Warm Lenses & AWB

[drisley]

I recently purchased a Canon 28/1.8 USM lens, and I had been doing some comparison testing against the Canon 50/1.8II.

One thing I noticed is that the 50mm lens is slightly cool, while the 28mm lens is slightly warm. Side by side comparison of the same image shows a very noticable difference, even when using Auto White Balance.

Curious? Why doesn't the camera's auto white balance make up for any warmth/coolness of a lens? As far as the sensor or onboard computer is concerned, warmth should be neutralized, regardless of whether the source is from a light, or the lens.... Yes/No?

Anybody have any opinions?

[J Rabin]

Drisley.
I thought I was nuts for noticing differences; imagining it. I kind of like the differences. "Neutral" reproduction is a lie in photography any way.

At work, the 70-200 f/2.8L IS is WARM, golden even, (in addition to contrasty). Makes it attractive for long portraits, not just sports. I like the warm tone. It renders skin in midday like it's morning or evening! Who could ask for better? I've noticed the 70-200 f/4 shares this.

At work, a new 60mm EF-S macro is COOL, not blue cool, but Kodachrome 25 or 64 of old cool. This lens is also contrasty, while neutral, which strengthens the attraction of its images. Again, I like it.

For personal travel use I bought the controversial 70-300 DO. This lens is definitely "bluish" cool, not neutral cool like the EF-S 60mm. It misses auto white balance often.

But, this lens has fabulous contrast AND THE MOST AMAZING LACK OF CHROMATIC ABBERATION. It took me a week to learn how to use it effectively. E.g., yesterday evening, sun moving lower, kids playing basketball in the yard, I spot metered this weed leaf with sun straight behind it:
http://postit.rutgers.edu/uploads/Leaf%20Silhouette%2002%2Ejpg
With the IS, I handheld this at 300mm 1/30 from 2 meters! NO POST PROCESSING!

http://postit.rutgers.edu/uploads/Leaf%20Silhouette%2001%2Ejpg
This one was 1/125

Where's the purple CA on all these edges? Absent! It was light outside, but spot metering the leaf dropped the forest background out.

My 17-40mm f/4L is neutral to pinkish warm.

I think these good lenses have higher contrast, which must have something to do with our perception of color. I'm not a physics/color theory person. Just take pictures.
J

[drisley]

J, thanks for your input. It is weird isn't it? I could see the lenses warmth affecting film, but I wouldn't think it would affect digital, atleast not Auto White Balance.
Nice leaf pictures too. I assume that there is no purple CA because you stopped down to F9.
I don't think there should be any at that aperture.

[Citizensmith]

yeah interesting point. I wonder if the variations are just not significant enough to trigger the camera into jumping to a different white balance. At a guess custom white balance could cure it.

[drisley]

Yes, the good thing is that I shoot Raw, so I can correct in post.
The difference is quite noticable. If I shoot in tungsten lighting with the 50/1.8II, the resultant white balance temperature is 2850 (normal for tungsten).
However, to get roughly the same white balance using the 28/1.8, I would have to select a white balance temperature of about 2700. That's quite signifant. You would think the AWB would adjust for this. Even the Auto WB setting in Photoshop CS2 Camera Raw selects 2850.
Again, it's not really a big deal, and I prefer the added warmth of some of my lenses (28/1.8, 135/2L), but it's rather perplexing.

[Jon]

Yes, the good thing is that I shoot Raw, so I can correct in post.
The difference is quite noticable. If I shoot in tungsten lighting with the 50/1.8II, the resultant white balance temperature is 2850 (normal for tungsten).
However, to get roughly the same white balance using the 28/1.8, I would have to select a white balance temperature of about 2700. That's quite signifant. You would think the AWB would adjust for this. Even the Auto WB setting in Photoshop CS2 Camera Raw selects 2850.
Again, it's not really a big deal, and I prefer the added warmth of some of my lenses (28/1.8, 135/2L), but it's rather perplexing.
That's about a 20 Mired difference, which isn't all that much. It's around the same range you can expect to see in bright, sunny conditions between 5500 and 6500 K. It's also the shift an 81A filter would give you. Still, I'm surprised

That Canon would release two lenses with that much colour transmission difference and
That AWB wouldn't be able to handle it.

[DavidEB]

I've also been baffled by auto white balance, but I'm not sure it's entirely due to the lenses transmission characteristics. I've seen similar kinds of things in a series of potraits taken seconds apart, identical conditions, with the only difference being the focal length setting on my zoom lens. It was most obvious in pictures of my son standing in front of trees. I decided that as the subject (skin tones) took up more of the frame, the AWB seemed to shift to cooler levels, and as the background was more prominent (shorter focal length) the camera saw more overall blue-green, and the white balance shifted slightly warmer.

As a result, whenever I have the time I set a fixed white balance (sun, cloud, etc...) for a series of shots. Then the shots all look similar. If I change lenses I can see a small difference, but it's not as great as the difference made by focal length and composition using AWB.

David

[drisley]

I have a hypothesis.
Could the 28/1.8 and lenses like it actually be slightly COOL and the AWB actually overcompensates? And vice versa with the 50/1.8II? Just brainstorming.

[Jon]

Shouldn't think so. I like DavidEB's theory on the different framing better. AWB's just going to smear the whole pic together and try to make it average 18% grey. A wide outdoors is going to have more blue (sky/skylight) to accomodate than a narrower AoV lens, so AWB will have more blue to cancel out by juicing the amber side of things.

[drewmk2]

Do we all know how AWB works? If not, how can we judge. Try the same comparison on your EOS film body, then we'll know.