Home · Register · Join Upload & Sell

Moderated by: Fred Miranda
Username  

  New fredmiranda.com Mobile Site
  New Feature: SMS Notification alert
  New Feature: Buy & Sell Watchlist
  

FM Forums | Canon Forum | Join Upload & Sell

  

Archive 2015 · Measuring field curvature: your input

  
 
Brandon Dube
Offline
• •
Upload & Sell: Off
p.1 #1 · p.1 #1 · Measuring field curvature: your input


Hello all,

My time in Memphis at OLAF Optical Testing is quickly coming to a close and I am now putting the finishing touches on the field curvature procedures and chart output.

Here is an example model:

http://i.imgur.com/YStdzk7.png

http://i.imgur.com/xudpyTf.png

The spatial frequency can be anything. I selected 30lp/mm pretty arbitrarily; I think it provides a good DoF "thickness." For teles this will shrink some. Would you guys rather 20lp/mm or 50lp/mm or something else?

My software supports averaging multiple copies, how much would you care about that? We can use OLAF itself to select the least tilted copy and use that to represent all of them which is what we are leaning towards since it takes about 50 minutes to run one of these and a failed measurement is fairly possible. In the example case, the tilt shift, I would imagine this is completely worst case for the residual tilt.

In terms of a clean plot of the curvature, I intend to standardize f/8 because it it balances minimizing spherical aberration and coma (they will be virtually zero) to clean up the plot and measurement time, which grows exponentially as one stops down. The focus range can be expanded to compress the view, but a "dirty" plot because of aperture-dependent aberrations is not something that can be cleaned up.

I look forward to hearing your thoughts,

Brandon



Aug 12, 2015 at 05:04 PM
Monito
Offline
• • • • • •
Upload & Sell: Off
p.1 #2 · p.1 #2 · Measuring field curvature: your input


I'm seeing rainbows. I'm not sure what to think because I don't know what the charts mean. Perhaps you are assuming we know some things (from previous posts of yours?) not stated in your original post.



Aug 12, 2015 at 06:09 PM
rw11
Offline
• • • •
Upload & Sell: Off
p.1 #3 · p.1 #3 · Measuring field curvature: your input


and I wonder why? doesn't LR/PS/etc. correct for this?


Aug 12, 2015 at 06:18 PM
Brandon Dube
Offline
• •
Upload & Sell: Off
p.1 #4 · p.1 #4 · Measuring field curvature: your input


This is a measure of MTF vs Field vs Focus and may be simplified to field curvature as well as astigmatism. field curvature is something no processing software can correct; if something is out of focus there is nothing you can do to make it as sharp as what is in focus.

Essentially, the shape of the chart tells you how the lens' field curvature is (at infinity). In this case the tangential field is forward-curving and the sagittal field is generally backward curving, but is wavy and when shifted will end up forward curving.

This is information most useful to landscape photographers.



Aug 12, 2015 at 06:30 PM
Monito
Offline
• • • • • •
Upload & Sell: Off
p.1 #5 · p.1 #5 · Measuring field curvature: your input


How am I supposed to decode it?

What does red mean? Good, bad? Blue? Good, bad?



Aug 12, 2015 at 06:39 PM
Dawei Ye
Offline
• • • •
Upload & Sell: Off
p.1 #6 · p.1 #6 · Measuring field curvature: your input


Field curvature is something that is not documented well enough in general. Many lens reviews have given erroenous conclusions about sharpness due to failing to account for field curvature.


Aug 12, 2015 at 09:51 PM
snapsy
Offline
• • • • • •
Upload & Sell: On
p.1 #7 · p.1 #7 · Measuring field curvature: your input


Very cool. Is there a way to interpret these graphs to know what the center resolution would be when adjusting focus to bring the edges out of curvature and back into focus? IMO that would be a very useful data point for those evaluating a lens for landscape use @ infinity.


Aug 12, 2015 at 10:13 PM
scalesusa
Offline
• • • •
Upload & Sell: Off
p.1 #8 · p.1 #8 · Measuring field curvature: your input


Brandon, the way I read this is that Tangential focus is right on, the center of the red field is pretty much at 0.0.

But Sagittal is back focused a little.

Is this due to astigmatism or tilting of the lens elements?

Than, I see the field of curvature as varying from front to rear focus in the horizontal red band.

My thought is that its too much into the innards of lens performance for most of us, we would like a higher level diagram of field of curvature that combines the two to give a average..



Aug 12, 2015 at 11:30 PM
Brandon Dube
Offline
• •
Upload & Sell: Off
p.1 #9 · p.1 #9 · Measuring field curvature: your input


Snapsy, the color represents the MTF at 30lp/mm (see the colorbar), so for example if you focused 0.25mm closer in terms of the image plane (the lensmaker's equation can give you what this is in the scene) the tangential MTF would be about 0.65 or so. The sagittal plane would be similar, but falls out of focus faster - this indicates some axial chromatic aberration in the lens, though this model is hardly bad.

We could easily provide sag/tan/average if that would be helpful.



Aug 13, 2015 at 07:09 AM
Milan Hutera
Offline
• • • • •
Upload & Sell: Off
p.1 #10 · p.1 #10 · Measuring field curvature: your input


rw11 wrote:
and I wonder why? doesn't LR/PS/etc. correct for this?


No, it cannot add DOF and/or loss of shaprness in the blurry corners, that were caused by field curvature. The only workaround is proper capture (proper placement of focus) or useage of a lens that does not have field curvature.



Aug 13, 2015 at 07:54 AM
Monito
Offline
• • • • • •
Upload & Sell: Off
p.1 #11 · p.1 #11 · Measuring field curvature: your input


Brandon Dube wrote:
Snapsy, the color represents the MTF at 30lp/mm (see the colorbar),


Oh, sorry, the charts were so huge that I didn't see that on the right. Thanks.



Aug 13, 2015 at 08:54 AM
Brandon Dube
Offline
• •
Upload & Sell: Off
p.1 #12 · p.1 #12 · Measuring field curvature: your input


These are the size the charts are drawn at - when they eventually go on TDP they will be 700px wide. This is about as big as they can be drawn on our monitor, this gives the best image quality.

All lenses have field curvature, though different lenses have minimized it to various degrees. This is what it looks like in the design software:

http://www.photozone.de/images/3Technology/lensdesign/image18.png

(from the article I wrote for photozone.de on aberration theory - http://www.photozone.de/aberrationsExample, see the companion "theory" page for basics).



Aug 13, 2015 at 09:36 AM
Toothwalker
Offline
• • •
Upload & Sell: Off
p.1 #13 · p.1 #13 · Measuring field curvature: your input


Brandon Dube wrote:
The spatial frequency can be anything. I selected 30lp/mm pretty arbitrarily; I think it provides a good DoF "thickness." For teles this will shrink some. Would you guys rather 20lp/mm or 50lp/mm or something else?

My software supports averaging multiple copies, how much would you care about that? We can use OLAF itself to select the least tilted copy and use that to represent all of them which is what we are leaning towards since it takes about 50 minutes to run one of these and a failed measurement is fairly possible. In the example case, the tilt shift,
...Show more


Looks good. A few remarks:
I would use 20 lp/mm because I think it is the most important frequency for normal viewing conditions.
Strictly speaking you are measuring astigmatism, not field curvature.
I have no strong opinion about the averaging.
Why does the measurement time "grow exponentially as one stops down"?







Aug 13, 2015 at 10:55 AM
Brandon Dube
Offline
• •
Upload & Sell: Off
p.1 #14 · p.1 #14 · Measuring field curvature: your input


Toothwaker,

Thank you for your input. The trouble of using a lower spatial frequency is that it basically 'fattens' the plot and when the lens gets wide enough the DoF will fill the entire range (e.g 11-24 at 11mm).

I would say strictly speaking this is a measure of MTF vs Field vs Focus (hence the titles on the plots) - but field curvature is the most valid name for it still. Astigmatism is purely the difference between the location of the sagittal and tangential planes, which we aren't measuring (but could compute). Of course we are not measuring petzval.

The measurement time grows exponentially as one stops down because of the loss of light. A lens with good transmission at f/2 uses about 1/40s shutter time, at f/2.8 it is about 1/20s, at f/4 about 1/10s, f/5.6 1/5s, and so on. A measurement of MTFvsFieldvsFocus requires 441 exposures, not to mention the time for the collimator to swing and so on. f/11 is approximately the border before we seriously approach the noise floor in the sensor, but it hasn't proved too big an issue yet for the stop down data.



Aug 13, 2015 at 11:16 AM
Toothwalker
Offline
• • •
Upload & Sell: Off
p.1 #15 · p.1 #15 · Measuring field curvature: your input


Brandon Dube wrote:
Thank you for your input. The trouble of using a lower spatial frequency is that it basically 'fattens' the plot and when the lens gets wide enough the DoF will fill the entire range (e.g 11-24 at 11mm).


Yes, but I think you don't always need f/8. F/5.6 will nicely clean up spherical aberration and coma with many lenses.

Anyway, 30 lp/mm is not a bad choice.


I would say strictly speaking this is a measure of MTF vs Field vs Focus (hence the titles on the plots) - but field curvature is the most valid name for it still. Astigmatism is purely the difference between the location of the sagittal and tangential planes, which we aren't measuring (but could compute). Of course we are not measuring petzval.


Let's meet halfway. You measure the curvatures of the sagittal and tangential fields.


The measurement time grows exponentially as one stops down because of the loss of light. A lens with good transmission at f/2 uses about 1/40s shutter time, at f/2.8 it is about 1/20s, at f/4 about 1/10s, f/5.6 1/5s, and so on. A measurement of MTFvsFieldvsFocus requires 441 exposures, not to mention the time for the collimator to swing and so on. f/11 is approximately the border before we seriously approach the noise floor in the sensor, but it hasn't proved too big an issue yet for the stop down data.


A-ha. I did not know that SNR is a serious issue with MTF benches.

Thank you for measuring all these interesting data.



Aug 13, 2015 at 05:09 PM
Brandon Dube
Offline
• •
Upload & Sell: Off
p.1 #16 · p.1 #16 · Measuring field curvature: your input


Toothwalker wrote:
A-ha. I did not know that SNR is a serious issue with MTF benches.



Indeed. They use a projector light bulb to feed the fiber source, just imagine a 150W (typical) iridescent bulb. Then you throw away almost 50% of the output for the color filter, and finally project a target through an f/6 (again, typical) collimator. There is not too much light going into the systems, honestly.



Aug 14, 2015 at 07:52 AM





FM Forums | Canon Forum | Join Upload & Sell

    
 

You are not logged in. Login or Register

Username       Or Reset password



This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.