Non-linear speech transition visualization
Klaus Reinhard · 1997
Modelling context effects and segmental transitions in speech recognition systems is very important. Explicitly modelling segmental transitions in a RNN framework would circumvent these problems. We present an interesting application of Principal Curves, an algorithm to extract a non-linear summary of p-dimensional data firstly published in 1989 by Hastie/Stuetzle. The algorithm can be used to visualize non-linear transient characteristics in speech. We will show that between-phone characteristics found within diphones can be used as discriminant information to distinguish ambiguous phones. The technique used is explained and illustrated on the examples /bah/, /dah/ and /gah/. INTRODUCTION Since speech is a complex time-sequential process, it is well established that particular phones vary acoustically when they occur in different phonetic context. In stateof -the-art systems based on short term spectral analysis this is achieved by enlarging the feature vector to include derivatives...