I am currently working on a project, which, in a nutshell. involves a computer/arduino interface that is capable of detecting (or hearing..) a specific sound by way of a microphone input. The specific sound is that of NYPD, NYFD, ambulance, and any other sirens that sound off within the microphone's auditory range. When the computer/arduino interface identifies such a siren sound it will output and instruction to switch on a large neon sigh which reds: WOOP WOOP :D
I've spent some time researching how to tackle the programming/sound detection issue and have come across three max objects which seem to be applicable to this project. These objects include bonk (from obi wannabe), Zsa.descriptors (from Mikhail Malt and EJ), and analyzer (from Tristan Jehan).
It seems that with bonk there is a distinct possibility of teaching the object to recognize a sound (or sounds) generally, but this would of course require making some modifications. Does anyone have experiences in dealing with bonk in this way?
I have not been able to find out as much about zsa.descriptors as with bonk though I get the sense that zsa is an object aimed more specifically at detecting sounds in instrumental music. Might I be able to use zsa to effectively detect siren sounds? What sort of modifications would be involved in teaching zsa to recognize these more general sounds?
Analyzer is yet another object that could be used for this project. With analyzer, I'm thinking I could establish a general description for each of the siren sounds picked up by the microphone and then modify the object to recognize these parameters when picked up by the microphone.
Does anyone have recommendations or strong feelings about which of these three objects I should try to use for this project? Any other ideas or thoughts on how to go about this? I appreciate any help! Also, here is a link to audio files of some of the sirens I'll likely need to pick up: http://www.nytimes.com/2007/06/14/nyregion/15siren_sounds.html