Nikon's handling of blues.

ian408

I've seen a couple of examples in the past of blues appearing purple with Nikon sensors. Especially night time shots. I'In this example; the car is all the same color, a deep blue.



However you can see a very distinct purple color on the quarter panel. Now I do know the paint will look different when viewed from different angles but I don't think that explains the purple.

This was shot with a d70 and a 14-24. Original was a raw image. Other shots I've see from P&S cameras look good color wise (same show).

Any hints as to why the purple? I'll re-shoot it at some point with a Canon just to see the difference.

Wil Davis

What a strange effect; I can't say I've ever noticed such a thing with any of the digital cameras I've had over the past 10 years (Nikon, Olympus, Canon). Is the colour pure, or is it a metallic or whatever they call it nowadays? I think it was called "opalescent" (mid '60s, UK). I've seen something similar especially noticeable on deep plum reds (metallic), where the colour looks quite different depending on the angle of the light and the viewpoint.

Curious…
- Wil

ian408

The paint is almost like Imron was 'back in the day' meaning it's a deep color with some metallic in it. I'm confused because the P&S shots from others look perfectly fine.

As I mentioned, I've seen this before and thought it had something to do with the sensor's ability to render colo(u)r.

Qarik

if I understand you correctly..the blue portion and the purple are to the nked eye exactly the same color? Then what explain the distinct line beween them?

My thoughts are..they are infact different paint colors and some trick of the light is making them appear the same...or paint thas some quality that looks different when viewed at different angles and blue portion is looking straight on for example and the purple portion is on a part of the car where the angle of viewing more oblique.

If it was a sensor issue wouldn't the blue also be just purple?

ian408

Sorry Qarik-car is all the same color in that area. The distinct line is a bend in the metal.

Qarik

ian408 wrote:

Sorry Qarik-car is all the same color in that area. The distinct line is a bend in the metal.

Then it appears that the paint changes colors when viewed from different angles pershaps due to different sources of light hitting it from different angles.

ian408

Qarik wrote:

Then it appears that the paint changes colors when viewed from different angles pershaps due to different sources of light hitting it from different angles.

Can you explain why pictures from a non-Nikon P&S don't exhibit the same problem?

ziggy53

Ian,

What color temperature are you using for the RAW conversion and what were the lighting conditions?

Qarik

ian408 wrote:

Can you explain why pictures from a non-Nikon P&S don't exhibit the same problem?

do you have a shot from P&S from a similar angle? Also was the lighting static at the venue? No strobes, colored lights, car wasn't on a spinning stage, etc? Also did you bounce flash from the SLR vs straight or no flash from the P&S?

brianb

Could be its weak IR filter?

http://www.pentaxforums.com/forums/pentax-dslr-discussion/2170-black-sweater-looks-purple-hot-mirror-filter.html

Brian

ian408

ziggy53 wrote:

Ian,

What color temperature are you using for the RAW conversion and what were the lighting conditions?

Lighting conditions vary. But most are outdoor, bright sunlight. Some are flashed all are AWB. I've tried auto correcting in PS as well as variety of different color temps in post.

ian408

brianb wrote:

Could be its weak IR filter?

http://www.pentaxforums.com/forums/pentax-dslr-discussion/2170-black-sweater-looks-purple-hot-mirror-filter.html

Brian

Thank you. I will read trough that.

ziggy53

ian408 wrote:

Lighting conditions vary. But most are outdoor, bright sunlight. Some are flashed all are AWB. I've tried auto correcting in PS as well as variety of different color temps in post.

If you wish to send me that original RAW file for the image that started this thread, I'll take a look and see if I can discover the cause.

ian408

ziggy53 wrote:

If you wish to send me that original RAW file for the image that started this thread, I'll take a look and see if I can discover the cause.

Thanks Ziggy. I may do that tomorrow.

ziggy53

OK, I think I have the most likely causes and best correction.

1. IR contamination

The Nikon D70 has a relatively weak IR cutout filter, and that appears to be a likely major cause. If the colors of the engine interior and fenders looked the same to the eye, then it is because the human eye has little sensitivity to IR.

Likely reasons for a difference:

The different painted regions probably have a different paint treatment, possibly related to the primer and number of top coats. Possibly related to pigment additives as well.

2. Wax or polycoat treatment of the exterior painted surfaces (but not the engine compartment components).

Both wax and polycoat treatments can change the reflective properties of the painted surfaces. Wax can form crystalline molecules in the wax coating and polycoat actually bonds to the painted surface through polymerization. Both processes can have an impact on color purity and tone shifts, subtle to the eye but amplified to a photographic sensor and subsequent image processing.

I suspect that either or both of these scenarios may apply in this case, with the IR reflective difference the most likely (IMO).

The most appropriate corrective processing seems to be in ACR, locate the HSL/Grayscale menu and move the Purple slider down to -75 or so. Both the problem and the correction are most obvious at colder white balance combined with more exposure. Warmer WB seems to reduce the effect of the color differences intrinsically.

In researching this issue, Nikon is not alone in the problem but as digital camera manufacturers started using stronger IR cutout filters over the imager the IR related issues relating to blue-purple color shift seems to have subsided.

Deep red colors with very subtle tonal gradations still occasionally confound imagers as well, although the problem manifests somewhat differently.

ian408

Thank you for your help Ziggy.

My approach to resolving this was similar to your suggestion. I ended up removing some of the Magenta as well--this helps on the flat quarter panel. It's not perfect but it will serve what I need to do.

Thanks again for all your help.

angevin1

Sounds like you got this solved somewhat, but what part of the car is this? Could the purple part be a plastic bumper cover and the blue part a metal component? Meaning of course that they absorb and reflect light differently also.

ian408

angevin1 wrote:

Sounds like you got this solved somewhat, but what part of the car is this? Could the purple part be a plastic bumper cover and the blue part a metal component? Meaning of course that they absorb and reflect light differently also.

It's the front quarter panel and under the hood. There's no plastic on the paint.

ian408

Here's the whole image.

angevin1

ian408 wrote:

Here's the whole image.

Looks like you fixed it! Sweet! Was this the only component area that acted up?

Pindy

I'm slightly relieved to see this thread—sorry Ian. I have noticed a blue-purple shift on the D700, but mostly as it pertains to the JPEGs and in-camera cooking. When ACR or LR process their initial default rendering from the RAW, the purple tends to almost always be replaced by a blue, if it's supposed to be blue. Often certain blues, in my camera anyway, tend to come out as a kind of cobalt color, in real contrast to the defaults in the Canons I've owned. I'll try to find an example. Ziggy's research seems solid, but i wonder if there's an actionable solution that can be done at the time of capture?