Smeared greens/yellows from K3


aliengrove

Link Posted 06/05/2014 - 00:11
Has anyone noticed the K3 tending to smear greens/ yellows when processing files heavily? I've seen this reported elsewhere (though only green was mentioned), and have come across it myself when processing aerial shots, which require expanding a very narrow histogram. So much so that a batch of shots I took over Turkey recently are unusable. With shots not requiring heavy processing, I never see it at all, except with a couple of crap lenses I own. I'm wondering if it has something to do with shooting through windows, using lenses that aren't sharp enough for 24mp. The biggest surprise has been my FA 24-90mm, which still works splendidly on my K5, but produces this effect on the aircraft at any aperture, and sometimes even on the ground when wide-open, on my K3. My Sigma 24-60mm doesn't seem to be affected.

I'd be interested to know if anyone else has noticed this, and also if anyone has any theories into what is going on.
Flurble

My Website
PPG
flickr
G+
Facebook
Last Edited by aliengrove on 06/05/2014 - 00:13

Smeggypants

Link Posted 06/05/2014 - 00:15
got any examples to post?
[i]Bodies: 1x K-5IIs, 2x K-5, Sony TX-5, Nokia 808
Lenses: Pentax DA 10-17mm ED(IF) Fish Eye, Pentax DA 14mm f/2.8, Sigma 17-70mm f/2.8, Pentax-A 28mm f/2.8, Sigma 30mm F1.4 EX DC, Pentax-A 50mm f/1.2, Pentax-A 50mm f/1.4, Pentax-FA 50mm f/1.4, Pentax-A 50mm f/1.7, Pentax DA* 50-135mm f/2.8, Sigma 135-400mm APO DG, and more ..
Flash: AF-540FGZ, Vivitar 283

richandfleur

Link Posted 06/05/2014 - 00:49
aliengrove wrote:
using lenses that aren't sharp enough for 24mp.

I'm dubious of this personally, believing that a crap lens is crap at all resolutions. The only issues I expect would be using on different sensor sizes, where the sensor 'looks through' differing areas of the whole lens. A sensor size (ie APS-C) looks through the same section of the lens surface regardless of resolution.

With regards to the colour issue, I found I got this happening years ago on a landscape mode of my K100d. I'm not in anyway saying they're similar cameras, but perhaps the processing applied to landscape shot photos is heavy on the yellows (in my case) or similar on your Pentax K3.

Are you shooting RAW?
Can you reduce the saturation of that colour specifically via Lightroom?
What 'mode' were you on when you shot these photos?

aliengrove

Link Posted 06/05/2014 - 00:56
I am shooting RAW. I'll post some examples tomorrow. I've just been looking through a few shots again, and I've realised that for some reason any slightly oof areas are much worse if they are bright green.
Flurble

My Website
PPG
flickr
G+
Facebook

McBrian

Link Posted 06/05/2014 - 06:58
This may be a long shot and way of the mark.

With aircraft windows being made of polycarbonate (Perspex, Lexan etc.) I'm wondering if they have some optical properties that can distort the the light waves and in effect filter some colour frequencies. I know some grades of polycarbonate have chemicals added to make them UV stable.

Have you found the same problem with images taken on Terra firma with only air in front of the lens?
Cheers
Brian.
LBA is good for you, a Lens a day helps you work, rest and play.

johnriley

Link Posted 06/05/2014 - 08:11
The aircraft window is not only part of the optical system, but also smeared, abraded and maybe or maybe not clean. It's not an optically perfect view and it's a bit like using a filter of unknown properties and quality on the front of the lens.

We know that some lenses don't like filters very much, so some will be better than others shooting through plane windows. Maybe the higher the resolution of the sensor the more it shows up these defects.
Best regards, John

aliengrove

Link Posted 06/05/2014 - 11:20
I usually get the windows cleaned, or do it myself, if I'm going to be taking pictures on the flight, and there's no abrasions on the flight-deck windows. Optically they are obviously not perfect, being of a sandwich construction and nearly an inch thick.

After doing some more research, I'm wondering if it's something to do with differing amounts of diffraction of different colours. What I don't understand is why the effect would be worse on the K3 than on the K5. Here's an image of diffration through a prism.




As I'm generally shooting at an angle, the windows probably act as a prism. I wouldn't notice blues so much as they are generally sky or water. On the ground I can also see the effect, though to a lesser extent, on my FA24-90mm at 90mm wide-open. (f4.5). It doesn't happen at all on the K5. On my old Tamron BBAR 90mm f2.5 lens, it doesn't happen at all at f2.8. I'm wondering if the K3 tends to use the red end of the spectrum more for phase-detection AF than the blue end, resulting in greens actually being oof and therefore smeared? I've just adjusted the focus on the FA 24-90mm and will try it out on my next trip.
Flurble

My Website
PPG
flickr
G+
Facebook

aliengrove

Link Posted 06/05/2014 - 11:48
An example-
Original image



100% crop



Flurble

My Website
PPG
flickr
G+
Facebook

aliengrove

Link Posted 24/05/2014 - 02:00
I've added the post I put to David Storm's query on my thread in the Classifieds here as it's probably a more appropriate place for it than the Classifieds section....


David, I'm still trying to get my head round some of it, and will post more detail when I have. I think I'm dealing with effects caused by shooting in a high UV environment, possibly aggravated by prismatic effects from the aircraft windows. I also think the effects may be worse with a higher pixel density, ie 24mp as opposed to 16mp on the same sized sensor.

An example of the stuff I'm trying to work out is here, the link provided by someone on another site (Z). Interestingly, this paper was all based on experiments with a K5. The findings here are mainly to do with low-key scenes, while I am dealing with low contrast (very low in a lot of cases), often with a very low dynamic range (if no sky is involved).

I'm going to experiment with using different white balances (even ridiculous ones) to see if a) I get better focus accuracy, and b) if a different WB might actually lessen the green smearing I sometimes get. However, my experiences in IR lead me to think I probably need a UV filter with a cut-off of at least 450nm instead of 420nm. A standard UV (420nm) filter makes no difference at all. I've not been able to find such a filter anywhere, probably as they will affect colour reproduction. My 665nm IR camera is affected very much less by atmospheric haze, and also achieves better focus than both my K5 and K3. My GR seems unaffected. My experiments in processing 665nm images makes me think I can deal with any colour shifts presented by images shot above 450 or even 470nm.

If anyone fancies going through the article and letting me know their thoughts, I'd be very interested to hear them. Meanwhile, I also have some other esoteric stuff to sift through and then if anyone's interested I'll post more stuff. Also, if anyone knows where I could find a filter that blocks everything below 450-470nm, I'd be extremely grateful for the info.
Flurble

My Website
PPG
flickr
G+
Facebook

aliengrove

Link Posted 24/05/2014 - 03:10
Meanwhile, a better example of the nature of what I'm talking about....100% crop




Initially I thought it might be defects in the windows; usually this occurs at the edges of the frame, where I've shot too close to edges of the window, but this is occurring in a lot of images all over the place, and only in areas of green. Images with lots of green generally have worse focus overall too. It also seems to more affect areas of green bordered by a different colour.
Flurble

My Website
PPG
flickr
G+
Facebook

kh1234567890

Link Posted 24/05/2014 - 09:29
How are you post-processing these ?

To me it looks simply like a bit too strong noise reduction.

Any chance of posting an example RAW somewhere ?

The few times I've tried shooting through plane windows at altitude the problem has always been low contrast. You then end up stretching the dynamic range in p-p with subsequent noise increase.

Why do you need autofocus at all ? Surely if your subject suddenly becomes anywhere else than at infinity then it is time to brace, brace ...
Flickr Stream

johnriley

Link Posted 24/05/2014 - 10:08
So we have a comparison benchmark, do you think you could post in the thread an example of a good shot from the K-5?
Best regards, John

davidstorm

Link Posted 24/05/2014 - 10:49
I haven't noticed anything with my K-3 Jon, apart from the way it renders greens and yellows which to my eye is more of a 'natural' tone than the K-5IIs gives. I often turn down the yellow saturation in images shot with the K-5IIs because grass and trees look a bit artificial, but I haven't had to do this with the K-3.

Having said this, I don't take any shots at 30,000 feet through airliner windows!

Regards
David
Flickr

Nicola's Apartments, Kassiopi, Corfu

Some cameras, some lenses, some bits 'n' bobs

ChrisA

Link Posted 24/05/2014 - 11:30
aliengrove wrote:
Meanwhile, a better example of the nature of what I'm talking about....100% crop





I don't really understand what's wrong with this. By 'smeared', do you mean 'lacking detail'?

Nothing in that crop looks all that sharp, so I'm not clear what the specific issue is. The buildings are more contrasty than the patches of green, so they'll look sharper - but I don't think they are.

Am I missing something?
.
Pentax K-3, DA18-135, DA35 F2.4, DA17-70, DA55-300, FA28-200, A50 F1.7, A100 F4 Macro, A400 F5.6, Sigma 10-20 EXDC, 50-500 F4.5-6.3 APO DG OS Samsung flash SEF-54PZF(x2)
.

Gravelrash

Link Posted 24/05/2014 - 13:05
I couldn't resist commenting despite knowing nothing.

I agree with Chris A that I see lack of detail rather than colour smear. I'm presuming that when a sensor tries to replicate the light concentrated on it by a lens it will average out over an area of limited definition.

You also talk a lot about 100% crop, a term I see a lot but do not fully understand in photo references.

The point I am querying, that may have some bearing, is...
You take two identical photos with the same lens but two different sensors. One sensor is 16mp the other is 24mp but they occupy the same physical area on a sensor. The 24mp are crushed together more tightly.

Therefore if what you refer to as a 100% crop is actually in effect a screen capture of say 1080 pixels by 764, or whatever, you would in effect be receiving a visualisation of a much more magnified area on the 24mp image than the 16mp.

There is a great deal of presumption on my part and I apologise for my ignorance if I'm barking up the wrong tree, but if by getting a K3 over a K5 you were hoping for a free ride I think you have met the limit of the optical properties regards air pollution, plane window and lens.

As far as the colours go, that will be an artifact of the sensor and or software algorithms, which is why I like the pentax colours over my brothers Sony colours. You may be able to find a software fix for this but may also find that you are better using the camera that you like the most.

I am sorry if I have completely missed the point. I was just fearing that perhaps we had failed to see the elephant in the room.
Steve

Sometimes I'm serious and sometimes not, but I consider sarcasm an artform. Which is it today?
Add a Comment
You must be registered or logged-in to comment.