Those who know me and my work know I like BIG photographs. Perhaps it is because I started in the world of painting and printmaking that I like large work that you can both stand back and take it all in or get in close and revel in the details and minutiae of the image. I love discovering the “Easter Eggs” sometimes hiding in the detail that, at the time of taking it, my own eyes could not resolve and they appear only now when frozen in time awaiting close inspection.
Paintings are easy to make big… you just prepare a big surface – canvas or panel – upon which to create your world. Traditional artists know that a given image has a proper “scale.” Some need to be small miniatures where viewers have to get in tight and interface with the delicate universe of the image. Others need to push you back to take it all in. That viewer response, both emotionally and physically, was part of your vision for the finished work of art and you did not dilute it by using an inappropriate size.
Photographers seemed to think, however, that scale does not effect the aesthetics of the art work; a bit of ignorance that stems from the ease of raising or lowering an enlarger. They ignore the inconvenient truth that resizing images, among other effects, also resizes the Circles of Confusion, the tiny circles representing rays of light that optically make up the image. That effects the depth of field and overall sense of focus and sharpness and can, in fact, quite noticeably effect the message of the piece.
Photographers in the film days also faced the distressing reality that extreme enlargement also enlarged any and all technical glitches so if you knew from the moment of capture that you wanted to make a large image from a particular subject, you had to shoot it on larger film. The concept is incredibly simple: the same image spread out over larger film meant there were more silver grains available for fine detail so the resolution increased. Medium format (MF) and large format (LF) cameras were a bit more expensive but not all that much; the real expense was in the supplies and the needs for equipment to process and print the film and final print along with the technical skill to use them.
That situation has translated nearly intact into the digital world though for technically different reasons. Larger chips produce superior images due primarily to larger photosites. For example a 16 Megapixel MF back will be greatly superior in image quality to a 16 Megapixel DSLR. But the bottom line remains, if you can increase the capture area, all other variables remaining equal, you will achieve more resolution assuming your lens can cover the new area. And that means you can make larger final prints with the increased resolution.
But medium format backs are anything but cheap and that would assume you already have a medium format camera capable of mounting the digital back. Dedicated medium format cameras have dropped in price in the last few years but if you are starting from scratch you will have no trouble dropping $20,000 to $50,000 for a camera and lens system. And large format digital backs, such as those made by BetterLight™ will take an $80,000 bite out of your bank account in addition to the cost of the camera body and lenses.
High end shooters who charge high end prices can afford them and in some ways have no choice since their clients demand the ultimate in quality on top of the ultimate in creativity. But that leaves students and entry level shooters as well as budget constrained amateurs of all types out in the cold.
Or does it? Are there ways the shooter on a budget can create images that will match those of the MF and LF digital shooters but on a hamburger and ramen budget? The answer is, “Yes” but there is a trade-off… there always is. If you are a shooter that just sets the camera on full automatic, holds down the trigger and hopes that somewhere in that flurry of shots something will be wonderful then this is not something that will work for you… but then neither is a MF, much less a LF digital camera system.
Those MF systems are designed for pros with the expectation that they know what they are doing and can shoot far more deliberatively. Once you wish to move into the large file/large image/high resolution world your shooting will have to become far more purposeful and thought out no matter which solution you start with. They do not have program modes or “do-the-thinking-for-you” settings. So before you seriously think about investing time and/or money in any of the solutions, make sure that on a frame by frame basis you can, at least technically, manually set the camera for the perfect exposure… every time just as if you were shooting film.
Conceptually the solution is simple: take the view you want to capture and then break it into smaller parts, each one of which is photographed with a longer focal length lens than would be used for the single shot. So, for example, if the scene you want to capture is best shot with a 28mm lens, then if you put a 150mm lens on the camera, you could break the scene into a “grid” and shoot each cell of the grid with that longer lens. If it took 10 shots with the long lens to cover the shot then instead of a single shot of, say, 20 megapixels, that same scene is now created with 10 x 20 or 200 megapixels. There is a lot of detail in a 200 megapixel capture!
The technique is to create a “Multi-Row Panorama” or what is often simply called a “Mosaic.” Conceptually this is quite easy to do; the actual logistics however require some effort. The major bug-a-boo of both Mosaics and standard Panoramas is an old photo problem common in Rangefinders and Twin-Lens Reflect cameras and known as Parallax. It is the spatial distortion and off-setting of near and far picture elements resulting from trying to “register” or line up two or more images shot from different viewing positions. The visual result is that foreground and midground and background elements no longer line up so stitching the frames together becomes a frustrating and tedious nightmare of editing torture.
The solution most often used is to use a camera mount called a “Spherical Panoramic Head” which holds the camera so that it can pivot both horizontally and vertically on the optical axis of the lens. These can range in price from about $100.00 for the Panosaurus™ to around $400.00 for one of the billet aluminum models from Nodal Ninja™ to nearly $1,000.00 for the Gigapan Epic Pro™ which is a battery powered robotic spherical panoramic head. With the increase in price comes greater precision so you can use longer and longer lenses for greater and greater sized photos. They all work and all work well within their precision constraints. The technique has been used for years most notably by NASA in its early shots from Mars and the Moon rovers by their EL Hasselblads.
To see how crazy you can get with this, you need to link to http://www.paris-26-gigapixels.com/index-en.html . The aesthetics of the photo leave a lot to be desired but the resolution is simply astonishing. The problem is it takes a computer to allow you to zoom in and see details in buildings nearly across town.
I have a shot of the Grand Canyon that was taken with my 1Ds MkII and the Nodal Ninja head using over 30 frames that has a native size of 160” x 80” and it is wonderful to check out the detail even down to the river. But when reduced to normal size, as it is for this blog/handout, it loses that effect and who, these days, has wall space for a print that big anyway? It turns out there are some practical limitations for a professional photographer unless you are shooting a museum diorama or similar special project.
So from a practical standpoint, what is really called for? A two row mosaic of 6-8 shots will yield practical sized prints up to 6’x4’ with no effort and minimal enlargement but even at that there are not a lot of homes whose walls can handle an image that size. A 3×2 frame mosaic from a full frame DSLR will give you twice the resolution of a higher end medium format digital back.
You can certainly do these with the spherical panoramic head but there are other options and those options can be important depending on the project in question, such as if you need to use optical movements for an architectural or product shot.
My first solution was to create what I call my “Digi-View” camera setup. It consists of an adapter back (obtained from www.fotodiox.com) that fits onto the Graphlok™ backs on my view and field/technical 4×5 cameras and allows me to take roughly a 4×2 frame mosaic while using the full optical movements allowed by the camera bodies and lenses. All of the final frames are taken from a single projected image since only the adapter back moves and not the whole camera or even the camera lens. Using the rear standard’s shifts and rises I could actually get more frames but really do not need them for all but the most extreme resolution shots that need the Nodal Ninja or Gigapan anyway.
You can search this blog with the keyword “digicam” to see those articles and samples.
But Fotodiox has created an even simpler solution for those more common shots where optical movements are really not needed. Called a “Rhinocam” it is a sliding plate that lets you take a 3×2 frame mosaic with a full frame DSLR using a medium format camera lens. The needed image coverage is about the size of 6×4.5 film so any of the 6×6 or 6×7 lenses it uses offer more than enough coverage. It also has a built in groundglass for focusing that simply slides out of the way. Originally made only for EVF/Mirrorless Sony bodies it now comes in mounts for Canon and Nikon DSLRs.
Since I have a good selection of Hasselblad lenses and Canon DSLR bodies, I got the version to connect Hasselblad V lenses to a Canon EOS mount. Because there is no distortion or alignment issues caused by the camera pivoting shot-to-shot, even Photoshop’s poor mosaic building capability works — sort of — though I still much prefer PTGUI™ or PTAssembler™ to assemble the final file. Those excellent dedicated stitching programs are less prone to alignment errors and therefore need less final layer touch-up. Using the DigiView rig and now the RhinoCam I’ve often needed to adjust/smooth a frame to frame intersection that was not quite lined-up using Photoshop but have never had any issues using PTGUI.
Result? My Canon 5D MkII on the RhinoCam yields a roughly 100 megapixel equivalent final shot and at the equivalent sensor size of a MF digital back. That translates into a larger than 300 megabyte RGB 16-bit file.
So how does it work? Let’s take a look. The image below, of an aloe in my back yard, was taken when the above shots of the Rhinocam were taken. Its native size, even after being cropped as shown, is 28” x 36”at 300 ppi without any enlargement. Here on the blog post you can enlarge it to full screen in one or two clicks to get a better idea of its clarity and detail.
Y’know what? I can live with that… The only down side of any of these solutions is the inability to take shots of moving objects. For a sports or fashion photographer that would be a deal breaker, but those are not the subjects or style of my personal or professional shooting so this suits me just fine. Now, except for motion capture needs I can create reasonably large high res nature/landscape shots for me or product/food/architectural shots for assignments, and not break the bank doing it.
With the options of using a spherical pano head, my digi-view rig, or now the RhinoCam as dictated by the specifics of the shot, there is really no such thing as a size or resolution restriction for my work.