View Full Version : Shortcut for testing bokeh ?

Ken Lee
1-May-2007, 11:17
I have tried this as a shortcut to shooting film. Will it lead me astray ?

Find a smooth white wall on a sunny day. Get a lens, remove the caps, open the diaphragm, and point it to a subject that contains a reflection of the sun. When in focus, the sun should be a very small white spot.

Move the lens further away from the wall, and closer to the sun. The image on the wall will get larger, and the spot will now be a circular disk. If the disk looks like a ring, or has a ringed edge, you have bad bokeh. If the disk is not circular, you've got coma. If the disk is uniform, then you have nice bokeh. If the disk has very soft edges, you have very nice bokeh.

Repeat the process by moving the lens closer to the wall than ideal focus. Now you have tested both regions of focus.

Try the same thing with the lens stopped down. See how much the diaphragm blades affect the shape of the disk. If the disk is now shaped like a polygon, you've got bad bokeh. If the disk is still fairly uniform, you've got nice bokeh.

This gives rise to another question: Given that a diaphragm with many blades is desirable for this sort of thing: is there a way to replace the diaphragm of a modern lens, mounted in a stock shutter - while retaining the modern shutter ?

Jack Flesher
1-May-2007, 11:24
I am sure many will disagree with me, but I personally do not think Bokeh and rendering of spectral highlights -- whther in focus or out -- necessarily always correlate...

Gordon Moat
1-May-2007, 11:31
I think one issue is shooting into the sun. It can be a generally bad idea, unless you really have on other way to get the image you want. Even legendary lens can show up harsh or bad looking defocus area distractions when pointed at a light source.

I think an easier test would be to set up objects with sharp edges at various distances relative to each other. Then photograph the various objects, and decide how you like the defocus rendition. You can then expand further by photographing the same objects at differing camera to subject distances.

My own way of doing this, in any format, is simply to use the lenses in ways I would normally shoot. Quite often that can be up close to mid distance, rarely infinity focus, and often near wide open aperture. I rarely ever shoot with point souce lighting in the background areas, nor towards the sun. Of course, other people might have different patterns of use than I do, so I might not be a good person to ask about lens recommendations.


Gordon Moat
A G Studio (http://www.allgstudio.com)

Ken Lee
1-May-2007, 11:49
Sorry if I was unclear. I'm not looking for a good lens to shoot images of the Sun.

I use the Sun, merely because it is small, bright, round (if you believe in that sort of thing) and readily available. :-)

I am trying to do all of this without having to shoot/develop/print/inspect an image. Even better, forget the camera altogether.

Michael Heald
1-May-2007, 11:50
Hello! You could always do a star test. Put the camera on a tripod and point it at a pole from 15 to 45 seconds depending on the focal length of the lens. The star images should be pinpoint thorughout the image for a perfect lens. Usually, there will be lens defects that become apparent at the corners. Stop down and repeat the test till all images are pinpoint. This is the optimum f-stop for the lens. You can repeat with the lens focused inside of infinity for our bokeh test as well, and compare the shots to your infinity focus test. Best regards.


Ken Lee
1-May-2007, 16:25
Hello! You could always do a star test. Put the camera on a tripod and point it at a pole from 15 to 45 seconds depending on the focal length of the lens.

Sorry, I don't understand: To what kind of pole are you referring - and how does that relate to stars, or star-shaped items ? Are you talking about the North or South pole-stars ? By 15 to 45 seconds, are you referring to a measurement of arc ?


Ted Harris
1-May-2007, 16:32
Ken, wouldn't be easier to just shoot some Polaroid? Whe I was writing the article on variable focus lenses a year+ ago that is what I did. I must have burned some 80 sheets of T55 and T72 while looking at the results, including the bokeh of course of 5 different lenses. Worked fine for me.

On the shutter blades thing, look for modern Compur shutters too .... they have a lot more blades and a much rounder opening ... but don't look too hard I need more :).

1-May-2007, 18:56
Just on shutter/apertures and bokeh, I have a 1900 something studio shutter that has a very "un" round opening yet the rendering of the oof areas is wonderful. In fact I would not be dissappointed if all my modern lenses rendered the backgrounds in the same way. Of course the actual glass plays a very large part in the final result.

Paul Metcalf
1-May-2007, 20:22
Not a bad plan, and would give you some insight into a given lens performance. My (limited) experience with using LF lenses wide open is that the distance to the object plane of focus makes a difference in how things look. Most (if not all) of my lenses (some moderen Japanese, some modern German, a few older Japanese, and a couple of very old ??? e.g. who made Conley agnastigmats) have good to great bokeh in the near distances from the object plane of focus, and a few have good to great bokeh in the distances just beyond the OPOF. Then there's a distance where things look not so good (especially if they are highlighted), and then things finally go very soft to non-distinguishable beyond that distance. So, as with most things of a technical nature, there are a lot of variables to try and get control of. I'm always very impressed with the images posted where these variables are under control (maybe intentional, maybe not). Good luck, don't let it keep you up at nights (unless you're into nighttime photography!). Paul.

Oren Grad
1-May-2007, 20:35
I think the glass is as important as the iris shape, if not more so. But we've had that discussion here before, with inconclusive result from a debating perspective because of insufficient objective evidence. I do have plans for a controlled comparison test of different modern plasmats in identical shutters/irises that should demonstrate how much of a difference the glass alone makes, but given other commitments it may be a little while yet before I can get around to doing it.

While looking at the rendering of point sources can help in understanding the optical principles at work, for selecting lenses to achieve esthetic goals there's no substitute for making real pictures. Subjects should be stuff you'd normally photograph where bokeh would be an issue for you. For me, the acid test is how a lens performs at middling aperture settings and focused at midrange, on a scene filled with trees that slowly fuzz out into the distance. The best lenses (for my taste and my purposes) turn the trees slowly into a silky, feathery blur that retains coherent form well into the OOF area; the worst render the OOF trees as a jangly double-line hash or as lumpy oatmeal.

For other people, the more important criterion may be how isolated, distant OOF backgrounds are rendered, be it in portrait or landscape work. Or how relatively close backgrounds are rendered in studio portraits. Or how OOF areas behave in close-up still life work. Or whatever.

Lenses that do well at one of these tasks won't necessarily be best for others. Also, I don't know of any lens, even among my favorites, that doesn't have some situations in which the image character has unpleasant elements; and conversely, even many lenses that I don't much care for overall, have situations in which they do just fine. So don't get carried away in pursuit of a mythical lens with perfect bokeh. Figure out what aspects of OOF character matter to you for the kind of pictures you like to take, and make your own judgments of fitness for purpose from test pictures taken in those situations.

Michael Heald
2-May-2007, 08:39
Hello! The star test is a very sensitive test for how a lens performs at infinity. The exposure runs 15 seconds to 45 seconds, depending on the focal length. The longer the focal length, the shorter the exposure has to be to ensure that there are no star trails. I used TMax 400 pushed to an ISO of 3200. The test can also be repeated with color filters, if there is a suspicion of chromatic aberration. I had to go through this procedure when I tested my 178mm Aero Ektar. It had a lot of chromatic aberration and coma, but when I stopped down from f2.5 to f5.6, the lens was perfect at infinity focus.
Reeve's book on astrophotography talks a lot about lens testing, the exposure times, etc.
The camera can be pointed at any part of the sky, but aiming it at the north or south pole stars increases the time of exposure that can be made before star trails appear.
Generally, wide open, every lens has defects, and the abnormal star shapes will be most apparent at the corners of the negative. Stopping down eliminates or minimizes many lens aberrations, and repeating this test at different f-stops will give the f-stop that produces the sharpest image at infinity.
I'm not sure about Bokeh testing. Since you suggested imaging the sun on either side of focus, you can do the same with the star test, except you'll know what the best f-stop is for your particular lens before you try it. Best regards.


Struan Gray
2-May-2007, 12:47
One good test for the sorts of image I like to make is a set of fairy lights or a string of Christmas tree lights in a line. Lay the string on the ground or a tabletop and focus on a middle light. The shapes of the out-of-focus lights give you an instant depiction of the blur in front and behind the plane of focus.

Were I to expose film in a test, it would be simple enough to rotate the camera and put different photos of the same string of lights at different aperture settings onto one negative.

Mentally translating from those 'spot diagram' shapes to pictorial effects comes easily to this imaging scientist, but might not be obvious to non-specialists. If you are happy with the ideas of convolutions in image processing you won't get any nasty surprises. If you cannot parse that sentence it might be best to take some real-world test shots.