Soundblogging by sounds

The real furore at the MobiSys 2009 conference was the new SoundSense program. Through the iPhone's built-in microphone, this utility continuously monitors background sounds and continuously tries to analyze and classify them. For all types of noise generated unique identifiers. For example, initially the program successfully defines categories such as voice, music, background noise. But the most interesting thing is that it can be tuned and trained on the sounds that accompany your and only your life! The video perfectly shows how simple and convenient it is done.



Thus, a simple program becomes a unique tool for liveblogging. For example, she can automatically tweet what exactly you are doing at the moment: driving, brushing your teeth, sitting in a cafe, watching TV, etc.

In principle, while improving the accuracy of work, this program will be able to determine by sound in which particular cafe or in which particular room you are sitting, because each room has its own unique acoustic picture.
')
The mathematical algorithms on which SoundSense software is based are described in this scientific paper .

via Technology Review , Slashdot