Dog Bark Identifier

[der_fisherman]

Does anyone have any idea as to if its possible with a PICAXE or PIC to "listen" to a microphone and recognize a dog's bark? and signal that.
I need to do it as I want to count just how many times a day my neighbour's dog barks...
I wonder if it is possible to say "record" 10 different barks and take an average of them in some audio manner, to allow other "Barks" to be identified and counted.....maybe using the ADC input of a PIC or PICAXE (would a PICAXE be quick enough?)
I don't want the counter to just count any loud noise.........
I will be grateful for any ideas, especially from someone who has already done something similar......
regards
Der Fisherman

 

[crutschow]

Detecting a dogs bark would require some complex analysis with a PIC.

As an alternate, there are dog bark devices available that generate a loud ultrasonic sound when they detect a dog bark, to curb the barking. If you did not want to use it for that purpose, you could disconnect the sound transducer and just count the signal pulses to the transducer.

 

[der_fisherman]

I already have that part, I wanted to automate it.....but thanks anyway.

 

[MikeMl]

Detecting a dogs bark would require some complex analysis with a PIC.

As an alternate, there are dog bark devices available that generate a loud ultrasonic sound when they detect a dog bark, to curb the barking. If you did not want to use it for that purpose, you could disconnect the sound transducer and just count the signal pulses to the transducer.

Our neighbor's mutt is wearing one of those. It is on the mutt's collar. I'm guessing that it just uses sound pressure level to trigger it, because it taught the mutt to bark quietly.

Might have to use a directional microphone to do this from afar.

 

[der_fisherman]

You are right, the sound pressure being close to the dog's head is easily measure. I am 40 meters away and want to distinguish from any other loud noises......so real time analysis is required.
I will have to set up and see what I can record......looks like it will be a first if I achieve anything.....
Regards

 

[HarveyH42]

My Lab puppy has different barks, for different reasons I suppose. It would be amazing if you could isolate just the barking, but got a hunch it's a little much to expect from a PIC (an AVR might ). Maybe the time between barks, since nuisance barking is repetitious, and pretty consistent. Anything else is worth checking out, since it's more like the kind of bark to alert you to trouble.

 

[flat5]

A bandpass filter, very directional mic, and proper gain setting might help to give a close count. It won't be perfect.

 

[birdman0_o]

Why are you counting how often your neighbours dog barks?
Perhaps you will have better success counting the squirrels that run accross his front lawn causing the barking

 

[kchriste]

A dsPIC would have the horse power to do a FFT of the waveform, analyze it, and guess if it is a bark or not. Microchip does have a **broken link removed** for the dsPIC but it is only for humans.
You'd need some pretty good programming skills to pull this off. It may be easier to try to get closer to the dog. Maybe put a directional mic on the fence that borders your property and set the VOX high enough that the dog is the only thing that triggers it +85% of the time. Then vet the recording (pun intended) before submitting it to the authorities.

 

[der_fisherman]

Well put

Thanks, I fear that you may be perfectly right......
regards
Der Fisherman