After my bio article on Petapixel, I had a bunch of people reach out to me and ask where I learned to do the optical stuff that I do. Books, links, that sort of thing. Unfortunately, I don’t really have a simple answer. I have done a lot of different things that enable me to think across different domains. What I can do is share some stuff that I find essential to thinking outside the box with regards to cameras and lenses.

There is one concept that I think really helps understand what lenses are: Lenses are the physical embodiment of a mathematical equation that project the scene in front of them into a space behind them. It is as if there were a miniature world behind the lens, as seen from the entrance pupil of the lens. There are a lot of specific statements we can make from there, but this thought exercise is really useful for thinking about how lenses work.

When you focus a lens, all that you are really doing is moving the intersection of the image sensor and the miniature version of the world. There is a plane in the miniature world that corresponds to infinity and when your image sensor is on that plane, you capture image content that is focused on infinity. In a tilt shift camera, you are moving the sensor in such a way that you capture different portions of the world that are non-parallel to the mount of the lens.

You can demonstrate this by taking a lens into a dark room that has a view to a brightly lit place and setting a lens on a table or holding it in one hand and finding the image it creates on a surface like a piece of frosted tape or even paper. As you move this surface, you see the image created by the lens in focus for different parts of the area in front of the lens.

Of the entire view that exists behind the lens, only what you capture with your sensor is “imaged” into a form you use.

As it happens, this world behind the lens has some peculiar characteristics. It is a mirror image of reality. The image is flipped and flopped so that up is down and left is right.  It also mirrors fore and aft. Things in the scene that are far away are represented closer to the lens and things that are closer to the lens are represented farther away. Ever use extension tubes or a macro bellows? They position the lens to allow different portions of that miniature world fall onto your sensor than what you would normally use.

The implication of this thought process is that, at least for me, is that it makes a tangible, physical representation of what the “image space” of a lens looks like. And once you know what it looks like, you can think about things that you can do with it.

For clarity, this physical embodiment of a mathematical transform (the lens) is not perfect, and the transform itself (the “prescription” for the lens) is not perfect. The transform is a bunch of tradeoffs to create desirable characteristics and limit undesirable ones. Flatness of field, distortion, chromatic aberration and more are designed into this transform and when you manufacture this lens, you induce statistically understood variations that have an effect on image quality. One lens that comes close to “perfect” isn’t really a lens at all, it is a pinhole of known dimension and position. It represents a simple mathematical model that can be easily realized with a physical object (the pinhole), that produces an image. Other simple examples are single optical elements (of a positive focal length.)

Having this bit of information in my pocket changed the way I realize what is happing when you capture an image. The next bit is how to actualize an actual device that affects change.

If you have a mirrorless camera, much preferably one you don’t use anymore, you can start experimenting with old lenses. I recommend mirrorless cameras because though their sensors are large, micro 4/3, APS-C or even full frame or larger, their corresponding flange depth is quite shallow. Flange depth is simply the distance from the mounting flange of the camera to the sensor.

An example would be A Fujifilm camera like an X-A3. Fuji has a flange depth of 17.7 mm. Next, you pick up a lens lie an old Nikon F mount. Nikon has a generous flange depth of 46.5mm. This means that when the Nikon is set at infinity focus, the image will be 46.5mm away from the mounting flange of the lens. This give you a generous 28.8mm (46.5mm – 17.7mm =28.8) of space to play with for doing things in.

I recommend a camera you don’t care much about, because cleaning sensors is a hassle, and a lot of your experimentation can expose your sensor to dust.

You can pit pieces of clear plastic in this space, small vessels of glitter and water, who knows. You can also try “Free Lensing” a process where you hold your lens in front of your sensor and tilt and move it about. In this case, you will benefit from a few things. A light baffle, something to keep stray light and dust out of your camera, and a sensor that is smaller than the image circle of the lens you are playing with. If your extra camera is APS-C or full frame, then you won’t get nearly as much “action” from a full frame lens as you would from something like a medium format lens.

Free lensing can be very rewarding. Old glass is cheap, and if you are free lensing, then the focus mechanism doesn’t even matter. This opens you up to lenses that may not even have a focus mechanism.  You can make bellows assemblies from sheets of opaque rubber cut into disks and glued together, or you can make bellows the old fashioned way with ribs and cloth.  The internet has lots of answers here.

The image circle is a property of a lens that will determine how much success you will have “free lensing.” A lens like a 80mm Hasselblad, Mamiya, Bronica or Pentax, which is designed for medium format coverage will have a large image circle giving you a lot of spare image area on even a full frame sensor but larger image circle lenses usually come at the cost of not having shorter focal lengths available. Anything wider than a 50mm lens in medium format is generally a fisheye, and there aren’t many choices. Go below 40mm and there are precious few options, severely limiting the ability to free lens with wide fields of view. It is always a tradeoff.

Going further behind the lens to the sensor, there are a few things you can do, but they are for the brave. You can remove the cover of your digital sensor and replace it with one that will pass infrared and ultraviolet light, allowing for interesting image creation possibilities. Ideally you replace the infrared/ultraviolet blocking glass with glass that is anti-reflective coated, so that your flange depth is maintained (glass slightly extends the path length of an optical system) and your sensor is protected. If you simply take it off you will have unpleasant unintended consequences.

Further madness involves manually removing the micro lens array and color filter array from the sensor surface. I have done this a few dozen times on multi thousand dollar sensors but it is scary. The benefit is that with special software you get the actual resolution of your sensor, in monochrome. You lose the sensitivity boost you get from the micro lenses, but you gain that back by removing the color filter array.

You get your resolution back because, unless you have a Foveon sensor, a photosite that has a color filter is combined with other photosites to create a pixel. In a monochrome sensor, each photosite is also a pixel.

There are services that will convert your camera to full spectrum, and I recommend that over having a specific infrared filter installed. A company called Maxmax.com will sell you a monochrome camera and convert camera to full spectrum. I do recommend this option over doing it yourself.

Imaging past the spectra of human vision is rewarding, but there are some pitfalls. Ultraviolet conversion on color cameras are not particularly awesome because the sensors themselves aren’t that sensitive to ultraviolet and the color filter arrays absorb quite a bit of UV as well. There is more sensitivity to infrared than ultraviolet, and there are a variety of filters to allow you to play in that regime.

I had a business selling infrared converted point and shoot cameras. I sold these with a set of colored acrylic filters. Acrylic is interesting because it does not absorb IR, but will transmit different colors. A blue acrylic filter, will pass the blue of the sky strongly, and pass the infrared as well, it will block red and green light from making it to the sensor.

I would shoot RAW via CHDK on the point and shoots, process in RPP and then in Photoshop I would edit in LAB color space and play with the A and B color channels by expanding their range via levels and then convert to RGB.

The “gotcha” with these conversions is that your normal lenses are not designed with these wavelengths in mind. The anti-reflective coatings can break down at certain apertures and yield lots of flare issues. You can research which lenses have proved better for people. There is a lot of good information on the web regarding this.

The next big pieces of the puzzle for me come down to making things. That is way beyond the scope of a simple article. There are some pieces that I find myself repeatedly using, and I will share those.

JB Weld 5 minute epoxy. The normal JB Weld stuff is great and a lot stronger, but the 5 minute stuff is good enough for me. JB weld by the way, not hobby store 5 minute epoxy.

Epoxy. For normal epoxy work I use Smooth-On brand.  By normal I mean epoxy that can be removed if necessary, but does not require heroic means like JB Weld does.

Silicone adhesive. I use 5 minute Silicone from Smooth-On as well.

Epoxy Putty. I have not found a bad brand.

Aluminum foil tape. Used in HVAC work. Opaque, strong, great for sealing up light leaks.

Stainless steel 0.002” shim stock. Cut with scissors, bond with Epoxy.

¼-20 marine nuts, washers, fender washers, bolts and set screws.

Any ¼-20 or 3/8-16 adapters for photography that you find on Amazon.

Step up and step down filter rings. Not only great for attaching filters of various sizes, but also great for making attachments for lenses that involve a lot of glue…

Male to male and female to female threaded filter rings. Great for attaching the front of one lens to the front of another lens.

Macro reversing rings. These allow you to attach the front of a lens to the body of your camera. Normally used for macro work.

Macro extension tubes. Buy metal ones. The mounts themselves are great for adapting, and the tubes can be glued to easily.

Gatorboard. Stronger than foam core.

Hot melt glue gun. Fast and quick setting for making things with Gatorboard.

Hand tools, all of them. A 1/4 -20 drill tap is a must.

That is about half the equation. The rest is electronics and software, but getting a handle on this will open a world of possibilities to you.