Zuiko 24mm 2.8 purple fringing at infinity

FM Forums \| Alternative Gear & Lenses \| Join Upload & Sell	Search Used
1 2 end

Archive 2009 · Zuiko 24mm 2.8 purple fringing at infinity

min lee
Offline
Upload & Sell: Off

p.1 #1 · Zuiko 24mm 2.8 purple fringing at infinity

Hi,
Long time reader, first time poster. I have a 5D2 and I recently picked up a multicoated zuiko 24mm 2.8, I noticed this haziness/purple fringing when I set the focus to infinity. I linked pics just before infinity focus and with the focus ring turned all the way. Is this normal? I thought it might be the thickness of my adapter (w/ AF chip from Big_Is on ebay) that's slightly off. Is there anything I can do about this? Thanks.

before infinity

This image is copyrighted by the owner

beyond infinity

This image is copyrighted by the owner

min

Oct 07, 2009 at 09:04 PM

cogitech
Offline
Dedicated FM
Upload & Sell: On

p.1 #2 · Zuiko 24mm 2.8 purple fringing at infinity

Can we please stop calling this "purple fringing", especially when it is not even purple?

This is lateral chromatic aberration, or "CA" for short (not LoCA, that's different).

The solution to your problem is to not focus beyond infinity.

If the adapter is too thin, which results in the lens focusing beyond infinity when it is dialed to "infinity", then try another adapter or adjust the infinity stop on the lens.

Edited on Oct 07, 2009 at 11:04 PM · View previous versions

Oct 07, 2009 at 10:41 PM

biotar
Offline
Upload & Sell: Off

p.1 #3 · Zuiko 24mm 2.8 purple fringing at infinity

SOME purple fringing should be visible at 100% with this lens wide open. LoCa's exist in front of, and at the back of the focus field at wide open.

Since your zuiko 24 isn't close range corrected, you can easily recalibrate infinity.
- FIRST you remove this ring, try to heat it up somewhat with a hair dryer so that the rubber expands without stretching and thereby doesn't loosen afterwards.
- After removing the rubber focus ring you need to loosen the screws underneath the rubber focus ring.
- Now, check when the lens is at its most ultimate sharpness at infinity and align the ring with this point.
- Then it just comes down to tightening the screws again and putting the rubber focus ring back on again by heatening it (somewhat) and minding not to stretch it.

-edit: OK. according to paul it isn't LoCa. Whatever it is, it turns up before and beyond the focus point in non APO lenses

Oct 07, 2009 at 10:45 PM

Jonas B
Offline
Upload & Sell: On

p.1 #4 · Zuiko 24mm 2.8 purple fringing at infinity

I thought that was LoCA. Confusing stuff... all this about all sorts of CA.

I always thought of simple plain CA as the color(typically red and green, sometimes blue and yellow) bands to the (typically) left and right of high contrast edges in the focus plane (or at least very well inside the DOF area). As usual I missed something.

Oct 07, 2009 at 11:16 PM

cogitech
Offline
Dedicated FM
Upload & Sell: On

p.1 #5 · Zuiko 24mm 2.8 purple fringing at infinity

Come to think of it, if the lens is focused so far beyond infinity that the house/trees are in front of the focal plane, then that would be LoCA (AKA defocus CA).

Sorry for the confusion

Oct 07, 2009 at 11:31 PM

jcolwell
Offline
Dedicated FM
Upload & Sell: On

p.1 #6 · Zuiko 24mm 2.8 purple fringing at infinity

No worries. Discussions like this help to make it more clear for everybody. Myself included.

Oct 07, 2009 at 11:35 PM

olyacme
Offline
Upload & Sell: Off

p.1 #7 · Zuiko 24mm 2.8 purple fringing at infinity

cogitech wrote:
Come to think of it, if the lens is focused so far beyond infinity that the house/trees are in front of the focal plane, then that would be LoCA (AKA defocus CA).

One niggle: Longitudinal CA only refers to what happens at the focal plane.

Longitudinal CA is failure to bring colours to common focal depth, but says nothing about what path those colours take to reach that depth. As light travels through a typical lens, its components repeatedly separate reassemble as the various elements and air gaps do their jobs. If the lens design is such that, as they travel toward the sensor, some colours come in from left field, then, even though those colours may arrive at common focal depth, they're still divergent above and below the focal point. "Defocus CA" is slowly gaining acceptance to describe this colour spread, that many lenses show outside of the focal plane.

A lens can show no LoCA, yet still have defocus CA. Evident LoCA almost guarantees defocus CA.

/Acme

Oct 07, 2009 at 11:52 PM

min lee
Offline
Upload & Sell: Off

p.1 #8 · Zuiko 24mm 2.8 purple fringing at infinity

Sorry about the terminology. Thanks biotar, I'll have to give that a try when I get a chance.

Oct 08, 2009 at 12:14 AM

cogitech
Offline
Dedicated FM
Upload & Sell: On

p.1 #9 · Zuiko 24mm 2.8 purple fringing at infinity

Ah, I see

Thanks!

Would an absence of defocus CA indicate an absence of LoCA, then?

Oct 08, 2009 at 12:16 AM

Jonas B
Offline
Upload & Sell: On

p.1 #10 · Zuiko 24mm 2.8 purple fringing at infinity

My brain hurts from all this CA stuff.

Now, if LoCA only refers to what happens at the focal plane - then what is lateral CA?

I admit that I haven't read all the more theoretical texts about CA but just sloppily taken in what I have seen at different forums (fora?) and lens reviews. In reviews longitudinal CA certainly often is used for this, to me relatively new term, defocus CA.

Any clarification is welcome, and, from now on, only APO lenses for me...

Oct 08, 2009 at 12:55 AM

olyacme
Offline
Upload & Sell: Off

p.1 #11 · Zuiko 24mm 2.8 purple fringing at infinity

cogitech wrote:
Ah, I see

Thanks!

Would an absence of defocus CA indicate an absence of LoCA, then?

Great question, and for the most part yes. Defocus testing actually makes it easier to discern certain aberrations than observation at focus. In astronomy forums you'll find expert observers discussing the quality of inside and outside focus star tests of their 'scopes for just this reason.

Oct 08, 2009 at 01:14 AM

olyacme
Offline
Upload & Sell: Off

p.1 #12 · Zuiko 24mm 2.8 purple fringing at infinity

Jonas B wrote:
My brain hurts from all this CA stuff.

Now, if LoCA only refers to what happens at the focal plane - then what is lateral CA?

Lateral CA is failure to bring all colours to common magnification at the image plane. In longitudinal CA some colours are displaced longitudinally (along the longitudinal axis of the lens), arriving at different focus depths, while in lateral CA some colours are displaced laterally, arriving at common focus but sagittally separated.

That note about magnification is a big hint as to how "Lateral CA Correction" software works.

/Acme

Oct 08, 2009 at 01:50 AM

Jonas B
Offline
Upload & Sell: On

p.1 #13 · Zuiko 24mm 2.8 purple fringing at infinity

Thank you Acme.

That explains why not all CA seen at objects seemingly in focus can be corrected by software then.

If not all colors are magnified the same (lateral CA) that's a way to say the focal length is different for different colours, no?

It is probably easier to correct the error with longer focal lengths. I wonder why all (?) APO lenses are longer than normal lenses. I would love a relatively fast prime to be an APO, for example a 50/2 or maybe even a 50/1.4 if possible.

Oct 08, 2009 at 03:17 AM

olyacme
Offline
Upload & Sell: Off

p.1 #14 · Zuiko 24mm 2.8 purple fringing at infinity

Jonas B wrote:
Thank you Acme.

That explains why not all CA seen at objects seemingly in focus can be corrected by software then.

Other reasons are the limitations of sampling a full colour spectrum through just red/green/blue filters, and the likelyhood of several sources of CA being mixed together, or being jumbled up by other aberrations.

Jonas B wrote:
If not all colors are magnified the same (lateral CA) that's a way to say the focal length is different for different colours, no?

It is probably easier to correct the error with longer focal lengths. I wonder why all (?) APO lenses are longer than normal lenses. I would love a relatively fast prime to be an APO, for example a 50/2 or maybe even a 50/1.4 if possible.

Less curvature (in the glass) is needed for longer, slower lenses, which introduces less aberration to start with. A 400/4 doesn't have to work as hard as a 300/2.8 for the same level of correction.

Short-ish focal length, fast, nearly diffraction limited lenses do exist, but not in a general purpose sense. They're usually built to work in monochromatic light (which tosses CA out the window), and at a fixed magnification.

A general purpose 50/1.4 that was truly Apochromatic wide open would be terribly expensive, and a bit pointless anyway as there aren't any large sensors available or on the horizon that could make use of its output.

Oct 08, 2009 at 01:16 PM

Jonas B
Offline
Upload & Sell: On

p.1 #15 · Zuiko 24mm 2.8 purple fringing at infinity

olyacme wrote:
(...)
A general purpose 50/1.4 that was truly Apochromatic wide open would be terribly expensive, and a bit pointless anyway as there aren't any large sensors available or on the horizon that could make use of its output.

I realize I probably step right into some trap here... but anyway: Why would the lens be terribly expensive and just not unusually expensive for a 50/1.4? And why aren't there any large sensors available for it?

I mean, Panasonic sells their Summilux-D 25/1.4 that is pretty expensive (USD900 or so) for a fast normal (or fast... it behaves like a 50/2.8 on a FF camera). Other fast normals cost less, but like the panasonic they all suffer from quite some CA of different sorts. I for one would be prepared to pay more for a good fast normal if it was at least APO enough not to show any defocus CA.

Are there terribly technical problems making such a lens? I find ordinary sensors make good use of other APO lenses...

The shortest APO lens I know of is the expensive and slow Coastal Optics 60/4. I guess there is a reason for this, but what is it? From 75 and 90mm APO lenses seem to be fast, small and available.

Oct 09, 2009 at 01:16 AM

Spyro P.
Offline
Upload & Sell: Off

p.1 #16 · Zuiko 24mm 2.8 purple fringing at infinity

Guys, is there any decent reading online on common lens faults, preferably with real-life example photos?

Oct 09, 2009 at 02:05 AM

cogitech
Offline
Dedicated FM
Upload & Sell: On

p.1 #17 · Zuiko 24mm 2.8 purple fringing at infinity

Spyro P. wrote:
Guys, is there any decent reading online on common lens faults, preferably with real-life example photos?

I've always loved Paul Van Walree's articles on lens aberrations.

Very good reading (even if he does use the annoying term "purple fringing")

http://toothwalker.org/optics.html

Oct 09, 2009 at 02:16 AM

StevenPA
Offline
Upload & Sell: On

p.1 #18 · Zuiko 24mm 2.8 purple fringing at infinity

Canon Lens Work III has some good info as well.

http://www.canon-europe.com/Support/Documents/digital_slr_educational_tools/en/ef_lens_work_iii_en.asp

Oct 09, 2009 at 02:31 AM

olyacme
Offline
Upload & Sell: Off

p.1 #19 · Zuiko 24mm 2.8 purple fringing at infinity

Jonas B wrote:
Are there terribly technical problems making such a lens? I find ordinary sensors make good use of other APO lenses...

Besides the growing difficulty of maintaining the same level of correction as the glass grows more curved (as the lens gets faster), the bar for true Apochromatism also slides further away.

This is because the faster the lens, the larger its aperture (the smaller the Airy disk), and hence the higher its potential resolving power, and because Apochromatic is (formally) defined as:

Chromatic Aberration for three widely spaced colours held to within the Airy disk.
Spherical Aberration for two widely spaced colours held to within the Airy disk.
Correction for Coma held to within the Airy disk.

There are some other aberrations photographers should care about, but this definition basically boils down to being diffraction limited.

Most of the lenses marketed as APO aren't anything close when shot wide open. They may be CA (and, hopefully, SA and Coma) free when shot with a current sensor, but if one put more resolving power behind them one would find all the familiar aberrations present. Some manufacturerS even use APO when marketing lenses that visibly poorer wide open than stopped down even on current sensors. Stopped down, to lower the bar by increasing the diffraction limit, these lenses probably do eventually become formally Apochromatic. The manufacturers just don't tell you at what f-stop, which makes the marketing term almost worthless.

So, the upshot is that a 50/1.4 which was marketed as being as well corrected wide open as at f/2.8, and Apochromatic at f/2.8 on a 135 sensor, would be a great beast of a lens, suitable for capturing 200MP images.

Oct 09, 2009 at 03:05 AM

Spyro P.
Offline
Upload & Sell: Off

p.1 #20 · Zuiko 24mm 2.8 purple fringing at infinity

Nice, thanks guys

Oct 09, 2009 at 04:55 AM

melcat
Offline
Upload & Sell: Off

p.1 #21 · Zuiko 24mm 2.8 purple fringing at infinity

All the OM Zuikos focus past infinity unless someone has been in there fiddling. I've had this lens for more than a decade and used it on half a dozen OM bodies, and, like all my other Zuikos except one, the focus stop is about 0.5mm past infinity measured on the barrel. There's a particular tech in this town who likes to "fix" this and I suspect he's had his hands on that one lens sometime in the past.

I'm looking at a shot from my archive taken with my copy of this lens on a Canon 5D. There's lateral chromatic aberration towards the corner of the frame: magenta to the left and green to the right (centre) of some rigging-like spaceframe architecture. It was a very bright day and I remember discarding several other shots taken together with this with another lens that couldn't handle the flare. There's spherical aberration and some coma. All of this is visible in Bridge's loupe tool but it appears I decided not to bother correcting it as it would not be visible at any normal print size. I've also seen lateral chromatic aberration with this lens on slide film - under a microscope.

The Zuiko 24mm f/2.8 MC wasn't APO and made no claim to be.

But there's no green opposite from the magenta fringe on the gable of the OP's shot, so it isn't *lateral* chromatic aberration. And, as I said, this lens's lateral chromatic aberration is much less severe than what we see here.

Oct 09, 2009 at 12:50 PM

Jonas B
Offline
Upload & Sell: On

p.1 #22 · Zuiko 24mm 2.8 purple fringing at infinity

olyacme wrote:
Besides the growing difficulty...
(...)
So, the upshot is that a 50/1.4 which was marketed as being as well corrected wide open as at f/2.8, and Apochromatic at f/2.8 on a 135 sensor, would be a great beast of a lens, suitable for capturing 200MP images.

I need some time to digest all this. I think I can get used to call what usually has been, and is, called longitudinal CA for defocus CA instead. I understand the differences between the old plain CA, or lateral CA, and the "new" longitudinal CA.

For an APO and relatively fast 50mm lens, well, maybe a great beast of lens isn't on any manufacturer's agenda... Oh well. I'm tired of defocus CA which in my opinion made us all used to see ugly green stuff in the background OOF parts of images, or strange tone transitions in B&W images. (Maybe the Leica 90Cron AA is the fastest and shortest APO lens I've learned to know. It wasn't perfect though and I sold it for a mix of reasons.)

I'll try to relax.

Thank you for all comments and answers!

Oct 09, 2009 at 02:00 PM

Toothwalker
Offline
Upload & Sell: Off

p.1 #23 · Zuiko 24mm 2.8 purple fringing at infinity

cogitech wrote:
I've always loved Paul Van Walree's articles on lens aberrations.
Very good reading (even if he does use the annoying term "purple fringing")
http://toothwalker.org/optics.html

A purple fringe is a purple fringe, isn't it? Is there a good reason not to use the term "purple fringing"?

Oct 09, 2009 at 03:11 PM

cogitech
Offline
Dedicated FM
Upload & Sell: On

p.1 #24 · Zuiko 24mm 2.8 purple fringing at infinity

Toothwalker wrote:
cogitech wrote:
I've always loved Paul Van Walree's articles on lens aberrations.
Very good reading (even if he does use the annoying term "purple fringing")
http://toothwalker.org/optics.html

A purple fringe is a purple fringe, isn't it? Is there a good reason not to use the term "purple fringing"?

I think the term is too broad and misleading. It seems to me that "purple fringing" can be caused not only by a combination of red and blue chromatic aberration, but also by many other possible factors, including:

* Chromatic aberration in each CCD cell (microlenses)
* Image processing and interpolation artifacts (almost all CCDs require considerable processing)
* Stray ultraviolet light
* Stray infrared light
* Image bloom from overexposure
* Leaks between cells of the CCD

So, CA can be purple, but "purple fringing" is not necessarily always attributable to lens CA. The terms are not synonymous. For this reason, I prefer that the use of the term "purple fringing" not be used when referring to lens CA which just happens to be purple due to a combination of red and blue CA. If the fringing is purple due to red/blue CA in the lens, then it should be referred to as CA. Calling it purple fringing opens up a can of worms that leads to misunderstanding and confusion on the part of those who are not aware of the many other likely causes of "purple fringing". I've seen too many people conclude that a lens has CA because they see "purple fringing" in their photos. Often enough, the "purple fringing" they see is not an attribute of their lens at all.

This is just my relatively uninformed opinion, of course. I hope you understand that I meant no offense to you for using the term. The documentation you provide on aberrations is the best I have read.

Oct 09, 2009 at 03:56 PM

olyacme
Offline
Upload & Sell: Off

p.1 #25 · Zuiko 24mm 2.8 purple fringing at infinity

cogitech wrote:
I think the term is too broad and misleading. It seems to me that "purple fringing" can be caused not only by a combination of red and blue chromatic aberration, but also by many other possible factors, including:

I'm ambivalent on the term, as it pretty much boils down to being a subset of LoCA.

cogitech wrote:
* Chromatic aberration in each CCD cell (microlenses)

Microlenses are one per sensel so CA in them shouldn't have much effect. The very nature of a microlens works against this, as they are designed to capture light that lands anywhere on their surface and concentrate it upon the photosite that lies under their (approximate) centre. Even if it did, there is no clear reason why they'd scatter light only into neighbouring red and blue sensels, and not green sensels, to create a purple fringe.

cogitech wrote:
* Image processing and interpolation artifacts (almost all CCDs require considerable processing)

Again, there's no clear reason why interpolation artifacts should be purple and not some other colour, so I'd discount this as a cause. Processing can make an existing purple fringe worse, but unless it's broken it won't create it.

cogitech wrote:
* Stray ultraviolet light
* Stray infrared light

These are much reduced by typical cameras and lenses, but even if they get through, and happen to be interpreted by the sensor as purple, I'm OK with calling them purple fringes.

cogitech wrote:
* Image bloom from overexposure
* Leaks between cells of the CCD

Blooming is charge leaking across cells, but, yet again, there's no obvious reason why charge would prefer to leak to photosites capped only with red or blue, but not green, filters. I'd very strongly discount blooming as related to purple fringing, which negates the concern about term ambiguity.

cogitech wrote:
So, CA can be purple, but "purple fringing" is not necessarily always attributable to lens CA.

It's not a perfect term, but the vast majority of the time it's going to be LoCA at fault. It'd be a good thing if, when people used "purple fringing", they also understood that LoCA was the underlying cause though.

cogitech wrote:
The terms are not synonymous. For this reason, I prefer that the use of the term "purple fringing" not be used when referring to lens CA which just happens to be purple due to a combination of red and blue CA. If the fringing is purple due to red/blue CA in the lens, then it should be referred to as CA.

Red/Blue CA would show up far more strongly, as these two together are a much bigger component of white light than violet. It's not impossible to find a lens that leaked these, but it would be a very poor performer. Even still, I'd be OK with calling the result "purple fringing".

cogitech wrote:
Often enough, the "purple fringing" they see is not an attribute of their lens at all.

I've yet to find convincing examples where this was the case. It can probably happen in exotic circumstances, but it's certainly not even the cause of a significant minority of "purple fringing" cases with the cameras and lenses we're using.

Oct 09, 2009 at 04:39 PM


FM Forums \| Alternative Gear & Lenses \| Join Upload & Sell 1 2 end