|
||||
Speech synthesis samplesThese samples are the output of a statistical parametric synthesizer built by combining HTS and AHOcoder, the vocoder I am developing in AHOLAB. So far, AHOcoder has two main analysis modes: the accurate one (default) and the efficient one (90-95% faster). We may choose one or the other depending on the application. Female voice in Basque
Natural speech:
HTS + AHOcoder:
HTS + efficient-AHOcoder:
Male voice in Spanish
Natural speech:
HTS + AHOcoder:
HTS + efficient-AHOcoder:
|
||||
|
(c) Daniel Erro
AHOLAB Signal Processing Laboratory, UPV/EHU, Alda.Urquijo s/n 48013 Bilbao, Spain Phone: +34 946017245. Fax: +34 946014259. Email: derro(a)aholab.ehu.es. |
||||
