Speech - Music Discrimination Demo

[yiannakop]

Hi everyone. We have developed a demo for Speech - Music Discrimination, using Bayesian Networks and Hidden Markov Models. The demo opens a .wav file and segments the audio stream into homogenoous segments of Speech and Music. Anyone who is interested can visit the link here (http://cgi.di.uoa.gr/~sp_mu/). As input you can for example use an recorded audio stream from online radio stations. The methods have been described in 3 Conferences (more details in the site). The demo needs Matlab libraries to run, so if you don't have matlab installed, you may need to install a Matlab Component Runtime installer.
I hope this may interest someone.

Best Regards,
Theodore

[yiannakop]

Sorry for posting again, but its been almost a year since my first post on the Speech Music Discrimination Demo, but I've seen there's been some interest on the demo. Some bugs have been fixed and we have also added (after the suggestion of some users) the ability to store the segmentation results in .mat files, for future use.
For any suggestions / questions / bug reports, please let me know.

Best Regards,
Theodore :wave:

[yiannakop]

I've just uploaded an open source (Matlab) DEMO for Manual Annotation of Audio Streams. It also has a couple of audio features implemented. It could be used by anyone who wants to set a "ground truth" for evaluating an audio segmentation - classification system.

For more info: www.di.uoa.gr/~tyiannak (http://www.di.uoa.gr/~tyiannak)

Regards,
Theodore