[Calmsea]

Hi all. I'm looking for a sanity check and some guidance re my next steps in my battle against the dreaded (by me anyway) purple fringes that I have been witnessing since getting my K5 some 18 months ago. I need to know whether these are being created by the lens(es) or, as is becoming increasingly likely, the K5 itself. I have attached two examples - they are to my mind extreme.

The fringes are most noticeable when taking shots involving areas of detail with high contrast (typically leafless trees against bright skies). Whole trees take on a purple aura, which whilst quite interesting, definitely wasn't visible to my eyes when I took the shot. The first photo below provides a good example of how contrast (between black and white in this instance) triggers the effect, with many of the white window frames of the Buxton pavilion actually turning purple (they are white - honest). Zoom in on any of the detail in the second photo to see how highlights such as sunlight reflecting from cars creates the same effect. Even the road markings in this picture have turned purple, and there is of course the omnipresent purple tree on the left

Whilst evident on the DA kit lenses that came with the K5 (18-55 and 50-200) the fringes are most noticeable when using my Tamron SP Di II 10-24. The Tamron also consistently delivered poor sharpness (to me) anywhere away from the centre of the image, hence I thought that it (i.e. my particular lens) was prime suspect for the excessive fringes also. So, after some toing and froing involving the UK Tamron distributor, their European parent in Germany, and even Tamron Japan, the lens was completely rebuilt and tested. The distributor advised that 'as a precautionary measure we replaced all element group's within your lens and re aligned the light axis to suit, After a Full check and test of your lenss we found it to be within Manufacturers specification. So armed with my newly re built lens, and new hope, I went off to test it in the very rare but welcome bright sunshine that we have been having over here in the UK recently. The results were shocking! The pictures below are two of them. I sent these to the distributor. They have advised me that these are by far the worse examples of fringes that they have ever seen. They added that in their opinion, these are indicative of an issue with the camera itself, not the lens, and suggested that I contact Pentax to have my K5 examined, and corrected.

So, whilst I know that I can use the software (Pentax Digital Camera Utility 4) that came with the K5 to greatly improve these images by correcting the lateral colour aberration (I've tried it, and it appeared to work), what I don't know is whether I have to accept that owning a K5 means getting used to reviewing shots in purple, prior to correcting those that are worthy of saving, or whether what I am seeing is not normal, and hence I should look to get the K5 examined and hopefully fixed. (It will be out of warranty, hence this will be at my expense).

Since I took the two photos attached, I have updated the firmware to 1.14, run the Pixel Mapping feature, and taken a few shots of a couple of green trees against a bright sky with the 18-55 Pentax DA lens, various apertures. Guess what colour they turned out? Yep, purple!

Any help or guidance would be appreciated. In my head, if there is an issue with the K5, it has to be software/firmware related. Is there a way to perform a factory reset perhaps, and see whether that convinces it that white objects against a black background, or vice versa, don't need a purple highlight or overlay?

Help! Please!

Attached Images

View Picture EXIF

File Date

 

View Picture EXIF

File Date

 

[JinDesu]

That purple does seem very unnatural. I haven't seen that when I used my Sigma 10-20mm on my k-x.

Do you have any other lenses? When shooting with your DA 18-55, what focal length did you shoot? Do you have any samples of those?

Can you take your Tamron or your DA 18-55 and put the camera on a tripod, set to F8, and manually focus in Liveview - until distant objects look sharp, and then take a few backlit shots and show us what happens?

[adpo]

according to my rough understanding of physics, purple fringing occurs when the lens inadequately corrects for different spectra of light, resulting in some colors to be a tad out of line with everything else. So, assuming I'm not horribly misunderstanding the issue at hand, purple fringing is a lens-related issue that is inherent in the lens' optical design. Looking at various lens reviews of the Tamron, it seems to be a common issue with this this lens, with the worst purple fringing at it's widest setting. It seems to simply be something about the optical design lens itself that's causing the effect. With that being said, I've typically found this rather simple to fix in post (one click with lightroom) and you also have the added benefit of in-camera correction as well.

[dadipentak]

You don't mention the aperture setting:.fwiw, stopping down usually helps..

[boriscleto]

You don't mention the aperture setting:.fwiw, stopping down usually helps..

The EXIF on the first says f/9, f/5.6 for the second.

[snostorm]

I'm no optical expert, but in my experience, lateral CA is mostly a lens phenomenon, but the degree of purple fringing can be sensor dependent to a significant degree. Of the Pentax DSLR bodies that I've used (DS, K100DS, K10, K20, K-7. K-5, and K-5IIS), the K10 and the K-5 have shown the most PF, and strangely, since it supposedly shares essentially the same sensor as the K-5, the K-5IIs shows the least.

I shoot birds, so there are a lot of shots where there are branches against a bright sky in high contrast. I consider PF something that I have to work around and have developed a PP process to deal with it. With fast ultra and super teles costing in the multiple thousands of dollar range each, and with the limited availability of many of the best of these produced by and for Pentax mount, I found that it's just impractical to look for better glass or examples of the lenses, and I chose a long time ago to work around the problem.

Lenses with at least one Low Dispersion glass element (various trade terms are used by lens makers to indicate this -- Pentax = ED, Sigma = APO/ SLD/ FLD, Tamron = LD, Tokina = SD) will usually help with all chromatic aberrations, but these are rarely found in consumer class lenses because the LD glass is relatively expensive -- some noticeable exceptions are the DA 50-200 and DA 55-300. The best lenses I own for CA/PF control are the FA* 300 f2.8 and the Sigma EX 180 f3.5 APO DG Macro, and both display some PF in high contrast situations.

Scott

[hoanpham]

These PF are extreme. Pentax lenses do have PF, usually not so visible. Depends on the scenes, some are difficult to remove in post.
Otherwise, I don't use zoom as much.

[Calmsea]

That purple does seem very unnatural. I haven't seen that when I used my Sigma 10-20mm on my k-x.

Do you have any other lenses? When shooting with your DA 18-55, what focal length did you shoot? Do you have any samples of those?

Can you take your Tamron or your DA 18-55 and put the camera on a tripod, set to F8, and manually focus in Liveview - until distant objects look sharp, and then take a few backlit shots and show us what happens?

Thanks for the rapid responses. The only other lens I still have (other than the kit ones) is a Vivitar Series 1, 70-210. Manual focus. Used to use it in the days of film, and couldn't bring myself to part with it. Only shots I have from that recently are of the moon. No fringes there come to think of it, but not worthy of review.

With the 18-55 DA the fringing was most evident at 18mm, and wider apertures, though still there at longer focal lengths on objects framed against an overexposed sky (I wasn't being very scientific I'm afraid - I had already tried the Tamron with the new firmware and established that this had had no noticeable effect on the fringes hence was a little trigger happy with the DA). Will attempt to add an example. Will also add a previous 'tree' example with the 50-200 DA, taken in the first month that I got the camera, back in Jan 12. I contacted Pentax at the time and their advice was to use software to correct.

I will have a go at the tripod/liveview test tomorrow - 11pm here now and dark.

I hear and accept that adding the 'lightroom' step will greatly improve the results, but that kind of detracts from one of the enjoyments we get as a family, immediately sharing our efforts via the TV after a day out (currently my wife's smartphone frequently outshine my efforts with the K5 - not a fringe in sight)

[Calmsea]

I'm no optical expert, but in my experience, lateral CA is mostly a lens phenomenon, but the degree of purple fringing can be sensor dependent to a significant degree. Of the Pentax DSLR bodies that I've used (DS, K100DS, K10, K20, K-7. K-5, and K-5IIS), the K10 and the K-5 have shown the most PF, and strangely, since it supposedly shares essentially the same sensor as the K-5, the K-5IIs shows the least.

I shoot birds, so there are a lot of shots where there are branches against a bright sky in high contrast. I consider PF something that I have to work around and have developed a PP process to deal with it. With fast ultra and super teles costing in the multiple thousands of dollar range each, and with the limited availability of many of the best of these produced by and for Pentax mount, I found that it's just impractical to look for better glass or examples of the lenses, and I chose a long time ago to work around the problem.

Lenses with at least one Low Dispersion glass element (various trade terms are used by lens makers to indicate this -- Pentax = ED, Sigma = APO/ SLD/ FLD, Tamron = LD, Tokina = SD) will usually help with all chromatic aberrations, but these are rarely found in consumer class lenses because the LD glass is relatively expensive -- some noticeable exceptions are the DA 50-200 and DA 55-300. The best lenses I own for CA/PF control are the FA* 300 f2.8 and the Sigma EX 180 f3.5 APO DG Macro, and both display some PF in high contrast situations.

Scott

Thanks Scott. Good to know and very helpful - will help manage my expectations I think.

I really don't recall any issues with CA when using my old 24-70 (again a series 1) on my P30, back in the days when computers filled rooms and cost a lot more than any camera. The results were good, (or seemed it then) even when blown up to the size of the wall (transparencies). I thought (hoped?) that the K5 and Tamron would exceed those results, easily, but I'm beginning to reconsider that. Appreciate there are so many pluses (I would not be shooting off test images, posting and getting such useful immediate feedback if I were still using film..). One thought I did have, prior to getting the Tamron overhauled = ditching that and getting the 15mm Prime. SMC Pentax-DA 15mm F4 ED AL Limited Reviews - DA Prime Lenses - Pentax Lens Reviews & Lens Database
But that was when the lack of sharpness was my main bugbear. From what you and others have said, I'm guessing I would still be haunted by the purple trees with that though. Ho hum.

[stevebrot]

Welcome to digital photography. As you noted, PF (not the same as LoCA) was not and is not a concern when shooting film*. Some lenses are more prone than others as are some subjects and sensors. Conventional wisdom is that you can expect better results with digitally optimized lenses, though I have to admit to not having a huge problem with most of my vintage glass.


Steve

* Field curvature in combination with LoCA can produce something that looks a lot like PF, but close examination will show that the false color is localized to a radial band around the center of the frame (usually towards the margins). The Zuiko 35/2.8 on my Olympus XA has this fault.

[BirdDude007]

I have a lens that is nicknamed the Purple Monster, and even though I have used it several times( for the reach, it is a 70~300 zoom I got on the cheap, just using it till I can afford better zooms) and many times I thought to myself I should just toss it in the garbage, it takes decent shots that I can live with, but once I saw the purple fringing I was just about to give up.
Enter-Lightroom 4, once I located the "lens Corrections" menu I became a big time believer in Lightroom 4, although I am used to using Lightroom 3 it just never did very well when it came to purple fringing.
Now when I input the lens model and fiddle a little bit with the fringing sliders, I manage to totally remove the fringing in 99% of my shots!
Now that lens is still a keeper at least until I can afford a better Zoom.
I get some fringing in my other pentax lenses but not no where as bad as that Tamron, however, when I put my 50mm M lens on my K-5, I have yet to ever find a shot with any fringing on it at all....lol.....love that ol'lens it just keeps chugging along and producing awesome image quality.

[rawr]

I'm with BirdDude007. Purple fringing (PF) happens, and is easily eliminated, especially with Lightroom, but also with DxO Optics Pro and other tools.

My worst PF (and green fringing) lenses are my Tamron 70-300 and Pentax K 300. I also had a bad copy of the Tamron 17-50 that was a un-correctable PF monster, but the 2nd copy I got has no PF at all. So I think PF is 99% a lens issue, not a camera issue.

I also have to ask OP: were you using any filters on your lens(es)? If yes, take the filters off and just use a hood. I wouldn't be surprised at all to see a bad UV filter introduce PF into a lens.

[stevebrot]

when I put my 50mm M lens on my K-5, I have yet to ever find a shot with any fringing on it at all....lol.....love that ol'lens it just keeps chugging along and producing awesome image quality.

The same here, only on my K10D. I just took my M 50/1.7 out into the yard and was not able to get any purple fringing. A very, very little LoCA on OOF branches, but that is all.


Steve

[crewl1]

The purple is not what I usually see in even my most PF prone lenses.
In some of your images part of the tree exhibits purple and the other part of the tree in similar light and contrast situation does not.
This would make me think it is a camera/sensor problem, since most lens errors will be uniform across the image or otherwise not have such abrupt changes in behavior across an image.
Does the purple change as you change focus? If yes then it is lens. If not then I believe it is camera.

[rawr]

This would make me think it is a camera/sensor problem

This got me to thinking about whether there are any settings that the OP may have made to their custom image or JPG settings that may be playing a role too. It's almost like a camera filter is active on high-contrast edges. Maybe a reset of camera custom functions and menu settings (pp 335-336 of the K-5 manual) to their defaults is worth a try.

Do these fringes show up when opening a PEF or DNG in PDCU4, or just JPG's out of camera, by the way?