angle definition

Hi Tim,
On our Krios I have performed the check for Mag distortions with gold images according to your paper and got the following result:

Distortion Angle = 19.6
Major Scale = 1.005
Minor Scale = 0.995
Stretch only parameters would be as follows :-
Distortion Angle = 19.6
Major Scale = 1.011
Minor Scale = 1.000
Corrected Pixel Size = 0.529
The Total Distortion = 1.10%

According to your paper the parameters given here are the ones needed for correcting the distortion, so the actual distortion angle would be reciprocal to this, so the the angle of the major axis of the distortion would be about 110 degree (or -70 degree), correct?

Now, I also have data taken at the same mag on the system and analyzed in a similar way as shown by Zhao et al. 2015 JSB 192, 209-215 by CTF fitting with CTFFIND4. The data has virtually no objective lens astigmatism, so all the astigmatism detected through CTF fitting should be a result of magnification distortion. In these data the angle of the major axis of astigmatism according to CTFFIND4 is very robustly around +70 degree.

This only seems to make sense if the definition of a positive angle is opposite in the two programs mag_distortion_estimation vs CTFFIND4 (e.g. mag_distortion defines positive angle as clockwise vs. CTFFIND4 defines it counterclockwise). I have looked through the papers but did not find any statements on the exact angle definition. Is it possible that they are opposite in the two programs?

Thanks a lot
Ulrich

Hi,

Yes this is very possible. One further complication comes depending on whether you are talking about Fourier space or real space, i.e the axis with the largest scale in Fourier space will be the smallest in real space.

You can check the angle a little better by using the mag_distortion_correct program. If you use it to impose a distortion at a couple of different angles it should be clear what the convention is.

Cheers,

Tim

In reply to by timgrant

Thanks for your quick response. I will test the angle convention as you suggest with mag_distortion_correction - that is a good idea.
Ulrich