Uniform colour casts

Pekka Buttler, June 2021

Introduction:

Colour casts can mean two different things: The first (and more common) issue is a significant colouration of the entire resulting image, typically due to lens elements that change the colour of light coming through them. The other issue is typically associated with specific types of lenses on most types of sensors and mostly affect the off-centre areas of images. These are physically very different and will be treated separately here.

This article is part of a JAPB series of articles on the optical flaws of lenses and you can find the index of the series here.

The physics of uniform colour casts

First off, the article stated above that uniform colour casts are caused by “lens elements that change the colour of light coming through them”, and that is technically not true. Because what actually happens is a bit more complicated.

When you have a pane of transparent orange glass or plastic (I should say: ‘orange -looking’), then the view through such a pane (such as in orange-tinted sunglasses) is more orange not because that orange pane would magically change the colour of the light passing through the pane, but because the orange pane absorbs some wavelengths of visible light (such as blues and violets), making the resulting view more ‘orangey’. Incidentally, this absorption of some wavelengths is also why that pane looked orange to you to begin with.

So, when you have a lens that produces an overall bluish image, then this is not because the lens would magically make non-blue wavelengths of light into blue wavelengths, but by absorbing a share of non-blue wavelengths, thereby making the share of blues higher. Likewise, when you hear/read someone stating that a specific lens produces reds with a beautiful punch, then this does not mean that that lens would do something to the red wavelengths (in fact, it leaves the reds wavelengths alone, but does weird stuff to the other wavelengths).

Importantly, while glass is intended to be transparent, all types of glass tend to absorb some light. When designing a lens, designers will primarily be selecting the glass types for the various lens elements based on the glass types’ refractive index1 and dispersion characteristics2, and not based on a balanced absorption. This is especially true for really old lenses (when only B&W mattered). In any case, these differences as designed are mostly slight, and can often be compensated for by using coatings (chemical treatment of lens surfaces) that not only minimise reflections and improve light transmission but also balance out the glass-caused absorption.

Even so, it is quite typical for one series of lenses (manufacturer & era) to offer an internally similar colour rendition profile while differing slightly from the colour profiles of others. This is one central reason to why people trying to build a set of compatible lenses for ambitious uses (e.g. cinematography) tend to stick with one line of lenses (e.g. early Takumars, Contax-era Zeiss lenses, Leica R lenses, Canon FD lenses, OM Zuikos etc.).

Side note: Just as lens elements can produce colour casts, also filters can do so. Considering that many kinds of filters are intended for this specific purpose (ranging from skylight filters used to balance an overly sunny sky, to yellow filters used to give more punch to B&W imagery, to hard-gradient partial filters used to make for spectacular sunset imagery), this should come as no surprise. Problematically, some filters that are not supposed to affect colours do lead to a significant colour cast / discolouration (e.g. some neutral density filters).

Extreme, inadvertent casts

But such minor differences are not flaws, unless you expect your lens to always perfectly reproduce the subject spectrum with supreme accuracy. Instead, we’re focusing on an altogether more massive shift in colouration. These shifts are most often caused by the use of thoriated lens elements, but may also occur with non-radioactive lenses

Thoriated lenses

Four otherwise identical images. All shot at f/4 ; ISO 100 on a Sony a7R2, all with WB 4900K. Left to right, top to bottom:
1) Canon FD 35 mm f/2 Chrome Nose (thoriated): 1/400 s
2) Minolta Rokkor MD 35 mm f/1.8: 1/640 s
3) Nikkor Ai 35 mm f/2: 1/640 s
4) Carl Zeiss Jena Prakticar 35 mm f/2.4: 1/640 s

The use of radioactive thorium in photographic lenses was quite common for some time (especially 1950’s to 1970’s), but has since been curtailed – not so much because that the use of such lenses would be hazardous, but because the manufacture of thoriated lens elements most certainly was either hazardous or (to avoid the hazards) expensive. (Read more here)

While thoriated lens elements (when new) did not lead to any discolouration, decades of slow radioactive decay has lead to that the thoriated lens elements have changed colour (typically shifting into the yellow-orange-brown spectrum), thereby also causing the ensuing picture to show a clearly changed overall colour profile. Another side effect of the discolouration (discolouration being an effect of selective absorption) is reduced light transmission, leading to longer shutter speeds for the same exposure. Thoriated lenses can be treated, which temporarily removes the yellowing. On the other hand, some say that the yellow cast produced by an untreated thoriated lens makes the lens especially suitable for B&W photography. (Again, read more here)

Below some more examples.

Common settings: Sony a7R2, ISO 100, f/4, WB 4650K
Left: Konica Hexanon AR 57 mm f/1.2 @ 1/640 s
Right: Canon FL 55 mm f/1.2 @ 1/800 s
Common settings: Sony a7R2, ISO 100, f/4, WB 4650K
Left: SMC Takumar 50 mm f/1.4 @ 1/500 s
Right: Olympus OM Zuiko Auto-S 50 mm f/1.4 @ 1/640 s

Colour casts with non-radioactive lenses

While thoriated, yellowed lenses are the most common reason for colour casts, other lenses may also exhibit a very similar behaviour. One such example I have is a Jupiter-12 -lens. My specimen is in M39-mount, designed for use with Soviet FED and Zorki rangefinder cameras, and it exhibits a very thoriated-like colour cast (together with the extreme vignetting and corner colour cast typical for rangefinder wide-angles on some mirrorless sensors, read more here). The colour Jupiter 12’s uniform colour cast is so much like that of thoriated lenses, that I was surprised when the Jupiter did in no way excite my Geiger counter.

The picture pair below – taken within minutes of each other, with the camera settings unchanged (ISO 100, self-timer, WB Sunny) – succinctly show the effect. Therefore, while not all radioactive lenses show a yellow colour cast, also not all lenses that show a yellow colour cast are radioactive.

Are uniform colour casts a problem?

That depends. Personally, I’ve mostly liked the colours produced by lenses that have a strong yellow/orange cast, but I’ll be the first to admit that they are akin to wearing coloured sunglasses: The world might look nicer, and especially foliage will benefit, but not because the world suddenly got any better.

More problematically, correcting for a uniform colour cast is not easy. Theoretically you could go into the RGB levels/curves in your post-production suite and tweak for some minutes, then save that tweak as a preset for future use, but it will never be quite perfect. Also – depending on the stage of the lens’ yellowing – you might be losing a fair share of light (between 1/3 and a full of a stop with my lenses), and that is non-recoverable (before the yellowing is treated). Considering that many large-aperture lenses are worst afflicted, it leads to a conundrum: While an f/1.2 thoriated lens still offers a very narrow depth-of field and allows fantastic background separation, the lens’ ability to gather light and (ceteris paribus) allow photography in low light might be worse than even that of an f/1.8 lens.

Footnotes

1 Every medium has a refractive index, that describes how strongly that medium bends light. But as the medium itself does not bend light, but the bending is instead achieved at the interface/boundary from one medium to another, the strength of bending at an interface/boundary is determined by the respective refractive indices of the two media.

2 dispersive characteristics describe how strongly different wavelengths of light disperse when being bent. See more about this phenomenon in the article on chromatic aberrations.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.