Fabrice Malfrère

Phonetic alignment: speech synthesis-based vs. viterbi-based

In this paper we compare two different methods for automatically phonetically labeling a continuous speech data-base, as usually required for designing a speech recognition or speech synthesis system. The first method is based on temporal alignment of speech on a synthetic speech pattern; the second method uses either a continuous density hidden Markov models (HMM) or a hybrid HMM/ANN (artificial neural network) system in forced alignment mode. Both systems have been evaluated on read utterances not part of the training set of the HMM systems, and compared to manual segmentation. This study outlines the advantages and drawbacks of both methods. The speech synthetic system has the great advantage that no training stage (hence no large labeled database) is needed, while HMM Systems easily handle multiple phonetic transcriptions (phonetic lattice). We deduce a method for the automatic creation of large phonetically labeled speech databases, based on using the synthetic speech segmentation tool to bootstrap the training process of either a HMM or a hybrid HMM/ANN system. The importance of such segmentation tools is a key point for the development of improved multilingual speech synthesis and recognition systems.

Speech synthesis for text-to-speech alignment and prosodic feature extraction

The aim of this paper is to present a new and promising approach of the text-to-speech alignment problem. For this purpose, an original idea is developed: a high quality digital speech synthesizer is used to create a reference speech pattern used during the alignment process. The system has been used and tested to extract the prosodic features of read French utterances. The results show a segmentation error rate of about 8%. This system will be a powerful tool for the automatic creation of large prosodically labeled databases and for research on automatic prosody generation.

New techniques for the compression of synthesizer databases

Most state-of-the-art TTS synthesizers are based on a technique known as synthesis by concatenation, in which speech is produced by concatenating elementary speech units. The design of a high-quality TTS implies the storage of a high number of segments. To facilitate the storage of these segments, this paper proposes a very low complexity coder to compress these databases with a toll quality. A particular interest has been taken in the databases of the MBROLA synthesizer constituted of fixed size frames. The developed coder uses this special characteristic to reach compression rates from 7 to 9 without degradation of the speech quality produced by the synthesizer.