Share your Coin Photography Tips & Tricks!

[dougsmit]

I am not suggesting anyone run out and do what I just did buying a new lens for, among other things, coin photos.  So far I have only shot a few 'old favorites' but am happy with results from my Laowa 85mm f/5.6 macro.  There is no auto-focus, there are no contacts to transfer EXIF data ---- just sharp glass in a solidly build and very small package. I am most impressed with what it did for my 4.5mm, 0.1g silver from Phokaea.  

Next is another old favorite Athenian 10x7mm obol from the early period when the tail was shown with three feathers.  These are not common but available as tetradrachms but this is an obol not seen every day.  To show the tail, I blew up that part more than a little bit.  These images were focus stacks assembled from about 20 input images. 

It seemed good to try a few more normally sized coins so out came a Corinthian stater, denarius of Septimius Severus and very darkly toned AE2 of Magnentius.  I was satisfied.  These three were single shots of each side combined as I always do into one file before reducing to 2048 pixels wide (from the original 9000x6000).  All of these were shot with a mixture of ring and directional lights adjusted in balance as seemed right at the moment.  Compared to my earlier cameras and lenses, these images show more texture details and every little scratch.  These are not images for people wanting to fool people on eBay that the coins are 'proof'.  

 

I bought this lens mostly for chasing bugs in the yard.  Shooting coins is a bonus for me.  Now I have to decide what coins to reshoot using the Canon R7 and Laowa 85mm and which to say they were good enough shot over the last twenty years.  I doubt I will be using my old macro rigs much anymore. 

 

 

[Valentinian]

Most helpful hints in this thread are far beyond what I do. I am only going to remind us of the importance of the angle of lighting.

In 2018 I bought this coin from this photo:

When I got it I was disappointed because it did not look so clear and sharp in regular light. When I photographed it at the time (with light from overhead or at most 45 degrees away from overhead) the best I could get (admittedly, with merely an old iPad) was this:

That's a lot like it actually looked/looks like. However, as you can see, that's a pretty bad photo if the coin really looks like the previous photo.  However, it didn't--until yesterday.

Yesterday afternoon just before the sun was going down I thought "It's sunny! I'll look at some coins in the light of the sun!" I did and noticed that in extreme raking light that coin looked really great. I guess I hadn't thought of using raking light from a low angle. So, this morning, the sun was out and 20-25 degrees above the horizon. I got out that same old iPad and took this photo.

If had used the photography trick that eliminates shadows, the photo would look much like the seller's photo, and it does look just like the coin with lighting at that angle. I had not thought of viewing it with strong light at that angle.  So, the coin is as nice as the seller's photo, if you look at it right (which I hadn't until yesterday).

Learn the lesson:  Pay attention to the angle of lighting!

[dougsmit]

8 hours ago, Valentinian said:

If had used the photography trick that eliminates shadows, the photo would look much like the seller's photo, and it does look just like the coin with lighting at that angle.

No.  The seller's photo differs by having the light come from directly above the head at 12 o'clock while your new photo is lit from behind the head about 10 o'clock.  Your first attempt has the light closer to 1 o'clock but is also overexposed washing out detail.  A slight wiggle of a coin can make quite a change in the 'look' of the coin but there is no photography 'trick' here.  It is not the job of the photographer to take a photo to make the coin look bad but I would call the seller's image rather poor with too much contrast and poor focus.  I have never taken a photo of a coin with an iPad and have no idea what to suggest to improve your image.  I might try something closer to the angle of the seller photo but with the tonality and sharpness of yours.  When you are bored, reshoot the coin again and again paying attention to what differences you get rotating and changing the elevation of the light. Practicing with this coin will make it easier to decide what to do the next time you have a similar coin. 

[Kaleun96]

Haven't shared anything photography related in awhile but have been busy with some small improvements to my flash adapter. Might post about those improvements soon if I get the time.

One thing I wanted to quickly mention, and that I think I started this thread with, is about the mirror-as-a-background technique. It's still serving me well, albeit with its limitations (if you want to tilt the coin, you must tilt it independently of the mirror), but I noticed one thing that's worth mentioning.

So I originally said that the mirror will give you a perfect white background that separates the coin from the background. Well, that is not entirely true. It turns out the mirror is just reflecting the light back from my custom flash adapter (3D printed in white plastic) that is mounted around the front of my lens, and it is this white light reflected off the flash adapter that gives the background a white appearance. So for anyone else not using a ring light, having a mirror as a background will probably result in a very dark background instead - though, still with good separation between coin and background for easy removal.

Since my coin tilting device tilts both the coin and the background, I've had to forgo tilting my coins for photography when using the mirror background. This is a bit of a pain so I've again been looking at alternatives. Earlier in this thread I went through some tests where I tried illuminating the background independently, using very white backgrounds, or even using convex/dome shapes to help reduce any shadows and create more separation.

I had a random idea to try retroreflective tape since the idea of a retroreflector is to direct light back to exactly where it came from. This might avoid the issue of white backgrounds creating a "halo" around the edges of the coin from stray/scattered light being reflected off the background at an angle. It also might help make the background brighter by maximising how much light is returned directly back towards the lens. Unfortunately, I didn't see much improvement with retroreflective tape over a normal white background, and even using a clear retroreflector disk for a bike didn't work that well either. The hunt continues.

[leeshiel]

Great thread good information setups coins and pictures,

I keep things quite simple i use a Olympus Tough TG-6 with a light diffuser attachment, just a quick shot freehand i could improve on this with a tripod and using photo stacking but pictures like this are good enough for my needs.

 

[Kaleun96]

I'm back after doing a lot of testing the last few weeks around optimising the background to make it easy for later removal in Photoshop. I started this thread with using a mirror, and more recently updated with that while it works well, it prevents me from tilting the coin. It also introduces a weird chromatic artefact around the edge of the coin, only a few pixels wide, but that was still a bit annoying to remove in some cases. I'll mention that more later.

I might break this into a couple of posts, let's see how I go. So I sought the help of the folks on photomacrography.net for help with the following issue:

• I want a white background that is easy to remove
• The background needs to have near-perfect contrast with the coin, that is the background cannot bleed into the edges of the coin
• The background needs to be reasonably bright and consistent enough that coins with very light or dark edges pose no problems
• I want to do everything in-camera, that is I don't want to take two sets of photos such as a silhouette mask of the coin and an actual photo of the coin.

The main recommendations were to try and ensure the light rays from the background were parallel to one another and perpendicular to the lens. In other words, whether you're using a light source on the background itself or just reflecting light off of the background, the light should be going past the edges of the coin parallel to one another and perpendicular to the coin (and thus the lens).

I've made a quick diagram below as to why this is important. In the example on the left, an LED is used to provide a bright white background illumination, but the uncollimated light of the LED spreads out across a wide angle (probably ~120 degrees or so). The issue with this is that light is hitting the edges of the coin at a low angle, and while the coin is a solid object, this light reflects off the edge and makes it softer, the edge is then a mix of bright white light and actual coin - this is what I call "light bleed" on the edge.

In the second example, a collimating lens is used to collect the angled light rays and convert them to parallel rays. These light rays pass the coin parallel to the edge and thus are not (theoretically) reflecting off the edge of the coin. They form a perfect silhouette outlining the exact edge of the coin.

So armed with this knowledge, I decided to tackle the problem as one of an optical comparator. An optical comparator is a device used for inspecting the profile of an object for measuring manufacturing tolerances and the like. It's also called a profile projector. Essentially it is used to produce a perfect profile or silhouette of an object, the same problem I'm trying to solve!

My tests then focussed on trying to perfect the silhouette of a coin and worrying about the rest later on - if the silhouette was not perfect, the rest did not matter.

Here's such a silhouette of a coin, showing virtually no light bleed around the edges.

A coin is perhaps not the best object to test the efficacy of an optical comparator setup because the edges vary from one coin to the next and a coin blocks most of the centre of the image - if there is stray light where there shouldn't be, it may not be visible when using a coin of a certain size.

So I changed to using a threaded bolt as a test subject. This is because a threaded bolt has many angle surfaces, can extend the length of the frame, but is not too wide. To get the parallel light rays, I was using a fresnel lens, which is a very smart but simple lens design to collimate light from a single point into parallel rays.

Here is a photo showing the profile of a threaded bolt hovering a few cm above a fresnel lens. What you can notice is that the silhouette is very good at the top and bottom of the thread, a near-perfect black outline. However, near the center of the bolt, you can see a bright diffused area and light hitting the edge of the thread profile. This indicates non-parallel light rays are hitting the bolt and bleeding into the profile of the thread. This is exactly the sort of problem I need to avoid.

Looking at a zoomed-in image of the middle and top of the thread, you can actually see a small amount of light bleed on the right side of the thread profile near the top but this is mostly manageable, much more so compare to the bottom part of the image that shows the bright spot in the center of the thread.

 

So some things I deduced from this test:

• A fresnel lens works pretty well for generating parallel rays, but it may not be perfect, especially...
• when using a very bright single light source like an LED COB, the majority of the light is in the center of the lens, creating a very bright spot and perhaps bleeding light out at other angles
• this also means the areas outside the center of the lens are less bright, the background is not of a consistent brightness and the sectional rings of the fresnel lens are quite visible

But the biggest problem is that using a background light system that relies on parallel light rays has the exact same issue as with the mirror: if you tilt the background with the coin, relative to the lens, the solution does not work as the majority of light is now directed away from the lens. This is not a problem if you only tilt the coin and not the background but my current setup does not allow for that.

Another problem it shares with the mirror method is a weird chromatic artefact that forms of the edges of the coin in the final stacked image, shown as this weird green fuzzy area in the image below. The second image below shows where this comes from by looking at the silhouette of the coin. As the parallel light goes around the edge of the coin, it forms a bright out-of-focus ring that the focus stacking software seems to detect as part of the edge. This seems to be because as the camera moves relative to the subject, this fuzzy ring shrinks or grows and the software identifies it as detail that needs to be kept. The author of the focus stacking software has commented on this issue I faced here.

 

So in the end this parallel light method was not working out for me. It was too bright in some areas and not bright enough in others. It also had the same issue with tilting that I faced with the mirror, as well as the weird chromatic artefact formed around the edge of the coin.

The next method to try was a more simple diffused light background. This is similar to the method on the left in the first diagram I shared above, so how can it be a solution when I've already dismissed that approach as less than ideal? Well, it seems if you can get enough distance between the diffused background light and the coin, much less diffused light is bleeding into the edge of the coin (inverse square law) and may be a negligible effect if the image is exposed correctly.

I'll get to those tests in my next post and end things here with some photos of the parallel light setup I had tried:

This one shows an early test of the setup. The fresnel lens is the clear plastic bit in the middle with the three "legs". The coin platform sits above it. The LED light is actually inside the black rotation stage (the thing with degree markings on the outside). The rotation stage has a 25mm diameter hole in the center, which was perfect for placing a small LED COB and allowing me to keep the height of the total setup fairly low.

Eventually I 3D printed a shroud for the fresnel lens to help ensure no stray reflections were messing things up. You can see it as the black plastic object below the coin platform from the photo above.

And here's the test with the threaded bolt sitting above the fresnel lens light background.

Edited December 24, 2022 by Kaleun96

[expat]

99% of that is way beyond my capabilities, but absolutely fascinating. I have a new model smartphone which I currently use, as my Sony DSLR broke. What I find great is that the details you, and others, provide, can be used to modify somewhat my current setup with, for me, dramatic improvements in the quality of images I can achieve. Albeit still nowhere near the quality you guys are producing. I look forward to more and thankyou for taking the time to create these informative threads for a neophite.

Edited December 24, 2022 by expat

[Kaleun96]

Following on from the discussion here:

My axial lighting test rig finished printing (18 hours!) so I ran a few tests today. In general, I quite like the design of the rig as it leaves me a lot of flexibility in being able to use my existing coin platform with rotation and tilt stages and backlight illumination. It's also relatively small, simple, and robust.

However, after flocking the inside of the rig, I'm still getting a lot of non-axial light illuminating the coin from the flash. I can test this by removing the lens filter, which is acting as the angled glass between light and lens, and when taking a photo I found I still get a pretty well-exposed photo of the coin. This doesn't seem to be ambient light, only light coming from the flash. It's possible it's still managing to reflect enough light off of the inside of the housing, even with the felt flocking.

If so, that is fairly problematic. One solution is to make the housing larger and use a bigger piece of glass, then less light can reflect off the 45 degree supports that holds the glass in place and down into where the coin is. But that seems like an imperfect solution to me. It's also possible the light is reflecting off the flocked "chamber" the coin is inside but I think if this were the case it would be more obvious in the direction of the shadows.

edit: I realised I forgot to flock the inner part of the "wall" that the flash attaches to. Hopefully I see an improvement after doing that.

Here's a test stack, it turned out mostly OK but it is clear there's some non-axial light leakage and the sharpness of the photo is absolutely trash. I know this is partly caused by the lens filter, as when using some beamsplitter glass the sharpness increased substantially (though still noticeably below what I get from my pseudo-axial setup).

Showing the lack of sharpness:

 

The setup:

Edited December 29, 2022 by Kaleun96

[Kaleun96]

So turns out I had much more non-axial light leakage than I expected! I was wondering why it didn't seem to take much flash power to get a properly exposed image as usually axial lighting requires A LOT of light power.

After flocking the inner wall where the flash mounts to the housing things improved significantly but I was still getting some leakage. In the end I needed a bit of a "fence" between the flash and the lens filter, only about 1cm high, at the edge of the circle that allows light to go down to the coin. I had figured having the coin about 1.5cm below this opening would be enough to stop light from the flash making its way down there non-axially but apparently not. After putting up the "fence", leakage of non-axial light was practically zero.

 

Where I started off after my post above - this is way too much leakage of non-axial light, the picture should be black.

 

After flocking the inner wall where the flash mounts, things improved significantly but still some leakage hitting the bottom of the coin.

 

And after adding the small "fence" to block direct light from the flash making it down to the coin chamber below the housing I achieved the desired result.

 

And here's a non-edited photo (not a full stack like in my previous post) showing the familiar harshness of the axial lighting effect. You do get the surfaces and toning to "pop" like no tomorrow but in addition to the harsh contrast it also draws attention to surface defects (e.g. between the slinger's legs). This is typically why people will change the setup slightly to introduce some non-axial light or otherwise diffuse the main light to lessen the effect.

If I recall correctly, this was at 1/32 on my flash (normally I'm at ~1/128), 80 ISO, 250 shutter speed, and f5.6 at 1x magnification...so yeah it requires a fair bit of light!

 

Unfortunately the sharpness is still trash so I might once again embark on trying to find the perfect glass for axial illumination. @HipShot Photography I can't recall if you've mentioned before but what do you use as the glass/beamsplitter for your setup?

Edited December 29, 2022 by Kaleun96

[kirispupis]

@Kaleun96you've probably tried this already, but I've been using a Lee filter. Keep in mind that I didn't buy it for this purpose, but I needed something and happened to have a bunch of Lee GND filters. I hadn't used them in years because improvements in HDR has made them mostly unnecessary, but I couldn't bear to sell them for next to nothing. 🙂

The major issues I've run into with it is a) ensuring that only the clear part of the filter reflects light into the coin and b) ensuring my lens and the filter are on the same plate (since I shoot at an angle for the reflection).

However, these filters are designed to be shot through with minimal loss of details. I imagine a Lee UV filter would be the most ideal, since there wouldn't be the graduation to deal with.

I took a quick look and didn't see any square/rectangular UV filters from Lee. However, Singh-Ray does make them (though they're pricy like usual) - https://singh-ray.com/shop/hi-lux-protective-warming-uv-filter/

I would think someone makes a cheaper option, though I didn't see one on 2filter.com.

[HipShot Photography]

3 hours ago, Kaleun96 said:

@HipShot Photography I can't recall if you've mentioned before but what do you use as the glass/beam splitter for your setup?

You might be disappointed to learn I use a simple pane of glass from a quality 8x10 picture frame. I built a custom holder for it. It works very well and accommodates coins and exonumia up to 4" in diameter. I position my light source 14" away from the coin, and it has a 40-degree honeycomb grid to focus the light on the glass pane. I also have a few other little tools for controlling the light depending on the coin.

[Kaleun96]

2 hours ago, kirispupis said:

@Kaleun96you've probably tried this already, but I've been using a Lee filter. Keep in mind that I didn't buy it for this purpose, but I needed something and happened to have a bunch of Lee GND filters. I hadn't used them in years because improvements in HDR has made them mostly unnecessary, but I couldn't bear to sell them for next to nothing. 🙂

The major issues I've run into with it is a) ensuring that only the clear part of the filter reflects light into the coin and b) ensuring my lens and the filter are on the same plate (since I shoot at an angle for the reflection).

However, these filters are designed to be shot through with minimal loss of details. I imagine a Lee UV filter would be the most ideal, since there wouldn't be the graduation to deal with.

I took a quick look and didn't see any square/rectangular UV filters from Lee. However, Singh-Ray does make them (though they're pricy like usual) - https://singh-ray.com/shop/hi-lux-protective-warming-uv-filter/

I would think someone makes a cheaper option, though I didn't see one on 2filter.com.

Interesting, was it a GND filter that you used? Ideally they'd have something quite neutral with not much in the way of light loss like a UV filter as you mention but I'm also not finding anything like that. Perhaps something like the low contrast filter would even help with the harshness of axial lighting. I guess the B+W UV filter I'm currently using may be similar in terms of how a Lee UV filter would perform if they had them, but it could be that all the coatings of the B+W filter are working against me here, e.g. I think both sides have AR coating when I really only want one side to be AR coated.

Would be more useful in a rectangular form like the Singh-Ray ones you found but given the massive drop in sharpness I experienced with the B+W is putting me off trying something similar. I have previously done tests where I've found the same piece of glass shot through when laid flat to the lens has no noticeable effect on sharpness compared to at 45 degrees and with axial illumination. Whether it's the mere fact of the glass being at 45 degrees that causes the biggest drop-off in sharpness, or the axial illumination itself, I'm not sure.

30 minutes ago, HipShot Photography said:

You might be disappointed to learn I use a simple pane of glass from a quality 8x10 picture frame. I built a custom holder for it. It works very well and accommodates coins and exonumia up to 4" in diameter. I position my light source 14" away from the coin, and it has a 40-degree honeycomb grid to focus the light on the glass pane. I also have a few other little tools for controlling the light depending on the coin.

Thanks for the info! Only a little disappointed in the sense that there doesn't appear to be a "right" answer to the glass question, just more trial and error. I have some picture frame glass that I used for my first tests a few years ago but I must look for something smaller - I'm enjoying this new small setup I've been testing today.

Ideally I'd also have my light a bit further away, currently I'm using a 50mm fresnel lens placed about 20mm from the flash to help collimate the light into parallel rays but it's only another 60mm from the fresnel lens to the glass. I have noticed that using the fresnel lens does seem to help somewhat though. 

[HipShot Photography]

On 6/13/2022 at 11:32 AM, dougsmit said:

I believe 2-3 hours once a week is plenty to devote to learning what you need to know about photography if you do it 'right'. 

I realize this is just your opinion, but in my opinion, this is an underestimate, though I suppose that depends on what the expectations are. Photography is a lot like music. You may learn the notes and scales in 2-3 hours a week, though they will hardly become ingrained. Regardless of what style you play, be it jazz, rock, or classical, you can't play it properly, or even adequately with 2-3 hours of practice a week. This will only lead to the frustration I see in so many of the photography-related posts I read here. There's a saying in photography; your first 10,000 pictures are just practice.  Again, I guess it depends on what the expectations are, but if you want great coin pictures, just like assembling a great coin collection, you have to put in the work and learn the craft. Otherwise, you will always get out of it what you put into it. 

[HipShot Photography]

28 minutes ago, Kaleun96 said:

I'm enjoying this new small setup I've been testing today.

You may enjoy it but I believe you're shooting yourself in the foot with its light control limitations.  It doesn't allow you to feather the light across the coin or control the density of diffusion. A Fresnel lens is hard light when, IMO, coins want soft light with highlights. Of course, my photographic goals are not scientific, but esthetic. I want to create an inviting coin image; one you feel like you can almost reach out and touch. I use light control to achieve this. Perhaps this is not a shared goal. So if not, please carry on. 

[Kaleun96]

12 minutes ago, HipShot Photography said:

You may enjoy it but I believe you're shooting yourself in the foot with its light control limitations.  It doesn't allow you to feather the light across the coin or control the density of diffusion. A Fresnel lens is hard light when, IMO, coins want soft light with highlights. Of course, my photographic goals are not scientific, but esthetic. I want to create an inviting coin image; one you feel like you can almost reach out and touch. I use light control to achieve this. Perhaps this is not a shared goal. So if not, please carry on. 

I don't exactly hide my dislike of the raw harshness of pure axial lighting and I largely shoot pseudo-axially with custom flash heads and various reflectors, modifiers, and diffusers (you can see some custom ones stuck to the whiteboard and on the table in this photo) to get more control over the light for the aesthetic I like - so yeah, it's a shared goal if that wasn't clear from my previous posts.

This setup is just for testing. What I've hated about other setups I've tried for testing axial lighting is the size (particularly when using something like a pane of glass from a photo frame) and the difficulty in controlling all the variables. With this little rig, I can hone in on the problems and iterate from there to fix them and also to allow for more aesthetic control. This is what I love about 3D printing - I can go from prototype to prototype, testing small variations and changes, and end up with a more refined finished product that is built to my needs. But big setups take longer to print and of course use more material, so best start small and go bigger later once I've addressed the major problems and limitations.

But the more important hurdle at the moment is the sharpness. I'm at least not getting any noticeable ghosting but the sharpness is a concern. The beam-splitter glass performed better than the filter but still not good enough for me to consider using this setup as an alternative to my pseudo-axial setup. Tomorrow I will try some other UV filters I have and also see if I can track down a small photo frame with a glass cover (2.5 x 3.5" would do).

While I've heard opinions to the contrary on various forums, I've still actually yet to see a test demonstrate that axial illumination doesn't result in a noticeable drop-off in sharpness, so perhaps I'm trying to fix a problem that can't be solved. If a 1.7mm thick optical-grade 50/50 beam-splitter glass with AR coating on one side doesn't give me the results I need, I'm not too hopeful about other options.

[kirispupis]

1 hour ago, Kaleun96 said:

Interesting, was it a GND filter that you used? Ideally they'd have something quite neutral with not much in the way of light loss like a UV filter as you mention but I'm also not finding anything like that. Perhaps something like the low contrast filter would even help with the harshness of axial lighting. I guess the B+W UV filter I'm currently using may be similar in terms of how a Lee UV filter would perform if they had them, but it could be that all the coatings of the B+W filter are working against me here, e.g. I think both sides have AR coating when I really only want one side to be AR coated.

Yes, it was a Singh-Ray GND that I happened to have around. It's interesting that you're running into issues with the UV filter, since I would expect it to be a better solution. I do think that rectangular filters are easier to deal with here than circular ones, though.

I must admit that I haven't gone into depth on sharpness tests with my setup so far, mainly because I haven't reached a setup I'm happy with yet. From my perceptions, loss of sharpness may be an issue, though it may vary with the angle of the lens against the filter. Clearly these filters are made for shooting through, but as you pointed out that's not what's happening here. It's well-known that one should never shoot through glass at an angle, so I would imagine that you'd see some degradation regardless of the thickness.

For my purposes, though, the loss in sharpness seems manageable. As you know, we have different goals. In my case, I just want an artistic shot that shows off the beauty and details of the coin on a black background. I'm not looking to allow someone to zoom in on details. Therefore, I just need the shot sufficiently sharp for the page.

There are two possibilities that may work in terms of recapturing some details, though of course neither technique is a replacement for pure optical sharpness.

1. Combine images at the same focus. This is common in astrophotography, but should work here too. I used the technique in this shot.
2. Topaz Sharpen. It's incredible what AI can do now. I use some form of it on most of my photos now.

As for me, I'm done experimenting for the next week. Going forward, I'll need to plan out a stage, though due to budget that may take some time. In my next setup, I'll probably try a more permanent setup in the garage with my strobes. I like @HipShot Photography's recommendation for a honeycomb to limit the beam.

Lastly, while there's nothing wrong with 3D printing, I've found a lot of flexibility in Legos. They've enabled me to try something out and make small adjustments as necessary.

[Kaleun96]

15 hours ago, kirispupis said:

Yes, it was a Singh-Ray GND that I happened to have around. It's interesting that you're running into issues with the UV filter, since I would expect it to be a better solution. I do think that rectangular filters are easier to deal with here than circular ones, though.

I must admit that I haven't gone into depth on sharpness tests with my setup so far, mainly because I haven't reached a setup I'm happy with yet. From my perceptions, loss of sharpness may be an issue, though it may vary with the angle of the lens against the filter. Clearly these filters are made for shooting through, but as you pointed out that's not what's happening here. It's well-known that one should never shoot through glass at an angle, so I would imagine that you'd see some degradation regardless of the thickness.

For my purposes, though, the loss in sharpness seems manageable. As you know, we have different goals. In my case, I just want an artistic shot that shows off the beauty and details of the coin on a black background. I'm not looking to allow someone to zoom in on details. Therefore, I just need the shot sufficiently sharp for the page.

There are two possibilities that may work in terms of recapturing some details, though of course neither technique is a replacement for pure optical sharpness.

1. Combine images at the same focus. This is common in astrophotography, but should work here too. I used the technique in this shot.
2. Topaz Sharpen. It's incredible what AI can do now. I use some form of it on most of my photos now.

As for me, I'm done experimenting for the next week. Going forward, I'll need to plan out a stage, though due to budget that may take some time. In my next setup, I'll probably try a more permanent setup in the garage with my strobes. I like @HipShot Photography's recommendation for a honeycomb to limit the beam.

Lastly, while there's nothing wrong with 3D printing, I've found a lot of flexibility in Legos. They've enabled me to try something out and make small adjustments as necessary.

I found a great thread on photomacrography.net from 10 years ago where the sharpness issue was brought up: https://www.photomacrography.net/forum/viewtopic.php?f=8&t=15820

According to rjlittlefield (Rik, creator of Zerene Stacker), the thickness of the glass won't effect how likely you are to see ghosting but it will effect the amount of spherical aberration you see, resulting in sharpness loss. So thinner seems better in this regard. Angling the glass is also worse as it increases the "thickness", so as you say the loss in sharpness would vary with the angle. But it's surprising it changes so much based on the angle, given that adding a filter usually has a negligible affect on image quality.

The exception is if the glass is placed in the "infinite section of the optical train", where thickness would then have no effect on aberrations. The simplest way of doing this would be to place it between an infinity-corrected objective and the tube lens, but I don't want to buy a 1x infinity corrected microscope objective.

Could also maybe be done with a coupled/stacked lens setup but you'd need to find a good pair of lenses where the front lens works well focussed at infinity, as IIRC usually only the rear lens is focussed at infinity in such a setup. Another option would be putting it between lens and camera but that's not fun either.

But going back to the lens filter. Let's say a typical UV filter is 1mm thick, is it then right to assume that as long as you have a piece of glass of similar quality that when angled at 45 degrees is <=1mm thick, that the loss in sharpness would be similar to placing a filter on your lens? That would equate to a piece of glass nominally 0.7mm thick or so and this filter would supposedly fit the bill: https://www.jjc.cc/index/goods/detail.html?id=158

"this filter features an incredibly slim design that the glass measures just 0.03" (0.7mm) thick"

Something like this Thorlabs beamsplitter would probably cost the same but is 1mm thick, so would be 1.41mm at 45 degrees: https://www.thorlabs.com/thorproduct.cfm?partnumber=EBS2

I did get another idea from photomacrography.net and that was to use a pellicle mirror. Surprisingly they're not insanely expensive given they're 2 microns thick but (1) they're still expensive and (2) extremely fragile: https://www.thorlabs.com/thorproduct.cfm?partnumber=BP245B1

I would love to try one out though, surely 2 microns can't result in too many aberrations. Maybe one could be salvaged from an old DSLR for cheap.

 

Btw cool astrophotography! I'd like to think that's what I'd be doing if not for macro. I've also seen some macro setups with a surprising amount of legos too. Before I was 3D printing I was also using this stuff a lot with reasonable success: https://www.amazon.com/Polly-Plastics-Moldable-Pellets-Resealable/dp/B01C559GC0

 

[Kaleun96]

The JJC filter with a supposed 0.7mm thickness was only about $20 so have ordered one to test. I'm not hoping for miracles though, I can't imagine my B+W filter is hugely thicker. A quick measure with my digital callipers suggests it's about 1.5mm thick.

Though, I know that the beamsplitter glass I mentioned earlier, which improved sharpness considerably, is about 1.2mm thick, so perhaps going down to 0.7mm will make a big difference after all.

[Kaleun96]

I ran some quick tests today changing two variables. My Laowa 100mm 2x comes with a lens filter pre-installed because the elements move back and forth inside an exposed tube and since the tube is greased, it would attract a lot of dust if a lens filter were not used. So I was curious if this filter was degrading the image IQ much. The second variable was the filter I was using as the "mirror" in the axial setup, I had a 49mm filter with 1.2mm thick glass and compared this to the 1.5mm of the 67mm B+W filter.

Photos:

1. 1.2mm thick "mirror" without lens filter on the Laowa

 

2. 1.2mm thick "mirror" with lens filter on the Laowa

 

3. 1.5mm thick "mirror" without lens filter on the Laowa

 

4. 1.5mm thick "mirror" with lens filter on the Laowa

 

Some quick conclusions:

• The Laowa filter only seems to reflect some additional light, brightening parts of the coin and reducing contrast. Effect on sharpness is minimal.
• Huge difference in sharpness between the 1.2mm and 1.5mm thick filters ("mirrors"). But still, the sharpness from the 1.2mm filter is still too low for me to consider using as an alternative method to my current setup.

Note I wasn't aiming for any aesthetics in the test, the photos are quite horrible and harsh. I just needed a simple test to check the sharpness and nothing more. Photos are unedited aside from a slight colour and exposure correction. These are also stacks of multiple images, the DOF from a single shot is much lower than as seen here.

Edited January 4, 2023 by Kaleun96

[Kaleun96]

The 0.7mm UV lens filter came in and I did a quick test. It definitely seems sharper than both the other lens filters I tried. The downside is that the anti-reflective coating on both sides is quite good and relatively little light is reflected onto the coin.

Here's a comparison between the 0.7mm filter and the 1.2mm filter from above.

0.7mm filter

1.2mm filter

[Kaleun96]

To continue documenting my findings (at the very least for myself in 12 months time when I inevitably revisit axial lighting for no apparent good reason), here's some quick comparisons of different lighting setups I have used vs. axial lighting. These are all unedited photos, straight from the camera, and I didn't bother to dial in the white balance or even get focus perfect, but nonetheless they're illustrative.

"Pure" Axial Lighting

"Pure" Axial Lighting w/ Diffuser on Light Source

Pseudo-Axial Lighting (my normal setup)

Pseudo-Axial Lighting w/ LED Ring

 

The "pure" axial photo is my least favourite, no surprises there, due to the harshness, contrast, and "busyness". The LED ring does pretty well, this is a Laowa LED Ring that I used to use for all my photos about 2 years ago and it was a solid performer but worked better on some coins than others. It picks up more of the luster and toning than the current Pseudo Axial adapter I used (with camera flashes instead of LED) due to the diameter of the LED ring.

The axial photo with diffuser is not too bad, somewhat close to the LED ring photo but with more toning and highlights from the luster. It's still a little too "busy" for me but I think it shows potential given all I did was stick a diffuser in front of the light - there's more I can do to modify the light to get a better photo.

But there's also the question of whether messing about more with axial lighting is worth it, given the pseudo-axial setups I have do basically what I need to do and it seems I'd just be trying to make the axial setup more similar to the pseudo-axial photos. The axial setup is also more complicated and finnicky and noticeably reduces sharpness. Nonetheless, I'm going to continue playing around with it to see how good I can get it - it'll be nice to have it as an available method when needed for the right coin.

And on that note, what I've realised after my tests with the 0.7mm filter is a few things:

1. Thinner glass is better

2. Filters are not ideal due to often being coated on both sides with anti-reflective coating. Very little light is reflected onto the coin.

3. Filters also seem to favour some wavelengths over others in their light transmission, meaning you often get a coloured hue unbalancing the photo.

4. A 67mm filter is barely big enough for large coins. The circle of light it reflects downwards is equal in diameter to the side opposite the hypotenuse of a 45-45-90 triangle, so about 47mm diameter. In reality, you probably aren't able to use that full diameter for various reasons, meaning the actual useable diameter of the "circle of light" is probably closer to 40mm, not leaving much room for coins 35mm in diameter or so.

So I've picked up an 80mm x 80mm 50/50 beamsplitter that is 1.1mm thick. It has anti-reflective coating on one side (the side facing the camera) and reflective coating on the other (the side facing the coin). The thickness is not ideal but it's difficult to find thinner beamsplitter plates that don't cost a fortune. I think in the end if this still produces unsatisfactory spherical aberrations that decrease sharpness, the only alternatives are a ultra ultra thin pellicle mirror (2-5um thick!) or an infinity setup with the mirror between the infinity-corrected lens and tube lens.

The former, buying a pellicle mirror, is expensive (~250 euro), and the latter is impractical. I don't think there are any macro lens meant to work in an infinity-corrected setup, only microscope objectives. Microscope objectives at 1x aren't ideal for macrophotography, they likely perform worse than a good macro lens, plus it would be expensive to buy one just for an axial setup.

Edited January 4, 2023 by Kaleun96

[Kaleun96]

I'm back with some more axial updates! I ended up finding a near-perfect beamsplitter plate to use for an axial setup on eBay. It's 80x80mm and 1.1mm thick with a 50/50 reflectance: transmission ratio and anti-reflective coating on one side and reflective coating on the other. This is about as ideal as it gets.

Someone from a macro photography forum I frequent has practically been doing the same tests in parallel and we happened across similar solutions and I think we're also using the same, or very similar, beamsplitter too. I'd recommend reading his thread that he posted on coincommunity here. He runs a macro photography business so perhaps he'll look to sell his contraption at some point. Either way, I think he's on to something with his more generic setup that can be used with mobile phones or DSLRs/mirrorless cameras.

As for the sharpness and detail, my first test has been quite successful. This is the first time ever that I've shot "pure" axially and not been immediately disappointed with the sharpness. There is still a reduction in sharpness but it's only noticeable when you zoom all the way in. It's so good, that I think it would not even bother most photographers.

N.B. I accidentally sharpened the "axial" photos when making these comparisons, so they might appear sharper than the "non-axial" photos but they definitely resolve less detail even after being sharpened.

I used a diffuser over the "light aperture" in my axial setup to reduce the contrast and harshness of the axial effect a bit. I'm quite pleased with how it turned out. As usual, it highlights surface issues but it also does a much better job of depicting depth to the features, particularly on Herakles' cheek. The non-axial setup has a tough time with this in the centre of coins because I use a ring flash and the middle of the "ring" of course has no lights, so the light hitting the centre of the coin is not axial and more likely to be diffused and indirect so you don't get the axial effect.

Axial Setup

Non-Axial Setup

 

Left: Axial | Right: Non-Axial

Edited February 5, 2023 by Kaleun96

[Severus Alexander]

Nice!  As expected, the axial setup does a better job on toning as well. I’ve ordered the beamsplitter and will give it a try!

[Kaleun96]

28 minutes ago, Severus Alexander said:

Nice!  As expected, the axial setup does a better job on toning as well. I’ve ordered the beamsplitter and will give it a try!

Great! When you get it, to find the side that should face the light place the end of a toothpick, or something else that won't damage the coating, on the glass and see if the reflection is directly beneath the toothpick or offset by the thickness of the glass.

The side where the reflection is immediately below the toothpick is the side to face the light as it has the reflection coating. The side with the offset reflection faces the camera. 

 

[kirispupis]

I also appreciate the recommendation of the beamsplitter, and I ordered three different ones to experiment. I'm still using the GND, but these look to be much better alternatives.

I've made some alterations to my setup as well. The biggest one is I switched my speedlight for an Einstein 645. I was a bit surprised that the light output is weaker, though, and I believe it's because the speedlight had a much smaller head compared to the Einstein and therefore provided more directed light even though it's much less powerful. I've attempted to address that by ordering a different reflector that adds a stop of light, along with a honeycomb grid as mentioned by @HipShot Photography. I've also thought of using an optical snoot, but since it has the same budget issues as the stand, I'll see what the honeycomb grid can do first.

I stopped with the reflections, since I felt they were limiting too much and the feedback I received was mixed. This enabled me to put felt over the Legos in order to prevent light reflecting back and to make the background easier to replace. Unfortunately, I received a setback when my wife sold every single one of my Legos. I'd given the okay to sell most of them, but they're all gone now save the ones I used here. The set I used for parts is no longer manufactured, so I'll have to order individual bricks because I want to make some modifications to my stand.

The two main problems I have now are that I'm shooting at a higher ISO than I'd prefer and that my focus is off in parts. Both can be solved by using a stand instead of shooting handheld. I already know roughly what I need there, but am waiting a bit for the budget to clear up after a number of expenses.

Here is an example from the current setup.