"Scanning" negatives with digital camera on a light table

[MattKing]

I don't think anyone who has ever taken a photograph has been concerned with the photograph precisely representing reality.

I guess you haven't spent much time around medical, lab and science photographers .
I have a friend who has photographically documented large numbers of dental procedures for demonstration and educational purposes. Some of that work can also be beautiful and powerful too.

 

[dokko]

I don't think anyone who has ever taken a photograph has been concerned with the photograph precisely representing reality. Or at least not twice.

Photographs ideally represent the experience of the photographer, or a third party, or an imagined entity.

No one cares whether "these are the right colors" or "these are the power lines in my viewfinder." What people want are "this looks like I remember it."

that's a good point, I often say a photograph is a transformation of reality into a personal feeling about it. same as with a painting.
with b/w for example, it's obvious that we're not concerned about how reality looked like (makes me think of Calvin and Hobbes though)
https://www.reddit.com/media?url=https://i.redd.it/mm1qp7rwniw71.jpg

but obviously people do care how a certain film renders colors, otherwise we wouldn't have these long debates of Kodak vs Fuji, or Portra vs Ektar etc.
In my experience, these days the biggest factor of the look of a digital representation of an analog image is the scanning system, and not the film emulsion.

Like, a Kodak Portra 400 and Fuji Superia 400 scanned on a Frontier System will look more similar than a Portra 400 scanned on the Frontier, vs on an Epson V850 with the Epson Software, vs on an Epson V850 with Silverfast, vs a camera scan with NLP etc.

so what I meant is not so much that a scanner should capture the colors like in reality, but in a pleasing way. now, what pleasing means is very subjective. personally I can't stand the colors of a Frontier scanners for example, but a lot of people seem to be very happy with it.

 

[Alan Edward Klein]

I shoot Velvia 50 and like the colors. However, when I scan, I don't try to match the colors and adjust to fit what my eyes find the most pleasant. I never than check against the original film shot to see if they match.

 

[Romanko]

I tested more light sources that I can remember (halogen with different filters, different CRI white LED, different narrow spectrum RGB LED, RGB OLED) and looked at the design of many commercial scanners

Are there any commercially available narrowband RGB LED light boxes or did you have to make one?

 

[dokko]

Are there any commercially available narrowband RGB LED light boxes or did you have to make one?

I had to make one (or two or three). I also tried some commonly available RGB light sources (like for film lighting for example), but the problem is usually that the red spectrum is not deep enough.
on top of that the blue and green lights need to have right spectral peaks, which ideally would be slightly different for different emulsions.

but even with optimally matched RGB bands, it's quite difficult to process the channels so that the colors don't look strange. it's definitely not something I would recommend unless you feel comfortable in advanced color science.

 

[J N]

I had to make one (or two or three). I also tried some commonly available RGB light sources (like for film lighting for example), but the problem is usually that the red spectrum is not deep enough.
on top of that the blue and green lights need to have right spectral peaks, which ideally would be slightly different for different emulsions.

but even with optimally matched RGB bands, it's quite difficult to process the channels so that the colors don't look strange. it's definitely not something I would recommend unless you feel comfortable in advanced color science.

That's not really the concern, though, is it? Unless you are doing photography with some kind of quantitative goal, what you are looking for is "good" color or color that suits your visualization. Even planetary geologists doesn't care about whether an image is eyeball-accurate. They only care that it fulfills a quantitative objective, e.g., accurately captures reflectivity in a particular wavelength or band.

You'll never have a workflow that matches the combination of the color response of the human eye (which differs significantly among individual humans, notably between men and women) multiplied by the spectrum of the light source multiplied by the reflective frequency response of the image subject, all of that processed by the unknowable and situational mechanics of human vision. You can only have something that, through a mixture of technology and human effort, creates an image perceived as "right" or "good."

Even things as simple as contrast and color temperature are confounded by human perception. What color temperature is right for a photograph? What amount of overall and local contrast control matches the human perception of a scene?

Color perception is a qualitative process, although obviously some intentional and repeatable quantitative process can be a big help in making that happen.

But there are physical limitations regardless. We've all seen the careful process fail, e.g. had the lavender flower that photographs blue.

By the way, it drives me crazy to hear chemical/film photography described as "analog." "Analog" does not mean "not digital." "Analog" means, broadly, "using continuous functions." Analog television is analog imaging. Film photography is essentially a digital chemical process, as grains are either on or off.

 

[J N]

I guess you haven't spent much time around medical, lab and science photographers .
I have a friend who has photographically documented large numbers of dental procedures for demonstration and educational purposes. Some of that work can also be beautiful and powerful too.

In forensic and scientific photography, color is really low on the list of things that need to be precisely perceived, if that's even possible or definable. How yellow is a tooth? Yellower or whiter than a different one. A quantitative measurement of how yellow won't match human perception, nor will human perception be quantitatively accurate.

 

[dokko]

That's not really the concern, though, is it? Unless you are doing photography with some kind of quantitative goal, what you are looking for is "good" color or color that suits your visualization.

sure, I totally agree here.
I didn't want to imply that the colors should be as close as possible to "reality" (otherwise, why would we need different film stocks?).
what I was trying to communicate are the technical problems that can arise when we try to get "good" colors or color that suits my visualisation.

what I meant is, that I found this quite tricky, and that in my attempts, using RGB narrowband light sources didn't make this process any easier (actually quite a lot harder). but if somebody tries it and likes the results than that is all that counts.

 

[Romanko]

it's definitely not something I would recommend unless you feel comfortable in advanced color science.

No, just started my fall into this rabbit hole. I want to understand the effect of the light source on the conversion process. I have a Raleno video light with a decent CRI and want to try a narrowband source as an alternative. I could use filters but good narrow-band filters are expensive.
What was the design of your light source and how many LEDs did you use?

 

[MattKing]

In forensic and scientific photography, color is really low on the list of things that need to be precisely perceived, if that's even possible or definable. How yellow is a tooth? Yellower or whiter than a different one. A quantitative measurement of how yellow won't match human perception, nor will human perception be quantitatively accurate.

Depends on the application. In some cases, the colour is critical. In other cases, the colour accuracy may be less important than other measurement accuracies. But in almost all lab/scientific/demonstration usage, the communication of accurate and realistic information is critical.

 

[koraks]

I also tried some commonly available RGB light sources (like for film lighting for example), but the problem is usually that the red spectrum is not deep enough.
on top of that the blue and green lights need to have right spectral peaks

Funny. Those were also my observations when I constructed an RGB LED light source (well, many versions in fact) for my color enlarger. I found that 660nm red performed a lot better than 620nm, but I think part of this is really down to the secondary emissions around 520-550nm that many high-power 620nm red ones gave. I also found that the generally used blue (often 465nm) is too close to green to maintain hue purity and royal blue (around 450nm) gave a better (more pure) yellow on RA4 paper.

My testing wasn't quantitative, but mostly qualitative - in part I compared color checkers that I photographed and printed, and mostly, I just used a given version of a light source until I hit its limitations (i.e. I ran into color rendition problems that didn't suit my needs, a.k.a. "surely, I can do better than this!?") at which point I dug deeper. This also illustrates that while I can see @J N's point, even without getting overly scientific we sometimes just run into practical limits that force us to solve a problem. In the art world, there are loads of examples of people delving into color issues incredibly deeply without much of a scientific approach to it (i.e. there's nothing "quantitative" about it) just because they struggle to get what they had imagined. I'm not an artist, but I can certainly related to that aspect.

 

[J N]

Depends on the application. In some cases, the colour is critical. In other cases, the colour accuracy may be less important than other measurement accuracies. But in almost all lab/scientific/demonstration usage, the communication of accurate and realistic information is critical.

If I wanted to record the "color" of something forensically or in any laboratory setting, I'd use a spectrophotometer, not a camera.