Lens design, filters and flare

[snusmumriken]

I'd have thought this subject must have been done to death somewhere on Photrio, but I can't find anything.

To what extent does the front surface of a lens contribute to flare in contre-jour shots? And if that shape is critical, does the propensity to flare increase when we add a flat filter in front of it?

 

[loccdor]

Filters will at best have no perceptible effect on flare (Hoya multicoated are better at this than many others) and at worse add a lot of flare and internal reflections/ghosting.

Larger front elements can have more potential for flare, as can convex ones, but it also depends on the ability of light outside the frame to hit that element, which is impacted by how recessed the front element is, the hood, and any other objects creating shadows from stray light on the lens. That is a complex physics situation so it's hard to make too many generalizations.

It's not only the front element, but number and position and size of the other elements as well, and what coatings they have.

This is an instructive test on filter performance that I found useful: https://www.lenstip.com/113.1-article-UV_filters_test_Introduction.html

 

[DREW WILEY]

That's a well done test, and the conclusions match my own experience. It would be nice if a parallel set of tests was done for skylight filters, rather than "colorless" UV types, and if they included the Singh-Ray brand too.

What isn't mentioned is how uncoated filters collect grime and condensation a lot more easily, and therefore have to be cleaned more frequently, or else they will contribute significantly to flare.

 

[DREW WILEY]

Flare is less related to the size of the front element than the angle of view, along with the specific lens construction and number of elements, as well as the nature of the coatings. Wide-angle lenses present the biggest problem because they're difficult to properly shade.

Modern manufacturers have done a remarkable job trying to improve lens transmission. But still, it's a much more difficult problem maintaining optimal transmission in something like a zoom lens with a considerable number of elements, versus something simple like a tessar formula lens which only has six air/glass interfaces. The highest performance lens I ever had for any camera in terms of contrast and hue purity was a multicoated Kern Dagor with only four air/glass interfaces; but that kind of lens is no longer made.

Yet another factor, particularly applicable to view cameras, is just how much surplus light gets taken in due to the larger than normal "image circle" in relation to internal "bellows flare". Once again, the effectiveness of the lens shade is key. And often people fit their view camera with a compendium shade, itself highly adjustable.

 

[loccdor]

That's a well done test, and the conclusions match my own experience.

Yes, it's a 15 year old test but the most thorough I've seen on UV filters. I use Hoya HMC UV-0 by default on most of my lenses now.

 

[DREW WILEY]

I only used the HMC UV-0 in relation to relatively recent Ektachromes - from E100G onward. The Hoya 1B did better with Fuji chrome films. My tests were conducted at around 12,000 ft elevation, where the effect of UV on color film is more noticeable than at lower elevations. In relation to color neg film, however, particularly Ektar, the 1B is the weakest filter I carry; more often a Singh Ray KN proves most versatile, unless I need even more color temp correction from an 81A etc. Ektar is fussy, but the closest thing ever to the look of a chrome film by a color neg film.

 

[snusmumriken]

OK, so I understand that there are good and less good filters; and that even the best filters add two extra surfaces that will contribute to flare. But what about the shape, the flat filter versus the convex shape of most front elements? Is flat worse (or better?) for flare than convex, or doesn’t it make any difference? If I’m shooting into the light, am I more likely to get veiling flare with, or without?

 

[DREW WILEY]

An optical engineer really needs to chime in on that question. But there are undoubtedly other things at play, including glass type and internal arrangement, and the specific lens coatings themselves. Front position lenses ordinarily have a convex surface, filters are flat. In combination, very little changes if both are of optimal quality. So my hunch based on practical experience is that it makes very little difference. The effectiveness of lens shading is far more important. You never really know your limits shooting into the sun unless you try. You can only get away with a certain degree of that anyway.

 

[Doremus Scudder]

It's not really necessary to know the answers in detail if you're just looking to reduce flare. Just use best practice: multi-coated lenses and filters and good lens shading and try to keep light sources out of the image. Nothing else helps much.

Doremus

 

[xkaes]

OK, so I understand that there are good and less good filters; and that even the best filters add two extra surfaces that will contribute to flare.

You, like many others, have blown this out of proportion. You'd need some sort of scientific test to determine if a filter SIGNIFICANTLY contributes to flare. Of course it will, but does a 1% increase matter to you? How about 2%?
You need to find something more important to worry about.

 

[snusmumriken]

You, like many others, have blown this out of proportion. You'd need some sort of scientific test to determine if a filter SIGNIFICANTLY contributes to flare. Of course it will, but does a 1% increase matter to you? How about 2%?
You need to find something more important to worry about.

I don't see how asking my question blows anything out of proportion. Out of proportion to what? Can I not just be curious to know?

 

[BrianShaw]

I don't see how asking my question blows anything out of proportion. Out of proportion to what? Can I not just be curious to know?

Of course you can!

Ignore the first sentence and replace “SIGNIFICANTLY” with “NOTICABLY” and it reads a bit easier.

 

[snusmumriken]

Of course you can!

Ignore the first sentence and replace “SIGNIFICANTLY” with “NOTICABLY” and it reads a bit easier.

Maybe we are confusing amount with probability?

 

[xkaes]

It's simple enough to run a test for anyone interested & with time on their hands. I know, running an actual test is more work than simply believing what someone else says about it.

 

[DREW WILEY]

Just how many hundreds of lenses are you willing to test, and at different angles of light incidence? Just making a "A" versus "B" generic comparison out of it would be no better than hearsay. On the other hand, one can experiment with just those lenses they personally intend to use, to find out under which conditions flare is most likely to occur in a practical sense. Trying to precisely quantify that is another matter, which few if any of us are equipped for.

 

[snusmumriken]

I vaguely hoped there might be a lens design expert on the forum. If not, then let's end this thread.

 

[MattKing]

There are a couple of lens designers who check in here from time to time, so I recommend patience

 

[DREW WILEY]

I think Jason Lane belongs to this forum.

 

[MattKing]

I think Jason Lane belongs to this forum.

He does, but for the life of me I can't remember his screen name .

 

[DREW WILEY]

I can't either. He uses a different name on the LF Forum (Nodda Duma). He does sometimes advertise his pre-coated glass plate services, Lane Dry Plates. Every time I've been in a discussion with him, it's been about neither flat nor convex surfaces, but aspheric designs. But I don't think he would give a simplified answer either; there are just too many variables.

 

[MattKing]

Thanks.
Contacting @Nodda Duma - your name is being used (not in vain)

 

[reddesert]

Flare typically comes from multiple reflections between surfaces or scatter of non-image forming light that is insufficiently baffled. There is no easy generalization like a relation to the amount of convexity of the frontmost surface. In fact, the front most surface might be sometimes less important because light reflected from it goes forward out of the lens and stops causing problems.

The relatively obvious issue is that flare is often observed in wide angle lenses because they are hard to baffle or shade from light, and wide angle lenses typically have highly convex front surfaces. But they're not flaring just because of that convexity, but because of all the light bouncing around inside.

 

[xkaes]

Yup, a correct (the KEY word) lens shade is the best way to control flare. Someone who spends more on expensive filters, but doesn't use a lens shade (MUCH cheaper), is missing the target.

 

[snusmumriken]

Yup, a correct (the KEY word) lens shade is the best way to control flare. Someone who spends more on expensive filters, but doesn't use a lens shade (MUCH cheaper), is missing the target.

Goes without saying, surely? Let’s not get distracted into lens shade design, but I’d just like to point out that finding or designing the ideal lens shade for a lens is a nightmare in itself.

 

[DREW WILEY]

For tripod mounted cameras, adjustable compendium shades are ideal. It's almost impossible to make fixed shades work properly in every situation; but they work best on long telephoto lenses with narrow angles of view.