Changing the Autocorrelation time

[Cameroon]

Hi there -

I want to run an autocorrelation for 1 second, not 0.2 seconds as the autocorrelation VI only lets me do. How can I change this?

Thanks,

Cam

[JLS]

Hello,

Which version of LabVIEW are you working in?  If you are concerned that a particular part of your code is running too quickly for your application, you can always add wait statements to slow it down.  It sounds like in your case you'd like to add an 800ms delay - see the embedded screenshot below.

Best Regards,

JLS

Message Edited by JLS on 06-22-2006 11:37 AM

[Cameroon]

Hi there - thanks for your suggestion but I think you might have mistakenly misinterpreted me on the unpreciseness of my question:

What you suggested only delays the autocorrelation from happening. I need the actual autocorrelation function to mathematically compute longer than it is doing. As it stands now, with a sample rate of 1 KHz, I am, for some reason or another, only acquiring data for 0.2 seconds in my autocorrelation. But I need to acquire data for at least a second, possibly 2 or 3 seconds sometimes. Can I possibly do this without changing my sampling rate. I see nothing in the autocorrelation VI's that would allow me to do this? I am using Labview 8.0 by the way. Is this any clearer?

Cam

[johnsold]

I think the autocorrelation function accepts the array you send to it. It sounds as though you need to change your data acquisition to acquire more samples, thousands rather than hundreds.

Lynn

[JLS]

Hello,

That is correct.  As I explained in an email, the autocorrelation is a convolution - you'll see that you end up with an array about twice the size you input.  There is nothing wrong with processing your entire array if you'd like, or sampling your data if you only want to process 200 points for some reason.  Keep in mind that the autocorrelation is a computation - how long it takes to compute is dependent on the size of the input array, and the computer you use to execute the operation.  If you are concerned with processing a second's worth of data, that boils down to using your sampling rate to compute how many samples you'd like to process.

I hope that all makes sense.  I have the feeling that there is either a misunderstanding here, or that you have a more fundamental problem to address, and the question you're asking is only about the method you've chosen to solve that more fundamental problem.  If you do have a more fundamental question which this one depends on, feel free to ask it here and perhaps we can attack it in a different way!

Best Regards,

JLS