Scanlight RGB, Spectracolor, and Flash for Camera Scanning

[silvergelatin]

I have been camera scanning for couple of years now using a 95 CRI panel, and the results have been mostly ok, though I often have to spend a lot of time color correcting to get things just right on more difficult color images (Ektar, etc). Since I have mostly been shooting b/w, this was not a big deal, but I’m back in a color phase again and want to optimize a bit.

I tried a few rolls with flash, and the results were improved rather dramatically. Colors much cleaner, NLP/Negmaster inversions nailed on the first try, etc. I have to decide whether to continue with flash, which requires extensive workflow modifications (and a need for a modeling light to focus), or find a constant light source that can give me the same quality.

I have been reading about Jack’s Big Scanlight and the Cinestill Spectracolor panel. No real conclusive comparisons out there so far, so I’m hoping someone has been down this rabbit hole already. I’m curious if either of these options would match or exceed my flash results.

 

[koraks]

If I were to set this up, there would be no doubt whatsoever in my mind how to approach it - a RGB light source, not broad-spectrum.

Jack’s Big Scanlight

That one looks promising to me; I see he's evolved it from the initial version that was presented about a year ago.

 

[4season]

Until seeing this video, I had assumed that if I had lighting with ~5000 degree Kelvin color temperature, and the highest possible Color Rendering Index, I'd be set: Wrong!

 

[koraks]

That was an amusing video and there's very little to argue about the content (if only because that would take the fun out of it).
But - this is about scanning film. And we should take into account this basic fact. When scanning film, esp. color negative film, we're somewhere in the middle of the imaging chain. We're not recording, and we're not projecting. We're digitizing and the requirements on that process are not necessarily (and arguably, really not) the same.

 

[silvergelatin]

Now I want that Sekonic meter.
I already understand why typical LED panels are not ideal, but am more interested in learning about how these newer tuned RGB lights interact with neg film dyes and camera sensors, as opposed to a smooth spectrum source like flash (which clearly beats the typical LEDs by a lot).

 

[koraks]

am more interested in learning about how these newer tuned RGB lights interact with neg film dyes and camera sensors

Take these three plots and put them on top of each other:
LED emission wavelengths
Film dye absorption spectrum
Camera sensor spectral sensitivity
The first two are generally fairly easily found in the datasheets. The third is trickier since camera manufacturers are often not very liberal in providing this data, but sometimes you can find it. And otherwise you can approximate it by having a look at other, similar CMOS sensors.

PS: modern film scanners for e.g. motion picture (Hollywood) scanning generally use RGB light sources instead of white light with sequential exposures, one for each color channel. Apparently the industry decided that this is somehow the best approach and I'm sure it's not a productivity consideration (esp. the 3 separate scans per image).

 

[photoyoda]

Just went through this same thought process and picked up a SpectraColor to start with since it'll swap in simply to my BlackScale Lab Holo setup. Will follow-up when I'm able to put it through its paces.

 

[silvergelatin]

Take these three plots and put them on top of each other:
LED emission wavelengths
Film dye absorption spectrum
Camera sensor spectral sensitivity
The first two are generally fairly easily found in the datasheets. The third is trickier since camera manufacturers are often not very liberal in providing this data, but sometimes you can find it. And otherwise you can approximate it by having a look at other, similar CMOS sensors.

PS: modern film scanners for e.g. motion picture (Hollywood) scanning generally use RGB light sources instead of white light with sequential exposures, one for each color channel. Apparently the industry decided that this is somehow the best approach and I'm sure it's not a productivity consideration (esp. the 3 separate scans per image).

I don’t see myself buying a monochrome camera and doing it the hard way, so hopefully the combined RGB light is still a big step up. According to others who have used it, it certainly is, but I’ve yet to see anyone compare it to flash (my gold standard so far).

 

[silvergelatin]

Just went through this same thought process and picked up a SpectraColor to start with since it'll swap in simply to my BlackScale Lab Holo setup. Will follow-up when I'm able to put it through its paces.

I imagine it will be step up from the old CS Lite. I ordered the Scanlight, but it will be a few weeks until I get it.

 

[koraks]

I don’t see myself buying a monochrome camera and doing it the hard way

You wouldn't have to; you can take an RGB sensor and only expose it with one color, then discard all other color information from the digital file. It's a quasi-monochrome camera that way. But personally I'd just start with the 3 colors at the same time to save time.

I’ve yet to see anyone compare it to flash (my gold standard so far).

My experience in color work, including color enlarging, is that depending on the process you follow, the results are always different, but the question always remains unanswered which approach is objectively better.

 

[silvergelatin]

I received the Scanlight yesterday and tried it out on a super contrasty Ektar neg. As expected, Negative Lab Pro did not do a very good job with it, but manual inversion and correction was a little easier. I dialed in reduced blue and red intensity to cancel out the base color and bring the RGB peaks together in the histogram, and this gave me much better results. I still have a lot of experimenting to do, but looks promising so far.