Daran wrote:
You could shine a stable continuous frequency light source through the lens and measure how much the lens cuts from the various frequencies. While perhaps not overly interesting it would tell you how different lenses actually are (AFAIK mostly due to their coating). Meanwhile a more telling test measures the overall response to a rainbow of frequencies in the RAW image: www.dpreview.com/forums/thread/4315294. Of course you could swap lenses for this test, too.
I don't know how to do this from a software point of view.
The spectro emits its own light (known light source) and is calibrated in a dock with a patch of known color.
It is then pointed at a patches printed off a printer, and software is run to read the patches as you slide the spectro across the patches.
Maybe you need a different kind of spectro... the light source needs to be on a wand or something
In theory, if you profile a light with a good CIR rating, and a lens has a color tint, you would see it when you compare the color response to the light with no lenses in front of it.
The one I linked to is a handheld, but an EE I worked with years ago had one that was a lot more flexible and would plug into his computer.
Daran wrote:
"I have no idea why this happens. I hope to learn more as opportunities arise to talk to the optical team behind these lenses."
[taken from the web page]
Seems honest enough.
Excellent example. You observed something valid, but due to limited understanding of how things work, your guess as to the cause just happens to be nonsense. Since there was little consequence in misunderstanding this, you just never noticed your interpretation was wrong.
How does this work then, you ask? Handling of clipped highlights is happening entirely in post, so when your RAW is developed. The lens has no influence on this whatsoever. Nor does the camera, other than yielding the hardware limits of where each color channel will clip. If you look at the image data in the RAW file (tricky, I know) you'll find that the data is simply clipped, with no graceful "handling" applied at all. Similarly any non-linear brightness responses (the "flesh tone brighter and shadows more open" from the web page) also come entirely from software, be it color profiles, gamma curves, LUTs or whatever else is done during RAW development.
What all this means is that if you wana improve on how this works for you, you need to look at your RAW development software and/or it's settings. Certainly not at your lens. Good news is that you can still fix this for RAW images you took with lenses you thought to yield poor color. They didn't. Your software did....Show more →
Are you suggesting then that the type and quality of glass has no effect on colour, contrast and brightness of the final image ?
duncangr wrote:
Are you suggesting then that the type and quality of glass has no effect on colour, contrast and brightness of the final image ?
I am not denying an effect of glass & coating on the passed color. I am suggesting that this difference between lenses can be compensated by a color profile adequate for the lens & sensor to a residual degree that is no longer humanly distinguishable.
Contrast is mostly determined by (lack of) internal reflections (baffles, coatings) and not the glass. "Brightness" would be the transmission efficacy, which for modern coated lenses should be high enough not to matter (>90%).
Daran wrote:
I am not denying an effect of glass & coating on the passed color. I am suggesting that this difference between lenses can be compensated by a color profile adequate for the lens & sensor to a residual degree that is no longer humanly distinguishable.
Contrast is mostly determined by (lack of) internal reflections (baffles, coatings) and not the glass. "Brightness" would be the transmission efficacy, which for modern coated lenses should be high enough not to matter (>90%).
In my experience there is no way to make my 200-600 G images look like they came from a 600 GM or even the 100-400GM for that matter.
As far as I know, Capture One has had a feature for some time now that transfers the look of any photo to any other photo. It seems to work quite well.
I don't think anyone here denies the fact that contrast and color transmission varies across lenses. What seems to be the point of contention is to what degree these differences can be eliminated in post-processing.
I think it should be quite obvious that it depends on the skill of an operator. There's a huge gap in capabilites between a causal Lightroom user and someone who lives in Photoshop for a living and uses specialized tools like LUT and color profile builders.
Those who know what they're doing can simulate the appearance of a specific C41 film emulsion like Portra printed on a given RA4 paper like Maxima, to an extremely realistic degree, so it's not controversial that they can easily wipe out relatively minor color casts caused by lens difference.
Interestingly, I stumbled upon that Lens Rentals back in 2018 did some spectrometry test on cine lenses and believe that you can see the color differences between lenses / brands in the results:
Unfortunately, it seems like they never went further with the testing, but it does validate the idea. At a high level, looking at the data, you can probably use HSL adjustments to flatten the response a bit, but it will probably never be perfect without some type of more advanced adjustment, since some of the color bands have a curve in the response.
I know this topic of color science gets debated to death. All I know is that my experience is that there is a significant difference between the output of these two brands. A shot like the one that started this thread is never going to show that difference. Go outside near a lot of green foliage with a variety of lighting on various parts of the scene. In a really basic way, the way that sony cameras process greens is fundamentally different from nikon cameras. Same for skin tones. Yes, a lot of that comes down to the profiles used to process the raw files, but that is part of the equation, too. Shoot a complex outdoor scene and use a color checker, calibrate both, and they are still going to look very different. The files between the two brands process very differently, too. It really just comes down to preference. If I'm shooting an indoor concert, I generally prefer Sony. If I'm outside I almost always prefer Nikon. But that's my preference for my editing style and aesthetic. But, all the test shots in the world can't speak to the real world experience of photographers using these cameras in the field and editing those files. My experience is that there is a big difference between the images out of each camera.
Nikon wins hands down... when I compare to my Sony RX100 7
I used Nikon, Canon, and Sony Rx100 ( 3 and 7)
All brands has very specific colors, but as someone mentions before, different lenses produce different results in some situations (this are my observations after taking many thousands of photos around the world in many very different situations)
Overall if light is nice all photos looks great, if light sucks camera brand doesn't matter much - photos sucks
Greens will usually be the major bone of contention because:
'As green and red make up so much of our sense of brightness, it (sic) is the brightest color that we can see. If red is 21% of our sense of brightness, and green is 71%, then together they make up 92% of our sense of brightness. It’s hard to beat that in saturated colors.'
And don't forget that one in twelve readers don't know what is going on at all with these colors, and one suspects the ratio might be a fair bit higher higher among photographers.
'Red-green color blindness is a genetic condition that affects males..It can cause difficulty distinguishing between red and green hues and is most commonly inherited.'
The idea of an experiment is a good one, always. Get two copies of a lens that is known for color - any Zeiss, Voigtlander or Simera will do - that is available in E and Z mounts, and photograph a garden path with plenty of greenery in soft light. Process identically and observe.
The entire underpinning of statistical research is to control for what we call confounding independent variables, so you must use the same lens/scene/lighting/technique/WB/exposure. There is so much complexity going on within our lenses that directly affect color that discussions with the camera-first tech people go nowhere useful. I didn't even mention using better color spaces and gamut handling, for this reason.
philip_pj wrote:
Greens will usually be the major bone of contention because:
'As green and red make up so much of our sense of brightness, it (sic) is the brightest color that we can see. If red is 21% of our sense of brightness, and green is 71%, then together they make up 92% of our sense of brightness. It’s hard to beat that in saturated colors.'
And don't forget that one in twelve readers don't know what is going on at all with these colors, and one suspects the ratio might be a fair bit higher higher among photographers.
'Red-green color blindness is a genetic condition that affects males..It can cause difficulty distinguishing between red and green hues and is most commonly inherited.'
The idea of an experiment is a good one, always. Get two copies of a lens that is known for color - any Zeiss, Voigtlander or Simera will do - that is available in E and Z mounts, and photograph a garden path with plenty of greenery in soft light. Process identically and observe.
The entire underpinning of statistical research is to control for what we call confounding independent variables, so you must use the same lens/scene/lighting/technique/WB/exposure. There is so much complexity going on within our lenses that directly affect color that discussions with the camera-first tech people go nowhere useful. I didn't even mention using better color spaces and gamut handling, for this reason. ...Show more →
"Process identically" - is this possible? Do the color profiles, e.g. in LR, have been made to give "identical" results for different cameras?
DWOfPaul wrote:
Interestingly, I stumbled upon that Lens Rentals back in 2018 did some spectrometry test on cine lenses and believe that you can see the color differences between lenses / brands in the results:
Unfortunately, it seems like they never went further with the testing, but it does validate the idea. At a high level, looking at the data, you can probably use HSL adjustments to flatten the response a bit, but it will probably never be perfect without some type of more advanced adjustment, since some of the color bands have a curve in the response.
You should not need to do that as that is (supposed to be) done by the color profile for your lens. After which different lenses should produce practically identical responses. Which is vastly simplified since neither the sensors nor our eyes can "see" all the various frequencies separately, all we see is three differently weighted summations over the frequencies.
Daran wrote:
You should not need to do that as that is (supposed to be) done by the color profile for your lens. After which different lenses should produce practically identical responses. Which is vastly simplified since neither the sensors nor our eyes can "see" all the various frequencies separately, all we see is three differently weighted summations over the frequencies.
As far as I know, LR lens profiles and DXO lens profiles do not remove color cast. If you have any links on profiles removing color cast, I would definitely be interested in reading them.
DWOfPaul wrote:
As far as I know, LR lens profiles and DXO lens profiles do not remove color cast. If you have any links on profiles removing color cast, I would definitely be interested in reading them.
I'm using Capture One, which seems to do a pretty good job, at least for the lenses I have. Otherwise there are tools around to measure your lens yourself, generating profiles exactly matching your lenses (at least under the light you used). Otherwise I'm sure you could complain to the Software companies for no effect whatsoever...:-D
PIOK wrote:
Nikon wins hands down... when I compare to my Sony RX100 7
I used Nikon, Canon, and Sony Rx100 ( 3 and 7)
All brands has very specific colors, but as someone mentions before, different lenses produce different results in some situations (this are my observations after taking many thousands of photos around the world in many very different situations)
Overall if light is nice all photos looks great, if light sucks camera brand doesn't matter much - photos sucks
The A1 wins hands down compared to the RX100... =x
The regular moves (I hope) most know how to make in the software's panel: WB - macro exposure - check BP and WP for RGB values being correct - check highlights don't clip and retain gradation - raise shadows to 'just right' - add sharpness settings - add contrast to tone segments that need it (I use luminosity masks) - shift BP/WP if the histo needs it.
That kind of protocol is formulaic and can be applied to all images. I don't use profiles apart from 'Neutral' in ACR, so I only need the s/w to recognise the camera. I want to be in charge and see what each move achieves. It's why I use lenses that do fine without distortion and CA correction like they do in the cine world. These aberrations are 'optical fails' designed to chain you to computer 'technology'. Burden shifting is not my go. I've never downloaded any 'firmware'.
Nothing but nothing shows the differences like doing these side-by-sides. Only well made comparisons show what a lens or system has, and that, friends, is why there are so few of them being done. There is always a winner and a loser. Using the same optically sound lens on two systems will show what we are talking about here.
What is coming up, sooner or later? The continued infiltration of light cine 'hybrid' into stills photography is bringing with it 'interchangeable mounts' so users can use their same favorite lenses on different cameras. Some even provide the mounts and shims (or provide/sell them). We are not talking backyard outfits either, but Zeiss Nano, Cooke SP3, Leitz Hektor, Sirui VP-1.
2-3 minutes and it is done. DZO/Thypoch are using M-mount lenses for adapting their Simera-C lenses to all the mirrorless mounts plus ARRI. Not much sensor optimisation going on in the real world, for $5000 lenses no less!
same for the others, Cooke even sells mounts for M cameras too
