Expensive lenses and "color"; any real basis for gushing praise?

[ScottyP]

I have been reading up on 50mm primes. Obviously Canon has about 4 current offerings, but then you see Zeiss has two. Manual focus of course, and apparently made out of the same stuff as Wolverine's claws if you believe their user comments, but a lot more than the Canon lenses.

User comments talk a lot about how the colors are extra saturated, etc...

Is there anything to this? Is it anything you could not achieve with mainstream glass and a decent lens hood?

One of the two Zeiss 50's is not entirely out of my price range, if there was a real reason. It would do poirtrait duty on my crop frame, at least until I might ever go FF. Maynbe I could will the thing to my grandkids or whatever.

 

[Mt Spokane Photography]

Not really. A digital camera photosite only sees black and white. There are color filters over the photosites, and a color image is produced by what amounts to the most probable guess by the software.

As long as you have reasonably white light (there is nothing you can do if one color is missing), the colors can be adjusted in post processing to anything you'd like. Adobe Lightroom and others let you set color sliders to get a image color as you like it, and then every image you import will have those adjustments. You can even make different settings for different lenses. After all, its only a best guess as to the real color.

Then, after you import images, you can further tweak the color if it does not look to you to accurately reflect what you saw.

Some people like warmer images, some cooler, its a personal preference, so the photographer should do it his way.

 

[PhotoCat]

It is mostly in the coating and the construction of the lens. A good lens will have higher contrast meaning it is "leaking " vy little stray light into the lens. ie shadow area will hold more details and the color more saturated. Analogy is that u can see a projected slide show with saturated color and good shadow details in a darkroom but u will see washed out color and reduced shadow details from the projector screen, when the darkroom is leaking in some stray light. Similar to the darkroom, a poor lens will "leak" more stray light into the lens and "contaminate" the shadow details/color saturation of the lens, creating washed out color.

I once thought that photoshop could bring back the color saturation and shadow details of an image captured by a poor lens.
I was almost right and photoshop indeed can do mostly that. The catch is that you will have more noise
in the shadow details of the image, after u boost the shadow details... Worse yet, u may actually lose some shadow details, since the stray light inside a poor lens has "overpowered/wiped out" some deeper shadow details.

Someone with more technical background pls feel free to jump in, as I only remember the basic principle...

 

[Loswr]

Mt. Spokane is right...and, IMO, wrong. It's true that a photosite just counts photons of any wavelength, gated by the RGB filters of the Bayer mask, then interpolated to assign a color to each pixel. Likewise, hue and tone can be adjusted in post. But, different lenses transmit color differently, and with some lenses that color is 'richer' and 'more vibrant'. Those are subjective, of course. Can that be replicated in post? Yes, but it's not so straightforward. An analogy is the Vibrance adjustment - that's an adjustment of a subset of color tones, and replicating that effect using curves in Photoshop (for example) is difficult and time consuming. IMO, similar to getting the exposure right in-camera rather than fixing it in post (which can negatively impact IQ), it's better to get more pleasing colors as the image is recorded. An example I'm personally familiar with is the 85/1.8 vs. the 85L - the former is an excellent lens, and a great value, but compared to the 85L the images are a bit flatter and have less 'pop'.

 

[Policar]

Lenses can have better color rendition than other lenses beyond what can be corrected for in post; sensors can have better color rendition than other sensors beyond what can be corrected for in post. Whether the people posting these comments are right or not, in theory their ideas are sound.

Sensors can have stronger or weaker (or more or less IR sensitive) filters in front of their photo sites. The more frequencies the color filter rejects (the deeper it's tinted), the more vibrant the colors the camera can reproduce--and the less sensitive the camera is to light. DXOmark's "portrait" rating measures for this exactly (in theory).

Lenses can be more or less corrected for aberrations, and not just garden variety CA; apochromatic and superachromatic lenses should have dramatically reduced longitudinal chromatic aberration (any lens with ED glass or ULD glass or fluorite should have LoCAs substantially mitigated) and that means as the image goes out of focus its out of focus areas take on magenta/green tints. A properly corrected lens will not have this! And that means everything, in focus or otherwise, should be almost perfect in terms of color, and areas that are just out of focus but still smaller than the circle of confusion look better, too...so really everything looks great.

I used to shoot Fuji 4x5 and 120 Velvia. I stopped shooting film a year or so ago and miss it tremendously. Velvia has color rendering that, however inaccurate, looks like nothing else. If you look at Velvia's sensitivity curves they're incredibly narrow and reject most light. And so photos taken with Velvia have amazing saturation and vibrance. Just the richest, most outrageously deep colors (and contrast; it had extraordinary d-max). I also shot with a set of apo-sironar-s, with ED glass, and the lenses and the IQ is intense. Here's an image shot on 4x5 velvia with such a lens:

4x5

Can your dSLR produce colors like that? I doubt it.

But it can get close. Optically, better. The Coastal Optics 60mm f4 is corrected for visible and IR/UV light and produces the most beautiful colors, better corrected than any other lens I know of. I've seen images taken with it and a Nikon D3 that have amazing color, noticeably better than most even in web size.

Coastal Optics

 

[peederj]

Personally I find at least 80% of the sex appeal of L glass to be their unique color rendition.

You're not going to find a particular mix of color very easily in post. "Can't get there from here" is of course a strong statement...you could, perhaps, but the question is whether you would.

 

[pwp]

I have been reading up on 50mm primes...

I'm with Mt Spokane here regarding colour. Provided you're shooting RAW of course, and why wouldn't you?

The Zeiss lenses will likely deliver brilliant optics and other desirable characteristics as a more noticeable and relevant difference to the Canon offerings than any colour perceptions. If you're shooting portraits though, unless you've got the best hand/eye reflexes on the block you'll get a lot more keepers with an AF lens.

Having churned through a few 50's over the years I've settled rather comfortably on the Sigma 50mm f/1.4. It's big, heavy and bulky, it's by no means totally perfect but it delivers some exquisite files. Include the Sigma in your research.

Paul Wright

 

[epsiloneri]

I think it's a good question, ScottyP, I have also wondered about that. I have always assumed that photographers really mean better contrast when they refer to better colour, since the former will imply the latter and I don't understand how a lens otherwise could affect the colour in any major way (except biasing it, which would be obvious and easily correctable).

The purpose of a lens is to take an incoming photon and put it on a desired location on a sensor. A lens never changes the energy (colour) of a photon. The photon can merely be absorbed, scattered or transmitted.

If the location to where a photon is transmitted depends on energy, we have chromatic aberration. This indeed changes the colour rendition, but I don't have the impression that this is what is meant by colour quality of a lens, and is usually pretty obvious.

If the photon is scattered, it can end up anywhere on the sensor. This gives an additional "haze" over the image that lowers the contrast and makes the colours less "poppy". Since the haze is not constant over the sensor (it really depends on the scene) it can be difficult to fix in post, but often one can improve things by plainly subtracting a constant from all pixels (or a smooth function). There are also other PP contrast-enhancing techniques like "unsharp mask" (essentially a way to subtract a scence-dependent haze from the image).

If photons of certain wavelengths are preferentially absorbed that would result in the photons reaching the sensor having a different balance from the ones entering the lens, giving a colour bias, or tint. (this happens with e.g. warming filters).

In summary, if you want a lens that gives great colours I think you should look for a lens with great contrast (= clean glass and good anti-reflective coatings) and low chromatic aberration.

 

[Loswr]

I don't understand how a lens otherwise could affect the colour in any major way (except biasing it, which would be obvious and easily correctable)... A lens never changes the energy (colour) of a photon. The photon can merely be absorbed, scattered or transmitted.

If photons of certain wavelengths are preferentially absorbed that would result in the photons reaching the sensor having a different balance from the ones entering the lens, giving a colour bias, or tint. (this happens with e.g. warming filters).

The thing is, lenses aren't equally transmissive across the visible spectrum. The elements are composed of different materials with different transmission curves, none of them completely flat. So, while you're correct that the wavelength of an individual photon will not be altered, some wavelengths are, indeed, preferentially absorbed (not completely, but the transmission curves have small peaks and valleys. So, there's a slight effect on color fidelity. Having said that, I agree that contrast will have a bigger, more evident impact.

 

[epsiloneri]

The thing is, lenses aren't equally transmissive across the visible spectrum. The elements are composed of different materials with different transmission curves, none of them completely flat. So, while you're correct that the wavelength of an individual photon will not be altered, some wavelengths are, indeed, preferentially absorbed (not completely, but the transmission curves have small peaks and valleys. So, there's a slight effect on color fidelity.

Yes, that is precisely what I meant by:

If photons of certain wavelengths are preferentially absorbed that would result in the photons reaching the sensor having a different balance from the ones entering the lens, giving a colour bias, or tint.

(sorry for using a confusing language). I argued that in contrast to the contrast, this effect is relatively easy to correct in PP by giving the basic colours different weight. But if there is a lot of structure in the transmission curve, however, then it will be diffcult to correct in PP (because the tint will be colour dependent, i.e. you cannot adjust the colour weights for all pixels at once). So perhaps this is indeed what photographers mean by how well a lens renders colours.

 

[AmbientLight]

I believe it is necessary to avoid oversimplification here. As Neuroanatomist has pointed out there is always some absorption leading to a loss of color fidelity based on the materials used in any given lens, which is the key to understanding what is actually going on. Here's an article on the subject for further reading:

http://en.wikipedia.org/wiki/Absorption_%28electromagnetic_radiation%29

We are using different lenses, which have different properties as to how they affect light (= electromagnetic radiation visible to the human eye) passing through them. In general we may assume/hope that more expensive lenses offer better images, because the more expensive material used in them has better characteristics for providing what appears to be "better" images.

Now what is the effect for a photographer?
Using some of the most highly prized lenses (as in most praise, not as in most costly) you will find that you can take shots that look similar to what you have seen, while with worse lenses you will end up with what is usually described as a flat look. Somehow, something is missing and although you look at what may pass as an adequate image of the scenery, you are missing something. Of course you can't point out that some of the wavelengths have been affected adversely by your lens, because this would be too blunt an effect, but you will see the effect nonetheless.

It is important to note that you cannot fully compensate for such effects in post-processing, because the software you use will never know just exactly what and how much of what has been lost due to the lens. Adding colour in post-processing can only compensate for a more general effect, such as for example a sky not blue enough or less vibrant green leaves compared to what you remember. This will enable you to make general corrections, but you will not be able to fully recompose the exact mood of the scenery as you have seen it before. Granted some post-processing artists may come very close.

I just think about the time lost in post-processing as opposed to time spent actually taking pictures. You decide what is preferable to you. Although I am quite happy with Adobe Lightroom I rather take pictures .

 

[epsiloneri]

As Neuroanatomist has pointed out there is always some absorption leading to a loss of color fidelity based on the materials used in any given lens, which is the key to understanding what is actually going on.

I agree that wavelength-dependent transmission in principle could change the detected colours, what I would be interested in learning is if this is at all a significant effect for modern lenses. I suspect not, but would be interested in being proven wrong. As stated in my first post, I think it is the better contrast of highly prized lenses that gives them the better colour. To see the effect of contrast on colour, the following comparison might prove helpful: http://www.cambridgeincolour.com/tutorials/camera-lenses.htm

 

[AmbientLight]

The link shown explains that lens errors such as chromatic aberration or misalignment may cause a reduction in image quality. Those errors are rather obvious causes for trouble, but in most modern lenses these factors are pretty well controlled. If we assume that given two different lenses both have no such obvious problems, what you are left with are the absorption characteristics and of course pincushion and barrel distortions, but those are not color-related.

This is an area where we have an obvious effect in physics, which does not translate easily into human understanding. People may think that minute differences in transmission characteristics for different wavelengths will not have any appreciable effect on a picture, but yes they do, because our eyes can see it.

Just keep in mind that those differences are subtle.

Imagine you look at two snapshots of a clown holding bright balloons you will see basically the same picture, even if one picture has been taken using high end equipment and the other using an entry level camera with kit lens. Exact color rendition won't have a real effect on what you see. People may say: That clown looked funny, but will anyone comment on the color of the balloons being something special? Probably not.

If you look at some of the really good landscape photos or portraits (avoiding obviously photoshopped ones) you may ask yourself: What did that photographer use? Often enough you will find that the glass used has been pretty good and you won't be able to find similar pictures shot with an entry level kit lens or something simlar.

That is not to say people cannot produce great photos with entry level gear, but these will be different great photos. It depends on the environment and circumstances, if such subtle effects will be meaningful or not.

I hope this helps.

 

[ejenner]

I agree it's subtle. There is no way any of my family who are not photographers would notice it and with modern PP it can become even harder to detect after good PP.

But it does become apparent when you have experience in PP shots with different lenses. I must admit I never paid much attention to it until I got my two primes - sig 1.4 (more contrast) and TS-E 17mm (so different across the board).

 

[peederj]

When you are shooting video and swapping out different makes of lenses and cutting scenes between them you will probably see the color effects of given lens coatings and formulae. L lenses are far warmer toned than the non-L Canons, L's do have higher contrast generally but I do see it as something in the coatings that makes colors pop in a certain way.

Contrast, saturation and luma vs. chroma transfer functions shouldn't be conflated, they are separate concepts just as they are separate controls in post. That's not to say they don't strongly interact...just that they aren't one and the same.

 

[epsiloneri]

The link shown explains that lens errors such as chromatic aberration or misalignment may cause a reduction in image quality.

Yes, the page I linked is about various lens effects, but I was specifically referring to the comparison of how the loss of contrast can impact rendered colours.

I hope this helps.

Thanks for trying. Googling a bit, I found the following interesting text where they look at transmissions from binoculars and lenses: http://www.lenstip.com/index.php?art=129

Glass in itself has a very flat transmission curve in the optical, it's in the coating the magic happens. The conclusion from the linked webpage (jump to the final figure) is that antireflection coatings have advanced leaps in the past decades, as even a cheap Samyang lens has coatings that rival the most expensive lenses from a few years back.

Given that the transmission curves for lenses seem smooth, albeit with a lens-dependent slope, I would expect that any tints introduced by the non-flat transmission would be easily corrected in post (by e.g. giving more weight to the red channel, if the transmission in red us lower). I.e., properly white balance.

L lenses are far warmer toned than the non-L Canons, L's do have higher contrast generally but I do see it as something in the coatings that makes colors pop in a certain way.

The warmer tones of L lenses are likely due to their different coatings, as you suggest, but that should be easily compensated in post by using appropriate white balance, right? (well maybe it's technically more tricky in videography, I have little experience with that) The contrast loss on the other hand, you cannot compensate so easily in post.

Contrast, saturation and luma vs. chroma transfer functions shouldn't be conflated, they are separate concepts just as they are separate controls in post.

I speculate that's part of the confusion - when a lens is claimed to render nice colours, I propose it's actually the contrast that is good.

 

[Mt Spokane Photography]

Its true, lenses have coatings which can't help but influence the color of light hitting the sensor. However, the camera body tries to correct this based on guessing which are the grays and color correcting to make them gray. This may result in different tones and hues. howeer, they are subtle and easily fixed in post.

Lens contrast is a different matter, you can't fix poor contrast in PP, you can improve it, but making a poor lens into a great one cannot yet be done.

Here are som images of lenses I own and the color reflected by the lens coating.