[LensBeginner]

Well, I just discovered that a UV filter can solve fringing issues almost completely.
Well, it doesn't "solve" them, because the glass in the lens will still bend different frequencies in a different way, but at least it will eliminate some problematic wavelength (i.e. extreme purples and UV).

Many people suggest a 2A filter.
Problem is, I only have "UV" (God knows which kind, it's an off-brand), 1A and 1B filters (=1A plus warming).
Is there a way to measure at which frequency they start their cutting action and how the curve is shaped?
I took some photo but they all work almost the same, with minimal differencies, plus the warming on some makes it difficult to judge that exactly.

Other types of filter that work well against fringing? I must say that I'm pretty satisfied in the way my "UV" cheapo reduces fringing wide open on a SMC-M zoom, but still... always the perfectionist...

BTW, who was that said that UV filters are useless?

[vonBaloney]

Are you sure you aren't just making your lens less sharp with the filter? That can get rid of some fringing along with some resolution...

[mrNewt]

The A filter are actually introducing a small "hue tint" that is supposed to warm the pictures.
You camera will detect it and attempt to remove it so you might not get proper white balance.

These filters does nothing in removing your problem. It just slightly degrades the quality (especially the very cheap ones) to the point where it makes you think something was slightly fixed.

[JeffB]

I agree, UV filters (good ones) are definitely useful. I think your own subjective judgment as to the differences between images with and without the filter is really the only that matters. If it solves your problem than go for it.

The only way to really measure what you are looking for is with a spectrophotometer. Unless you have one lying around I'd look at this well done test.

The best approach using your own camera would be using controlled test shots and analysis software such as Imatest . That would give you real answers as to what your own camera + lens + filter combo is doing.

[LensBeginner]

Are you sure you aren't just making your lens less sharp with the filter? That can get rid of some fringing along with some resolution...

Uhm... I can't post pictures now but I've shot shots with and without a filter in the past and when I was home I couldn't tell the difference at 100% on a 40" TV...
The SMC-M 80-200mm f/4.5 resolves a lot of detail, but is prone to fringing even at f/11 in some (extreme) cases... so the tradeoff might have a sense.

The A filter are actually introducing a small "hue tint" that is supposed to warm the pictures.
You camera will detect it and attempt to remove it so you might not get proper white balance.

These filters does nothing in removing your problem. It just slightly degrades the quality (especially the very cheap ones) to the point where it makes you think something was slightly fixed.

WB doesn't vary all that much (it was the first thing I checked: same temperature +-2 on tint as seen in Camera RAW, IIRC, easy to bring the two pictures up to the same standard for comparison).

I wouldn't call this example from dpreview "slightly fixed":
UV 2A filter makes purple fringing go away: thank you Anders W!: Micro Four Thirds Talk Forum: Digital Photography Review
Albeit that's with a 2A...
I did some test shots a couple of days ago, but I must have misplaced them...

I agree, UV filters (good ones) are definitely useful. I think your own subjective judgment as to the differences between images with and without the filter is really the only that matters. If it solves your problem than go for it.

The only way to really measure what you are looking for is with a spectrophotometer. Unless you have one lying around I'd look at this well done test.

The best approach using your own camera would be using controlled test shots and analysis software such as Imatest . That would give you real answers as to what your own camera + lens + filter combo is doing.

Nope no spectrometer, and those are really old and wonky filters (Cozo, Itorex, Naigon, a Hoya)...

Will redo the test as soon as I can, aiming at a white car in full sunlight.
In the meantime thanks all for your replies.

[JinDesu]

If the UV filter removed the purple fringing as part of its design, then it would also render other purple objects as black. That would be a pretty easy test.

[LensBeginner]

If the UV filter removed the purple fringing as part of its design, then it would also render other purple objects as black. That would be a pretty easy test.

Well... not necessarily...
I believe there's a difference between UV light that I believe is still picked up by a camera sensor (just like you can "see" the IR led on your remote with a camera) and purple = red + blue.
The fact that UV is rendered as purple is just because it's the nearest visible color on the spectrum, I think.

---------- Post added 02-11-15 at 08:14 PM ----------

...which brings the next question...
Is this Purple Fringing or Blooming?
There are few (255, 255, 255) spots (one in the middle of the engine), and none in the problematic areas.
Pic is exposed properly, RAW with no PP, 100% crop (plane's quite small).
It's f/11 so, yes, I stopped down and yes it's been the only time I've experienced something that bad with that lens.
Plane was far enough that I could use the infinity stop and it's not "color bokeh" or what's called, since it would have been green and not purple if I were near-focusing.

---

Last edited by LensBeginner; 02-11-2015 at 12:14 PM.

[mrNewt]

WB doesn't vary all that much (it was the first thing I checked: same temperature +-2 on tint as seen in Camera RAW, IIRC, easy to bring the two pictures up to the same standard for comparison).

I wouldn't call this example from dpreview "slightly fixed":
UV 2A filter makes purple fringing go away: thank you Anders W!: Micro Four Thirds Talk Forum: Digital Photography Review
Albeit that's with a 2A...
I did some test shots a couple of days ago, but I must have misplaced them...

You do know that most of the sensors have multi-coats and one of them is for UV light as well? If UV filters will work against your issues, it should already be fixed.
In the example you give me, look also at how much detail was actually lost. IQ and sharpness were affected quite a lot as well... since the lower PF. Don't lie to yourself, look at the overall. There is no hidden magic going on...

Anyway, I am not trying to convert to anything... if you think it works for you, then all the marry. As long you are happy, who cares?

[vonBaloney]

It's interesting, the tech guys always say it is impossible because digital sensors are supposedly totally insensitive to UV light and so it is not a matter of blocking any particular wavelength -- the fringing is an artifact of the wavelengths not being focused at exactly the same areas and/or a quirk of the sensor/photosites/microlenses. Whenever I've tried filters in the face of such problems, they always add aberrations, not take them away. Is it possible a filter could change the angle of the light enough to make a difference? (Some fringing on lenses with large front elements seems to be caused by the oblique angle of the light coming in from the edges and the photosites and their microlenses don't like that.)

[LensBeginner]

You do know that most of the sensors have multi-coats and one of them is for UV light as well? If UV filters will work against your issues, it should already be fixed.
In the example you give me, look also at how much detail was actually lost. IQ and sharpness were affected quite a lot as well... since the lower PF. Don't lie to yourself, look at the overall. There is no hidden magic going on...

Anyway, I am not trying to convert to anything... if you think it works for you, then all the marry. As long you are happy, who cares?

The example was without a filter.
It's what prompted me to actually start looking for a solution.

[mrNewt]

The example was without a filter.

On the second image it said this: "14mm f/2.5 @ f/6.3 with Tiffen Haze 2A filter".
So I would assume there was a filter going on... and the IQ and sharpness as far as I can see on the second image, was affected quite a lot.

Btw, I am talking about the dpreview link you provided in the text I quoted...

[LensBeginner]

On the second image it said this: "14mm f/2.5 @ f/6.3 with Tiffen Haze 2A filter".
So I would assume there was a filter going on... and the IQ and sharpness as far as I can see on the second image, was affected quite a lot.

Btw, I am talking about the dpreview link you provided...

Sorry, I thought it was about the plane picture.
Yes, true, some detail was lost, but the fringing went almost completely away!
You can't correct PF in post (it's a pain and it's a long process, and you almost always lose color. Camera RAW sliders are a joke IMHO - OTOH the CA checkbox works like a charm), so it could be an ok tradeoff.

[JinDesu]

"A strong UV filter, such as a Haze-2A or UV17, cuts off some visible light in the violet part of the spectrum, and has a pale yellow color; these strong filters are more effective at cutting haze,[3][4] and can reduce purple fringing in digital cameras.[5] Strong UV filters are also sometimes used for warming color photos taken in shade with daylight-type film."

From Wiki - so it does make sense that it would cut off purple fringing - at the cost of violet in the visible spectrum. Wiki also mentions that purple and violet are distinctly different, where violet is the color on the spectrum and purple is a combination of red and blue light to our eyes. Given this distinction, objects that reflect violet wavelengths will be rendered black or without light when viewed through a 2A filter. Purple objects (reflecting some combination of red and blue wavelengths) will not be affected.

[VisualDarkness]

As I've understod it purple fringing got nothing to do with violet light per se entering through the lens, hence filters won't have anything to with it. The problem is that red light and blue light are both undercorrected in the lens resulting in a blend of those two colors, resulting in the purple.

[JinDesu]

As I've understod it purple fringing got nothing to do with violet light per se entering through the lens, hence filters won't have anything to with it. The problem is that red light and blue light are both undercorrected in the lens resulting in a blend of those two colors, resulting in the purple.

On wiki it lists purple fringing as:

"Lenses in general exhibit axial chromatic aberration in which different colors of light do not focus in the same plane. Normally, lens designs are optimized so that two or more (at least three for apochromatic lenses) wavelengths of light in the visible range focus at the same plane. Wavelengths very different from those optimized in the design process may be severely out-of-focus when the reference colors are in focus; this axial chromatic aberration is usually severe at short (violet) wavelengths."

So it may just be violet light.