[Gwyddion-users] profile not matching 'read value' data

[Luke S. <ls...@sa...>]

I am using Gwyddion 2.9 (I built via macPorts, which has not  
incorporated the more recent versions)

I have some data I wanted to read off and examine offline. In  
particular, I wanted to extract the values along segments of  
individual scan lines. To do this I made minimum-width profiles along  
the scan lines.

There were several oddities about this.

First, on the list of profiles prior to extraction, the table gave the  
left ends as being at (12,5), and (9,10); both proceeded straight to  
the right. However, when I generate the profiles and export the data,  
the starting points of the profiles correspond to the points (11,5)  
and (8,10).

Secondly, as I trace along these profiles, even after accounting for  
this, the values do not match. Many are close, looking like rounding  
errors. But not all.
For example, the point at (12,10), comes up in 'read value' as 8.19  
nm, but in the profile data, comes up as 8.55.
(13,10) continues as before, reading as 12.34 nm vs 1.24542*10^-9 m.
The deviations are thus irregular and are not the result of some low- 
order curve fit difference.

The only explanation I can think of is that even at the minimum  
thickness, 1, I am still getting averaging with neighboring lines.
That would be extremely counterintuitive.
Note that I have cropped and blown up the image so as to be able to  
deliver pixel perfect selections. Does the profile tool defeat this  
attempt by interpreting more finely than a single basic image pixel? I  
can see how this might explain both phenomena.

I can send a file featuring these errors to anyone who will help me  
resolve it.

Thank you,
Luke Somers