View Full Version : Focusing technique for very wide angle lenses
I have recently been shooting architectural interiors, and more often been worki ng in very cramped, quarters with either a 47mm (XL) or 75mm lens. With both le nses (the 47mm in particular), the light drop-off is dramatic towards the edges. The interiors are typically not that bright to begin with which further compou nds the dimming effect. More often than not, the near and far critical focus poi nts (I use the 2 point method) are at these dim edges of the image. My camera i s setup with a Boss screen which I find better for focus, and fairly difficult f or visualization due to the inability to view the whole image without moving you r eye at an angle to the screen around the outer edges of the image.
In order to focus at points near the edges of the screen, you must orient your l oupe at an angle to the screen to be in-line with the direction of the light ray s. I'm finding tilted loupe focusing to be somewhat error prone.
I am considering another focusing process with these lenses and would like other people's feedback. I know what I'm proposing may be somewhat non-pure from a m athematical standpoint, but would be interested if others think it is of practic al use.
With these lenses the center of the image is quite bright and very sharp on the screen. I am still proposing using a 2 point method, but reorient the camera on the tripod so it's directed at the near focus point and then do the same for th e far focus point. After arriving at focus, then frame the picture for the shot .
This certainly makes focus *much* easier with these lenses, however is there som e fundemental flaw with this approach?
I'm not sure that I understand your method. I think you mean: (1) point the camera so that the near point is at the centre, (2) get that point in focus, (3) re-point the camera so the far point is at the centre, (4) tilt/swing to get that point in focus, (5) re-point the camera for the framing you want.
That method won't work. Steps (3) and (5) will, in general, take one or both points out of focus.
This is how I focus the 47 XL. I use a dark cloth, of course. I also use a strong pair of reading glasses (approx dioptre 5, f=250mm) for evaluating composition, and an 8x loupe for focusing. For the centre of the image, I have the loupe flat on the glass. For off-axis, I reverse the loupe, allowing me to have it at a decent angle.
1. I decide what plane in the subject I want in focus. Almost always, one point in that plane will be in the centre of my negative.
2. I visualise where that plane intersects my lens plane, and I tilt the back as appropriate for Scheimpflug. Wearing the glasses, I check the overall composition and focus, adjusting if necessary.
3. I focus the centre of the image, with the loupe.
4. I check the focus as far up the image as I can go. The back tilts on-axis, so tilting to correct this point doesn't affect the centre much. But it does a little, so I re-focus the centre.
5. Repeat (4) for a point as far down as I can get.
6. Repeat for points to the sides.
7. Repeat steps 3 to 6 as necessary.
I tilt and swing the back, because it's easier on my camera, my photos are not architecture, and I'm not bothered about the very small amount of false perspective this gives. For architecture, you may want to ensure the back is parallel to your object, and tilt/swing the lens.
I shoot a lot of interiors with those lenses as well as the 65. What I've found is this...they are wide enough to give a large amt of depth of field IF they are correctly focused. The cramped quarters (usually backed up against a wall, or something) and the light fall-off, (or loss if I'm using a center filter) make it an exercise in futility to use conventional focusing methods. What I do is..1. evaluate the scene, what do I want sharp..and not sharp. (Often I want the immediate foreground to be out..never the background). 2.Choose a shooting angle, and straighten the back so the verticals are straight. 3. Focus on a point approx 1/3 into the 'zone of sharpness' (if you will) with an 8x loupe. 4. Shoot a Polaroid test. I exclusively use type 55 PN. I allow it to develop enough to give me a full negative. I use the neg, with my 8x loupe to check exactly what is sharp and not (place the loupe on the back to avoid the goo), then re-adjust the focus point, if necessary. I always use an aperture between 11.5 and 22.5 as this this is the optimum working zone for my lenses..minizing edge fall off and flare. If I find it necessary to use a front tilt or swing..it is VERY slight. With lenses this wide, camera movements are tiny. 5. The PN print will give a good composition, lighting, exposure, and corner fall-off check. I sometimes find it necessary to add some additional light to the bottom to compensate for the lens fall-off. Generally, I'm looking at ceiling lights on top..so the top fall-off isn'y usually a problem. Remember that the PN material has a different ASA for tungsten and flourescent (32) than daylight (50)
Re: previous two responses. Thank you, I now understand why my proposed method is inaccurate. Thank you for some of other tricks on how to approach this problem. I agree that in general, conventional focusing techniques are a problem with ultra wide angle lenses.
Since there isn't often an object in the center of the image 1/3 the distance into the scene, I guess a mini-lightstand with a focusing target attached which could be placed at this point would be useful.
The previous response brings some other questions to mind. I have started using Polaroid Type 55 and find it very useful for focus. Where I find most Polaroid materials funky for exposure "proofing" for interior photography is with it's reciprocity characteristics. Fairly long exposures (up to 1 minute) are also common with this type of work. Any advise on how you rate Type 55 when exposures get long?
The 47XL Center Filter also seems to be a mixed blessing at times. I'm finding strobe fill in these cramped quarters can sometimes create unusual bright spots in the image when using the CF. Removing the CF makes them go away (but then you have to manually compensate through your lighting).
I would second the suggestion to use a light stand with target as a focusing aid. This works very well. I also have used a 1k tungston spot to blast light into regions of the scene as well. Have an assistant move the light as you view different areas of the scene, then don't forget to kill it before the exposure [uh, personal expereince speaking...oops!] The use of polaroids is with-out question your ultimate answer though.
The reciprosity charisteristics of Type 55 are available in a tech data sheet from Polaroid..But, as I recall, the different ASA seems to take care of that. The lighting proficency comes with experience..but a tip....I switched my Copal #0 shutters on my WA lenses to the Copal "Press" type which does not require cocking between exposures. This allows me (assuming a solid tripod) to "build up" the exposure with multiples reliably. Unless the interior is mixed daylight with artificial..I generally light with Lowell tungsten lights, and filter the camera (or gel the lights as needed)
Another thing...experience has dimmed my previous enthusiasm for super wide lenses. Even though they're super corrected for barrel distortion..etc...the perspective is innacurate..especially in tight quarters. The knack, is to represent the spirit of the scene (by not necessarily showing every inch) with a more normal lens..a 90 or even a 75 which renders a scene closer to what we see. The problem is, of course in the translation. Our eyes build up an image in our brains by looking @ a series of points..with the ability to zoom. The lens renders a flat scene all at once...so we have to translate this with trickery and magic
I find that the reversed lupe trick does a good job but as you have found the angle to the groundglass/fresnel combination is pretty extremely acute in the far reaches of the ground glass so I check it about a third of the way in, and then shoot a type 55 and check the entire negative. A little messy but it works. Of course, you have to make sure that your Polaroid holder, your film holder(s) and your groundglass align with each other, but that is a pretty simple test.
I find the simplest way to focus in low-light situations with wide angles is to carry two small quarzt-halogen flashlights with me. I put one at the near focus position within my photo, turned on and facing nearly directly towards the lens, and the other at the far focus point, facing back also. Then with a lupe I focus on close and then far light pinpoints, getting a feel for where each point is on the rack-and-pinion, and then settle at the one-third/two-thirds point. I carry a BTZS exposure computer, and I use it to figure depth of field taking into account what lens I am using and the two focus distances (how far the two flashlights are from my camera). It is good only where you have the ability to walk in and out of the scene pre-shooting, but it works time and time again.
C Matter: when you say "perspective is innacurate" for superwide lenses, do you mean that the perspective is somehow wrong, or merely that we tend to observe the final image at the wrong distance to give natural perspective?
Where can I get eyeballs that zoom? :-}
I like the flashlight technique. I must try it out.
Re: Rob Tucher
"Then with a lupe I focus on close and then far light pinpoints, getting a feel for where each point is on the rack-and-pinion, and then settle at the one-third/two-thirds point."
The Rodenstock calculator suggests that you pick the point along the focus track which is exactly half way between the near and far focus points. I've always assumed that this point actually represents the 1/3 in front - 2/3 behind point. When I feel the need to be extremely accurate, I use a slide caliper to measure the relative displacement of the rear standard along the track. Is this half-way point method indeed correct?
The flashlight trick sounds very promising.
Through trial and mostly error I have also found that the "halfway in between point on the focus track" above mentioned method works best.
Larry: yes, exactly halfway is the usual technique for getting equal unsharpness for the two points. To be precise, it isn't exactly halfway (you would have to look at a diagram to see why), but it is close enough.
The one-third/two-thirds rule-of-thumb is for subject distances, and only applies to 'normal' distances, i.e. neither infinity nor macro. The half-way rule applies to the lens-to-film distance, and is the reason why DoF markings on lenses are symmetrical.
re: "Larry: yes, exactly halfway is the usual technique for getting equal unsharpness for the two points."
In fact I just completed an emperical experiment on this issue. My experimental setup is as follows: Shot Type 55 (4x5) with a 210mm Rodenstock Sironar-S. Test target was an 8.5 x 11 sheet of paper with repeating text (Times-Roman 12pt) from top to bottom (46 lines). Target was positioned on a stand with an angle of 39 degrees. The furthest ways lines (lines 1, 2, 3 etc.) were at the top, and the closest lines were at the bottom (lines 44, 45, 46). Reproduction ratio was about 1:1.5. All exposures taken with strobe at 1/125 sec. Camera leveled, and pointing squarely at the target (to a good approximation). Type 55 negs examined with a 10x hastings loupe. All focusing done with rear standard. Camera was a Canham DLC45.
Since I seem to be having endless critical focus issues, I first focused on line 24 and shot it at f=5.6. I wanted to see if there is some problem with my 545i holder. Test negative was right on the money. Only line 24 was sharp.
Next experiment was to use the two point method mentioned in previous posts. First I focused on line 13, and then line 27, measured the rail displacement (14.7 mm) with a caliper, and set the rear standard half-way. Incidentally, line 20 was in focus without the lens stopped down at this rail position. I then consulted my handy Rodenstock calculator, and determined that f=45 would be required (even though the experimental setup is a perfect situation to use front tilt, I didn't tilt because I wanted to limit the experiment to a smaller set of variables). The negative from this setup was interesting. Under the 10x loupe, I wouldn't rate either lines 13 or 27 as being in good focus. The range of good focus was closer to lines 16-17 at the far end, and lines 24-25 on the near end. Obviously, for my taste the Rodenstock calculator must assume a C of C larger than I'd find acceptable. To Allan's point, you could easily rate line 20 as being the sharpest, and everything on either side was to some degree a compromise in sharpness. I knew this was the theory, but was suprised that I could actually observe this fact. Doing some simple trig it also appeared that the zone of sharp focus was pretty much symmetrical around the center point (not following the 1/3 - 2/3 rule).
As an aside, I put a scrap Fuji Quickload in the 545i and pulled the darkslide. It appears that the 545i keeps the film flat everywhere except on edge near the rollers. The design of the 545i (and 545) doesn't appear great in that respect. I haven't used the Fuji mechanism, so don't know if this issue is only when using the 545i.
Now, if I could only solve my overall focusing problem. When using the two point method and the Rodenstock calculator, more often than not, my focus is excellent in the foreground and slightly out of focus in the background. Typically, the range of focus I'm trying to achieve is fairly large (requiring f=22 to f=45). I've checked the position of my Boss screen and film holders, and everything seems OK.
To those who read this far back. I use a 80mm enlarging lens as a loupe to check my focus on ground glass. Theyn are inexpensive ($15) and short. They can be held at an angle to check for focus at the edges of the ground glass and still be sharp, unlike the expensive schnieders and toyos. Try one and you'll dump your expensive glass.
Powered by vBulletin® Version 4.2.0 Copyright © 2013 vBulletin Solutions, Inc. All rights reserved.