Focusing accuracy vs baselength

<p>So, I have two lenses that I like, a 50mm Collapsible Summicron and a 90mm f4 Elmar. I am looking at options for a faster 90ish lens. Also, I have been looking at different rangefinder bodies and playing with the equation for estimating focusing accuracy:</p>

<p>Max Aperture = (Visual Acuity*Focal Length Squared)/(Baselength*Circle of Confusion)</p>

<p>For 20/20 vision (natural or corrected) one can use 0.0003 mm for visual acuity and the Circle of Confusion value depends on how big you intend to make your prints. For 8x10s 0.03 works. For 13x19s it is more like 0.018. </p>

<p>Using an M3 and shooting for 13x19 prints, it looks like am M3 can handle a 50mm f1 lens with margin and barely manage a 90mm f2. <br>

<br />A leica with 0.72 magnification lets you stick to the 50mm f1 but at 90mm you need to go to f2.8 for 13x19s. </p>

<p>Using a Leica Cl it looks like it can handle a 50mm at f2.8 and a 90 needs to be stopped down to f8 or you need to stick to 8x10s.</p>

<p>For the Bessas the R, and R2s get you 50mm f2 but at 90mm you need f5.6 for big prints. </p>

<p>The Bessa r3 gets you to a 50mm f1.4 and a 90mm f4.</p>

<p>It is worth noting that for 13x19 prints it is a good idea to be around f4 to f5.6 to let the lens get to full power. </p>

<p>This matches my experience with my R2. I loved the handling but shooting my 90mm at f4 was dodgy. I am looking at getting a new Bessa R3 or an old M3. The Bessa R3 looks like it can handle my lenses, which is heartening. </p>

 

<p>I guess these calculations are important when the lenses are used wide open or nearly so. My own practice is to focus on the most important areas of the three dimensional scene and then set the DOF scale to encompass what I want in that scene to be reasonably sharp. I then often close down two more stops beyond the DOF reading. It usually allows at least a 13 x 19 print to be made, provided other factors are not controlling the sharpness.</p>

<p>In that regard, camera shake is a most important variable and why, when practical to do so, a monopod or tripod is used, or a high shutter speed implemented. Often overlooked also is the optimum aperture of the optic, which is usually about 2 stops down from wide open and is especially relevant for the corners of the image.</p>

<p>Some lenses work best with landscapes or distant architecture at the infinity setting. I learned this the hard way with an otherwise good lens, the 21mm f2.8 aspherical. By sacrificing a bit the resolution at moderate distances, the rendition of far off objects off axis (Ex. fine branches and sprouts of a tree) improved greatly by setting the lens at infinity and compensating by closing it down slightly for the moderate distance subject matter.</p>

<p>The equations show the best you can expect with 20/20 vision, a stationary subject, everything calibrated perfectly, high resolution film, perfect lenses, and no camera shake. As I look at some of the assumptions I wonder how accurate the 0.0003 mm acuity number is for different rangefinders, as that will be affected by the quality of the rangefinder optics. A Leica rangefinder's double image is much sharper than a FED's. </p>

<p>For landscape photography and other areas outside of close up focusing wide open these equations are less important. </p>

<p>The equations do explain why Olympus put the nearly token rangefinder on the XA (focusing a 35mm f2.8 lens requires about 8 mm) and why my tiny little Olympus RC makes such nice prints. It is also interesting to see clearly why I have been able to nail the focus where I want it on a fast 50 mm lens with my Leica iiif better than I can with an SLR. </p>

<p>Apart from the quality of the rangefinder mechanism (including mechanism play, and getting different results when focussing on the subject from opposite distant points), which varies from RF camera to RF camera, the effective base length is important (think of a triangulation diagram). Your example of different cameras suggests this point. Some manufacturers (inbuilt on the short RF base IIIf) offer supplementary focus magnifiers to enhance the precision of focus. I believe that stopping down a few stops after adjusting the DOF position is the best way to ensure high quality large prints. </p>

<p>The camera does not focus the lens: its aids make it possible for the user's eye to achieve focus. R/F base length is important, as I discovered when briefly I used Bessas (R2 and R2A): but perhaps it is not so important as we might think. I had two M3s for 19 years in all, but I find (even today, when I am rather over 60) that I focus as accurately with 0.72 bodies (M6 and M2) as I did with an M3. I use vision correction lenses at the view-finder.</p>

<p>I agree, and from personal experience <em>age is indeed a factor</em> in our ability to discriminate the perfectly focussed and the only nearly perfectly focussed RF image, thus reducing the importance of the M3 compared to the 0.72X viewer of the M6. In a way, this is sort of like extremely high notes in music that we once heard well, and now hear in the form of only the lower harmonics of the sound. </p>

<p>In the real world you can ignore all of that math. A SLR will allow you to take better shots wide open because you see just what the lens sees, but at slow speeds you'll need to lock up the mirror to avoid vibration. Some people can get sharp shots at 1/30 with ANY lens, the rest of us mortals usually pull a tripod out. Still, mirror shake, a restless subject, a nervous photographer, and many, many other things will determine whether or not you get a good image. I use a 35mm rangefinder for quick shooting (as well as scale focus cameras), the odd landscape in 35mm, etc. For 35mm portraits I use Nikon SLRs w/ Leica R glass, and if I want a big print I use the correct type of camera....a medium format camera w/ a big neg. Works a treat.</p>

At one stage in my life I lived in a rented room with no spce for a darkroom and bought a 35mm camera and used Kodachrome exclusively and viewed them projected on a screen. At that size even a slightly blurred subject was very obvious and objectionable. A self imposed rule of never setting a speed of less than 1/250 brought an immediate increase in the number of keepers. I have also compared depth of field tables from various lens manufacturers and found quite a lot of disagreement. I think Leica may be the most conservative in their recommendations and as Arthur suggests it may be wisweto stop down even further. These rules may make some subjects and lighting conditions impossible but so be it.

<p>In his book "The Camera" Ansel Adams says on page 116: "...using a normal lens with a hand-held camera, the slowest shutter speed that ensured maximum sharpness was 1/250 second." I will accept that advice.</p>

<p>Fair enough, for those who want the highest possible technical quality in a photograph treated as an artefact. For my part, a photograph is a representation of reality, and I shall take risks in order to get it. Today I made a dozen hand-held exposures at f/2 (50mm and 35mm lenses) and speeds of 1/15 sec. and 1/8 sec. I know that two at least will have to be thrown away because of subject movement. But the others, if they are good, should be valuable because of their content. Content matters more to me (and to many others) than form does. We are prepared to accept considerable technical imperfection if we see the possible rewards as substantial.</p>

<blockquote>

<p>Using an M3 and shooting for 13x19 prints, it looks like am M3 can handle a 50mm f1 lens with margin and barely manage a 90mm f2.</p>

</blockquote>

<p>That is my impression also. And in-between those extremes, it can also handle a 75mm f/1.4 equally well. Another thing: the camera and the eye behind it form an optical system. They work together as a team. A well corrected eye that is fresh and rested can focus the camera better than a tired one, or one that needs a visit to the optometrist!</p>

<p>Quit master bating about the lens theory, and try taking pictures to see if it works or not.</p>

<p>Thanks for sharing. Two very minor points.<br>

There is an error in your value for visual acuity. This is a dimensionless or angular measure. If you don't believe me that look at the equation you give. Since an F number is a ratio of dimensions it too is dimensionless. So the units on the righthandside should cancel out. If you make acuity a measure of distance they don't.<br>

<br />Your basics premise is the combination of depth of field and rangefinder error cancels. Basically as the rangefinder error increases with distance the depth of field compensates. While oft repeated I find this to be troubling for three reasons. One, in practice the details of such don't matter to most photographers. Two, it takes a simplistic view of depth of field which though very common isn't always so. Basically one needs to consider object space as well as image space. Three, it makes no allowance for infinite distance. The rangefinder error at infinite distance can't be a portion of infinity. So a more complicated theory is needed. But it isn't see One. </p>