Automated Speech Processing of Bengali using SPPAS software

Moumita Pakrashi, Shakuntala Mahanta, Indian Institue of Technology Guwahati and Brigitte Bigi, The Laboratoire Parole et Langage

SPPAS (SPeech Phonetization Alignment and Syllabification) is an open-source, multi-platform (Linux, Windows, MacOS) software package that has been developed to produce automatic annotations of speech and even perform some descriptive statistical analysis of the annotated speech data. Developed by Brigette Bigi at CNRS, in Laboratoire Parole et Langage, Aix-en-Provence, France; its main functions include the segmentation of recorded speech files and its corresponding orthographic transcription. With the minimum requirement of adding just a few linguistic resources in the package directories, this tool can automatically segment speech files into units of phoneme, syllable, word and utterance. The software has been currently implemented for languages like French, English, Italian, Chinese, Naija (Nigerian Pidgin). We are currently working on the implementation of SPPAS on Bengali language by developing the required linguistic resources