Need tips on squaring camera

[mdd99]

When shooting buildings straight on, I have trouble squaring the camera to the building--keeping the film plane parallel to the front of the building. Invariably, I get a slight ski slope at the top and bottom. Any tips you can offer to set me, uh, straight?

[Robert A. Zeichner]

When shooting buildings straight on, I have trouble squaring the camera to the building--keeping the film plane parallel to the front of the building. Invariably, I get a slight ski slope at the top and bottom. Any tips you can offer to set me, uh, straight?

If you happen to own an i-Pod Touch, you could download a program from i-Tunes called Inclinometer for .99. This will allow you to measure the angle very accurately by just leaning the i-Pod against the front and rear standards.

That said, you might actually want some foreshortening to avoid a situation where the building looks too wide at the top. This is the visual impression one gets when the sides of a tall building are absolutely parallel. Our visual point of reference causes us to expect the building to be narrower at the top when viewed from ground level. The opposite it true when viewing from a high vantage point. On close ups of sections of buildings, parallel might be desirable because there is no sense of where the ground is so the building assumes more of a 2-dimensional shape anyway and that often looks better when parallel with the edges of the print.

[Bill_1856]

Use a ground glass or plastic overlay with a grid on it, so you can see clearly if everything is square or not.
Incidentally, many buildings actually look more natural with just the tinyest bid of convergence at the top.

[David A. Goldfarb]

A level and/or clinometer will get the camera square with the ground and square with itself, but a compass will tell you if the camera is square with the building, presuming you have a flat surface on the building that is itself square with the facade. In practice, this isn't always the case, so a grid overlay is another method of checking that the building is visually square.

I use a Suunto Tandem clinometer-compass to check these things. It will turn any floppy wooden or ultralight camera into a fairly precise instrument.

[Dave Aharonian]

I have bubbles levels mounted on both front and rear standards and I have a gridded ground glass. I find these pretty much do the trick for me.

[Brian Ellis]

Levels, a grid ground glass, and a lot of time making sure you have everything right are the only solutions I've found unless unless you print digitally, in which case you can usually fix the problem you describe in Photoshop. Levels are a big help but they aren't a complete solution because sometimes the building itself isn't perfectly level or parallel.

[Leonard Evens]

When shooting buildings straight on, I have trouble squaring the camera to the building--keeping the film plane parallel to the front of the building. Invariably, I get a slight ski slope at the top and bottom. Any tips you can offer to set me, uh, straight?

I use a flash shoe level I got at B&H. Orienting it properly while placing it directly on the ground glass, I can be sure that surface is plumb to an accuracy of 1/4 degree or better. Simple calculations show that an error that large would not be detectable as convergence of truly vertical lines without making careful measurements of distances in a print. They certainly wouldn't be detectable to the normal human eye.

On the other hand, there is no way to use a level to avoid convergence of truly horizontal lines. A level on the top of the rear standard can be used to be sure the top and bottom of the frame are horizontal to a high degree of accuracy, but that won't tell you if the camera is square on a building facade. If the facade is aligned perfectly in a known direction, you could in principle use a compass to make sure the lens axis is perpendicular to that direction. I've never tried that, but I have my doubts that you could get the required accuracy.

I have a grid on my ground glass, which was made by Maxwell Precision optics, and I've checked that it is accurate, i..e, with the camera level, the vertical lines are vertical and the horizontal lines are horizontal. I use the grid lines, together with horizontal lines in the subject to check squareness to the subject. (Of course, I also use the vertical grid lines to check verticals.) If you use a magnifier, and you can't see any convergence on the ground glass, then you shouldn't end up seeing any convergence in the print, at least if the film lines up properly with the edges of the standard. The film could be skewed slightly in the film holder, which would rotate the image slightly. But you can compensate for that either when enlarging, by positioning the easel, or when scanning, by adjusting the position of the film in the scanner.

Another way to check parallelism of lines is to use a metric scale to measure the distances between the lines at different parts of the image. You have to be careful that the scale is aligned properly in each case, by using grid lines, for example. But it is hard to read such a scale to much better than 0.5 mm. A convergence that large would in most cases be visible without any measurements, so this shouldn't be used as a primary method.

Of course, it is possible the building is not plumb or that its edges, vertical or horizontal, are not parallel. Builders are usually pretty good about that, particularly for modern buildings, but settling or other issues can upset things. So, if you are sure the camera is perfectly level, and you still get convergence of verticals or horizontals, after carefully making adjustments, consider the possibility that the problem is in the building. In that case, a photo-realistic ethic would suggest just showing it as it is.

[Ralph Barker]

I use a small plastic triangle that I picked up at the hardware store for something like $5 that has bubble levels on two sides. It works nicely most of the time for both assuring that the GG and front standard are plumb and for squaring the standards. Plus, it fits nicely into the base plate of the camera, so it's always there.

But, as Brian mentioned, if the building itself is no longer plumb, you still need to make adjustments for the image to look "right".

[mdd99]

All good thoughts--thanks. I have an Arca with grid lines on the ground glass, as well as two levels. I do check them closely. Adding to the squaring-up challenge is trying to position the camera in the midpoint of the building.

Here's the problem: After squaring up the camera by checking the top and bottom grid lines, I check the sides. If they're not perfectly vertical (or as in the attachment, the amount of space on either side of the object I'm focusing on isn't equal), I adjust to square them up. But then the horizontals are thrown off. Presumably this means I need to reposition the camera at the midpoint. It ends up being trial and error until I get something close to square all around. I could back the camera up, so the imperfections aren't as noticeable, but I like to shoot tight.

See attachment: The top isn't square, and the space on each side isn't equal (I don't mind the convergence). How would you solve this problem?

[bklynjohn]

Here's the problem: After squaring up the camera by checking the top and bottom grid lines, I check the sides. If they're not perfectly vertical (or as in the attachment, the amount of space on either side of the object I'm focusing on isn't equal), I adjust to square them up. ... How would you solve this problem?

I'm new to this and I could be wrong, but it looks to me that you're not parallel with the building and wall face. Your camera is meeting the vanishing point on the left of the frame. In other words, the the left part of the film plane as you are standing behind your camera is ever so slightly closer to your composition. Having a geared tripod head that allows for three independent movements is not the sole solution but it is helpful in times like this.

However... it's tough to know without also knowing the actual shape of the subject. It's possible that a simple front or rear shift would fix the picture entirely.

[mdd99]

...it looks to me that you're not parallel with the building and wall face.

Yes, that's the problem I'm having--setting up the camera so it's square to the building.

[Brian Ellis]

All good thoughts--thanks. I have an Arca with grid lines on the ground glass, as well as two levels. I do check them closely. Adding to the squaring-up challenge is trying to position the camera in the midpoint of the building.

Here's the problem: After squaring up the camera by checking the top and bottom grid lines, I check the sides. If they're not perfectly vertical (or as in the attachment, the amount of space on either side of the object I'm focusing on isn't equal), I adjust to square them up. But then the horizontals are thrown off. Presumably this means I need to reposition the camera at the midpoint. It ends up being trial and error until I get something close to square all around. I could back the camera up, so the imperfections aren't as noticeable, but I like to shoot tight.

See attachment: The top isn't square, and the space on each side isn't equal (I don't mind the convergence). How would you solve this problem?

There is no simple solution. I'd solve it the same way I mentioned before - grid ground glass, levels, and a lot of time.

In addition to the problem I mentioned before of the building itself possibly being off, you have the problem you mention of the camera itself aimed every so slightly to the left or right of the building. The camera can be perfectly level but if the camera itself isn't positioned properly in relation to the building then you have a problem (which is why levels are helpful but not a complete solution). And when you look at the grid and see that things aren't quite right it isn't always easy to know where the problem lies, i.e. is the buidling itself out of whack or is your camera not perfectly level front to back and side to side or is the camera not in correct relationship to the building? That's where trial and error come in and why this isn't easy (at least for me, maybe professional architectural photographers have found better solutions that I have).

I've heard people suggest solving the problem of the camera not being correctly aligned relative to the building by measuring the distance from the left and right tripod legs to the front of the building. I've never tried this because I'm usually too far from the building to make measurements practical.

When photographing buildings, windows, doors, etc. I used to spend as long as maybe a half hour fiddling around with the camera, trying to get everything aligned perfectly. Fortunately I don't need to do that any more because there's an easy way to fix most of these problems in Photoshop, especially if you anticipate a possible problem at the time you make the photograph and leave a little extra room on the sides and top/bottom for a little cropping. But I still spend time because I try to get it as close as possible in the camera.

[Alan Davenport]

It looks like you had the back swung a bit to one side (relative to the building facade.)
Use a swing to straighten the top of the building, then use shift to center the image.

I carry a small (6") torpedo level in my camera bag and use that for leveling the camera. It's much more accurate than the tiny levels on the camera.

[Leonard Evens]

All good thoughts--thanks. I have an Arca with grid lines on the ground glass, as well as two levels. I do check them closely. Adding to the squaring-up challenge is trying to position the camera in the midpoint of the building.

Here's the problem: After squaring up the camera by checking the top and bottom grid lines, I check the sides. If they're not perfectly vertical (or as in the attachment, the amount of space on either side of the object I'm focusing on isn't equal), I adjust to square them up. But then the horizontals are thrown off. Presumably this means I need to reposition the camera at the midpoint. It ends up being trial and error until I get something close to square all around. I could back the camera up, so the imperfections aren't as noticeable, but I like to shoot tight.

See attachment: The top isn't square, and the space on each side isn't equal (I don't mind the convergence). How would you solve this problem?

I think you may be doing it in the wrong order.

In a case like this, you want the ground glass parallel to the front of the building, and any centering should be done by a horizontal shift of one of the standards. It has nothing to do with whether of not the sides are vertical, at least if we assume the building front is a rectangle with the top and bottom parallel to the horizontal.

In your example, the top of the bottom of the yellow trapezoid appear to be parallel as best I can tell. Let's assume that trapezoid is the image of the building you are talking about. Let's assume the top and bottom of that building are parallel in reality, which should almost always be the case. Let's also assume the sides of the building are vertical and parallel in reality. Then if you make sure the ground glass is plumb, it will be parallel to the front of the building, as noted above, and the images of the sides have to be parallel on the ground glass.

[Leonard Evens]

Apparently others are seeing something different from what I saw. I downloaded the image, but the resolution wasn't high enough to be sure. I thought the images of the top and bottom on the ground glass were parallel, but if indeed they aren't, there are two possibilities. They aren't parallel in reality, which I would find unlikely, or the back is not parallel to the building front. In the latter case, after leveling the camera, you have to turn it slightly until the top and bottom of the building appear parallel to the horizontal grid lines and to each other. Don't worry about centering in this process, as long as the sides remain visible on the gg.. You can address that later by using a horizontal shift.

If the front of the building is not actually a rectangle, then your best bet is to use a compass, with a flat surface. You would put it on the building, note the setting and then put it on the ground glass and turn the camera until the setting is the same, keeping in mind that the compass is now pointing in the opposite direction. If the camera is level, then you will get as accurate a rendition of the building facade as you can, and the images of the top and bottom or the the images of the two sides may end up not being. parallel.

Personally I doubt that the building front is not a rectangle. You are just having trouble judging when the top and bottom are parallel to the grid lines. To do this, use a large 2 X magnifiers so you can see better. Also, use rise fall to move grid lines directly over the top or bottom. (You can reframe afterward.)

It is of course remotely possible that your grid lines are not parallel, so you should check that. The best way to do that is to find a line you know is perfectly horizontal,and, with the camera perfectly level, adjust rise.fall and the direction of the camera until one grid line lies right on top of the image of the horizontal line and matches it directly. If necessary adjust the horizontal leveling slightly to get that right. Then, using rise/fall, move the camera so that the other grid lines also lie on the horizontal line. they should all match the horizontal line when one of them does.

[cjbroadbent]

Ralph's plastic triangle level held against the the GG + a GG grid for the camera, to start with.
I got laughed off the thread when I last mentioned how I get the camera centred with respect to a set, a background or a building. I'll say it again nevertheless.
I use a gardeners string from the corners of the whatever it is to mark a perpendicular line (a point on 2 arcs, another point on 2 more arcs further off and a line joining the points). I can move the camera up and down the line and always be in the centre. Then I place two little flashlights on the extremities of the whatever and align them on the GG grid. It takes a couple of minutes.

[Colin Graham]

I use a grid with a center bullseye. Easy to tick off grid marks on either side to assure symmetry, the rest is swing and shift. If your standards are machined precisely in relation the the gg, you can sight along them to find parallels in all planes when setting up. In other words, before even looking through the gg, use the standard planes to reference the corresponding subject planes, sort of like the winding sticks used in woodworking to see if a board is flat.

[Leonard Evens]

All good thoughts--thanks. I have an Arca with grid lines on the ground glass, as well as two levels. I do check them closely. Adding to the squaring-up challenge is trying to position the camera in the midpoint of the building.

Here's the problem: After squaring up the camera by checking the top and bottom grid lines, I check the sides. If they're not perfectly vertical (or as in the attachment, the amount of space on either side of the object I'm focusing on isn't equal), I adjust to square them up. But then the horizontals are thrown off. Presumably this means I need to reposition the camera at the midpoint. It ends up being trial and error until I get something close to square all around. I could back the camera up, so the imperfections aren't as noticeable, but I like to shoot tight.

See attachment: The top isn't square, and the space on each side isn't equal (I don't mind the convergence). How would you solve this problem?

Okay. I've downloaded the latest version, and I agree that the top of the building is not parallel to the top of the frame. There are several possible explanations for that.

1. The camera is not level from side to side. That should be easy enough to check using a level.

2. The top of the building in reality is not level. I think that is unlikely, and I addressed it previously.

3. The ground glass (and hence the film plane) is not parallel to the front of the building. Usually this is fairly clear because you have two horizontal lines in the subject, the images of which should be parallel, but instead they converge, and that helps us distinguish from case 2. But in this case, it is hard to identify a second horizontal line.. There is one below the grass at the bottom, but I can't tell whether that represents something in the subject or is just the bottom of the frame. The line for the lower roofs to the left and right also tilts to the left, but it is too close to the other line to tell if there is any convergence. If indeed we are genuinely in this case, and the camera is level, then the solution is to rotate the camera slightly to the right in order to get the top of the building parallel to a grid line and the top and bottom edges of the frame.

As I noted before, centering is not relevant. It can be accomplished by a horizontal shift any time one wants to. If the subject showed a building side in addition to the front, it might be preferable to move the camera.

[neil poulsen]

I have a 229 (formerly 3039) Bogen tripod head with adjustable levels in each of the two horizontal directions. I adjust these levels with the camera mounted on the head so that the ground glass is vertical and level. I rely on the camera detents to keep the front standard oriented properly with respect to the back.

With this setup, camera alignment is quick and easy.

[arca andy]

I used to photograph a lot of paintings for one of the large auction house in London and this is how we got a painting square. It works the same for buildings.
1 By eye roughly aim the camera at the centre of your building/painting.
2 Make sure you have all the building/painting in your ground glass screen (ggs) dont worry it won't be square
3 look at the top of the ggs (bottom of the building) and now level the grid on the ggs with the bottom of the building...so good so far.
4 look at the bottom of the ggs (top of building), the chances are the top of the building will not be level with the grid, don't worry...
5 Now here's the clever bit, depending on which side of the buildings centre you are on by moving your camera either to left or right and repeat the above process, you will eventually get the both the top and the bottom of the bulding level with your ggs grid..
6 Then deal with the verticals in the usual way by rear tilt and front raise

This only works if you want shoot dead square to the middle of the building...and of course if the building is square....Practice with a large square bit of card at home first...once you get it its easy.
Andy

[Henry Ambrose]

I walk around without a camera and find the spot to stand. When I get there I look at the subject carefully. I move steps to the side or back or forward and find "the" spot. At least that's what I consider it until I get the camera out. :D

Then with the camera "zeroed out" and with a gridded screen I just look at the image in the camera. Its right there -- so look at it. Study it. Now study it a bit more to be sure. Use the lines, that's what they're there for. Don't get lazy, move the camera until you are done.

If you're having a bad time, follow Christopher Broadbent's string method. Once you are near plumb and center of the subject building it gets easier.

Please note that you will rarely find a building that is perfectly square.

When the subject lacks perfection you might stand somewhere else. This especially when you have a series of surfaces or openings that aren't quite planar. Pick a place to stand that makes this obvious. A faint variation or slight misalignment most often looks wrong. Which I suppose is the subject of this thread.

[gevalia]

Buy a small carpenters level at a hardware or lumber store.

Yeah, got a small 6" level at home depot. I think they call the type a "submarine level". Anyway, I carry it all the time. My Chamonix has small levels on its standards but the fluid is very viscous and takes forever to move. The level I bought is cheap and works great.