Apurva sent me an email about integration trouble

[joshu...@gmail.com]

From Apurva:

Hi Josh,

I was playing with the chi integration yesterday and noticed something
funny. It seems to work well when the constraining Q range is small,
but gives the wrong answer when the Q range is large.
The only way I can visualize this to happen is if you are doing the
polarization correction AFTER integration. If so, that is wrong
because PC is a function both chi and q(2th) so can not be brought
outside the integration block, right?

I think, it maybe better to see integration as the step after caking
instead of independent block, because by creating the cake you have
done 90% of the work necessary for doing the integration right.
The Q and Chi integrations are just vertical and horizontal averages
of the polarization corrected and masked region excluded caked pattern
(as Sam pointed out last time we met).
So I think, you need two additional steps after caking,

1) Correct the cake intensity pixel by pixel by the polarization
correction.

2) and then average - not sum - the intensity in the vertical or the
horizontal range as requested by the integration modules, you will
have the correct integration required without any addition
corrections.

Let me know if this makes sense...

cheers

Apurva

[joshu...@gmail.com]

Hey Apurva. I moved this discussion onto the discussion group so that
it gets archived. I hope you don't mind.

I am sorry I have delayed working on this for so long. I have been
working on my other work and have not made this a priority, but I am
doing my best to get on top of things. I fixed the bug that Sam found
in the code and rolled out version 1.0.5 of the code (with the fix)
and also a new beta of the program with the fix:
http://areadiffractionmachine.googlecode.com/files/Area%20Diffraction%20Machine%20v2_beta_248.zip

Can you run your same analysis with this new code and see if you have
the same trouble. If so, we will have to think about this some more.

The new beta of version 2 also lets you set the scale factor or set
the min / max intensity. It should be pretty easy to figure out how
this works, and it is also documented in the new version of the manual
I just uploaded. There are no preset scale factors like you asked for
because it seemed like too much work, but it should be easy enough to
just set the scale factor by hand. If you want the scale factor to be
persistent, you can always use the ``set as initialization'' option to
do so.

Also, I don't think I ever told you guys, but awhile ago I decided to
add a new feature ``Select the Center'' which you can use to click on
the diffraction image to pick the center of the image. You can find it
in the calibration menu. It might come in handy.

I have started writing the paper and would like to talk to you about
it at some point, but I should probably put some more work into it
first.

Cheers,

Josh