Identifying Sounds: Apparatus Needed?

  • Thread starter Thread starter callumbrad128
  • Start date Start date
AI Thread Summary
The discussion centers on creating a device that activates a light when it detects the specific sound of a blender, while ignoring other noises. Recognizing the complex sound of a blender, which consists of multiple frequencies, poses a significant challenge and may require machine learning techniques for effective pattern recognition. A simpler approach involves using a microphone and an operational amplifier tuned to a specific frequency, making it easier to detect single tones. The conversation also references existing devices like the "Clapper," which activates based on sound recognition, as a potential inspiration. Overall, the project is feasible but requires careful consideration of sound recognition technology and apparatus.
callumbrad128
Messages
10
Reaction score
0
Is there such a part that when it hears a particual sound, that it only recognises this sound.

let me explain...

I am trying to conduct an experiment, and have limited knowledge in this field! Basically my project is to make something, so when a blender is turned on 'my project' hears this sound, and a light comes on (prefer a light, but it could be anything which indicates the noise has been heard). To make things harder for me, is this possible for the light to come on when only a particular noise is heard! If this is possible, could someone explain the theory behind this and the type of apparatus need to make this!

I hope this doesn't come across as if it were "a homework project"! I am infact 22 and just an experiment I want to carry out for my own pleasure! (hopefully everyone can see, that by my lack of knowledge it would def not be anything to do with a homework)

Any help would be muckly appriaciated
 
Physics news on Phys.org
Tricky
Recognising a single tone is very easy, recognising the pattern of almost random noise from a blender is much harder.
It becomes a pattern recognition/machine learning task (see neural nets, Haar classifiers etc)
It depends on how disciminatory you need to be, do you need to be able to tell the difference between a blender and a dishwasher, or a lawnmower outside, or a concrete mixer? Or do simply need to respond to any noise like a blender.
 
Just need to recoginise one noise! A blender was an example! So that it knows the sound of a blender, but when it hears a "cement mixer" sound, nothing would happen! Please please help!
 
Like MGB was trying to get across is that a blender has a sound that is made up of many individual tones, i.e. random noises. If you were to look at a spectrum of the noise from a blender it would have many peaks at many frequencies that would comprise the overall sound. That would be a difficult task to recognize without something like a spectrum analyzer or the like to recognize the overall spectrum. Now, if your sound source was a single, pure tone, i.e. one frequency, then it would be a lot easier to do.
 
Ok, I understand what u are saying! If it was a one consistent noise with one frequency! What appuratus would be able to only pick out that, apose to other noises!
Thanks
 
Hello callumbrad-
It is very straight forward to set up a microphone and an operational amplifier tuned to a very specific audible frequency, such as C below middle C for example (~256 Hz).
Bob S
 
There is a device they use to sell a few years ago and may still sell it called the “Clapper”
By clapping your hands you could turn on the lights. This device may give you some insight on what you are looking for.

https://www.amazon.com/dp/B0000CGKLR/?tag=pfamazon01-20
 
Last edited by a moderator:
Here's a link to a pretty neat thread where the OP is trying to recognize the two tone signal created by telephone number dialing. You can get an idea of just how difficult it could be to recognize a signal with hundreds or thousands of components!

http://www.vbforums.com/showthread.php?t=325457
 
Back
Top