Color Negatives and Colorspaces

[Zygomorph]

After quite a bit of trial and error, I have discovered that I must convert from my scanner's color space to my working color space (Adobe or ProPhoto) only after I have inverted it and set the black point to the orange mask of the film. If I do the conversion before these steps have been completed, I invariably get all sorts of wacky colors. Can anybody explain why this is so?

 

[timparkin]

After quite a bit of trial and error, I have discovered that I must convert from my scanner's color space to my working color space (Adobe or ProPhoto) only after I have inverted it and set the black point to the orange mask of the film. If I do the conversion before these steps have been completed, I invariably get all sorts of wacky colors. Can anybody explain why this is so?

Yep - I recently discovered you can change things quite dramatically by altering the working space before setting black and white points (and gamma).

All different aspects of colour space change things too - try changing to a custom RGB space and changing gamma, white point and primaries. I've found using CIE RGB, gamma 2 and temp of 5500 works fairly well..

As for why, different spaces will have different gammas and hence change the saturation in highlights or shadows when contrast in each channel is increased. I'm still looking at other aspects

Tim

 

[Zygomorph]

Hmm, I guess if I ask a purely technical question, I should expect a purely technical response... ;-) I think I understand now... allow me to paraphrase? Certain functions, such as per-channel changes in contrast, only have a linear response (not sure how else to say that...) in the colorspace in which the image was first conceived.

I guess I'm mostly concerned about finding the most efficient way of ensuring that my colors don't get hashed up any more than they need to in the conversion from scannerspace to standardspace. I remember thinking it was perfectly normal, not so many years ago, to manually correct for color casts introduced by an uncalibrated scanner! And now I see people newer to the game than I am posting their examples of how wonderful a new film is, and all I can think is "Trust me, it gets better."

The more I figure out color management, the more like wet printing this whole thing becomes. E.g. you put up a negative shot in daylight, put a piece of Kodak paper under it (RIP), dial in the recommended filter pack, and you're 95 percent there. Though, I never did actually had to roll my prints around in tubes while handling toxic chemicals, so... whatev. I do, however, remember getting dizzy from fumes, waiting for my prints to come out the other end of the processor. So, yay technology!

 

[timparkin]

Hmm, I guess if I ask a purely technical question, I should expect a purely technical response... ;-) I think I understand now... allow me to paraphrase? Certain functions, such as per-channel changes in contrast, only have a linear response (not sure how else to say that...) in the colorspace in which the image was first conceived.

I guess I'm mostly concerned about finding the most efficient way of ensuring that my colors don't get hashed up any more than they need to in the conversion from scannerspace to standardspace. I remember thinking it was perfectly normal, not so many years ago, to manually correct for color casts introduced by an uncalibrated scanner! And now I see people newer to the game than I am posting their examples of how wonderful a new film is, and all I can think is "Trust me, it gets better."

The more I figure out color management, the more like wet printing this whole thing becomes. E.g. you put up a negative shot in daylight, put a piece of Kodak paper under it (RIP), dial in the recommended filter pack, and you're 95 percent there. Though, I never did actually had to roll my prints around in tubes while handling toxic chemicals, so... whatev. I do, however, remember getting dizzy from fumes, waiting for my prints to come out the other end of the processor. So, yay technology!

Now I'm not completely sure however I'm using Aztek DPL which on my Howtek drum scanner which allows me to set the per channel white/black points as part of the scanner hardware (reprogram the op-amps behind the PMTs is how I've been told it works). The results of this give me fabulous colour so I'm going to compare this process with some manual processes (i.e. scan raw and then process in photoshop) and I'll come back and report.

 

[Ben Altman]

Now I'm not completely sure however I'm using Aztek DPL which on my Howtek drum scanner which allows me to set the per channel white/black points as part of the scanner hardware (reprogram the op-amps behind the PMTs is how I've been told it works). The results of this give me fabulous colour

Hi Tim, I'm doing the same thing (DPL for color negs) and would be interested to compare methods. I'm traveling at the moment so don't have it all in front of me, though. Maybe PM me and we can discuss when I get back?

Best, Ben

 

[Zygomorph]

That sounds great, I was just thinking that there might be a way to do something like this with a CCD scanner for which the per-channel analog gain/exposure time can be controlled. I guess to set both black and white points before the A/D would require some sort of signal stretching circuit... which, obviously, my lowly scanner does not have and which your drum scanners do.

Do you think this would preclude the need for color profiling? It is my understanding that changes to the analog gain in scanners require a re-profile.