Sunday, May 22, 2011

A quick word about MTF

A recurring theme on this blog is the subject of optics. That’s for the very simple reason that all image processing and analysis begins with an image. And since it’s not possible to put information in to an image, you’d better start by getting the very best one that you can. But when you’re shopping for a lens how can you tell what’s good and what’s not so good?

The stock answer is to look at the MTF graphs. Well I’ve done that, and they made about as much sense to me as ancient Greek. But now, thanks to Vision System Design’s Andy Wilson and a piece he wrote way back in 2009, (“Match Game”) I’ve found a website that explains the whole MTF thing in terms that even I can follow.

The article you need to read is “Modulation Transfer Function - what is it and why does it matter?written by Bob Atkins and published on the photo.net website. Bob is writing more for photographers than for machine vision folks, but that doesn’t matter: he provides some very clear guidance on how to interpret those MTF graphs.

Now you have no excuse for letting the lens be the weakest link in your system.

4 comments:

Andy Wilson said...

Dear Dr. Grey:

It's OK to read MTF graphs but try finding a lens vendor that provides them! You will have a very hard time!

All the best,

Andy

Andrew Wilson
Editor
Vision Systems Design magazine
E-mail: andyw@pennwell.com
The Magazine On-Line:www.vision-systems.com
Tel: (800) 225-0556 x9115
Fax: (603) 891-0574
Skype: andywilsonatskype
LinkedIn: www.linkedin.com/in/andycwilson
My Machine Vision and Image Processing Blog: http://machine-vision-news.blogspot.com/
My Infrared Technology and Applications Blog: http://infraredtechnologyandapplications.blogspot.com/
YouTube: http://www.youtube.com/user/AndyCWilson
Mail: PennWell, 98 Spit Brook Road, Nashua, NH 03062-5737, USA

Support the World Community Grid at: https://secure.worldcommunitygrid.org/reg/viewRegister.do?teamID=LZXVV18ZS1

B Grey said...

I have to agree with you Andy, though I'm pretty sure Linos, sorry, Qioptic, provide them on their web site.

Spencer Luster said...

Regarding MTF curves, Light Works is happy to provide these for all of our telecenetric lenses upon request. We'll do them for various F-stops, wavelengths, out of focus conditions, etc. We don't provide them as a regular part of our website or other documentation because frankly most people don't care much.

I agree that MTF comparison, under the same conditions, is the best way to compare lenses. Our experience with our customers, however, remains along the lines of: "Will I be able to see X?" An MTF specification doesn't help 98 out of 100 people asking the question.

On our website we've provided what I hope are some helpful comments about MTF, along with four bar target images showing different conditions. [http://lw4u.com/index.php/machine-vision-help/glossary/ --bottom of the page] The intention is to help customers get a better intuitive feel for MTF numbers.

At the end of the day, however, customers still want to know, "Will I be able to see X?" That's why part of our SOP is to do free sample imaging of their parts when appropriate.

Spencer Luster

Brian Durand said...

My experience has been similar to that of Andy Wilson. For the most part, only high-end lens manufacturers will provide MTF data. Even then, the graphs seldom allow for apples-to-apples comparison. This is no accident.

Try comparing standard 2/3" 16 mm lenses from Tamron, Fujinon, and Navitar. You really have to buy each lens to compare value. One exception I'm aware of is Kowa, which has MTF data on their website, even for lower end optics.