Canon 100mm f2.8 Macro with Nikon CFI 10x/0.25 on FF body?

Have questions about the equipment used for macro- or micro- photography? Post those questions in this forum.

Moderators: rjlittlefield, ChrisR, Chris S., Pau

Asha
Posts: 70
Joined: Tue Apr 09, 2013 8:08 pm

Post by Asha »

Has anybody tried the Canon 100mm f2.8 Macro with Nikon CFI 10x/0.25 on a full frame body?

I have 5D MIII and the 100mm/f2.8, and love the camera/lens combo, but am looking for a bit more magnification. I don't have a particular magnification goal, but want good image quality. Cropping the edges/corners off is ok with me too. When stacking, the outer edges end up being unusable anyway because I think the lens is not telecentric.

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

Asha wrote:Has anybody tried the Canon 100mm f2.8 Macro with Nikon CFI 10x/0.25 on a full frame body?
Yes, somebody somewhere has tried that, though I don't have a link handy. The edges and corners will definitely go bad when it's pushed that far. If they didn't, then the lens would be covering an image circle on subject of over 8 mm diameter, but its actual coverage with high quality is more like 6mm.
I have 5D MIII and the 100mm/f2.8 ... When stacking, the outer edges end up being unusable anyway because I think the lens is not telecentric.
I'd be interested to hear more about this problem. Summary is that the 100 mm f/2.8 is definitely not telecentric. But neither are most other lenses, and they all hold up well to stacking unless you're running into the issue described HERE. In that case probably a change in your method of focus stepping would solve the problem.

--Rik

Edit: to remove a bit that would be confusing now that this thread has been split off to be stand-alone.
Last edited by rjlittlefield on Sat Apr 18, 2015 10:48 am, edited 1 time in total.

Asha
Posts: 70
Joined: Tue Apr 09, 2013 8:08 pm

Post by Asha »

Thanks Rik--I tried searching the forums but came up empty on full frame discussions.

Regarding the focus methodology, I typically do 1:1 macro, and switched from rail to focus ring methodology based on your tutorial. I still get bad edges--worse on my Panasonic u43 with 45mm 1:1 than it is on the Canon FF with 100mm 1:1.

ChrisR
Site Admin
Posts: 8668
Joined: Sat Mar 14, 2009 3:58 am
Location: Near London, UK

Post by ChrisR »

Regarding the focus methodology, I typically do 1:1 macro, and switched from rail to focus ring methodology based on your tutorial. I still get bad edges--worse on my Panasonic u43 with 45mm 1:1 than it is on the Canon FF with 100mm 1:1.
That's for another thread really, but "you would". Choice of focus method is more about view angle than subject size. Longer lenses are generally easier, and focus at 1:1 can be "odd".

Asha
Posts: 70
Joined: Tue Apr 09, 2013 8:08 pm

Post by Asha »

Thanks to all for your repsonses!

Here is an example of a stack I did sometime last year--it is an antique antler with some fur attached to it. I used this subject so I could ascertain the quality of different kinds of texture, especially "hairy" ones.

As you can see, around all four edges, the assembled stack looks "smeared". I spot checked some of the individual slices, and saw that there was a significant difference of content near the edge of the frame. I thought this might be due to the fact that my lens is not telecentric. However, I'm not ruling out other possibilities, especially given that this stack was relatively deep. Perhaps it is a combination of factors.

For this stack, I used my Canon 5D MIII with 100mm/f2.8 USM II (1:1 mag). I can't remember if I used the focus ring directly, or if I used the EOS software to control the focus ring. In either case, the stack was acquired using focus-by-ring method as shown in the previous link/tutorial.

Image

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

Asha, thanks for sending me the source stack to look at.

What you're running into is the classic "edge streaks" issue that happens when a stack gets processed starting at the wide end, causing the output image to include regions that were outside the frame of one or more source images. In some of those regions, the focused image content that you might hope to appear in the output was simply never seen by the camera.

There's a description of the underlying problem in one of the Zerene Stacker FAQs, "My output image has a band of streaks on one side. Why is that?". That description addresses the case where the framing problem is due to alignment or sag, but forcing a stack to be processed starting at the wide end causes the effect to be seen on all four sides.

The output shown here looks like it was produced by Helicon Focus, not Zerene Stacker, but the underlying problem appears with all stacking programs and the only question is exactly how they deal with it.

I should have time later today to show you in more detail what I mean.

--Rik

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

Forgot to mention...

The framing issue is discussed in some detail in one of the ZS tutorials, Tutorial #2: Using a Macro Lens on a DSLR. There's a section at the bottom, "Summary: Use either method but be sure to check framing".

If you look higher up in the tutorial, you can find illustrations of how framing changes from front to back, for the particular optics used in that tutorial. But the details vary widely between lens systems. Even when you're focusing by ring, the wide end of the stack can be at either the front or the back, depending on which lens or lens combo you happen to be using.

--Rik

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

Sorry for the delay...

Here is some imagery that may help in understanding what's going on.

To begin, here are the first and last frames in your stack, shown at half-size based on the 720x480 versions that you sent to me.

Image Image

It may not be apparent from that representation that there's a framing issue.

But here's what appears when we pull the stack into Photoshop using File > Scripts > "Load Files into Stack...", then have Photoshop do an Edit > "Auto-Align Layers..."

Image Image

Notice that Photoshop has made the first frame larger and the last frame smaller. The checkerboard region around the last frame indicates the parts of your stack that are outside the frame of at least one source image.

In that checkerboard area, there is missing information. Any composite image that includes that area is likely to cover some parts of the subject that the camera simply never saw focused. Once this fact is recognized, it should be no surprise that those regions may be rendered badly.

Now let's look at how Zerene Stacker handles this situation.

The first and last frames as input to Zerene Stacker are just the same as we saw earlier:

Image Image

When we process the stack using default parameter settings, Zerene Stacker automatically figures out that the last frame has a narrower field of view. So it reverses the order and starts processing at the narrow end. The frame at that narrow end retains its original size, and all other frames get expanded and cropped to align with it.

Here are the same frames after this default treatment of alignment:

Image Image

Within this crop area, full information is available and the stacking process can produce a clean result.

Here is the Zerene Stacker PMax output with default alignment settings, shown now at the original resolution you sent me.

Image

On the other hand, one can turn off the "Automatic order" option and force Zerene Stacker to process the stack starting at the wide end.

In that case, here are the first and last frames after alignment ("Show as adjusted"):

Image Image

Notice now that in the troublesome area where Photoshop showed a checkerboard of "no data", Zerene Stacker has instead produced edge streaks. This is due to the interpolation process described in the Zerene Stacker FAQs, HERE.

In this case, with processing forced to start at the wide end, Zerene Stacker is very likely to propagate those edge streaks from this or other narrow frames into the output. For this stack, some of the intermediate frames produce very nasty streaks indeed:

Image

I think that the image you showed was made by Helicon Focus, which in recent versions does a better job of handling missing data than any other software I know. Unfortunately it cannot create correct data from thin air, so even the best job that can be done is still not very good -- as you noticed. As with Zerene Stacker, the default settings in Helicon Focus would cause those bad edges to be automatically cropped off, but that behavior is simple to change by deselecting "Sort automatically".

Bottom line, the way to get stacks with clean edges is to set up the shoot so that everything you care about is inside every frame, then let the automatic analysis process in the software figure out how to avoid the bad edges.

Thanks for the discussion. I've updated the description in the Zerene Stacker FAQs to do a better job of covering this and one other source of edge streaks that have come up in recent discussions.

--Rik

Asha
Posts: 70
Joined: Tue Apr 09, 2013 8:08 pm

Post by Asha »

Rik, thanks so much for the detailed explanation! You are correct--that stack was processed with Helicon. I do not remember the settings as it was too long ago (and I didn't take any notes). I suspect I overrode the automatic settings thinking that I could avoid the software making cropping decisions for me. I see now that since the only useable area is inside where the software crops the final stack, it only serves to create extra work if I decide to do things manually.

I had noticed this same issue when I did my very first stack with a different camera and lens combination. After that first photo session, I started to frame the object with adequate buffer around it, to account for the change in perspective.

Back to the issue of edge/corner falloff when using a microscope objective on a full frame sensor (assuming Canon 5D MIII with 100mm/f2.8 USM II)--I think it would not work because there would be falloff in all the frames, even the ones on the "short" end. This would result in an even smaller area than I am already working with in the 1:1 setup.

I've taken your advice and ordered a diopter lens. I'll see if that gives me enough magnification while maintaining good quality. If it doesn't, I'll consider the Canon MP-E 65mm.

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

This is one of those fairly frequent cases where I'm struggling to understand what somebody needs to know.

As I understand the situation, you have a Canon 5D Mark III full-frame camera and a Canon 100 mm macro lens. That combination naturally goes down to imaging a subject size of 36x24 mm, and now you're trying to figure out a good way to image subjects that are smaller than that.

For flexibility in this situation, it's hard to beat the MP-E 65. It covers full frame by design, so just by mounting the lens and turning the ring, you can image subjects from 36x24 mm down to about 7x5 mm.

You've asked specifically about pairing a Canon 100mm f2.8 Macro with a Nikon CFI 10X NA 0.25 objective. That combo definitely will not cover your full-frame sensor with high quality image. It may not cover your full-frame sensor with an image at all. As you can see at http://www.photomacrography.net/forum/v ... 2051#92051, that combo vignettes a little bit even on an APS-C sensor. I don't have equipment to test it, but I'd expect it to vignette quite badly on full frame. This is yet another entrance pupil issue, not unlike some others that you and I have discussed off-line. At least on the Canon 100 mm f/2.8 lens that I have, the entrance pupil is far enough back that it blocks light rays getting through the objective from reaching the corners of a full-frame sensor.

Even without vignetting, there's the issue of field diameter. The Nikon CFI 10X NA 0.25 has a high quality image circle, on subject, of perhaps 6 mm. (Some would say smaller than that.) So, there's a big gap between the 43 mm diagonal subject space that is naturally covered by your 1:1 macro lens on a full frame sensor, and the 6 mm diagonal subject space that might be covered by the CFI 10X NA 0.25 with suitable relay lenses.

At this point I think it would be really helpful if you stepped back for a moment and gave more thought to what you really want to do. If you're interested in imaging a field diagonal of 6 mm or less, then microscope objectives are the way to go and we should talk about how to make those play nicely with your full-frame sensor. But if you're interested in imaging field diagonals that are significantly larger than that, then good solutions (in addition to the MP-E 65) include enlarging lenses on bellows, extension tubes behind your current macro lens, and closeup lenses like the Raynox.

As for the framing issue that we've been discussing, I think it's important to keep that in context. You seem to be thinking of it as "Good grief, I'm losing so much field!" But another way of looking at it is that you're just losing a bit of resolution over part of the background. Consider again how images are processed by Zerene Stacker with default options. As you've shot this stack, areas in the foreground are propagated to output at full resolution, very close to what was originally shot by the camera. Areas farther back are expanded before they propagate to output, but even at worst the expansion is less than a ratio of 4:3. So despite the dramatic appearance, what's actually happening is that you're losing at most 1/4 of your resolution over part of the subject. It's maybe not so scary expressed in those terms, and certainly it's far better than anything you could do without stacking, by just stopping down.

BTW, you've mentioned elsewhere about telecentric lenses. I'm not sure which Raynox you've ordered, or for that matter exactly which Canon 100 mm lens you have, but at least for the lenses I have, it's possible to make a nicely telecentric combo with Raynox DCR-250 as described HERE, for certain magnifications. Using a full frame sensor there would be some other restrictions, but it might be worth investigating. Note that in order to exploit this telecentricity, you would have to switch to focusing by adjusting subject-to-lens distance. Focusing by AF motor won't work so well because then you would still get the same sort of "focus breathing" that is causing the framing issues we've been discussing in this thread.

I hope this helps...

--Rik

Asha
Posts: 70
Joined: Tue Apr 09, 2013 8:08 pm

Post by Asha »

Rik,

I don't have any concrete magnification goals--I just want a tad more than I currently have because I haven't been getting the kind of edge-to-edge composition that I want. Thanks for the reminder to use the smallest limiting slice--it is true that is the full frame of my camera, so I should be able to get a good idea about composition that way.

The diopter I ordered is the Canon 250D. My lens is the 100mm/f2.8 mark II, and I was afraid that would push the Raynox to the limit. The front lens element is right around 43mm clear aperture (same as the Raynox), and the filter ring is 67mm on the inner thread (Raynox max is 67mm, don't know if that is inner or outer). It would be interesting to see whether I can get the telecentric behavior with this lower power diopter, assuming I also can get the magnification I need. I have a cheapie manual rail that I can try.

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

Asha wrote:The diopter I ordered is the Canon 250D. My lens is the 100mm/f2.8 mark II
..
It would be interesting to see whether I can get the telecentric behavior with this lower power diopter
Probably not, though it depends on details of the 100mm/f2.8 mark II that I don't have in hand.

As always, the key thing for going telecentric with an add-on lens is that the add-on has to be located at its own focal length in front of the main lens's entrance pupil. Unless the 100mm/f2.8 mark II has its entrance pupil located unusually far back in the lens, that won't be doable with a 250 mm focal length add-on.

In your case I recommend to not spend much time worrying about telecentricity. Telecentric optics are wonderful for a few special applications. In this community, the big one is stack-and-stitch at moderate magnification, say 5-20 mm field width per frame. For that, they make life much simpler. But other than that, their drawbacks in cost, complexity, and sometimes loss of image quality make them more trouble than they're worth.

--Rik

Post Reply Previous topicNext topic