Cut a continuous signal into frames

Hey everyone,

I have a continuous signal with size 1x400000 (analog data that corresponds to a 20 min signal) and I would like to frames with 1 min each and do some signal processing on each. I wrote this code and tried it for a small matrix but when I increased the size of my matrix matlab was busy for a very long time. So, what is wrong and what can I do to improve it.

The code: t= [1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 17 17.01 17.02 17.03 17.4 18];

for i =1:length(t)

if t(i) <= 17

T(i)= t(i);

else

break;

end

end

plot(T, D(1:length(T))) %matrix D corresponds to the amplitude of my signal

when I used t= [1*ones(1,400000) 17 17 17.01 17.02 17.03 17.4 18] with the same code the program just like hanged!!!!

Any help will be appreciated.

[Luffy]

 t= [1*ones(1,400000) 17 17 17.01 17.02 17.03 17.4 18];
 T = t.*(t<=17) + 0*(t>17);
 This generated T,but when i wanted to see it in variable editor it got hanged,however it ws able to plot plot(T,T),any other plot it showed out of memory

[Image Analyst]

What does 17 have to do with anything? Is that where you detect the separation of one "frame" from the next "frame"???

[Yasmin Tamimi]

@Luffy Thanx for the code but I want to cut the signal and not introduce zeros!! The plot was wrong,, but thanx anyway.

[Yasmin Tamimi]

@Image Analyst It is only a number where I want to cut the signal. In my real scenario I want to cut the signal into frames at multiples of 60.

Answer 1

[Wayne King]

I'm guessing by your description, 400000 samples in 20 minutes that your sampling interval is 3 msec. So I'll use that value which gives 20000 samples in 1 minute.

You can just use reshape()

   % just making up some data the length of your signal
   x = randn(400000,1);
   xnew = reshape(x,20000,20);

Now each column of the matrix xnew is a segment (frame) of your signal 1-minute in length, you can now do your processing on the columns of the matrix.

If you have the Signal Processing Toolbox, see the help for buffer() also, which enables you to do the same thing but works when you want some overlap between the columns, or you do not have exactly the number of samples required for reshape.

[Yasmin Tamimi]

Thanx it really helped. I have tried another way:

t(t>60)=[];

As for buffer() i'm still trying it to make it work!