More bokeh? 200mm @ f/2.8 or 100mm @ f/2.0?

[Chuck Lee]

William Castleman has done some testing and written an article about Bokeh and Background Blurring with Canon Lenses. See


http://www.wlcastleman.com/equip/reviews/aps-c_port/bokeh.htm
<div style="CLEAR: both"]</div>

That's a good read Johnny. I've seen that before.


His comparison of the 50 1.4 vs the 50 1.2 is really good as well.


Thanks for the link.

[Colin]

Something that bryan pointed out about bokeh and focal length is that telephoto lenses may or may not have less blur, but because of their field of view, given the same framing on the subject, they will have less background detail to start with, and as such the blur will seem like more. In other words, longer the focal length, there's less to blur out. Even if you could adjust aperture to match depth of field, the blur will be qualitatively different between focal lengths. longer focal lengths will likely have a more uniform blur with less variation in color/contrast.


Ignoring the qualitative impact of the aperture shape and other variables I can't recall...

[Jon Ruyle]

Which would have more bokeh (granted you had the same framing)?

I think it depends on what you mean by "more bokeh". One can quantify "depth of field", but how does one quantify bokeh? This isn't just a rethorical question... I would be interested if anyone has an answer, but I suspect that "more bokeh" is something that depends on perception.


It seems to me that an important step in answering this question is to ask what happens to bokeh when you crop. Does a cropped picture appear to have more, less, or the same bokeh as the original? This is the same question as "does a lens with a given aperture produce more bokeh if it is slow or fast?" (or equivalently, "long focal length or short focal length")?


If you say a cropped picture has the same bokeh as the original, then only aperture matters. So 200mm f/2.8 has exactly as much bokeh as a 100mm f/1.4 (and thus it has "more bokeh" than a 100mm f/2). If we accept this, we can quantify the bokeh a lens has very simply (the "amount of bokeh" a lens has is simply its aperture).


My impression is that bokeh probably stays the same or increases when you crop. (If it increases, the 200mm f/2.8 is even more favored over the 100mm f/2).

[Jon Ruyle]

Calculating OOF blur is much less common than calculating DOF. The only calculator I know of is Bob Atkin's:

Interesting, but... does he say *what* he is calculating? How does he define OOF blur? Do all these numbers come together to make a single measure of bokeh?

[Daniel Browning]

One can quantify "depth of field", but how does one quantify bokeh? This isn't just a rethorical question... I would be interested if anyone has an answer [...]

Jon, you are using a very wrong definition of the word bokeh. I think you are doing it intentionally, since others in this thread have been doing it, but I think something should be said.


Bokeh has nothing at all to do with this thread. In his original post, powers_brent used the word "bokeh", but what he meant was "background blur". There is no such thing as "more bokeh" or "less bokeh". It's an attribute like someone's face. You don't say that someone has "more face" or "less face", but an "ugly face" or a "pretty face" (or specific qualities about their face, such as symmetry, feature size, etc.). In the same way, we talk about specific qualities of bokeh, such as highlight edge softness and overall comparisons like "uglier bokeh" and "prettier bokeh".


However, it does make perfect sense to talk about "more OOF blur" and "less OOF blur". Most photographers only care about background blur, since they don't use foreground blur in most of their compositions. So usually, what they really mean by "bokeh" is just "quantity of background blur".


So there are three very different things here:

• Quality of the OOF blur (bokeh)
  
• Quantity of the OOF blur
• DOF

You can have any combination of the three:

• Deep DOF
• High quantity of OOF blur
• Poor quality of OOF blur (harsh bokeh)
  

Or:

• Thin DOF
• Low quantity of OOF blur
• Good quality of OOF blur (smooth bokeh)
  

In this thread, the poster asked about bokeh, which is completely separate from DOF and quantity of OOF blur. But what he meant was quantity of background blur, not bokeh.

It seems to me that an important step in answering this question is to ask what happens to bokeh when you crop. Does a cropped picture appear to have more, less, or the same bokeh as the original?

The bokeh is the exact same, but the background blur (if any) changes in relation to the subject.

This is the same question as "does a lens with a given aperture produce more bokeh if it is slow or fast?" (or equivalently, "long focal length or short focal length")?

You can answer that question definitively for background blur (or OOF blur), though even that will vary based on proximity to the hyperfocal distance, but not for bokeh.

Interesting, but... does he say *what* he is calculating?

Yes, it's all on his web page.

How does
he define OOF blur?

Diameter of a point source specular highlight at a given distance from the plane of focus.

Do all these numbers come together to make a
single measure of bokeh?

No, they're separate.

[Jon Ruyle]

Thanks Daniel. Your post (and Bob Atkins page) helped me understand some stuff.

Jon, you are using a very wrong definition of the word bokeh. I think you are doing it intentionally, since others in this thread have been doing it, but I think something should be said.

Yes, I was using the word in the way I thought Mr. Powers Brent intended, even though it isn't how I would have used the word.

How does
he define OOF blur?

Diameter of a point source specular highlight at a given distance from the plane of focus.

Okay, that makes perfect sense. Of course, if you really want to quantify blur with a single number (and thus answer which lens has more) you have to specify which given distance you care about . (Correct me if I'm wrong, but I think if that "given distance" is close to infinity, only aperture matters. If the "given distance" is close to zero, only f number matters.)


In addition, I guessed that by "bokeh" the OP wanted to also take into account another feature of long lenses: they magnify background details more. I *thought* this contributed to our perception of a blurred background (quite separate from how big the OOF disk of a background point will be), giving an additional advantage to the long lens. Maybe I'm confused about this as well, though.

So there are three very different things here:

• Quality of the OOF blur (bokeh)
  
• Quantity of the OOF blur
• DOF

According to Bob Atkins (not that I nececarily would take his word over yours []), bokek is a combination of quality and quantity of OOF blur. This is what I thought the word meant, too, though I never tried to check a definitive source.

[Jon Ruyle]

Correct me if I'm wrong, but I think if that "given distance" is close to infinity, only aperture matters. If the "given distance" is close to zero, only f number matters.

No, you idiot. 1) This is only true if the subject is far from infinity. 2) there is diffraction (which favors the longer lens, aperture being the same).


Think before you post next time, loser head.

[Daniel Browning]

Thanks for the kind response, Jon.

Yes, I was using the word in the way I thought Mr. Powers Brent intended, even though it isn't how I would have used the word.

When in Rome. []

Okay, that makes perfect sense. Of course, if you really want to quantify blur with a single number (and thus answer which lens has more) you have to specify which given distance you care about .

You're quite right.

In addition, I guessed that by "bokeh" the OP wanted to also take into account another feature of long lenses: they magnify background details more.

I think so too.

I *thought* this contributed to our perception of a blurred background

Agreed.

(quite separate from how big the OOF disk of a background point will be)

I think one of the ways in which the increased magnification of the background affects our perception of background blur is by increasing the size of the OOF disk on the sensor (though in a very different way from which f/number affects the OOF disk.) One other way that someone else mentioned in this thread is by having less background to begin with.

According to Bob Atkins, bokek is a combination of quality and quantity of OOF blur.

That's not the impression I get. Bob says:
<p style="padding-left: 30px;"] "While the quality of blur - bokeh - is in itself a "fuzzy" concept and
something that's quite difficult to predict or control, the quantity of blur can be
calculated quite easily and it's something over which the photographer has control through
choice of focal length and aperture."


To me it seems like he treats bokeh as quality only.


Thanks again, Jon.

[Jeff Lucia]

I wish I could keep up with this thread because it's a fascinating subject. Maybe one of you guys who understands it in detail could post some images that illustrate your points? Speaking on behalf of the "no question too stupid" contingent [], it would be much easier to understand if we could see it.


Thanks and I hope someone can come up with a way to illustrate this.


Jeff

[powers_brent]

[View:http://farm3.static.flickr.com/2682/4234224849_ff11394af8.jpg]


XS with Canon 100mm f/2.0 @ 2.0, 1/200 sec, ISO 100. Kodak P20 shot thru @ 1/16 camera right and Canon 430EXII bare @ 1/64 left behind snowman. (http://www.flickr.com/photos/brentpowers/4234224849/in/photostream/)





XS with Canon 300mm f/4.0 @ 4.0, 1/200 sec, ISO 100. Kodak P20 shot thru @ 1/4 camera right and Canon 430EXII bare @ 1/16 left behind snowman. (http://www.flickr.com/photos/brentpowers/4234227817/in/photostream/)





If the framing is not exact does the test not count? They are framed almost the same, just slightly different. The OOF blur seems more diffused with the 100mm lens. Or is it just me? What may I have done wrong? Or is that what I should expect?