Adapting spectral sensitivity to imitating film look in digital

[David R Munson]

Mods: I really wasn't sure where to put this, so please move if necessary.

I honestly don't even know what to call what I'm thinking of trying to do. I'm positive it's been done before, but again if you don't know what to call it it's hard to look up info...

Anyway, let's say I love the look of Kodak Tri-X Pan 320 and I have digital work I shot along side film and I want to make the digital shots resemble the film as much as possible. I can play with curves, I can tweak it in Photoshop in arbitrary ways, but if I wanted to go about it in a more academic, more accurate method based on the spectra sensitivity curve of the film in question how would I go about doing it? That is, how do I take a graph of spectral sensitivity for a film and translate that into a particular way of converting to a B&W file to match it?

Maybe I'm just really tired, but at the moment I just can't make it work.

Ideas? Links? Point me in the right direction if you can.

Many thanks in advance.

 

[wintoid]

I use ICM profiles from http://www.jfilabs.com/ with Capture One. Not sure it's really worth it though, to be honest. Spectral sensitivity is one thing, but the texture is defined by the grain, and so the profiles don't hit the spot in that regard.

 

[dan denmark]

try using film.

 

[rogue_designer]

Your biggest problem is, the film holds a wider dynamic range than the digital does. So you're not matching apples to apples.

That said, alot of people seem to like the DxO filmpack for imitating film characteristics with digital shots.
http://www.dxo.com/us/photo/filmpack

 

[David R Munson]

try using film.

I *am* using film. Film as well as digital. It's a matter of getting the two groups of images to look a bit less dissimilar as final products.

I am well aware of the differences in dynamic range, grain, etc but frankly dealing with those differences is something I have a better understanding of than matching the sensitivity characteristics of a particular film.

 

[Al Kaplan]

There's more to it than the lack of dynamic range or spectral sensitivity. Film doesn't have a linear response to light. It has a close to "straight line" response in the mid tones but flattens out in the "shoulder" region (high lights) and "toe" (shadows). The exact "characteristic curve" varies from film to film and is affected by development time, agitation, and even by the exact developing chemicals used to formulate the developer, dilution, how long since a powdered developer was mixed, the ph of the water, other disolved minerals in your local water supply...it's a long list.

Brief exposure to low light during development can produce "mackie lines" and the "Sabbatier effect", and they'll vary depending on when, during the development, the exposure takes place, but this would only be a problem if you develop by inspection.

Then there's something called "bromide drag" caused by lack of agitation.

Color film has three or four layers, each sensitive to a different part of the spectrum. I doubt if Fuji and Kodak use exactly the same dyes.

So that brings up the question of "What exactly is the look of film?" I developed my first roll in 1961. I can't answer the question.

No two films have the same spectral sensitivity, probably because it's of little concern to anybody using it. You quickly learn that a blue filter imitates the older orthochromatic emulsions or a yellow, orange, or red filter will darken a blue sky. Nobody much cares that "medium yellow" ain't exactly the same yellow for Hoya, Leitz, Kodak, Tiffen, Ednalite, etc.

You're chasing a will o'the whisp but you'll never catch it. Have fun trying. It's right up there with the lens conundrum ~ is sharpness king? Bokeh? Lack of flare? Another long list, and they probably affect film differently than a digital sensor.

 

[mfogiel]

I think, that you would like to see a film like curve in your digital images - well the problem is as far as I understand it intuitively, in the bit depth, especially in the highlights, where the relative amount of shades of light gray differs dramatically. So if you wanted to match e.g. Tri X developed in a pyro type developer in that highlight range with your digital, you would probably have to underexpose your digi image terribly, and then push all the tones a few stops up. This would give you some detail in the highlights, but you would have nothing else, no midtones, shadows, etc. My take is, if you want a quality digital B&W image, shoot grey walls or wait 10 years, and shoot the film in the meantime - maybe at that point we will have 48 bit procesing...