5D blue cast especially in the shade areas

cabinetbuff

I just purchased a 5D kit and noticed a strange blue cast that I was able to consistently reproduce this evening by moving a camera slightly in relation to a shot that did not generate the blue cast. Here are the two shoots and their EXIF info:


Normal image - EXIF




Blue Cast image - EXIF



What do you think is causing this? Wrong WB being picked by the camera? Is it a camera defect?

cabinetbuff

Since I took these images in RAW I tried to adjust the WB (using the click tool in DPP) on the image with the blue cast but I wasn't able to adjust it to an acceptable level where the blue cast would be completely removed so it must be something else at play or a combination of things not just WB ...

ziggy53

Photographing lights or colored lighted displays can affect white balance so I suspect the displays had something to do with it. Also, your camera has metering "zones" which are weighted in measuring the overall scene and moving the camera slightly could affect how those zones are impacted. (At least I don't think the 5D has an external WB sensor.)

You best bet in scenes like these is to set your WB yourself and not depend upon Auto WB. Of course if you shoot RAW you can determine and set WB in the RAW converter and set consistant WB through a series of similar images.

ziggy53

cabinetbuff wrote:

Since I took these images in RAW I tried to adjust the WB (using the click tool in DPP) on the image with the blue cast but I wasn't able to adjust it to an acceptable level where the blue cast would be completely removed so it must be something else at play or a combination of things not just WB ...

Since you shot in RAW you should be able to manually set the WB in DPP to at least match one image to the other.

cabinetbuff

ziggy53 wrote:

Of course if you shoot RAW you can determine and set WB in the RAW converter and set consistant WB through a series of similar images.

Thanks for your thoughts on this ... I also thought it had something to do with WB and since both of these pictures were taken in RAW I just tried to PP the one with the blue cast by using a WB click tool in DPP but I failed to remove the blue cast -- it just becomes subdued but remains visible ...

pathfinder

If these were shot in RAW, indoors, how do the RAW files compare, when both are set to Tungsten light balance in the RAW converter? If the lighting was not tungsten, what was it? Fluorescent or skylight?

cabinetbuff

pathfinder wrote:

If these were shot in RAW, indoors, how do the RAW files compare, when both are set to Tungsten light balance in the RAW converter?

They both look horrible in Tungsten: the image without blue cast has blueish tone throughout; while the one with the blue cast to begin with now has it's blue cast shade areas amplified.

pathfinder wrote:

If the lighting was not tungsten, what was it? Fluorescent or skylight?

The lighting was just natural light coming int though windows overhead and on the right side of the frame. There was no artificial lighting involved at all.

Scott_Quier

I don't really think you're looking at a camera defect.

I'm guessing these were shot AWB. If so, then about the only reliable way to get the color temperatures to match is to make another shot (at the same shoot) of a known gray and use that to standardize the color temperature of the group. In this case, since you didn't get a gray card shot, to get the color temperature the same, pick a spot (I would suggest somewhere on the case of the VCR) and take color temperature and tint readings. Load up the other and enter those values into the WB areas of ACR.

The over-all difference in contrast may be contributing to the effect as well.

pathfinder

If this was shot by skylight coming in through a window, then it was in shade, was it not?

When I measure the whitish wooden front to the cabinet, it is not blue, but yellowish - typically 240,238, 218 in the first image and the second as well, but the electronic faces are definitely blue..

Did you try rebalancing the White balance in the RAW converter by clicking with they dropper on the black ( dark gray ) of the disk player? I believe it should be neutral dark gray.

cabinetbuff

pathfinder wrote:

Did you try rebalancing the White balance in the RAW converter by clicking with they dropper on the black ( dark gray ) of the disk player? I believe it should be neutral dark gray.

This sounds like a good test ... Do you know which RAW converter will allow me to do this? I've only used DPP and I can't seem to find a way to do that there -- it only allows for a "Click white balance" adjustment where I have to find a "true" white survace to click on in the image.

Thank you.

cabinetbuff

Scott_Quier wrote:

... In this case, since you didn't get a gray card shot, to get the color temperature the same, pick a spot (I would suggest somewhere on the case of the VCR) and take color temperature and tint readings. Load up the other and enter those values into the WB areas of ACR.

The over-all difference in contrast may be contributing to the effect as well.

Thanks Scott -- I'll try this -- I just need to figure out how to do this in DPP.

pathfinder

cabinetbuff wrote:

This sounds like a good test ... Do you know which RAW converter will allow me to do this? I've only used DPP and I can't seem to find a way to do that there -- it only allows for a "Click white balance" adjustment where I have to find a "true" white survace to click on in the image.

Thank you.

Either Adobe RAW converter 4.3 with PSCS3, or Lightroom allow clicking on what should be a true neutral tone to correct for white balance. I cannot be sure that either the "white cabinet" or the "black" cases are true neutrals, but the cassette cases almost certainly are a "black"

In Lightroom or ARC you just click with the Color Balance eye dropper on the gray tone that should be neutral to correct the color cast. This can be very helpful if you can find a known neutral in your image.

Are you certain there was no light except from the sky coming through the window?

I shoot with a 5D frequently also, and like it a lot, but Canon's AWB is not that good, especially in low light. I was most impressed by the Nikon D3's ability to set white balance via AWB in low light, but the 5D does not match it in my opinion. The 5D is a great camera, but I find I need to ride herd on the AWB setting. YMMV!

Robinivich

I have to admit, at first glance it looks like flare, is it possible that this is what we're looking at?

The bottom left corner seems nearly identical in both images, whereas the top right is night and day, smells like light from an open window or other light source glancing off the front of the lens

LiquidAir

Robinivich wrote:

I have to admit, at first glance it looks like flare, is it possible that this is what we're looking at?

The bottom left corner seems nearly identical in both images, whereas the top right is night and day, smells like light from an open window or other light source glancing off the front of the lens

I agree. If what you are worried about is the bluish cast at the top of the frame, that looks like flare to me too. To me it looks like you are picking up a reflection off the screen onto the lens. By any chance, do you have a UV/protector filter on the lens? If so, try taking it off and see if that cuts down on the problem. Also make sure the front element of your lens (24-105?) is clean.

cabinetbuff

I've tried everything I could imagine within the realm of DPP's abilities to adjust WB but I cannot remove the blue cast from the second image.

I think you are right -- it is flare.

What I don't understand is why? The lens is brand new 24-105, I also have another 24-105 that produced exact same results and I even tried 24-70 2.8 which resulted in the same bluish cast.

I used lens hood in both cases -- the shot with the blue cast was aimed closer to the active LCD screen so is it possible that the reflection of the screen somehow caused a lens to flare? Notice though that the screen was not in the shot it was just a little closer.

cabinetbuff

pathfinder wrote:

Either Adobe RAW converter 4.3 with PSCS3, or Lightroom allow clicking on what should be a true neutral tone to correct for white balance ...

Thanks for the detailed information. I've tried to stick with DPP but It looks like ARC might have some things I am currently missing with DPP.

pathfinder wrote:

Are you certain there was no light except from the sky coming through the window?

The only other light source was from the screen itself but as you can see -- it never entered the frame directly it was just closer on the "blue-cast" shoot. I never imagined that an LCD screen proximity can cause such a havoc.

pathfinder

I think the second shot has flare also.

I took the liberty of downloading both screen shots here on dgrin and opened them into PhotoShop via Adobe RAW converter. My goal was to use the WHite Balance eye dropper and see how it would work on these images.

Interestingly, when I clicked on the black components I got different results all over their fronts. I got even different results from the gray DVD cases at the lower right of the image. The highlights in the first image were over exposed and I had to use the Recovery slider to bring it back, and then white balancing went a little better.

Here is my result with the first image. I never did get absolutely neutral readings with R,G,B equal on the component's front panels.

pathfinder

The second image had more flare and was even more resistant to correcting in ARC 4.3 with the White Balance eye dropper. I used the same settings in the Raw converter for the second image that I used for the first. But the blue cast is still present, tho to a lesser extent. You images were 16 bit, and I reduced my edits to 8 bit to allow uploading as attachments.

The 24-105 lens can have problems with flare sometimes, particularly with back lighting or shooting into the light. Mine does too, but I have not noticed it as much as I used to, maybe I am just more aware of it and watch for it now. It is the lens I use more often than all the rest combined.

I wonder if the bright, yellow LCDs played a role in the white balance issue.

Are the cabinets tan or yellow? My inclination is to think they are off white. The blue cast can be killed by a yellow filter ( yellow will kill the blue and helps the electronic panels get closer to nuetral ) but that makes the cabinets a definite yellow that does not look right either.

Not sure this is even an improvement over your second image, either.

cabinetbuff

pathfinder wrote:

The second image had more flare and was even more resistant to correcting in ARC 4.3 with the White Balance eye dropper.

pathfinder wrote:

Not sure this is even an improvement over your second image, either.

Thank you for taking a shoot at it -- I really appreciate it! I agree with your assessment that there's no improvement. I saw the same results in DPP.

pathfinder wrote:

I wonder if the bright, yellow LCDs played a role in the white balance issue.

I am 100% sure now that LCD had something to do with it.

pathfinder wrote:

Are the cabinets tan or yellow?

They are more yellow than tan. It's birch wood.

pathfinder wrote:

The 24-105 lens can have problems with flare sometimes, particularly with back lighting or shooting into the light.

I researched the flare that was associated with the first batch of 24-105's and it looked a lot different from what I am experiencing here. And I was also able to reproduce this with 24-70 that's why I started to think it had something to do with the 5D and the LCD "glow" ...

LiquidAir

cabinetbuff wrote:

I used lens hood in both cases -- the shot with the blue cast was aimed closer to the active LCD screen so is it possible that the reflection of the screen somehow caused a lens to flare? Notice though that the screen was not in the shot it was just a little closer.

Note that hood for the 24-105 is designed for 24mm, but the exif for your shot says 105mm. If you can see a bright light source in the frame at 24mm it can still cause flare problems if you zoom out to 105mm. The particular flare in your shot looks like a large bright light source above the frame; possible the reflection of a window in the LCD monitor. You can see it as a loss of contrast across the top of the frame. Since the blacks are taking on a blueish cast it make me think that the light source causing the flare is blue. It is possible that the light from the LCD panel itself is causing the flare and I'd say its worth trying the shot again with the panel turned off.

Robinivich

cabinetbuff wrote:

I researched the flare that was associated with the first batch of 24-105's and it looked a lot different from what I am experiencing here. And I was also able to reproduce this with 24-70 that's why I started to think it had something to do with the 5D and the LCD "glow" ...

I think it bears pointing out that flare isn't a manufacturer's defect per se, but an unfortunate drawback of modern materials and technology. Every lens in existence can be forced to flare if you try hard enough, until you get materials with true 100% transmissivity and 100% absorption (ie, coatings on the elements and flocking in the barrel) it will ALWAYS, in some way, or to some degree, be possible to force your lens to flare. Some lenses are quite good at resisting flare (simpler designs, better coatings and flocking, suppresion of internal reflections), others quite bad. Filters make a difference too, since they add a flat, big, light catching, reflecting aspect to the equation.

Bottom line? Until you notice that real shots are being hurt by the presence of flare, and when your normal shooting environment causes it, I wouldn't worry about it. And if this IS the case, then removing any filters, and adding a nicely fitted lens hood, are the two easiest steps, way easier than the retouching that is so difficult here.

cabinetbuff

LiquidAir wrote:

Note that hood for the 24-105 is designed for 24mm, but the exif for your shot says 105mm.

Thank you -- I didn't know this detail -- good to know.

Robinivich wrote:

Bottom line? Until you notice that real shots are being hurt by the presence of flare, and when your normal shooting environment causes it, I wouldn't worry about it.

I agree -- in fact I've never noticed this prior to this. The only reason I am being a little paranoid here is because it's a brand new camera (5D) and I really didn't know the root cause of this. If it were the camera I could still exchange it -- it's within it's return policy.

Robinivich wrote:

And if this IS the case, then removing any filters, and adding a nicely fitted lens hood, are the two easiest steps, way easier than the retouching that is so difficult here.

100% agreed -- I always try to glance at the shoot on the preview screen ... something like this would be noticeable and would typically force me to retake the shoot. I wouldn't mess around in PP to fix something so out of whack. In this case I was just trying to get to the root cause of this issue and originally it appeared as if it were a WB but after trying numerous times to adjust it I have to conclude that it's not a WB issue after all.

Thank you for your thoughtful analysis.

PS: so far I am really impressed with the IQ that 5D has been producing in relation to my 20D. FF makes a huge difference to the type of shooting I am into. All of a sudden it's like I've cleaned my classes everything is much more clearer, sharper and wider. I don't think I can use a crop body anymore I am always thinking in terms of how much "real state" I am loosing with my lenses.

Robinivich

Seems you've a great handle on things! Even the paranoia

And now I'm just left drooling over full frame, I lust after the viewfinder...