Large Format Photography Beginner's Notes

Some random thoughts on the topic, for anyone new to large format photography and thinking about taking it up...
  1. I recommend trying to find an old copy of "Photography with Large-Format Cameras", the fifty page Kodak Publication No. O-18, if you can, as it provides succinct but sufficient coverage of just about all the basic aspects of the field that are not obsolete. The 1st edition was dated 1-73 (January, 1973), and cost $2; Cat 152 7894. -And that is the letter "O" not a zero)

    [The small amount of information about specific products which are no longer available is easily read through, or over, or is translatable and adaptable into what would be applicable to currently available products. I've thought of scanning my copy into a PDF file, but it's a bit of work (especially since there are a lot of notes written in the margins) and doing so may violate the copyright anyway for all I know. I'm not sure how long Kodak put this out, but it could have been thirty-plus years, so there might be several slightly different editions. It covers just about every topic to some degree or another, including some stuff you'll never use. ]

    I actually didn't run across this until having done LF (and other) photography for some half dozen years, but it's been a valuable reference at times ever since. I learned mostly from Ansel Adams' famed "Basic Photo Series" (of five books), primarily the first two -- "Camera and Lens" and "The Negative" -- which are most applicable to LF film photography. The third, "The Print", is relevant even if you're not interested in printing by light onto the old silver-based papers, as it contains quite a lot of Adams' philosophy as well as information which is more or less independent of the medium.

  2. Not everyone makes the transition to working with the upside-down image on the ground glass. One tends to look *through* the viewfinder of most small format cameras, whereas you look *at* the image on the ground glass of a large format camera. This is fundamentally different in a subtle way that one may not have previously considered, though it is similar in some respects to looking *at* an image on the LCD screen of a digital camera.

    A good Ansel Adams tip was to make and use a simple cutout card held in front of the face to first compose the image by eye right side up, and then note where the desired edges and/or corners are, or what is on-center. This acts as both a compositional aid and scene analyzer, but it can also help determine exactly where to set the tripod up, and maybe also what lens to use (from how close the card is held to the eye - the closer, the wider the angle of lens needed). Playing around with this is a good exercise in noting how small changes in camera position can alter relative perspective of nearer-by objects and slightly/subtly rearrange components of the image.

    You want a minimum sized cutout of 4"x5" in an ~8"x10" card (black on one side and white on the other is best), which with a normal focal length lens of F.L. ~6" will need to be held about that same distance (6") from the eye to frame the field of view properly. That distance will be even less for shorter F.L.'s, so a 6"x7½" cutout in maybe a 9"x 11" card would be better. At times I've thought of inscribing lines on a piece of plexiglass, and trying that as a variation on the same idea, but have never gotten around to doing it.

    As an additional aid for B&W, Kodak had a dark greenish-amber Wratten #90 filter which visually mutes colors sufficiently so one could partially "see in B&W", so far as this is possible, by looking through it (not using it on the camera). It saved me the trouble of making a lot of not-so-good photos, while educating me over time what subject matter translates best into B&W. If certain image components look darker or lighter through the filter than expected or desired, it may also suggest filters to use on the camera during exposure to change their tonal values as desired. I'm not sure if #90 filters are still around, so it might be a rare find if you run across one. You should mount it in some sort of sturdy holder (like a big slide mount) so you can handle it without damage, since the gel filters themselves are flimsy and fragile.

    One other note: if you're much past forty years old and/or need reading glasses to see things up close, especially under dimmer light, you'll likely find you need a strong pair to be able to focus at the ~5-6" distance of the ground glass under the darkcloth of a LF camera. A small (~2" diameter) magnifying glass, I've found, is essential when focusing the camera. Mine is strung on a length of light picture hanging wire attached to the camera so it's always right at hand. It folds up into its protective case, the outside of which I've glued or taped a small piece of mirror to; this is so I can see the bubble levels on the top of the camera back without needing a step stool when the camera height is too much to be able to see them even tippy-toe, which tends to be the case when the center of the ground glass is at or near eye level, where it's the most comfortable to work when on level ground.

  3. If you intend to make exclusive use of the web and/or electronic displays to show your work, LF photography may be a bad mismatch.

    My 17" laptop has only a 1 MP display, and I don't think even the highest resolution office displays are much above 2½-3 MP. That is basically just a tiny thumbnail image for a 100 MP original LF scan.

    If an even smaller "catalog" image of maybe only ~½ MP (600x850) is what's being displayed, the medium mismatch is even more severe. All of what I call the stupendous information richness inherent in the medium of LF film photography can be lost in the huge downward compression needed to fit it into the relatively tiny peephole of the everyday digital universe, to the detriment of the medium. In other words, LF won't make your work instantly look better on small digital displays.

    Even 35mm film or APS-C digital can be overkill when paperless media are the principle means by which people are viewing your work. Depending on what you're doing, there may be no real point in going with large format except as a broadening and learning auxilliary experience. There are things you can do with LF which are impossible to do with other cameras, but they won't necessarily be immediately apparent at low resolution.

    Where large format is at its best IME is in making large prints. A 16"x20" print, whether made analog or digital style, is only about a 4x enlargement of the original negative or slide, which would be comparable to a postcard sized (4"x6") enlargement from 35mm film or common amateur digital formats, roughly speaking. For these, a 16"x20" print would be an ~15x enlargement, about the upper limit beyond which graininess and inherent sharpness limitations in the lens become apparent even for a carefully made photo using quality equipment, since at some point one merely enlarges the blur. Remember, a 12x enlargement for 4"x5" is about 4x5 *feet*.

    A technical explanation and somewhat lengthy discussion to make all this as clear as possible follows...

    At a scan resolution of 2400 DPI, my 4x5 based digital images start out at about 9,000 x 11,400 pixels, which is just less than 100 MP. This is a 10.6 micron pixel size on the film, and this is sufficient to extract almost all the information in the original negative or slide which was made with regular films and lenses.

    This is a scan resolution of 90+ pixels/mm. If you're familiar with sampling theory (and the Nyquist limit), you'll know this is sufficient to resolve a maximum of half as many spatial cycles or line pairs per millimeter, 40-45 cycles per millimeter in this case. This is good enough to exploit the approximate maximum resolving power of regular films with most lenses that produce a decent image on the ground glass.

    Though it depends on the subject contrast, film, development, etc., if you've looked at film technical specs you might have seen what's called the Modulation Transfer Function (MTF), which charts this film property. The generic curve for many regular modern films is falling pretty rapidly by the 40-50 cycles per mm area, meaning the film's ability to fully resolve such fine detail (and finer) is pooping out. It does record it some, but at a much lower (and falling) contrast. The MTF is usually given as a percentage of the input contrast, and it's an arbitrary choice as to what low percentage it has to fall to before the film's ability is effectively exhausted.

    Ignoring depth of field considerations for the time being, most modern lenses are capable of this level of resolving power or more, so the film is typically the limiting factor, but not by a huge amount. IOW, with a very nice lens you may see more fine detail on the ground glass when focusing than you'll end up with on the film after it's processed. (It can depend on how fine your ground glass is.) Don't be disappointed. Older lenses, say from the post-WWII era up to when computers started being employed by the late `60s and into the `70s, which resulted in much better lens designs (like by +50-60% in resolving power), are typically pretty well-matched to generic film technology of the same era (i.e., Plus-X, Tri-X). Such lenses may not have much going for them in terms of image circle size, but at a practical level their sharpness is sufficient to do lots of good work with.

    Because resolution in a chain of components adds reciprocally in quadrature, if a 45 cycle/mm capable lens is used with 45 cycle/mm film, you get an output with ~32 cycles/mm resolution when the two are used together. A sixty-fourth of a millimeter is 15.6 microns, so this is within the capabilities of the scanner (10.6 micron pixels) to resolve. This may sound very coarse compared to modern silicon sensors with 150-175 pixels (~6 micron) up to 500 (2 micron) or more pixels per mm, but if you've ever seen an Ansel Adams print at an exhibition you know that great work can be done in the 25-40 cycles/mm regime. This is because you have so many square millimeters of silver real estate at your disposal. (-More than 100 sq.cm's, ~115 to be exact, each one of which has about as much information -- a little less than a MP -- as an HD video still frame.)

    See Addendum A for more.

    Back to regular large format cameras, what can one see on the ground glass when focusing? If we take the eye's resolving power as being 1-2 arc-minutes, at 5" (127mm) distance this is 27-13½ per millimeter, which is at best still just slightly less than the typical lens+film combination is capable of -- hence the utility of a 2-3x magnifying glass for critical focusing. When trying to focus only by eye, there should be a small dead zone as you rack things back and forth where the focus doesn't appear to change much, and the best you can do is try to hit near the mid-point of it.

    The situation changes entirely when the thinking switches to being in terms of a non-fixed display (print) size. The eye's maximum resolving power corresponds to 1/1720th to about 1/3500th of a radian; thus, at a typical reading distance of 14", 1-2 arc-minutes is 1/125th to 1/250th of an inch (14" divided by 1720 or 3500). Thus a print made at 250 DPI (Dots per Inch) from a 2400 DPI scan (9.6x) will not show any pixellation or discreteness effects under even close casual inspection. Most scans of decent source material can be enlarged to about 15x, or 160 DPI before any deterioration in quality will start to become apparent when the enlargement is given critical close inspection. At 125 DPI -- pulp printed material like newspaper photos used to be printed from plates made with 100-125 lines per inch screens -- an enlargement may look okay from a distance but upon closer distances will hit a limit within which getting closer doesn't help.

  4. This tends to be really slow, deliberative photography, especially at the bottom of the learning curve.

    It takes time to set up and level a tripod, mount and level the camera, attach the darkcloth, choose and mount the lens, and then get started working with the image. After that's complete and you're ready to "fix" it, there's the entire subsequent process of metering the scene and determining an exposure setting, presuming you already have an idea which film you're going to use, and possibly which filter, and what your intent in making the exposure is.

    Actually inserting a film holder, withdrawing the dark slide, and clicking off an exposure with the cable release can almost seem incidental. It can easily go by unnoticed to a casual observer standing nearby. Then there's the entire camera/tripod disassembly process in reverse. (I'm assuming you'll be working in the field not the studio.)

    Adams uses the term "contemplative" to describe the contrast with the approach which comes naturally when using small format equipment, but this is maybe not always the case. There have been plenty of times when I've wanted to go through the setup/expose process as quickly and efficiently as possible to get a shot before the light changed or clouds moved, or maybe even the location required it. In college, I did a cross-country roadtrip in a van with my 4x5; it wasn't long before a request to stop so I could take a picture would elicit groans. Most people think taking a picture means jumping out at the view spot, while leaving the engine running, to take a snapshot proving they'd been there and seen that, and even though I got so I sometimes could set up, shoot, and tear down in only ~10 minutes, on that scale LF does take 10-20x longer than a snapshot.

    Often when I'd go out for a shooting session of a few or several hours I might feel okay about only making some four or five exposures when all was said and done, and anything up near double that or above was a good haul. When I'd go out where I could shoot almost all day, ten or a dozen (ok, fifteen) felt like coming back from fishing with a full creole.

  5. So LF photography is about making every exposure count, at least for what it is.

    Color slide film is now (c.2015) about $3.50 a sheet and $2.50 to develop, not to even factor in the time, travel, and equipment investments, so LF can be at the opposite end of the spectrum in terms of cost-per-shot from potentially throwaway, quick, operationally almost free, digital photography.

    [It was probably going out on multi-day and 1+ week wilderness backpacking trips in high school, with only 2-3 rolls of 36-exposure Kodachrome, which made me very economical with film usage (*), and made it natural to hone basic skills to the point where there weren't many wasted exposures. Each shot had to be a keeper, but because of the time delay involved with film in seeing the results (since you have to develop it), well, you had to get good enough to have the confidence to take a single picture. Much of the process has to mastered to the point where the individual decisions are simple. Being uncertain and bracketing everything is somewhat the antithesis of this approach.]

    * - There was someone who I went on a number of trips with a friend who knew my film frugality. He would still remember the time he'd thought I'd gone completely nuts since I was going through almost half a roll of film in about 15-20 minutes. (We were camped at a half-frozen tarn on the tundra in a place that was very difficult to get to, and one evening the sunset colors bouncing off the surrounding mountains and rock cliffs lit the entire place up spectacularly; it was difficult not to see one incredible photo after another, and the light was changing so slowly but steadily that I didn't even want to stop and change lenses.) You have to get while the getting's good sometimes.

  6. In order to be technically competent, as well as capable of exploiting what the format has to offer, one will likely need to familiarize oneself with several of the fundamental aspects of lens optics. No, I won't drag you through depth of field or any other equations.

    First, it's important to realize just how simple the basis is for almost all large format lenses...

    Start with a simple lens: ()

    Bisect it: (|)

    Separate the two plano-convex halves by a small amount: (| |)

    And then insert an aperture and shutter into the space: (|¦|)

    Voila!

    This is the simplest, most elegant design, also known as a symmetrical lens. It has survived all the way to the present. Even my modern, aspherical element, XL super-wide field lens is a derivative of this design lineage bearing the trade name Symmar, even though it's technically an asymmetrical design.

    The next step in evolution and sophistication was to make each plano-convex lens into an achromatic doublet: ((| ¦ |))

    This results in a 4-element lens which works amazingly well at apertures as wide as ~f/4.5-5.6, for normal and portrait focal lengths. It has a relatively wide good field (image circle size at the focal plane), and gives good enough color correction that it can be used with the then new orthochromatic and panchromatic emulsions (and color films). Sometimes a minor 5th element was added to one component to improve one or the other of the common aberrations and/or increase the image circle size a little.

    Simple designs of this sort dominated the first half of the film era because they were able to inexpensively provide sharpness comparable to or in excess of what the early LF sheet films were capable of recording. So, many of these lenses and their designs survived into the post-WWII era and up to the era of computer design.


    The Super Angulon exhibits its earlier lineage in dual plano-convex designs.

    In considering the properties of the symmetrical design it's sometimes helpful to think in terms of the converging power of a lens, rather than its reciprocal, the focal length. This is a change in perspective because we usually think of long focal length (telephoto) lenses as being "more powerful" because they produce a larger image. But they can also be thought of the other way around, as having less converging power. The total converging power of a compound lens (two lenses together) is shared equally by both halves in a symmetrical design, each giving half the total converging power. This means they each have a longer focal length as separate elements than together in combination.

    In the old symmetrical lens designs you could use either element by itself (usually behind the shutter), essentially giving you two lenses in one, one with a normal-ish focal length, say, and the other with a focal length maybe almost twice as much. By going to a mildly asymmetrical design, one could get three different focal lengths from one lens. In fact, one of Ansel Adams' favorite lenses for his 8"x10" camera was a Cooke Triplet lens which gave three different focal lengths.

    By convention a "normal" focal length for a given format is taken to be roughly equal to the length of the diagonal of the format. This condition gives a diagonal angular field of view of about a radian, or 57.3°, from corner to corner. For 4x5 this is Sqrt(41=42+52) = 6.40" or 162.6mm.

    The f/4.5 Graflex Optar lens with 6 3/8" focal length would be an (old) example fitting this criteria, in the simple 4-element design category. I have one in a Graphex shutter made by Wollensak (Rochester, NY), and it's a very good lens which I doubt cost me more than $50.

    If one takes into account a border or gate for a format, since sheet film needs to be held down by its edges in anything that holds it, or like a slide opening is slightly smaller than the size of the exposed film so the unexposed (clear) edges don't show, the number goes down a little bit. The scanned area of my 4x5's is about 3.7" x 4.7", which works out to a diagonal of almost exactly 6" (152mm). If the typical 1/8" film holder borders are figured in the gate is more like 3¾" x 4¾" Thus many take a 150 mm lens to be a "normal" focal length for 4x5.

    When you double every dimension to 8x10, the diagonal doubles also, and so the "normal" focal length goes up to 12-13" (305-330mm). For 35mm film, the number is around 42½mm, depending on the exact dimensions taken (this corresponding to a 23.45 x 35.45 mm's gate, or a roughly ½mm border. An APS-C silicon/digital format camera (~15x22½mm's) works out to 27.0mm, which is actually large compared to some video chip sizes or point-and-shoot camera sensor sizes with 8mm or smaller diagonals.

    As an alternative to a "normal" focal length lens, one might first consider a so-called normal-wide lens, one with a slightly shorter focal length. "Slightly" means something like 10-15%. As it turns out, there was a 135mm version of the Graflex Optar normal lens mentioned above, and this was my first and only lens for the first several years after I got my 4x5, while I was learning the basics. It is 10% shorter than a 150mm lens, and 5/6ths the focal length (17% shorter) than the longer 6 3/8" version of a "normal" lens. For doing a lot of nature and landscape, as well as for general subject matter, the 135mm was a good, all-purpose way to go. (When I got a second lens it was an even shorter, 90mm, very wide angle lens.) In the slightly wider than normal-wide category, it's worth mentioning there was also a 127mm (5") version of the Graflex Optar lens, though I've never seen one.

    Others may have different requirements in an alternative to a "normal" focal length. The regular normal-long option for 4x5 is 210mm focal length, which is 8-8½" for older lenses. This is ~30% longer than "normal", though an 8½" versus a 150mm (the most extreme comparison) is ~44%. These are fairly common lenses that allow a greater camera-to-subject working distance, and eventually just about everyone who does 4x5 will end up with one.

    For a long time, a "portrait" lens was one with roughly double the "normal" focal length -- 85mm for 35mm film, for example. The longer focal length gives a greater, more comfortable, working distance, less depth of field, and a flatter, more pleasing perspective. For 4x5, this would correspond to the focal length of a normal lens for an 8x10, roughly 12-13". However, such a lens would be very big and heavy, and it would unnecessarily cover a huge image circle (>8x10), so before we get up to that focal length special telephoto designs for the 4x5 format start to become advantageous. These will be more compact, meaning a longer rail or bed for camera focus travel is not needed to use it. They are still fairly hefty pieces of glass. The longest lens I have, in fact, is a 270mm (a Tele-Arton), which is only double the focal length of a 135mm normal-wide lens.

    The image scale on the film is a function only of the lens's focal length, not the format, so all the problems which arise using 300+ mm lenses with smaller formats, like camera/image stability, apply for LF, too. The only distinction is that the degree of enlargement is less for a given size print for LF, so it is more forgiving in that respect.

    I notice somewhat the same thing with depth of field: a normal-wide 135mm lens seems to have a lot of depth of field on the 4x5's ground glass, but with a smaller SLR/DSLR the same focal length lens it seems much smaller.

    While I've mentioned a lens's image circle before, it's worth giving it some extra elaboration at this point. In a camera where the lens's optical axis and focal plane are fixed with respect to each other, the lens only has to have an image circle diameter of good definition sufficient to just cover the corners of the format. This diameter is equal to the format diagonal. So, for a "normal" lens with a focal length this same distance, the image circle diameter is a little less than one radian (53.1° =2*Tan-1[½], or 92.7% of a radian). Image circle diameters are usually given in angular units, as seen from the center of the rear of the lens (technically the 'exit pupil'), or as linear units at the focal plane.

    A general principle, we thus see, is that as focal lengths increase, the requisite angular image circle diameter for a given format decreases (to keep the linear diameter the same), while for shorter focal lengths it's the opposite. The practical import of this is that short focal length, wider angle of view lenses are more difficult to design and construct, and generally to come by.

    The demands on the lens increase when we can tilt or swing the lens's optical axis relative to the focal plane, or shift either laterally relative to the other. A larger image circle is required than is sufficient to cover just the format corners in a camera where the optical axis intersects the focal plane at the latter's center, where these movements aren't possible.

    This gets to an important peculiarity of LF photography: the term 'wide angle' (or 'wide field') may refer to the angular size of the image circle cast by the lens, regardless of its focal length, and doesn't necessarily imply a short focal length for the format or a wide angle of view in object space. The term can refer to the image space side of the lens instead.

    As an example, a 12-13" lens of "normal" focal length for 8x10, when used with 4x5 as a "portrait" focal length lens, will have a relatively enormous image circle diameter (~12-13" at least) for the smaller format and might be considered a 'wide angle' lens, even though it is a longer not shorter focal length than "normal". Such a lens would allow very extreme tilts, swings, and/or shifts. This is the main advantage of 'wide angle' lenses in general: they allow a greater range of adjustments before one hits the edge of the image circle. Of course there are always trade-offs. Cost and weight are the main ones. A higher minimum f-stop would be another. The need for a center filter, since the illumination level decreases the further away from the optical axis one gets (per the "cosine to the 4th law") is yet another.

    By the principle developed above, shorter focal length lenses are more necessarily also going to be 'wide angle' designs. If the focal length is taken to be equal to the format's maximum horizontal dimension rather than its diagonal, this is ~4¾"=120mm for 4x5, which makes for a good example. For someone using a normal focal length lens of 150mm or 6" who wishes to skip over normal-wide lenses (135mm), it makes a good focal length for modest wide angle work, giving an object space field of view of 43.3° x 53.4°. It is 76mm from the format center to the corner on the film, so a 2*ArcTan(76/120) = 64.7° image circle diameter is needed to just cover the 4x5 format on-axis with this focal length.

    In practice, a lens with a coverage at least ~10% more (greater than ~71-72° in this case) is needed to provide even minimal adjustments with a view camera. For 4x5 this corresponds to a linear image circle diameter of ~173mm. Recall that the format diagonal is ~152mm (when 1/8" borders on all four sides are taken, as in a typical film holder), so with this size image circle on-axis the corners of the format are only ~20mm (<1") from the edge of the image in any direction in the corners. When working with the image on the ground glass it's helpful always being conscious of where the edge of the image is.

    There are two caveats: first, the corners of the ground glass on many cameras are snipped off, so air can enter and escape the bellows freely to keep the pressure inside and outside the camera equalized, so you may not be able to see the actual corners of the image. Second, the image coverage figures one might find published somewhere are typically taken at f/22, presumably because they are maxed out by about that f-stop as the lens is stopped down, and determination of where the image "goes bad" may be somewhat subjective, so they may be only indicative of where the image edge seems to be, both for focusing and for other f-stops that may be used for exposures.

    Staying with the 120mm focal length and moving to concrete lens examples, a Schneider Symmar S f/5.6 120mm lens, according to one online reference, has an image circle diameter of 173mm (71.6°), while for their Apo Symmar (also f/5.6) the number is 179mm (73.4°). By contrast, both the Schneider Angulon (f/6.8) and Super Symmar HM (f/5.6) lenses are listed at 211mm, which corresponds to a very generous 82.6° field, ten degrees larger. As it turns out, this is just enough to cover the 5x7 format on-axis, which has an ~210mm diagonal. Lenses for use with 5x7 with adjustments have image circle diameters of 290-312mm (100-105°), though they are all f/8 lenses. The higher number (312mm) is just sufficient to about cover an 8x10 on-axis. 120mm is an extremely short focal length for 8x10, corresponding to a very wide angle object space field of view of 78.7° x 91.8°.

    The general principle is that the linear image circle diameter for a lens can tell you what format it is made to be, or can be, used with. For 4x5 this is especially important for focal lengths roughly in the 60mm to 100mm range, as these may be fast, normal-ish focal length lenses made for use with various smaller, medium format cameras, and they may not be capable of covering a 4x5, even without adjustments.

    Few commonly available lenses with image circle diameters >105° are made. For example, a 10mm focal length ultra wide angle lens for the APS-C format (27 mm diagonal) needs to have a 107° diameter image circle to cover the format's corners on-axis. To reduce the focal length a little, to 8mm, would require 120° coverage.

    The next logical step down in focal length from 120mm would be to take the 4x5 format's minimum dimension rather than its diagonal. This is ~3¾", which is roughly 95mm. Perhaps because 4x5 has a relatively small aspect ratio compared to other popular formats (which have 4:3 or 3:2), this got rounded down to 90mm (which is closer to 3½"), though there are also some common older lenses of 4" (100mm) vintage.

    Using the same equation as before, it will take a lens with an 81.1° image circle diameter just to cover the 4x5 format on-axis at a 90mm focal length (~75° for 100mm/4") -- the old f/8 Goertz Wide Angle Dagor design (81.8°) would be an example. Something closer to 88-90° is needed in practice to provide for minimal view camera adjustments. Much better are the ultra-wide angle designs, like the Schneider Super Angulon or Rodenstock Grandagon, which will cover at least a 5x7 and are up in the 100°+ image coverage realm.

    People often take the very widest angle lens focal length that is practical as corresponding to half the "normal" focal length, or F=75mm for 4x5. It takes a 90° image circle diameter just to cover the corners on-axis. A 100° image circle is only 178mm linearly, so an ultra-wide angle design is mandatory at this short a focal length. At F=65mm (the next step down), a 105° image circle is needed to give ~170mm linearly, so such lenses are the widest made for use with 4x5; they have an object space field of view of 72° x 85 1/3° and can only be used with virtually no swings or tilts.

    I'll end this lengthy digression by once again emphasizing the value of using a center filter with lenses that have large angular image diameters. A 105° field diameter means the angle to the edge of the image is 52½°, the cosine of which is 0.61. When taken to the fourth power this means the illumination level at the edge of the image is only 13.7% as bright as on-axis (angle=0°). And this is at best, for an ideal lens. This translates to almost a 3 stop difference (2 6/7), which is a significant amount, especially for higher contrast materials like slide films.

    Even in controlled studio situations it is difficult to light things to compensate for this much unevenness. As well, sophisticated digital darkroom programs are generally unable to salvage a scan with this much of an over- or under-exposure in part of the image. And, remember, one usually doesn't even know where the optical axis (center of the image) is on the film to begin with, which is necessary in order to apply any sort of post-exposure radial correction. Simply put, a center filter is almost a vital component of a wide or ultra wide angle lens, which is why in general each lens needs the particular filter which goes with its design. This qualifies center filters as rather specialized equipment, so they are relatively expensive. Be sure to factor their cost into your lens purchasing budget.

    As per the above, as it turns out the Schneider center filter I have for my Super Angulon 90mm f/8 lens has a carefully calibrated filter factor of 2 5/6 stops (±1/6th of a stop). However, some center filters for other lenses have manufacturer listed filter factors of only about half as much (~1½ stops), which makes me suspect they are only able to provide partial correction, or represent a compromise because a filter factor much above 1½-2 stops makes a filter much less likely to be used. My 75mm Caltar lens is of a Grandagon design, and the Rodenstock center filter for it is of this type, with a 1½ stop nominal filter factor.

    As a final note, realize that you'll most likely want to install the center filter on the lens only after you're done working with the image on the ground glass (which is already dim enough) and are ready to make an exposure. Any subsequent filters one might normally use (like a UV filter) go on the outside of the center filter, and will thus be of very large diameter -- maybe up in the 80mm diameter and above range, though it varies from center filter to center filter. For gel filters, this is larger than the 75mm (3") size can accommodate. Short focal length lenses are commonly mounted in recessed lensboards -- generally the deeper the better -- and the center filter, with its large diameter, may be the limitation in how deep a recess can be used, and it will likely vary a little from camera to camera. You'll want to take this into consideration when choosing or designing and making a recessed lensboard.

  7. Okay, now it's time to get down to those depth of field equations. The following illustration of a simple photo-taking situation is intended to convince you that the view camera controls, and especially the front tilt, are essential because of the relatively long focal lengths of the lenses used.

    So consider a camera set up on perfectly level ground so it is level horizontally. It has been outfitted with an F=135mm lens and the camera back rotated to vertical or portrait orientation, so the long dimension of the film runs up and down. With the same sort of simple geometry used before, the vertical field of view is 48°, or 24° both above and below the horizontal. If we take the optical axis to be 5 feet above the ground, as would be typical for a comfortable tripod height, then more geometry tells us the point on the ground at the lower edge of the field of view (top of the film or ground glass) will be 11¼ feet away in the horizontal, and 12.3 feet from the camera, line-of-sight. A side view diagram looks like this:

  8. >> Page Still Under Construction <<

  9. One disadvantage of LF photography is that each lens needs its own shutter, unlike the situation where the shutter is part of the camera body. This was more of a problem in the past (as different shutters could vary somewhat), but all the post-WWII era LF shutters I've used have been consistent enough that I didn't feel like I ever needed to do tests and calibrate a given one to get good exposures.

    The one exception to this statement would be at the longest exposure times, like ½ or 1 second; some shutters seemed slow, or long, as judged using a stopwatch. But this wasn't exactly a strike against them since reciprocity failure (of the film) would call for a slightly longer exposure anyway. So having a slow shutter at the long exposure times was kind of like having automatic reciprocity correction.

    [Reciprocity failure kicks in for regular films at an exposure time of anything longer than about 1/10th second. When the film is very weakly illuminated, which it is when such long exposure times are required, as a storage device the film acts like a leaky bucket, requiring an even longer exposure time to achieve the same equivalent exposure.

    From 1/10th sec up to ~1/4th sec, one needs to add only ~¼ to 1/3rd stop more exposure. By a 1 second exposure time it's more like ½ stop, and over the 2 to 4 second range it goes up to about a stop more being necessary. By the time one gets up to a 5 second exposure it takes about a dozen seconds to get a corrected exposure time. At 8 seconds the adjusted exposure time is about 3x longer (25 seconds), beyond which further exposure becomes pretty much futile and you almost can't over-expose. Though films differ, this is typical for consumer films designed for bright light and short exposure use.

    ] Shutters from about the mid-70s on don't have this "feature" so you have to figure the reciprocity factor into your exposure determination. With the camera on a tripod you'll likely use these shutter speed settings at some point, so they're worth checking. Consistency is the main factor you want, since obviously a shutter which hangs or seems sticky in any way needs service and would be nightmare to try to use.

    With this exception, shutters are thus pretty much interchangeable, except that each one seems to have the various controls in a different location, and might attach to its accompanying lens board (to hold it in the camera) in a slightly different way.

    Also note that even if the lens threads fit a given shutter, the f-stop numbers may be all wrong if it's the wrong shutter for the lens. You can check this by measuring the aperture opening diameter and seeing if the shutter scale gives the correct f/ number. For example, an opening of ¼" (~6mm) on a shutter for an F=150mm lens should have the f-stop pointer reading about f/25.

    The one unavoidable consequence of the one-shutter-per-lens rule is that of weight: my padded lens case, with seven lenses, filters, attachements, etc. is quite a load.


Addendum A

By way of driving home the point from an entirely different angle, when the design for the 4-meter telescopes at Kitt Peak Nat'l Observatory and Cerro Tololo (Chile) was done c.1970, the angular scale at the prime focus on the 8x10 plates, where the instrument functions as a very large format camera with an f/2.8 telephoto lens (F=11,200mm!), was made ~18½ arc-seconds per mm. The then new fine-grained astronomical emulsions weren't structurally much different from regular B&W films [-except for being noticeably coarser and grainier -- and a good part of this is due to them customarily being developed to completion in high energy, very contrasty, D-19 developer.] 1 arc-sec is roughly the limit of resolution under the best conditions imposed on all ground-based telescopes due to turbulence in the atmosphere (twinkling, generally), so optics aren't constructed to be capable of much more, and this plate scale (~18½ arc-seconds per mm) was well-matched to emulsions capable of ~40 cycles/mm. One wants as small a scale (which is *more* arc-sec/mm) as possible to make the field of view as large as possible, to capture as much sky as possible with an exposure, and this also corresponds to a faster f-ratio, both of which make better use of the valuable telescope time -- but not so much that the recording emulsion's capabilities are exceeded and valuable detail present in the optical image is lost. By c.1980 the first computer-controlled scanner/digitizers became available. One could get a Ph.D. analyzing a single field - perhaps several plates taken with different filter passbands. Under generic good "seeing" conditions, 1 arc-sec star images will be a little larger than 50 microns across, so scans were made with a 50 micron aperture and step size of 20 pixels/mm (a relatively coarse 500 DPI). The image circle on the plate is ~7½" in diameter (8" minus two ¼" borders; this is also 190.5mm, or almost a full 1° on the sky), so when you work it out a scan would be ~11 MP. Well, during the end of the era of mag tapes and punch cards, a hard drive that size or a little larger (in bytes) was the size of a refrigerator, required an air-conditioned clean room, and cost quite a bit on top of all that. And it could only store a single plate scan, maybe two! Computer memory was measured in dozens of kilo-bytes then, so in practice the plate had to be scanned and handled in numerous postage-stamp sized pieces. A star in such a scan is then either contained completely within a single pixel, or (more likely) falls near the four-corner region of adjacent pixels. It is then easy in principle to tell a mere star from, say, a distant galaxy with a diameter of 5 arc-seconds. A galaxy like the Milky Way, if removed to such a distance that it was only this size, would be more than a thousand mega-parsecs away, and seen as it was roughly four billion years back in time. So this surveys a lot of space roughly a quarter or a third of the way back to the Big Bang. By shooting in different colors one is able to separate galaxies by color into various types and compare with the current, local population of galaxies.

For those interested in the technology, the scanner was essentially a microscope, with the plate attached to an X/Y stage driven by stepper motors. At the end of each scan line the motors had to be ramped down in speed, and the plate stopped before being ramped up in the reverse direction to do the next line. One could control both the distances and velocities involved. While it was possible to shake a plate loose by making things go too fast, it optimally took on the order of ~10 hours to scan a whole plate. The detector was a PMT (photo-multiplier tube), amplified and fed into a 10-bit A/D convertor. The electronics has a certain response time, and the PMT sees a rather dim signal in the high density areas of the plate, so sampling at only kilo-pixel per second rates was necessary to give a good S/N, beyond which going slower didn't provide much benefit. So each line of a postage-stamp sized piece would take about a second, and each piece on the order of eight or ten minutes; at about six per hour, a plate broken up into fifty such pieces would take about eight hours to scan completely. The scanner was not light tight and lived in its own darkened small room. The dedicated computer was a PDP-11 running a command line "OS" known as PIP (Peripheral Interface something). While there were two terminals, you could not use the second one when a scan was going. The same computer was used for processing the scans, via compiled FORTRAN code. Programs tended to do only one or a few operations on an image at once. Only very small (like 64x64 or maybe 128x128) images, or extracted sub-images, could be read in into programs in their entirety all at once. Otherwise one had to shuffle some smaller number of longer lines in and out. Some operations that are very computationally intensive, like median smoothing, which was used to measure the reference background empty sky level, were done on "skeleton-ized" versions of the original scan where, say, every tenth pixel from every tenth line was lifted out for processing, to gain a 100x speed advantage -- since resolution can be sacrificed in this instance, the density level not changing over distances on the order of ½-1mm.

Addendum B

Kodak Pub. O-18 -- addendum to section on Characteristic Curves.

The very earliest emulsions were not very fast and so they were developed just to, or almost to, completion in very active, high pH, sodium carbonate (or stronger) developers. Development can be looked at as an amplification process, and by cranking it up to a practical maximum one got contrasty negatives with what would be a high maximum density by later standards. This was especially the case with films which would be reversal processed for direct viewing or projection, into either B&W or color slides, where maximum densities >3.0 give the best results. As films got better, developers became less active ("gentler"), but the custom of developing negatives to rather high maximum densities (above 2.0, and perhaps closer to Dmax = 2½-3) persisted, especially in commercial and graphics arts applications.

Such high densities created a problem when these negatives were enlarged: the higher the maximum density, the dimmer the image on the print paper. One either requires a higher intensity light source in the enlarger, or a faster enlarger lens, or a longer exposure time, or faster print paper. All these are costs of some sort, so as machine processing and printing became the norm for mass consumer small-format photography during the `60s it became clear that printer throughput could be increased by about an order of magnitude if negatives were given less exposure and development, and their maximum density reduced by about 1 or a little more.

Enter the era of the thin, flat negative, with a maximum density of 1 1/4 or 1½. Such a negative is given the minimum amount of exposure necessary, which translates into the film being rated at a high speed, and the exposure is then given minimally sufficient development. Such a negative can print well on a high contrast paper developed just to, or almost to, completion in a very active, high pH, sodium carbonate (or stronger) fast developer.

It's important to realize this is the kind of negative portrayed in the generic characteristic curves shown in this section of the O-18 publication. Almost all the available headroom, both in exposure and development, on one side of the equation, has been pretty much used up. One could say it has much greater tolerance for over-exposure and/or over-development than it does to their opposites. Another way of saying this is that one is always in danger of under-exposing and under-developing following the film speed and development time recommendations Kodak was putting out then. When in doubt, give more exposure and development if you want to shave things in what is likely to be the safe direction.

Also keep in mind that this type of negative, it can be argued, is obsolete if one is going to scan and digitize it rather than make an old-fashioned silver print from it. This is because scanners are designed to be able to handle slides, with their high maximum densities, so there is no reason not to make negatives which are better matched to the available density dynamic range, negatives with Dmax more up in the 2+ to 2½ range, as before. By the older standards such a negative is given both very ample exposure -- perhaps a stop or more than is suggested by its rated speed -- and is then given rather full development. You've paid for the silver to be there in the emulsion, why not use it rather than dissolve it away in the fixer and then pour it down the drain?

©2015-17, Chris Wetherill. All rights reserved. Display here does NOT constitute or imply permission to store, copy, republish, or redistribute my work in any manner for any purpose without prior permission.


Tip Me Please!

 

Your support motivates me to add diagrams and illustrations!

 

 

-------------------------------

[ To VISNS Home ]