telecentric setup attempt: advice please

[austrokiwi1]

This post is from a complete novice so you may see some really silly questions comments. Please be patient and couch replies for a complete ignoramus:

My focus on macro photography is coins I usually work in low magnification ranges 1:2 up to a maximum( with my skill and knowledge set) of 5:1. recently in a coin forum a post was made (by Ray) about a visit to GIGAmacro. I had a good look at the Gigamacro site and noticed an unusual piece of equipment a "Telecentric adapter". I spent a week researching what that was. Being brutally honest; I understand the overall principle but not the indepth optical principles. the key point I picked up is that a telecentric system makes stacking and stitching simpler. In my research I found this thread:

http://www.photomacrography.net/forum/v ... php?t=1032

I am still getting my head around the optical principles. But in reviewing that thread I realised I had every thing needed to try the system out. Heres my first set up front lens is an olympus zuiko 135mm F4.5 bellows lens, the rear lens a minolta 100mm F4 bellows lens( these worked best out of what I had available to me). I adjusted the system as was recommended in the above thread



Note: many of you will see the weird extension between the front minolta bellows mount and the olympus 135mm. That extension is a personally assembled Olympus Om- Minolta SR adapter.

I seemed to get the telekinetic effect with that set up. My first stack was of an ancient axumite coin( forgive the lighting it was a test):




I was pretty chuffed with this attempt...but it was all monkey-see-monkey-do. Researching further I understood that a telecentric system is limited by the diametre of the entrance optic. My interest is in reasonably large sized objects( ranging up to 50mm in diameter).

A quick digression: for about 12 months I experimented with long lenses so as to increase the lens to subject distance so as to simplify the lighting system. One spectacular fail( it was fun trying though) was adapting a Schneider Kreuznach large format enlarger lens to my bellows. I had a local camera technician ( cheaply) make an adapter that allows the use of a 52mm reversing ring to mount the lens: a 210 mm F5.6. The adapter is brilliant the problem is with my bellows at full extension the front optic to subject distance ( around 800mm) was impractical. So I shelved the SK210.

Back on subject. my first trial was Ok despite the lighting but for my purposes the system was limiting me to small sized coins. thats when I remembered the SK210, this is the resulting "tuned"( is that the right word) system( the rear lens is the Minolta 100mm F4) Note this was the first photo of the set up and I hadn't mounted the bellows and arca rail accurately( hence the angle):



My first attemp[t was with a 47.66mm medal: Heres that attempt. the SK was wide open and the minolta was between F11 and F16( mistake!!!)



It looked pretty good to me until I looked at a 100 % crop:



So looking at reducing the diffraction I ended up finding the best I could do was to have both lenses wide open. Heres a 200%( yes 200% crop) from a 25 shot stack of the same medal( I have never managed to obtain such resolution before):




Now my questions( forgive my ignorance):

I am confident with that I did get a telecentric effect with the Olympus 135mm and minolta 100mm. I am not so sure about the SK and the Minolta particularly with the minolta wide open. Ray suggested I use a calibration slide to test. what sort of calibration slide should I use( are they all the same?)? Can I mount it immediately behind the SK?


lets say I haven't got the telecentric effect with the SK-minolta set up... Have I wasted my time?.......I don't think I have as I already noted I have never seen such high resolution results before?

I believe the minolta mounted behind the SK has reduced the working distance nicely. is this on its own( with out the telecentric effect) a good option? ( I think it is). If the only advantage is reducing the working distance can I improve the performance? I suspect the minolta is a limiting factor.
Here a silver dollar sized medal I tried( its very high relief so its a 25 shot stack):





I would appreciate any comments and suggestions..just be gentle with me and forgive me if I seem slow to pick up on what your saying. I prefer to ask stupid questions than to think I have got it when I haven't. Camera is a Sony A7r.

[Peter De Smidt]

Just to fill some space until Rik has his coffee, for those larger sized coins, are you stacking? My guess is that you don't have to for them. If so, then for them telecentricity wouldn't be important. It's helpful for stacking, especially if the subject has a fair amount of height.

[austrokiwi1]

With th e SK at 5.6 and the minolta wide open I am getting a very narrow depth of filed on that last medal photo theres just no way I could get the DOF to cover the full depth of the design. I suspect the real F value of the two lens combination is very close to, if not just over, the A7rs diffraction limit( F10). I had tried a stack of 10 first and it just didn't get the detail. 25 probably might have been too many. However it could also be my technique and too little knowledge.

[Peter De Smidt]

I had in mind using only one of the lenses for the larger coins, stopping down if it helps. Generally, starting with the simplest solution to set a benchmark seems like a good idea. That way you can tell if the added complexities are worth the trouble.

[TheLostVertex]

It looked pretty good to me until I looked at a 100 % crop:



So looking at reducing the diffraction I ended up finding the best I could do was to have both lenses wide open. Heres a 200%( yes 200% crop) from a 25 shot stack of the same medal( I have never managed to obtain such resolution before):

The first image here you say is at 100% scale, and the second image you say is at 200%. They appear to line up, so that would mean between these photos you changed the magnification you shot them at. You repeated 200% twice, so I assume it wasn't a typo. Can you clarify this?

The lighting also appears to have changed between the two photos. This can have a large impact on the appearance of details, especially on metallic objects.

lets say I haven't got the telecentric effect with the SK-minolta set up... Have I wasted my time?.......I don't think I have as I already noted I have never seen such high resolution results before?

What was the objective of trying to make your set up telecentric? Did you have practical purpose in mind, or is it just as an experiment? I do not think telecentric set ups would provide many benefits for stitching together coin images, since the subject is pretty flat.

[rjlittlefield]

Peter, many thanks for your thoughtfulness and for getting the discussion started. Both my stomach and my brain appreciate your time.

austrokiwi, your post involves several different phenomena that are coming together in a way I have not noticed before. I'll try to sort things out for us, but as I start to write this post, I don't know how well it will come together! Ah well, onward...

For starters, let me provide some more background and references.

The thread that you mention was only the first of three closely related threads that I wrote when I began to seriously study telecentric optics.

The full set of three are these:

• (Nov 19, 2006) "Stack & stitch experiment with telecentric optics" -- demonstrations using a) two lenses, and b) one lens plus an added aperture
• (Jan 01, 2007) "Stack-and-stitch: ordinary optics vs telecentric" -- further explication of the theory
• (Jan 07, 2007) "Telecentric optics, third round" -- emphasizing the idea of adapting an existing lens by adding another lens in front of it.

The folks at GIGAmacro learned how to make their "telecentric adapter" from the third of those threads.

More recent threads that are highly relevant here are:

• (Jun 21-22, 2011) Discussion with forum member danibo -- regarding his home-brew telecentric combo
• (Oct 03, 2012) "Telecentric combo at 0.8X to 1.69X" -- using Raynox 250 in front of Canon Macro Lens EF 100mm f/2.8L IS USM.
• (Feb 07, 2015) How to tell if your optics are telecentric -- by computing relative scale on both sides of focus (second half of that post).

Now, regarding why you might want to use telecentric optics in the first place...

In the world at large, telecentric optics are most commonly used for measuring and machine vision applications, because they separate magnification from focus and avoid perspective distortions.

In this forum, the most common compelling case is to enable parallax-free stack-and-stitch while using lateral shift instead of rotation around an ordinary entrance pupil.

There are other applications, such as uniformly illuminating a flat mirror (Theodore M. Clarke, "Brightfield Illumination of Large Field Sizes"). But these are much less common.

Note that telecentric optics are not required just to get good results with a single stack. In unusual cases telecentric optics can be helpful with a single stack, but almost always it is simpler and equally or more effective to use some other approach that keeps an ordinary entrance pupil in fixed position or close to it.

So now, to the questions that you've asked...

1. To tell whether your system is truly telecentric, see the procedure linked above (HERE).

2. Have you wasted your time? Absolutely not! But perhaps you've misinterpreted how you spent it well.

The improved resolution and absence of speckle shown in your 200% crop has nothing to do with telecentricity.

Instead, that's due to opening up your lens and making up the DOF by stacking.

A closely related aspect, which I think you have not touched on but need to, is that speckle is intimately related to illumination. The smaller and more directional your light source is, the more speckle you'll get, even with identical optics. For illustration and discussion, see "False color and detail/DOF from shiny metal". As Steven mentions, it looks like your two detail crops were done with different illumination.

So, to get best results from your coins, what I recommend is the following progression:

1. Mount your best single lens (not telecentric).
2. Start with very diffuse illumination. Use a light tent.
3. Run a series of test shots to determine what aperture setting gives the sharpest result with that lens. Probably this will be pretty wide -- only a stop or two down from full open.
4. Determine the step size needed for clean stacking.
5. Incrementally make your illumination more directional (less diffuse), so as to get the lighting effect you want but without introducing too much speckle.

I think you'll be very pleasantly surprised by how much resolution you can get from this approach.

However, if it's still not enough -- say you want that medal at 200 megapixels -- then it's time to start stitching, and for that you may want the telecentric optics also.

I hope this helps!

--Rik

[austrokiwi1]

Thanks alot for the comments I will take a day or so to digest them and I will follow up on the suggestions.....Yes I did use two different lighting sources. I had forgotten that! The first time was with Direct lighting and then the 200 % crop was from coaxial lighting ( an old edmunds 50/50 plate beam splitter; Titanium dioxide and no anti reflective coating)

Thanks again...... I have a lot of reading to do and a lot of practice after that.

Edit: I would just add the two crops are from very different photos I re checked the first is a 100 % crop of Direct lighting the second is a 200 % crop using axial. Just to double check 100% crop I zoom in on the image to a 100% then take the crop, and for the 200% I take the magnification to 200%.