The KTH text-to-speech system has been under development for many years (Carlson & Granstrom, 1975; Carlson, Granstrom & Hunnicutt, 1982). Different parts of the system have been given priority at different times. The segmental quality has been and still is an important area. Evaluation of this aspect has been carried out at regular intervals. The basic test has been a nonsense VCV test, presented to naive listeners after a standard training session. At the workshop, the last years' results from these nonsense word tests will be presented. The results will be compared to similar tests on other systems.
Cite as: Carlson, R., Granström, B. (1989) Evaluation and development of the KTH text-to-speech system on the segmental level. Proc. Speech Input/Output Assessment and Speech Databases, Vol.2, 11-14
@inproceedings{carlson89_sioa,
author={Rolf Carlson and Björn Granström},
title={{Evaluation and development of the KTH text-to-speech system on the segmental level}},
year=1989,
booktitle={Proc. Speech Input/Output Assessment and Speech Databases},
pages={Vol.2, 11-14}
}