Adding 2D waves with varied Y-scaling

Share icon

Hi all -
I need to add 2 2D waves with different y-scaling. I know how to do this with 1D waves:
Wave WaveSUM = wave1(x) + wave2(x)


But I can't find any way to do this with the y-scaling. My 2D waves have the same delta scaling in Y, but different offsets.

Any suggestions would be very welcome!
Your 1D expression is essentially executing a linear interpolation. In 2D you can use ImageInterpolate with the bilinear keyword.

A.G.
WaveMetrics, Inc.

October 27, 2015 at 11:12 am - Permalink

How about
Wave WaveSUM = wave1(x)(y) + wave2(x)(y)

in case the "+" is just an example and you actually want to do something else. Otherwise I also would go for ImageInterpolate.
HJ

October 28, 2015 at 03:50 am - Permalink

@HJDrescher: This will not work as Igor does not interpolate with multidimensional waves but returns the closest neighbour. See DisplayHelpTopic "Multidimensional Wave Indexing".

October 28, 2015 at 05:17 am - Permalink

Thanks all - I think I understand how to proceed. I will need to use the ImageInterpolate for both images to arrive at interpolated images with the correct y-scaled values for summing.

October 29, 2015 at 08:23 am - Permalink

@Thomas_Braun
To be precise: it depends. If the offset and the y-values are multiples of delta, it works fine. (Maybe "[]" 's are a better choice -- but not always)
It just depends on the data and what exactly is to be done.
HJ

October 30, 2015 at 04:54 am - Permalink

@HJDrescher: Could you post an example? All I can think of are situations where "()" and "[]" are interchangeable.

October 30, 2015 at 05:31 am - Permalink

I can generate one next week...
Here at least an idea of a case where you want the mentioned: poorly adjusted "2D detector" data sets, like ARPES data (2D) from a rapidly imaged charging sample.

The offset is adjusted to a certain feature in the individual spectra and afterwards the data is summed up to give better statistics.
Not a very common case I admit, and of course it also can be solved by "[x2pnt]" or "[d*p-shift]" constructs (Typically there is more than one solution).

But:
You don't want interpolation in this case. Your shifts and offsets are quasi-quantisized (like 5meV, due to detector resolution) and you only want that "middle" part of the spectra summed where they share the same energy range (x and y switched in this example but well..).
In this case, I would go for the w()() option.
HJ

October 30, 2015 at 09:52 am - Permalink