I thought I'd have a go at quantifying the colour casts given by lenses of different manufacturers by testing them all side by side against a known reference point using my 450D.

The reference shot used for setting the custom white balance was taken with no lens attached to the camera. Subsequent shots where taken with each lens at the same exposure.

The crop on the left shows the actual cast as you would see it against pure white. The crop on the right shows an exaggerated view of the cast, basically +80% saturation of the crop on the left.

You can't compare hues unless the shades are evened out. The only ones there that seem to have a really different hue are the Oly and the Leica. And even then, the color of the lighting situation you are shooting under will have MUCH more influence on the colors of the shot than that slight cast to the optics.

If you want to really compare colors, shoot a full spectrum of shades and hues along with neutral gray scale and even out the brightness, contrast and saturation before you compare them. Make sure the neutral gray scale is the same on all of them, THEN you'll see if there is a color difference.

Isn't white already composed of the full spectrum?

sworth wrote:
You can't compare hues unless the shades are evened out. The only ones there that seem to have a really different hue are the Oly and the Leica. And even then, the color of the lighting situation you are shooting under will have MUCH more influence on the colors of the shot than that slight cast to the optics.

If you want to really compare colors, shoot a full spectrum of shades and hues along with neutral gray scale and even out the brightness, contrast and saturation before you compare them. Make sure the neutral gray scale is the same on all of them, THEN you'll see if there is a color difference.

The test target was plain white paper, the lighting was constant and used to set a custom white balance with no lens attached.

The shots were processed by taking a center crop then doing Filter -> Blur -> Average to get the overall cast. Then converted to LAB and the brightness equalized by setting the white input point of the L channel to the far left (that leaves only the colour set against pure white).

Does anyone find any of the results surprising? It's been said many times over that Sigma lenses have a yellow cast and I can certainly see that here. Zeiss producing slightly warm colours and Leica slightly cool is also something I expected to see before doing this test.

doesn't seem surprising to me, different coatings on the glass will give slightly different results. And possibly the type of glass used too? not sure about that though.

Nice test, Andy.
I have noticed a slight yellow cast with Sigma's lenses, but nothing problematic
in practice. I expected a cool cast with Zeiss glass, as that's what I feel I get
with CZ lenses on my Canon bodies. Nice to see the Zeiss T* coatings
are delivering very neutral images.

trumpet_guy wrote:
Nice test, Andy.
I have noticed a slight yellow cast with Sigma's lenses, but nothing problematic
in practice. I expected a cool cast with Zeiss glass, as that's what I feel I get
with CZ lenses on my Canon bodies. Nice to see the Zeiss T* coatings
are delivering very neutral images.

+1 Nice test, and very accurate too.

Cheers

Andrew

Thanks for the test. I bet the SMC Takumar would have been a much more neutral hue when it was younger - many of the Taks (like my Super Takumar 50/1.4 on the right below) get a very yellow cast as they age - you can clearly see it. There are two ways to "fix" a yellow Tak (1) expose it to a lot of UV light (direct sunlight will do), or (2) http://www.hermes.net.au/bayling/repair.html.