Solved

Contrast Optimization and variable image size

  • 2 September 2022
  • 2 replies
  • 86 views

Userlevel 2

How do you cope with contrast optimization when optimizing on a system where the focal length is still to be determined (to be optimized within a range)?
In this case the MTF frequencies at which contrast is evaluated are constantly changing.

Say I have to meet a contrast specification at specific spatial frequencies defined at the object plane. 
Zemax calculates contrast at the image plane, therefore to use contrast optimization I must set fixed contrast frequencies. Fixing contrast frequencies at the image plane forces me to lock EFFL and magnification, but those parameters might have been exploited as useful degrees of freedom in the optimization.

This can be the case when, for example, one is optimizing an intermediate imaging stage and image size is not strictly bound to sensor dimensions.


I don't think there's a way to tell Zemax to update contrast frequencies at each optimization interation, even with API. If this is the case, should I simply stick to spot radius optimization?
Can you think at some workround? Not necessarily Zemax-wise. I’m thinking, more broadly, to the design workflow as a whole.

icon

Best answer by MichaelH 2 September 2022, 17:53

View original

2 replies

Userlevel 6
Badge +2

Hey Alberto,

I think it’s a good idea to take a step back and see what Contrast Optimization is doing.  The core concept for Contrast Optimization is minimizing the shifted wavefront in the Exit Pupil.  The diffraction limited MTF, which is the theoretical best MTF you can get across all frequencies, is a perfectly flat wavefront with 0 slope; when you autocorrelate this perfect wavefront with itself, the OPD vs Shifted OPD will always be 0, giving the best MTF.  Another way of thinking about this is by minimizing the slope of the original wavefront map, you maximize your MTF.  Since an abberrated wavefront can produce local minima/maxima in the MTF curve, once you get closer to a diffraction limited design, you want to input the exact frequency you care most about in the Merit Function Editor.  

Since Contrast Optimization is a form of diffraction optimization which requires a “stable” F/#, if your first order prescription data is changing, your system is not to a point yet where it makes sense to use Contract Optimization.  I would stick to geometric first order optimization (Spot Diagram) until you have a “stable” system and then you can switch over to Contrast Optimization.   

Userlevel 2

Hi Michael,

this insight about contrast optimization and its relation with wavefront aberration is totally clarifying to me. Even with the good theory in my backyard, sometimes I just forget it and get lost in the settings.

Thank you for bringing this again to my attention and for your streamlined reasoning.

Reply