cross correlation output format

[Rubicon]

Hi-

I am trying to do cross-correlation analysis of cardiac electrograms (egm's), but the result (y-value) has a format that makes it difficult to compare the different egm segments with respect to degree of morphological similarity. I guess this can be done by some sort of normalization, but I do not know how to do that.

Analyzing the attached files a and b containing one short (100 ms) and one longer (300 ms) segment of the same egm recording with the attached cross correl_egm.vi; produces a peak y= 750.000 at x=280. At this point segment a is identical to the corresponding length of segment b.

How can I transform the y-values produced into the range -1 to +1 to enable comparisons in other segments with other egm morphologies?

Rolf  

[egraham]

At  first I thought you wanted to shift the time axis of one of the signals for comparison with the other then I reread and it seemed like you wanted to scale the y-axis down... so I guess I'm not sure what you're trying to do. Below example takes the top array input, finds the maximum value and scales the entire array down accordingly.

If what you're trying to do is shift the x-axis you could just insert an initialized array of zeros to move one signal over to line it up with the other.

Eric

[Holmes]

Hi guys,

I am trying to do something similar as posted above but with audio signals. I have two microhpone that give analog channel.

I want to use the correlation function to find out the phase difference. This is with respect to the orientation of sound source.

I am unsure as to how to zero pad the signals. Any help will be appreciated!!

Cheers!