An Automatic audio segmentation system for radio newscast

Current web search engines generally do not enable searches into audio files. Informative metadata would allow searches into audio files, but producing such metadata is a tedious manual task. Tools for automatic production of metadata are therefore needed. This project describes the work done on the development of an automatic audio segmentation system which can be used for this metadata extraction. In this work the radio newscast are divided into segments in which there is only one speaker. Audio features used in this project include Mel Frequency Cepstral Coefficients. This feature was extracted from audio files that were stored in a WAV format, using CLAM. Model-Selection-Based segmentation is used to segment audio signals using this feature.