MPE-65 lens DOF, diffraction and sensor size

A forum to ask questions, post setups, and generally discuss anything having to do with photomacrography and photomicroscopy.

Moderators: ChrisR, Chris S., Pau, rjlittlefield

LordV
Posts: 1568
Joined: Thu Nov 22, 2007 10:28 am
Location: UK

MPE-65 lens DOF, diffraction and sensor size

Post by LordV »

Didn't quite manage to find any info via search on the interaction of these parameters.

I've shot for a few years with 1.6crop cameras (20D,350D,40D) with my MPE-65 lens and have settled on some "normal" apertures I use at different magnifications which give the balance I want between DOF and diffraction softening.
I typically used :- 1:1 F11, 2:1 F9, 3:1 F7.1, 4:1 F6.3, 5:1 F5.0.
I settled on these values after doing various test subject photos to get the level of detail I wanted. It is of course subject dependent plus I can often recover DOF using focus stacking.

I recently got a 5D mk2 camera body and am using that with my MPE-65 lens (I tend to use dedicated bodies with macro lenses to reduce sensor dust).
2 questions arise from this change from 1.6 crop to FF sensor.

1. Am I correct in thinking that I can probably use around 1.6 stop smaller aperture at the same magnifications above and get approx the same level of diffraction softening ?

2. Linked to the above question and assuming I am not making A2 size prints - If I want the same "print magnification" as the 1.6 crop body, am I better off shooting at 1.6* the magnification on the lens perhaps at 1.6X smaller aperture with the FF sensor or cropping the image at the same magnification by approx 1.6 X. Should add that cropping a 5Dmk2 image by this amount seems to give approx the same or better IQ than a 40D image.
By better off, Im thinking of balance of IQ/DOF and diffraction.

Brian V.
www.flickr.com/photos/lordv
canon20D,350D,40D,5Dmk2, sigma 105mm EX, Tamron 90mm, canon MPE-65

Charles Krebs
Posts: 5859
Joined: Tue Aug 01, 2006 8:02 pm
Location: Issaquah, WA USA
Contact:

Post by Charles Krebs »

Brian,

Before this spins off into the algebra... :wink: :roll:

(Just continue doing what you're doing... looks great :wink: )

Your values for the f-number used with a 1.6 body are about what I like as well. (Interesting that if you use the Effective aperture = marked aperture*(m+1) relationship you are pretty consistently around an effective f28-30). It seems this is a good compromise for a single image or a two or three image stack. If I know I'll be using a bigger "stack" I like a little larger aperture.

I don't have the Mark II, but with the 5D (full frame) I'm generally willing to go about a stop farther down for a similar use. (That would be 1.4X rather than 1.6X but it keeps it simple). With 24x36mm I usually try to avoid effective apertures smaller than f32 unless the DOF really over-rides resolution concerns.

Of course this will be fudged considerably depending on whether the image will be a "single shot" or the opportunity is there for a stack. (The larger the "stack" possible, the less inclined I am to stop down very far).

The interactive chart near the bottom of this page is useful:
http://www.cambridgeincolour.com/tutori ... graphy.htm

That's my $.02.... now I'll get out of here before the numbers start flying around.

rjlittlefield
Site Admin
Posts: 20840
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

Brian,

I saw Charlie's post while I was working on this one. But I wasn't planning to get into many numbers either -- for reasons that I will explain below.

Your second question is easy. You are better off shooting full frame whenever possible. Of course you will have to crop for small subjects that would require magnifications greater than 3.1X (5X/1.6) on the small sensor. But other than that, you can capture more light and get more pixels by using the full frame.

Your first question is either easy or hard, depending on exactly what the question means.

If the question literally means "What f-number should I set at any particular lens magnification", then the answer is easy: Yes, just go 1.6X smaller, in proportion to the sensor size.

But then of course you're not talking about equivalent images between the two cameras. At 3X, for example, the full-frame sensor will be shooting a subject field of 12 mm wide. The smaller sensor would require a magnification of only 2X to fill its frame with the same subject. The problem is, if you shoot at 3X on the larger sensor using the 1.6 rule for aperture setting, and you shoot at 2X on the smaller sensor using your current preferred settings, then you'll get final images that have different DOF and diffraction blur.

If you're really serious about getting equivalent images from the two cameras, then the calculations get ugly even though the concepts are simple enough. Let me walk through the concepts, summarize some results, and you can decide whether you want to pursue that route any farther.

First, a bit of background.

DOF and diffraction blur are determined by what I call the "entrance cone angle", that is, the vertex angle of the cone formed by the aperture and an in-focus point on the subject. A narrower cone gives larger DOF but more diffraction; a wider cone gives shallower DOF and less diffraction.

Assuming that you end up with the same "print magnification", then the tradeoff between DOF and diffraction has everything to do with the cone angle and nothing to do with the sensor size. If you set apertures so that two different lenses or sensors give the same DOF, then you are also setting the same cone angle, and thus you'll get the same amount of diffraction blur.

However, when you're using the same lens, the size of the sensor does affect what setting of the aperture is required to give that same DOF and diffraction blur. This is because you must use different subject-to-lens distances in order to fill different size sensors with the same size subject. When you change the subject-to-lens distance, you also change the entrance cone angle. The larger sensor will require you to move the lens closer to the subject in order to fill the sensor. If you use the same aperture setting, then moving closer increases the cone angle, which reduces DOF and gives less diffraction blur.

To retain the same cone angle (and the same DOF and diffraction blur), you must set a smaller aperture with the larger sensor. The question is, how much smaller?

In one sense the answer is simple: the aperture size just needs to be in proportion to the lens-to-subject distance so that the cone angle stays constant. Unfortunately, the lens-to-subject distance changes in a complex way with magnification and sensor size. At low magnifications, changing the sensor size changes the lens-to-subject distance by a lot, while at high magnifications, it's only a little. Sigh.

I made up a spreadsheet for this, using the thin lens formulas and casually assuming that an MP-E 65 really does have a fixed focal length and is "symmetric" meaning pupillary magnification factor P=1. (I don't trust those assumptions, by the way, and I don't have an MP-E 65 to check against.)

At the moment, it looks like the f-number adjustment for equivalent images ranges from a high of around 1.3 down to a low of 1.0 depending on magnification. But the formulas are messy and I would have to play with them for a while to figure out A) whether I have them correct, and B) how to make the calculations understandable. They're never going to be simple.

Personally, in this case I wouldn't bother messing around with anything more complicated than the one-stop-down rule that Charlie mentions. You'll figure out quickly enough whether you want to tweak the aperture one way or the other, just like you figured out that currently at 5X you prefer effective f/30, while at 1X you prefer effective f/22. In this case the math is just a place to waste time when you could be shooting pictures.

--Rik

LordV
Posts: 1568
Joined: Thu Nov 22, 2007 10:28 am
Location: UK

Post by LordV »

Thanks for the replies Charles and Rik - appreciated. Upto now I've tended to shoot with the same apertures or just 1/3rd stop down on the FF camera and have marvelled at the detail capture. I did try some crude maths myself but am never sure how precise the formula are, but came to the conclusion that any DOF loss on the FF compared to the 1.6 crop for equal FOVs (ie different mags) might possibly be made up by shooting at narrower apertures to an equivalent level of diffraction softening.
Fraid I still find the nomenclature of apertures confusing as to whether small or large aperture referes to the size of the hole or the F number :).

The query on image cropping at high mag was slight hangover from the 1.6 crop camera where sometimes I could not convince myself that at an image taken at 5X actually had any better detail than one taken at 3X and cropped when viewed at a reasonable size.

Anyway will try some one stop down apertures from my normal to see how they look :)

Brian V.
www.flickr.com/photos/lordv
canon20D,350D,40D,5Dmk2, sigma 105mm EX, Tamron 90mm, canon MPE-65

rjlittlefield
Site Admin
Posts: 20840
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

LordV wrote:any DOF loss on the FF compared to the 1.6 crop for equal FOVs (ie different mags) might possibly be made up by shooting at narrower apertures to an equivalent level of diffraction softening.
Yes, it certainly can. The exact setting to do this may be hard to figure out, but it's always true that same DOF = same diffraction softening, assuming you have the same print magnification.

--Rik

Charles Krebs
Posts: 5859
Joined: Tue Aug 01, 2006 8:02 pm
Location: Issaquah, WA USA
Contact:

Post by Charles Krebs »

Rik,
I made up a spreadsheet for this, using the thin lens formulas and casually assuming that an MP-E 65 really does have a fixed focal length and is "symmetric" meaning pupillary magnification factor P=1. (I don't trust those assumptions, by the way, and I don't have an MP-E 65 to check against.)
FWIW... A non-critical "hold it out at arms length" and look in the front and back shows very little (if any) size difference in the exit and entrance pupils. (Maybe a bit at 5:1, but certainly nothing major). And in use the lens does indeed seem to follow the simple f*(m+1) relationship as far as exposures are concerned.

This exposure information was also the info given in my (now long lost) instruction booklet that came with the lens.

rjlittlefield
Site Admin
Posts: 20840
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

Thanks -- good to know. I got seriously surprised by the Olympus bellows lenses, especially the 20 mm f/2, so I'm a little goosey on the subject.

--Rik

LordV
Posts: 1568
Joined: Thu Nov 22, 2007 10:28 am
Location: UK

Post by LordV »

Couple of test shots below taken around 2/3rds stop smaller than my normal.

Brian v.


Small muscid fly- 5:1 FF, F7.1 with some focus stacking

Image

Larger muscid fly 4:1 FF, F8 single shot

Image
www.flickr.com/photos/lordv
canon20D,350D,40D,5Dmk2, sigma 105mm EX, Tamron 90mm, canon MPE-65

DQE
Posts: 1653
Joined: Tue Jul 08, 2008 1:33 pm
Location: near Portland, Maine, USA

MPE-65 cross-section and other info - Canon web site

Post by DQE »

See this web page, and its built-in menu tabs.

http://www.canon.com/camera-museum/came ... 5.html?p=2

It appears to be quite different from the assertion (at other macro forums) that it's an ordinary macro lens with a built-in variable extension tube, or that it's a reversed wide-angle lens with a variable extension tube. Its design also seems to challenge the assertion that simple extension tubes are better than other options since they don't have any glass elements. Since I have a limited education in lens design, I can't offer much in the way of constructive comments.

Below is an excerpt from the provided URL that precedes it, but I'm not sure if this info is accurate.

http://www.mir.com.my/rb/photography/ha ... /index.htm

"The optical construction comprises of a 10 elements in 8 groups design which includes a three lens groups floating lens elements system that moves the first, second and third lens groups to deliver variable magnification as well as high quality images in the macro photography. The design effectively corrects the aberration fluctuation which accompanies changes in magnification."

Below is a link to a scanned PDF copy of the lens manual with what appears to be Canon data tables. I hope this helps Charles Krebs and others!

http://www.lensrentals.com/rent/canon-m ... /for-canon

http://files.getdropbox.com/u/754116/MPE-65manual.pdf

I hope this post is helpful and useful in some way. It's certainly a very interesting technical discussion, and relevant to my own attempts to learn macro photography (5DII, MPE-65, etc).
-Phil

"Diffraction never sleeps"

rjlittlefield
Site Admin
Posts: 20840
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

DQE, thanks for all the links.

One thing to watch out for with this forum software is that the url ... /url tags have to be properly matched or your whole posting disappears (prints as blank!).

This was happening to your post. I edited it to fix the problems, so now I think it looks the way you wanted. You can take a look at how the tags are used by clicking the Quote button (or Edit, but Quote is safer).

It's also a good idea, before hitting Submit, to Preview your posting to be sure that it looks like you intend. If Preview shows blank, then go looking for problems with the url tags.

Hope this helps!

--Rik

rjlittlefield
Site Admin
Posts: 20840
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

Aha! So I edited your initial post, then you submitted a different one with no url tags at all -- what glorious confusion! But at least the info is all there now and looks OK. It works fine to have no url tags, as long as all you want is for basic url's to be recognized.

I deleted the extra post that you asked me to.

--Rik

DQE
Posts: 1653
Joined: Tue Jul 08, 2008 1:33 pm
Location: near Portland, Maine, USA

Post by DQE »

rjlittlefield wrote:Aha! So I edited your initial post, then you submitted a different one with no url tags at all -- what glorious confusion! But at least the info is all there now and looks OK. It works fine to have no url tags, as long as all you want is for basic url's to be recognized.

I deleted the extra post that you asked me to.

--Rik
Rik,

Thanks for your help and useful posting advice. Sorry for the confusion.

Here's a link that you and others might be interested in checking out - it's apparently a full text PDF of a well-known textbook on lens design:

http://int6.net/ebook/laser/ldlai.pdf
-Phil

"Diffraction never sleeps"

Post Reply Previous topicNext topic