Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

sound recognition in matlab using neural networks

Status
Not open for further replies.

charub

Newbie level 3
Joined
Apr 8, 2012
Messages
4
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,281
Activity points
1,311
Hi.

My project is that I am importing the 10 sounds of gunshot and 5 different sound in Matlab. then I have to train the Neural Network with all the sounds. Now when I enter another gunshot I want my NN to recognize that whether the sound is of Gunshot or non gunshot sound. Can someone please help me in implementing this and also in the testing the network.

Thank you
 

BradtheRad

Super Moderator
Staff member
Joined
Apr 1, 2011
Messages
14,182
Helped
2,807
Reputation
5,622
Reaction score
2,744
Trophy points
1,393
Location
Minneapolis, Minnesota, USA
Activity points
105,784
The nn must distinguish any and all ways that a gunshot might sound different from other sounds.

I suppose your job is to teach it how to find ways to do this?

Such as:
* volume characteristics
* attack and decay envelope
* pitch and overtone characteristics
* echo characteristics.
Etc.
 

charub

Newbie level 3
Joined
Apr 8, 2012
Messages
4
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,281
Activity points
1,311
The nn must distinguish any and all ways that a gunshot might sound different from other sounds.

I suppose your job is to teach it how to find ways to do this?

Such as:
* volume characteristics
* attack and decay envelope
* pitch and overtone characteristics
* echo characteristics.
Etc.


Hi BradtheRad
Thanks you. but can you please help me with the implementation of it. I am taking the spectrogram of each sound wave. Is it right????
 

BradtheRad

Super Moderator
Staff member
Joined
Apr 1, 2011
Messages
14,182
Helped
2,807
Reputation
5,622
Reaction score
2,744
Trophy points
1,393
Location
Minneapolis, Minnesota, USA
Activity points
105,784
Yes, look at how the spectrograms differ, and what they have in common.

I don't know what it takes to implement a neural network. From what I gather, you must not tell it exactly what are the differences between the various parameters. You give it raw data, and tell it how to extract the parameters.

As for what will be raw data, I suppose it will be an image of each spectrogram. The program will examine pixel groupings, and see how they change over time, versus frequency / volume / overtones, etc.
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Top