A novel codebook-based excitation model for use in speech synthesis

Speech synthesis is an important modality in Cognitive Infocommunications. Statistical parametric methods have gained importance in speech synthesis recently. The speech signal is decomposed to parameters and later restored from them. The decomposition is implemented by speech coders. We propose a novel speech coding method with codebook-based excitation. In the analysis stage the speech signal is analyzed frame-by-frame and a codebook of phoneme-dependent, pitch synchronous residuals is built from the voiced parts. During the synthesis stage the codebook is searched for a suitable element in each voiced frame and these are concatenated. Our initial experiments show that in most cases the method can resynthesize speech to a similar quality than the original. This new excitation model fits well in the machine learning part of the statistical parametric speech synthesis framework.¹