the contents that should convey it and the functions that should perform it. This discrepancy stems from the fact that the present capabilities...
