Digitizing color slides with camera: backlight temperature and chromatic adaptation

[River Mantis]

I have an established workflow for black and white negatives but I have never worked with color slides before. So here are my thoughts. Ektachrome E100 is a daylight balanced film. When applied to a transparency it means, I guess, that the white card shot under daylight will have the color of the projector's light source when the slide is projected. At the same time slide is meant to be viewed in a dark room with no other light sources. So viewer's chromatic adaptation should theoretically mask any possible difference in sources' temperature. From this perspective the obvious choice would be to adapt the digitized image from the backlight to a standard whitepoint of the final space. In other words the white balance must be set by backlight as a reference. But this formally correct approach leads to a way too neutral and kind of flat result on a screen while when I see the slide over my CS-Lite backlight in its warm (aka "slide") mode the colors are so much more vibrant. So what's the right way of digitizing color slides?

 

[Chan Tran]

That was the right way. If you want your slide to look different then process a bit in post.

 

[MattKing]

Slide films were designed to be projected using tungsten (which includes halogen) bulbs.
Most importantly, try to use a backlight with a continuous spectrum.
Be cautious about using LED sources and relying on a quoted a "Colour Rendering Index" or "CRI". That CRI measurement standard is aimed at different criteria than those that may make a difference with digitization.

 

[NMSS_2]

The neutral backlight approach is “correct,” but yeah, it can look flat. Projection always felt warmer because our eyes adapted to tungsten light. Best compromise is scan neutral for accuracy, then warm it up a bit in post if you want that classic vibrant slide look.

 

[River Mantis]

The neutral backlight approach is “correct,” but yeah, it can look flat. Projection always felt warmer because our eyes adapted to tungsten light. Best compromise is scan neutral for accuracy, then warm it up a bit in post if you want that classic vibrant slide look.

I feel like it's the best approach

 

[loccdor]

There is always variance in color, it can be from film expiration, from a cast from the lens, from a reflective colorful wall or body of water in your photo scene, etc. In the end I found while using the base temperature of the CS-Lite to be a good first step, it's always good to tweak it a little afterwards until it looks the best. Luckily when using slide film you start off much closer to accuracy than when using negative.

 

[DDTJRAC]

I have an established workflow for black and white negatives but I have never worked with color slides before. So here are my thoughts. Ektachrome E100 is a daylight balanced film. When applied to a transparency it means, I guess, that the white card shot under daylight will have the color of the projector's light source when the slide is projected. At the same time slide is meant to be viewed in a dark room with no other light sources. So viewer's chromatic adaptation should theoretically mask any possible difference in sources' temperature. From this perspective the obvious choice would be to adapt the digitized image from the backlight to a standard whitepoint of the final space. In other words the white balance must be set by backlight as a reference. But this formally correct approach leads to a way too neutral and kind of flat result on a screen while when I see the slide over my CS-Lite backlight in its warm (aka "slide") mode the colors are so much more vibrant. So what's the right way of digitizing color slides?

Send in some samples of your scan work. These flatbed scanners are garbage for digitizing chromes. At least the ones I've tried.

 

[Alan Edward Klein]

Does anyone know the Kelvin ratings for the light bulb used in the Epson scanner models V600 and V850?

 

[Chan Tran]

Does anyone know the Kelvin ratings for the light bulb used in the Epson scanner models V600 and V850?

I don't know but my HP flatbed scanner I think is 5000K or higher. Perhaps 6500K

 

[River Mantis]

Send in some samples of your scan work. These flatbed scanners are garbage for digitizing chromes. At least the ones I've tried.

I'm using a camera

 

[MattKing]

Does anyone know the Kelvin ratings for the light bulb used in the Epson scanner models V600 and V850?

Don't forget that film scanners have the advantage that any necessary calibration adjustments for the peculiarities of the fixed, built in light source can easily be built into the firmware or software, and thus be invisible to the end user.

 

[Alan Edward Klein]

Don't forget that film scanners have the advantage that any necessary calibration adjustments for the peculiarities of the fixed, built in light source can easily be built into the firmware or software, and thus be invisible to the end user.

Good point. I wonder if there are changes to the lamp's Kelvin over time?

 

[MattKing]

Good point. I wonder if there are changes to the lamp's Kelvin over time?

Possibly, but that wouldn't matter a lot.
What would matter is whether there is any variation to the continuous or non-continuous nature of the source's emitted spectrum.
Particularly for those scanners making use of LEDs.

 

[isaac7]

This is probably the most comprehensive overview I’ve seen. Lots of theory and testing of various light sources. It’s a long thread but if you really want to know…

Suggested backlight sources for scanning film with DSLR

Nate Weatherly posted a pretty thorough reply to a post on Facebook with generic recommendations for light sources that would work better than LED plates such as tracing panels etc. I know @Richard1Karash has done a pretty detailed comparison of different light sources as well. Perhaps it would...

forums.negativelabpro.com

 

[Aidan Sciortino]

All of the documentation from Kodak I’ve read on duplicating slides indicates that it’s best done at D50 if possible. I use the same recommendation for scanning with a camera, the idea being that it’s probably a closer match with my sensor’s reference balance.

I know some companies advertise a special “warm” mode that emulates a projector bulb on their light sources, but I wouldn’t trust that, especially since these products are generally only rated with a CRI at a single specific temperature.

 

[River Mantis]

I know some companies advertise a special “warm” mode that emulates a projector bulb on their light sources, but I wouldn’t trust that, especially since these products are generally only rated with a CRI at a single specific temperature.

I have one. The problem is that the warm mode is about two times less bright and I'm starting to have a vibration issue on lower shutter speeds