摘要 |
In a voice pitch normalization device (Tr) equipped in a voice recognition device (VRAp) for recognizing an incoming command voice (Sva) uttered by any speaker, and used to normalize the incoming command voice (Svc) to be in an optimal pitch for voice recognition, a target voice generator (9, 11, 3, 5) produces a target voice signal (Svc(Ni)) by changing the incoming command voice (Svd) on a predetermined degree (Ni) basis, a probability calculator (15) calculates a probability (P) indicating a degree of coincidence among the target voice signal (Svc(Ni) and a plurality of words (Sr(m)) in sample data (psf), and a voice changer (9, 11, 3, 5) repeatedly changes the target voice signal (Svc(Ni)) in voice pitch until a maximum probability (Pmax) becomes a predetermined probability (Pth) or greater. <IMAGE> |