Scanning Color Negatives

[Deon Reynolds]

7 minutes ago, digitaldog said:

When you set the color space on the camera for Adobe RGB (1998) you are simply telling the camera and processor how to convert the raw into a JPEG in that color space. The camera still shoots raw (it has to) but if you tell the camera just to provide a JPEG, the raw is converted to that color space and Adiós raw data. Raw data has a color space too, this is a deep kind of geeky rabbit hole to go into. Native (raw) camera color spaces are almost never colorimetric, and therefore cannot be defined using primaries. Therefore, the measured pixel values don't even produce a gamut until they're mapped into a particular RGB space. Before then, *all* colors are (by definition) possible. 

Fascinating. Thank you for the education! I always shoot in RAW but had no idea what happens under the hood, I call it FM (F-ing Magic).

[conrad_hoffman]

You guys are all so far ahead of me it isn't funny. I've been trying to digitize some 1980s Kodacolor negs and have yet to be remotely satisfied with the results. Doing the mask subtraction and trying to get the colors right manually is near to impossible. I don't use PS or LR so the various plugins aren't useful to me. I just discovered that RawTherapee has a pretty good tool for the conversion but I haven't had time to fully explore it. It does seem essential to get as many bits as possible, so raw makes sense. 'Chromes have been much less trouble. Surprisingly, black and white negs can be as much trouble to get right, curves being essential. In either case I use a Z6 and a PS-4, lit by a Nanlite 5C and preset color balance. Anyway, more basic info on how you get your color balance right would be welcome.

[digitaldog]

First, its often hard to follow when some of the info provided isn't close to accurate. Enough said. 

Next, converting scans of negs positive isn't easy, never was dating back to the old 1990, Leaf scanner days. I've used and reviewed dozens upon dozens of film scanners including drum scanners costing as much as cars in the days. It is all about the software; some is far better than others. Inverting and removing the orange mask isn't trivial and the mask isn't consistent from shot to shot (exposure alone plays a role). 

At this point, the 'fix' is to find some software solution that works with modern scanners or after the scan (like maybe Lightroom Classic). 

SilverFast is pretty good if can drive your scanner with it.

NegativeLab is supposed to be a very good option (full disclosure, I've never used it, but those who use LR I respect, those that know what they are doing, have given it high marks and the hope is, such functionality will find it's way into LR/ACR 'natively' someday). You can try a demo:

https://www.negativelabpro.com 

[AlanKlein]

Here are my settings after calibration completed.  I usually calibrate only when I'm to do some editing or scanning. 

Edited December 12, 2022 by AlanKlein

[digitaldog]

1. What do you want me to say Alan? You are not calibrating to sRGB as outlined earlier.
2. Did you even read what I posted?
3. Or the manual with the page that explains what you're doing?
4. Do you feel calibrating every 173 days is a good idea? 
5. Does this have anything to do with scanning color negs? 

[rodeo_joe1]

On 12/11/2022 at 3:13 PM, digitaldog said:

Additional primaries is a straw man; the primaries ARE visible! Two of those in ProPhoto RGB are not.

Does it matter how the display space is widened? 

Whatever works, works. 

Since you can't have primaries that are more saturated than a single monochromatic wavelength, then other options need to be explored. Because it seems logical that only real and physical pigments, dyes and light sources give rise to the entire range of colour sensation experienced by humans, and that therefore those sensations should be reproducible by some system of light-sources, dyes and (but maybe not including) pigments.

Filling in the literal holes in RGB theory seems like a good way to do that. 

Edited December 13, 2022 by rodeo_joe1

[digitaldog]

It matters when people assume the a device such as a display emit invisible radiation when talking of their color gamut. It illustrates they don't understand what color is, that's all. 

On 12/11/2022 at 3:49 AM, rodeo_joe1 said:

Until we get display devices with the imaginary primaries of the Prophoto space. 

Edited December 13, 2022 by digitaldog

[AlanKlein]

15 hours ago, digitaldog said:

1. What do you want me to say Alan? You are not calibrating to sRGB as outlined earlier.
2. Did you even read what I posted?
3. Or the manual with the page that explains what you're doing?
4. Do you feel calibrating every 173 days is a good idea? 
5. Does this have anything to do with scanning color negs? 

1. What am I missing?  I select sRGB Emulation as the Target Settings in Spectraview and let the program do its work.  The resultant settings seem to match yours from what i can tell.

[digitaldog]

2 hours ago, AlanKlein said:

1. What am I missing?  I select sRGB Emulation as the Target Settings in Spectraview and let the program do its work.  The resultant settings seem to match yours from what i can tell.

You are missing the answers already provided.

And those are not MY setting! I'd never use them.

[JDMvW]

On 12/11/2022 at 4:49 AM, rodeo_joe1 said:

Ektachrome and Kodachrome are capable of some saturated colours outside of any real-world RGB space.

 

You can approximate slightly underexposed Kodachrome by darkening and by setting the saturation to ungodly levels.

Kodachrome 64

Edited December 15, 2022 by JDMvW

[glen_h]

On 12/7/2022 at 8:49 AM, Robin Smith said:

Add my agreement. There is no point in trying to make color negative film look like modern digital output. People are still shooting film precisely because they like the look of these emulsions. The shot you posted looks good. Although different to C41, I want my Kodachromes to look like Kodachromes when scanned.

Kodachrome has a unique look for a few reasons. One is related to the way it is processed.

After first development, it is exposed to red light from the back, and the red sensitive layer developed.

Next, it is exposed to blue light from the front, and the blue sensitive layer developed.

And finally, the green sensitive layer is chemically exposed, and developed along with

any undeveloped silver halide still left in the previous layers.

As the first two color developers aren't perfect, there is still some, and so some magenta dye in other layers.

This last one is unique to the Kodachrome process (also, as I understand it, used by Dynachrome).

If this last effect is linear in the exposure of those layers, it should be easy to emulate with a matrix.

If not, it will be a little harder, but might be possible. 

[invisibleflash]

On 12/11/2022 at 1:42 PM, AlanKlein said:

Before editing my scans, I set the NEC Spectraview II calibration software to calibrate my NEC monitor with the puck to sRGB.   I assume NEC program calibrates to appropriate standards for sRGB color gamut.  I'm happy with the colors and others can check my FLICKR site to see if they look OK with them.  

Here are my Velvia scans and edits both in medium format 6x7 and large format 4x5. 

 

Excellent job!

[invisibleflash]

On 12/12/2022 at 3:09 PM, AlanKlein said:

Here are my settings after calibration completed.  I usually calibrate only when I'm to do some editing or scanning. 

Alan, every monitor is different. Everyone sees them differently.

Have you printed anything? Happy with the prints from your scans? 

[invisibleflash]

On 12/12/2022 at 1:30 PM, digitaldog said:

First, its often hard to follow when some of the info provided isn't close to accurate. Enough said. 

Next, converting scans of negs positive isn't easy, never was dating back to the old 1990, Leaf scanner days. I've used and reviewed dozens upon dozens of film scanners including drum scanners costing as much as cars in the days. It is all about the software; some is far better than others. Inverting and removing the orange mask isn't trivial and the mask isn't consistent from shot to shot (exposure alone plays a role). 

At this point, the 'fix' is to find some software solution that works with modern scanners or after the scan (like maybe Lightroom Classic). 

SilverFast is pretty good if can drive your scanner with it.

NegativeLab is supposed to be a very good option (full disclosure, I've never used it, but those who use LR I respect, those that know what they are doing, have given it high marks and the hope is, such functionality will find it's way into LR/ACR 'natively' someday). You can try a demo:

https://www.negativelabpro.com 

I'd like to try it if they ever make a stand alone version. 

[AlanKlein]

9 hours ago, invisibleflash said:

Alan, every monitor is different. Everyone sees them differently.

Have you printed anything? Happy with the prints from your scans? 

I've printed very little since moving.  The small ones I have printed are OK.  If I start printing for large prints, I'll probably have them scanned and printed professionally.  I scan for the web and personal presentation on my 4K TV, monitor, cellphone and to makes video slide shows also for personal and family. 

 

Does the color in the scans on my Flickr site look OK to you? 

 

[digitaldog]

9 hours ago, AlanKlein said:

Does the color in the scans on my Flickr site look OK to you? 

The Kodachrome scans? 

How about a 'before and after" keeping in mind, GIGO:Garbage In Garbage Out: What you've again clipped (see overaly) in either capture or scan is gone.

 

 

KodachromeBeforeandAfter.tif

Edited December 20, 2022 by digitaldog
TIFF for download wasn't flattened

[AlanKlein]

9 hours ago, digitaldog said:

The Kodachrome scans? 

How about a 'before and after" keeping in mind, GIGO:Garbage In Garbage Out: What you've again clipped (see overaly) in either capture or scan is gone.

 

 

KodachromeBeforeandAfter.tifUnavailable

Andrew,  Thanks for spending the time on this.  I don't know if the original slides and/or my editing were clipped but certainly I could have done better on the editing.  These slides are 50 years old and are my wife's parents,  I don't know where they are right now and I think the original scans are on a DVD saved in storage, some where in the house.  I'll have to look for both. 

What interests me is the color change.  What did you do and how do you know the new colors are the correct ones, especially the building?  When I scanned them originally and edited, I don't think I compared my results to the original slides, only that the colors looked OK in the edit, if not perfect. 

 

[digitaldog]

4 minutes ago, AlanKlein said:

What interests me is the color change.  What did you do and how do you know the new colors are the correct ones, especially the building?  

What I did; white balance. A bit of editing on blacks and shadows.

How do I know it is correct? 

1. It looks better (unless you like magenta color casts). But this is of course subjective. 

2. Items which appear they should be neutral (like a white shirt) are neutral or very, very close viewing Lab values or RGB values. 

3. What the color of the building and other's fall into, after WB is what they are. 

To answer the question you asked us (and others may have an opinion): No! Hence the quick edits. 

Edited December 20, 2022 by digitaldog

[digitaldog]

21 hours ago, invisibleflash said:

I'd like to try it if they ever make a stand alone version. 

That's not going to happen. We can go into why if you wish. 

[rodeo_joe1]

6 hours ago, AlanKlein said:

.... how do you know the new colors are the correct ones, especially the building?

'Cos tarmac (blacktop) isn't usually blue/Magenta, and neither should the creases on a white shirt be.

There's usually something in a picture that should be neutral grey. If it isn't, then that's the place to poke the grey-point eyedropper. 

Edited December 20, 2022 by rodeo_joe1

[AlanKlein]

Thanks Andrew and Joe.

[Ed_Ingold]

If you get green grass, blue (not magenta) skies, and reasonable flesh tones in the same image, consider your negative "scanning" successful. Be grateful you're not grading video, where standards are set by lawyers and government officials, with all the consistency we expect from politicians 😉 NTSC = Never the Same Color.

 

[Robin Smith]

When scanning Kodachromes or any slides I find one needs to carefully compare your result with the original and try and match the best you can, being sure you have a calibrated monitor and also are viewing the slide with a good consistent daylight light source. A really good match is sometimes impossible in all details, a certain amount of subjective decision making is required. Like all color matching it is important to not allow yourself too much time immersed on one slide in any session because your brain tends to start compensating for poor color so much that you can stop noticing that the shadows are blue or that the grays are magenta. At least with a slide you do have a point of reference though, unlike negatives where you may have to rely a good deal on memory.

Edited December 20, 2022 by Robin Smith

[digitaldog]

Match or improve. A transparency with an ugly magenta color cast (perhaps as shown yesterday); no reason to match it. 

Rendering is always subjective, massively with negatives but also with transparencies and digital capture: http://www.digitaldog.net/files/Langs_Render_to_Print.pdf

[Robin Smith]

I would never have kept a slide with ugly magenta color cast, so doesn't apply to me in that case, but yes, you can improve color if it is off. The red cast under tungsten light is a case in point. If I was unable to reduce it satisfactorily I would probably just convert to black and white.

Edited December 20, 2022 by Robin Smith