"gauge" of lens affects bokeh?

[Nesster]

Over at the pentax forum, someone posted a 'which is which' test
http://www.pentaxforums.com/forums/pentax-slr-lens-discussion/28178-guess-lens.html

essentially a Vivitar 105 macro vs. a Pentax 100. The Viv is the faster lens and (I surmise) is therefore larger in diameter. (Or larger 'gauge')

I identified the Vivitar correctly because I see the bokeh is broader, more spread out, while the Pentax is more contained and tidy. In part, the tidiness I think is a Pentax house look, but I suspect there may be something else going on, namely:

A larger diameter lens, even stopped down to equal a small diameter / slower lens, will have different bokeh due to this fact. My thinking is, the aperture is a choke point, but this doesn't mean that light doesn't take all available paths through the lens - and the larger diameter lens has more room and glass for the light to go through, sideways. This then scatters out of focus light more, resulting in a broader bokeh.

I've sort of seen this with my lenses, however I don't have a pair of similar quality lenses in different gauges and the same focal length to really test it out. (I've seen similar with my fat Vivitar 28 vs skinny SMC-A 28, but it could just be the Vivitar is less controlled.)

Does anyone here have any insight into this, or the means to check out the theory? I'd be much obliged.

[Rob Leslie]

IMO ‘Bokeh’ is another of those strange and wonderful subjects many like to write ‘CRAP’ about. (Sorry for the descriptive language)
There are so many variables and tastes that no one person can be right about it.
Here is one good article. I think all those who consider themselves Boke experts should do a bit more reading about it.
I should add I am not an expert on it and will never claim to be. I do my best to get a decent background and resort to PS if I’m not happy with it. I would certainly never dismiss a decent lens just because I got the subject background distance wrong or the light wasn't right for decent background blur neither would I accept the trade-off between sharpness and good bokeh.
http://hobbymaker.narod.ru/English/Articles/bokeh_eng.htm

[LucisPictor]

Rob has a good point (and an excellent link, thanks for that!). "Bokeh" is imfluenced by so many different aspects that I think it is impossible to predict the bokeh a lens produces.
Try the lens and you will see the bokeh and if you like it or not...

[Nesster]

Thanks for the article, it is very interesting.

My intent isn't to argue about good or bad bokeh, or to base judgements of lenses on this.

It is just this observation: the aperture hole in different lenses can be a different size for the same f/stop, having to do with overall light transmission and the size of the lens barrel. I'm thinking, maybe the size of the barrel has something to do with how far the oof blur spreads out (bokeh in that sense) in addition to how spherical and other aberrations have been addressed in the design.

In the particular example, it is likely the Pentax is more (some might say over) corrected than the Vivitar. Perhaps. But also, the Vivitar being a f/2.5 and the Pentax a f/4, I was thinking the extra 'space' in the Vivitar was affecting the size of the blurred objects in the background.

[Orio]

Bokeh is simply a subjective evaluation.
There is no objective meter to evaluate a "good bokeh" or "bad bokeh".
There could be, at most, a more or less widely shared opinion that highly structured elements (in this case, a doughnut highlight) are psychologically more distracting that simpler element (in this case, "neutral" full circles).
But it's as far as it can go.
Everything else is just purely a matter of individual taste.

[Nesster]

LOL, I'll give up after one more try: I'm not after any statement on good or bad bokeh.

In the samples on the pentax board, to my eye the blurred background objects in sample 1 are larger, more spread out, more 'fleshy' if you will, than on the second sample.

The proposal is: given two otherwise equivalent lenses, the one with the larger barrel will probably give larger, more spread out oof blur.

Whether one prefers one or the other doesn't matter: only that there is a difference, and is this difference due to the physical diameter of the lens.

[Nesster]

ps. I suppose this will cause me to purchase yet more duplicate focal lengths, just so I can put this theory to test myself. Now that's a pleasant thought!

[Orio]

I am no lens engineer, but I don't think that the diameter of the lens might have a direct influence on the type of bokeh.
The perceived difference (indeed minimal, to my own observation) is probably to be ascribed to:

- the small, but yet existing focal lenght difference (100mm vs. 105mm), longer focal lenghts have shallower DOF and therefore smoother bokeh when compared to lenses of shorter focal lenght at the same distance from objects.

- a difference in the contrast rendition of the lens, probably due to some more internal flare in the larger diameter lens - and here's how it could be influencing the bokeh: not directly (any theory about a larger glass "spreading" the image more sounds weird to me), but indirectly - larger diameter causing more flare causing lower contrast causing smoother bokeh.

[Nesster]

Ah there you go, that makes sense...

[Rob Leslie]

I must admit that this is about the most sensible discussion of Boke I have ever read.
Nesster I don’t think you can put a simple formula or even reason to why your lenses are so different. You may well find that given a different subject to background distance or different lighting the reverse may be true. This IMO is why test of lens bokeh are often meaningless. Orio's answer also makes sense but lots of trials could be the only way to find out. Have fun.
There is also a good article on the subject at Luminous landscape. I believe you know your way round the site. I can't be bothered to look it up, I have to finish this bottle of wine!

[estudleon]

[quote="Orio"]I am no lens engineer, but I don't think that the diameter of the lens might have a direct influence on the type of bokeh.
The perceived difference (indeed minimal, to my own observation) is probably to be ascribed to:
- the small, but yet existing focal lenght difference (100mm vs. 105mm), longer focal lenghts have shallower DOF and therefore smoother bokeh when compared to lenses of shorter focal lenght at the same distance from objects.
- a difference in the contrast rendition of the lens, probably due to some more internal flare in the larger diameter lens - and here's how it could be influencing the bokeh: not directly (any theory about a larger glass "spreading" the image more sounds weird to me), but indirectly - larger diameter causing more flare causing lower contrast causing smoother bokeh.[/quote]
ORIO : I agree.
Because the small distance focal of the lens more depth of field, and this affect the bokeh.
And the flare is equal to less contrast and smoother bokeh because the internal surfaces of lenses (not MC specially) reflecting rays to the sdensor (film) and the borders of the reflected images become smoother.

[Nesster]

I'm really bothered by this all - I suppose I have to get Adams' Lens book next time I'm at the library.

There are things I know I could work out - or follow an explanation - but just seem strange when I think about them:

- closing down a lens increases performance to a point. This is because... you're using the 'good' center bits?
- but then, closing down also reduces vignetting! LOL there's a physics answer to that I'm sure

But back to the bokeh on those two examples (not my lenses by the way)

So the front elements collect light, invert it through an internal crossing point, and then the rear elements fix it up on the way to the sensor. When focused, a single line (say) sends out light rays that travel various paths through the lens, converging on the film plane. Presto: a focused image.

When those rays meet in front of the plane, they continue on, and you get in-front out of focus. Conversely, if they don't yet meet up at the plane, the sensor/film records them as scattered, ie. out of focus. (They would be focused behind the plane)

An aperture cuts down the possible light paths between the front and rear elements, thus reducing the total number of rays passing through - but not the total coverage, as the entire frame is still filled!

With a bigger diameter lens, the angles that need to converge at the film plane can be greater. Thus out of focus images could also be more spread out.

This doesn't mean the out of focus is SMOOTHER or BETTER, only that it could be larger. In the examples the guy posted, I don't know if I really prefer the fleshy Vivitar over the trim Pentax. I just noticed the difference in size of the oof blur, and figured it was the faster lens that allowed it.

LOL I'm SO FAR from knowing crap about optics, but I am getting better at noticing the rendition. I'll really have to read up on this.

So what two lenses do I use for my test? I have a 35/3.5 Takumar I can compare with a 35/2.8 Yashinon, if the two can be considered reasonably in the same class. Otherwise, do I like have to buy two 135's by the same maker, one fast and one slow?

[Rob Leslie]

Bother you have made me look it up.

http://www.luminous-landscape.com/essays/bokeh.shtml

Part of the article.

‘The camera functions very much like a painter having access to set of round brushes of all sizes. The camera chooses the size of brush based upon pure geometry: the effective brush size depends only upon where the object is, where the camera is focused, and upon the aperture of the lens. The camera then paints a fully detailed picture, using an amount of light corresponding to that for each detail, spread over a disk the size of the circle of confusion appropriate for each detail in turn.’

You will have to follow the link and read the full article.

I shouldn’t worry too much about it. This is my own very personal opinion but if you get carried away looking for perfect boke you will end up with a load of poor lenses. They may have fantastic out of focus backgrounds but they will also have out of focus subjects. The answer is always compromise or a choice of priorities.
My own preference is to choose the lens I know will give me the result I am imagining. One acquires those lenses in time.
Perhaps another point in relation to this is to be a bit discrimination in ones choice of lenses. It is pointless having half a dozen 50mm lenses that all perform in the same way. It is equally pointless to have 50mm lenses that offer you no characteristics you like. I will often reject very well regarded lenses because they offer me nothing I value. For example I know of one particular lens that is often praised for its boke but I just see a lens that is soft until stopped right down. I personally would rather have sharpness to any illusion or personal preference to the bokeh.

Adams won’t help you, he believed in everything being in sharp focus.

[Nesster]

I just keep puzzling this thing out - again, for understanding rather than value judgement!

I've looked at various ray diagrams online, and can't seem to find very many realistic ones that look like a real camera lens with the aperture in the middle. I'd link one in to illustrate if I could find one!

So imagine this:

The front elements collect the cone of light rays from a letter "A" and converge them. This light collection ability is related to the size of the front elements - the larger the diameter the more light is caught (for a given focal length).

Then the rear elements refract and converge these in order to provide even coverage of the full image frame (and to correct more distortions). The resulting image on the film plane is an upside down A.

The aperture is in between the front and rear elements.

The light from the tip of the A does not travel only one single path (like it would in an idealized pinhole) but rather ALL paths including the center and edges of the front element.

If light is impeded partially, the full image still appears, but with less light intensity. The aperture acts in a way that impedes the light rays from the edges of the front element, thus subtracting the contribution of these on the film plane. Since the edges produce the most distortion and suffer from light falloff, as the aperture gets smaller these distortions are subtracted: the image quality becomes better and more evenly illuminated. (then diffraction starts to take its toll)

The rear elements are still doing their thing: normalizing the cone of light and ensuring full coverage of the frame.

Basic geometry tells us that the A is in focus when all rays convege on the same spot on the sensor/film. If the point of focus is behind the film plane, the resulting image has not fully converged, and we have the circle of confusion, out of focus. The more out of focus, the larger the circle of confusion.

Now, keeping the distance from lens to camera constant, a smaller diameter lens collects less light, and also the rays converging on the tip of the upside down A come in from smaller angles. A larger diameter lens collects more light, and the rays converging on the film plane come in from larger angles. (This assumes that the lens design uses most of the available glass, rather than only the central portions)

When you stop down this lens, the cone of coverage gets LARGER, ie. more evenly lit within the frame. What this SEEMS to indicate to me is that the rays continue to travel through all paths in the rear element, rather than a smaller cone due to the stopping down.

This is why I think - and where I need to learn more - that two lenses, at the same f/stop and focal length, will exhibit different size circles of confusion: the larger diameter one will cause light rays to come at the film / sensor at a wider angle than the smaller lens.

I think I need to go look at the 4/3 system stuff, as the paradox there is that despite the smaller sensor the lenses have to be fatter. Maybe the explanation, and my error, is to be found there...

[LucisPictor]

All this really sounds interesting. As soon as I have a little more time and it is not that hot anymore, I will try to read your posts - and try to understand them.
Then, perchance, I will write an answer.

[estudleon]

Yes, Ansel was part of the F/64 group (a technical reaction to the mode of the pictorial's out of focus - soft - pictures).
But I think that his affirmation about the "previsualization" is applicable here.
You see the subject (flower, girl) And inmediately you imagine how do you want to see this subject in the picture (distance, ligt, background), you previsualize the final picture.
If you know the characteristics of the lens (what sharp or unsharp is at close focus, the results in the F/range, etc, and how this cuestions are receive for the sensor/film) then choose the right lens for your previsualization.
Yes, is right, there are more circumstances to consider, but is a good beginning, no?
And with the bokeh, I think that is a cuestion of personal taste, nobody can tell us which is the best bokeh for a specific subject in determined atmosphere. Regards

[Nesster]

I don't want to give anyone the headache I get puzzling this. I found a recent discussion here:
http://forums.dpreview.com/forums/read.asp?forum=1018&message=26892110
that actually says that the larger the front element, the larger potential size the fuzzy circle will be. I can't get myself registered at dpreview, so I can't ask the author directly: with his equations, and two lenses at the same stop, will the larger diameter lens produce larger blur.

Note: LARGER, not prettier, better, more artistic, etc.

The reason this is interesting: there would be another purpose in the arsenal for picking that f/4 or f/2.8 lens in a given situation. Namely, the SIZE of the bokeh (not the quality, prettiness, goodness etc, just the size of the blur)


Ps. I thought of Ansel because I figured he'd have the clearest practical explanation of how lenses actually work. Plus I'm preverse that way: if a guy can tell me exactly how to get maximum sharpness, I figure I just found a way to get maximum unsharpness