New Christopher Perez research on "bokeh" + more

[Ken Lee]

I found Christopher Perez's latest article to be very interesting and well-written. See Six Lens Matchup - an ambitious test (http://www.hevanet.com/cperez/test/BigMash210.html" target="_blank).



He compares 6 lenses of various design, all around 210mm, and draws some surprising conclusions. Emprical results can be persuasive :-)

[Andrew O'Neill]

Very interesting! The very last line: "Perhaps the best realization I had in doing this test is that aperture shape influences more than just out of focus highlight shape and rendition. It defines everything about out of focus area rendition, shape, and texture."
I've often wondered about this.

[Armin Seeholzer]

But I would like if he would test also the newest lensdesigns like Apo Symmars, Apo symmar L and APO Sironar S and he would then also com to the conclution thad this new glasses are best at around f11!

[Phong]

Most interesting indeed that the "aperture shape was the only influence on the rendition, texture, and shape of out of focus areas" !

[Andre Noble]

I have Bronica portrait lenses with only 5 aperture blades that give beautiful bokeh.

[William Mortensen]

Perhaps a more informative test, given the post-test statement on aperture shapes, would be to test one or two lenses at, say, f/11 and f/22, but unscrew the front element to replace the diaphragm with waterhouse stops in the shape of a triangle, square, pentagram, and circle, and compare these to what the factory aperture blades yield.

[Oren Grad]

unscrew the front element to replace the diaphragm with waterhouse stops in the shape of a triangle, square, pentagram, and circle, and compare these to what the factory aperture blades yield.

Mark, take a look at Harold Merklinger's article, in particular figures 2-5 and the associated discussion in the text. Not exactly what you're asking for, but very much to the point:

www.trenholm.org/hmmerk/ATVB.pdf (http://www.trenholm.org/hmmerk/ATVB.pdf)

[Ken Lee]

For those that are not familiar with "Waterhouse Stops", here (http://www.skgrimes.com/wat/index.htm" target="_blank) is a custom set of Waterhouse stops, crafted by SK Grimes.



Like so many modern "improvements", the adjustable iris, composed of blades, seems to have favored speed and convenience over quality.

[Mike Kovacs]

I've yet to the read the article in any great depth, but does anyone else think that the train image used to analyze bokeh isn't really the best choice? Its those backgrounds with lots of highlights and fine details like sun-lit trees that usually show the worst effects IMO.

Will have a good look though - thanks to Chris again for putting in the time and hard work.

[Christopher Perez]

Mike,

Here's my justification on subject matter: I wanted to observe not only the effect of aperture shape on highlights (easily accomplished with out of focus trees and sky intermixed), but on the overall tonality and rendering of the softer dark to light transitions. For this, the train worked well.

If you haven't already visited it, toward the end of the document is a link to my raw scans. They'll give a hint at the differences in texture and rendition.

I hope this helps.

[paulr]

i'm suspicious of these conclusions, based on more systematic examples that i've seen. there's strong evidence for the spherical aberration hypothesis. namely, lenses that generate "nice bokeh" ... disks with soft edges ... tend to do it either in front or in back of the plane of focus but not both. a lens with perfect correction (the laboratory ideal) will render an out-focus-point as a smooth edged disc in front or in back of the plane of focus ("neutral" bokeh). imperfect correction leads to undercorrection on one side of the plane of focus (nice and smooth) and overcorrected spherical aberration on the opposite side (hard and ugly). i don't see how aperture shape can influence this, although it well may influence other aspects of the unfocussed image.

[Jason Greenberg Motamedi]

Thanks for doing the work Christopher!

I think Paul is correct in writing that " ... imperfect correction leads to undercorrection on one side of the plane of focus ... ". This is certainly the case with many portrait lenses. It is worth noting that all of the lenses tested were much closer to perfect than imperfect corrections. Tessars and Plasmats are notoriously well corrected, and thus wouldn't tell us very much about under or over correction. While I realize that it wasn't the point of this exercise, it would be nice to redo this test including a Petzval, Rapid Rectilinear, and an Achromat.

As for aperture shape, the point being, at least in my experience, is that that the "disks with soft [or hard] edges" are no longer disks, but pentagons or hexagons with "soft [or hard] edges".

jason

[Oren Grad]

Tessars and Plasmats are notoriously well corrected, and thus wouldn't tell us very much about under or over correction.

That's not quite right. Late-model plasmats from different vendors vary in whether the bokeh looks smoother in front of or behind the plane of focus. The Nikkor W design, for example, gets frizzy in the backgrounds, a phenomenon that Harold M., in the article linked above, attributed to over-corrected spherical aberration.

About the test itself - I add my own thanks to Christopher for going to all the trouble and posting his findings in such detail that others can study them carefully too. I guess I have only one fundamental quibble, but it's an important one. The final statement in the discussion...

Perhaps the best realization I had in doing this test is that aperture shape influences more than just out of focus highlight shape and rendition. It defines everything about out of focus area rendition, shape, and texture.

...is untenable as a generalization. Aperture shape cannot define "everything" about the out of focus area, because different lenses in identical shutters can produce radically different OOF effects. That leaves the question of the relative effect of aperture shape vs optical design. My sense from experience is that the glass has a substantially larger effect, but the skeptical reader is certainly entitled to await further empirical evidence on that point before being swayed.