TR Forums

I was shooting spring foliage in morning light outside of the new office, and turned up an effect I'd never witnessed before:



Shooting east-ish at 8:45am, so the sun was to the upper left, in front of the lens. Note the concentric rings in the bokeh from the specular highlights of the fountain (center) and an apparent flare artifact of the same (upper left). Here are 100% crops from those regions, showing that the rings seem to follow the axis of incidence:





I can't find much about it in general searching, except (a) some speculation that a mixture of flare artifacts and reflections within the lens are being bounced into a "hall of mirrors" effect, and (b) the Tamron 17-50mm f/2.8 is one of several lenses that tend to strongly produce this effect when the lighting is right.

Anyone have insight on what's actually happening?

Does the number of rings in the bokeh match or have resonance with the number of lens elements?

[/total SWAG]

EDIT: On second thought, how many compound elements are there in that lens. I'm thinking refraction/diffraction events at the interfaces of compound elements with dissimilar indices of refraction.

Perhaps it's ghosts of out-of-focus parts?

Captain Ned wrote:

Does the number of rings in the bokeh match or have resonance with the number of lens elements?

That's a good guess, but if so, I can't find direct evidence of it in this image. The lens uses 16 elements in 13 groups.

There are all sorts of weird effects that occur when light bounces back and forth between semi-reflective surfaces. Perhaps a pattern somewhere in the image (though not necessarily in focus) interacted with the optics to create the effect that only shows up in the very out of focus areas. It looks like bright spots of water are the only areas with the effect, so perhaps the angle of the sun, the extreme brightness, and the optics interplayed in the most bizarre way to cause it. Maybe the fact that water polarizes light had something to do with it? Eh, now I'm just guessing...

I assume you checked, but are there any irregularities in the glass or the coating on the front element?

I did a bit of poking around and was able to find similar effects only in sections of reflections, or light coming through, water. Can you post the full ress, i kindof want to see it.

@Spot: I think the ones in the upper left are ghost artifacts from the fountain -- same pattern array, nothing reflective over there that I recall -- but interestingly, they have the same rings, with the alignment skewed along the axis of incidence.

@Heiwashin: Those crops of the bokeh are full resolution.

FireGryphon wrote:

are there any irregularities in the glass or the coating on the front element?

None that I'm aware of, no. I do have a UV(0) filter on the front for lens protection, but it's a Hoya Super HMC, and my online searching indicated that this also occurs on bare-front lenses, so it seems to be coming from inside the lens assembly.

Onion bokeh is a common failing of many zoom lenses. Generally, the cheaper zooms and especially superzooms (10x zoom) suffer the most from these artefacts, although there are several L lenses that exhibit the same behaviour. Also, primes, though usually less susceptible, are not exempt (I've used a few primes - usually the Cosina Voigtlanders - with fugly bokeh).

The characteristic of the 'onion' shapes also varies from lens to lens - in this regard, I think the Tamron is actually pretty decent, with regular concentric features. Some lenses will have blotches and/or irregular rings, which are a lot more noticeable and ugly.

The sad truth is that there are more lenses with bad bokeh characteristics than there are lenses with good all-around bokeh. That doesn't mean you can't take great pictures with the "bad" ones - I have seen some stunning results from lenses with "bad" bokeh, but it does also mean that bad bokeh can and will get in the way of some shots you want to take. As always, one has to learn the drawbacks and limitations of their gear and work around or transcend them. Sounds simple, not so easy in practice.

Lastly, "bokeh-spotting" is a common pitfall of photo geeks. Be wary of missing the forest for the trees - some people, myself included, can get too obsessed with bokeh details sometimes, so it's a good thing to be mindful of. Have fun shooting!

Voldenuit wrote:

Lastly, "bokeh-spotting" is a common pitfall of photo geeks. Be wary of missing the forest for the trees... Have fun shooting!

QFT.

PS Just like to point out that the Canon 17-55/2.8 also has onion bokeh (though not as defined as in the Tamron), so paying 2.5x more doesn't always get you a lens that's over twice as good.

Voldenuit wrote:

Lastly, "bokeh-spotting" is a common pitfall of photo geeks. Be wary of missing the forest for the trees - some people, myself included, can get too obsessed with bokeh details sometimes, so it's a good thing to be mindful of. Have fun shooting!

I generally lean toward the "shoot 'em all and let post-processing sort them out" school. That's when I discovered this bit of oddness going down -- I had about six different similar shots and was comparing details to see which was sharpest.

Heres a few good articles about how bokeh works, look especially at the two lense comparisons 2/3 rds down in the first one, and the mirror lense example in the second one. Getting onion/doughnut bokeh, while annoying in cases, is quite common, if not always pronounced enough to notice.
http://www.luminous-landscape.com/essays/bokeh.shtml
http://www.bobatkins.com/photography/te ... bokeh.html

FireGryphon wrote:

There are all sorts of weird effects that occur when light bounces back and forth between semi-reflective surfaces. Perhaps a pattern somewhere in the image (though not necessarily in focus) interacted with the optics to create the effect that only shows up in the very out of focus areas.
- - -

Just wanted to chime in with two images that shows what actually can happen if the circumstances are correct. Even though its not interference by adjoining optical surfaces, its a rather extreme example of flares due to internal reflection inside the lense.

The first issue of the Canon 24-105/4L IS USM which was later taken back and exchanged... IIRC the cause was lacking paint or internal flocking on the internal barrel.




A 300D with a kit-lense pointed at the sun, could be some refraction in the microlenses or something due to the angle of light from the lense, but thats just speculation, because I really cant imagene a lense that would go through Q/A showing that level of refraction.

Wow that mug shot is gorgeous.