MATLAB code help please....

Thread Starter

jennifer22cute

Joined Sep 18, 2006
2
Hi,
My name is Jennifer.Iam new to Matlab programming.Iam working on reducing the noise content in audio communication.Iam using Hamning codes for the same.I have a sample code with me,which iam uploading in a zip file.Iam unable to understand how it works.Can someone please explain me how the code is written,what it means and what the different plots mean and how to read them? I have trouble especially with the plots trying to understand them.

Secondly,I want to add a filter to the output of the code obtained from the code so that I can filter the noise out.An Adaptive filter is recommended but not limited to.As long as it can be filtered,that's fine.

Being a beginner,Iam having a lot of trouble trying to understand.Iam trying to read Matlab tutorials at the same time but I have to finish up this work soon before the deadline.So can anyone be kind enough to help me through this please,with atleast one of those requirements,most importantly,the filter thing? I would greatly appreciate all the help I receive. Thank you in advance.
 

Attachments

atferrari

Joined Jan 6, 2004
4,763
Hola Jeniffer,

After some testing, Hamming code proved to be very limited. I didn't want to believe but better you do.

Try some beter error detection/correcting scheme.

I can't help on Matlab anyway. Sory.

Good luck.
 

Papabravo

Joined Feb 24, 2006
21,094
Matlab's a bit pricey for most of us who hang here. I'm curious about the amount of noise in the original sample and how much reduction you are trying for. Is the audio signal voice or music?
 

Thread Starter

jennifer22cute

Joined Sep 18, 2006
2
Hi,
Audio voice signal is provided in the zip code above as attachment.But i want to be able to apply it to music too.The reduction is not limited.As much as possible noise has to be reduced.Thanks.
 

Papabravo

Joined Feb 24, 2006
21,094
The problem with all this is that filtering is inherently a frequency domain operation and it doesn't care if the the input is signal or noise. It treats them both the same. Where you have a chance at noise reduction is where the spectral characteristics of the noise are different than the signal.

For example most of the spectral energy of a voice is between several tens of Hz to about 3kHz. So low pass filters will preserve the voice information while attenuating the higher frequency noise. this doesn't work so well with music.

I'm still having trouble getting my arms around the point of using coding to reduce noise.

Edit:
After thinking about your problem for most of the afternoon, and pulling out my copy of Gallager, R.G.,Information Theory and Reliable Communication, I realized that you are doing something different that what I undestood from your original post.

You are not trying to reduce the noise in the original sampling of the audio signal. You are sampling and encoding an audio signal into a digital form. You then transmit the signal in digital form over some channel, such as an AWGN(Additive white Gaussian Noise) channel. On the other end of the channel, in your decoder, you attempt to recover the original signal by using the properties of the code to remove the noise introduced by the channel. This problem has absolutely nothing to do with any noise in the orginal sampling of the audio signal.

Is my revised understanding correct?

Since I can't run the Matlab programs, could you post one ot the plots that is confusing you?
 

Dave

Joined Nov 17, 2003
6,969
Matlab's a bit pricey for most of us who hang here.
Hey Papabravo,

You could look at FreeMat as a free alternative to using Matlab (which I agree is rather on the expensive side). FreeMat has the capability to read m-files which constitute the core format for Matlab, amongst many other features one associates with the Matlab environment.

Dave
 
Top