Compare arrays

[Cory_K]

---

Jeff Bohrer wrote:

Give the work to the processor- its bored

---

I'm sure that exactly what the processor wanted, a bunch of work on a Monday

[altenbach]

---

shenp wrote:

A nonlinear fitting process should be applied.

---

I am pretty sure that's not true for the data you gave us!

I made a quick example to optimize the nonlinear (quadratic polynomial) remapping using nonlinear least squares and the results are c=56.3239, b=1.00001, c=-1.2e-10 with a residue of=0.0013367. Doing the same with a linear fit gives c=56.3343 and b=1.00001 with the same residue. Thus, adding the quadratic term does not improve the fit!

Forcing b=1 and only fitting for the shift gives c=55.8808 and makes the residue only a little worse (0.00136855), well within the quality of the data.

A 56 point integer shift is all you probably need. 🙂

The picture show the difference between the aligned datasets. White is the original result from the crosscorrelation and red is the result after quadratic x-correction (offset +0.4 for better visibility).

Message Edited by altenbach on 04-06-2009 03:31 PM

[shenp]

altenbach,

Would please try these two sets of data. Thank you.

Paul