Using Color - At what point do Gels effect CRI

[Albion Hockney]

I generally shoot stuff that is pretty much with in the bounds of reality color wise. For the first time I shot something pretty stylized recently with heavy color. I had a interesting experience using colored light sources....especially monochromatic color. Camera and my eyes did weird things!

 

I'm curious how Gels effect CRI and how the CRI value of a source will effect the final look.

 

For example take two options.

 

Shoot a scene at 3200K with Daylight Balanced lights

 

Vs

 

Shoot at scene at 5600K with Full CTB on a daylight source.

 

 

Does FULL CTB ontop of daylight lower the CRI? it must right? ....the result of shooting the 2nd way must make the light appear more monochromatic? ....what do you think the difference between these options would be in look?

 

 

 

[Guy Holt]

I'm curious how... the CRI value of a source will effect the final look.

 

A major problem with low CRI sources is that gels (party, effects or color correction) do not produce the same repeatable result that you’d expect from gelling a high CRI continuous source like tungsten.

 

Take for example the discontinuous spectrum of even high CRI remote Phosphor LEDs like the Cineo Trucolor. As you can see by the spectral distribution graph for the Cineo with 3200K panel above, they have a spike in the blue range that does not exist in the continuous spectrum of a tungsten light source (green line.) The greater proportion of blue and green/cyan in the Cineo Trucolor 3200 will result in an unexpected and undesirable result if a color correction (CTO, CTS, CTB), or color effect gel (Congo Blue, Bastard Amber, etc.) calibrated for the continuous spectrum of tungsten light is used on this discontinuous spectrum of a 3200K remote phosphor LED.

 

The same is true of the Cineo Trucolor with the 5600K panel above. The quite prominent blue spike (black line) will likewise result in an unexpected and undesirable result if a color correction, or color effect, gel calibrated for the continuous spectrum of a continuous daylight source is used on the discontinuous spectrum of a 5500K remote phosphor LED.

 

Color correction and color effect gels are a part of a finely calibrated imaging system that involves a highly specific light receptor (film emulsion or video sensor), light sources, and color correction or effects gels calibrated for both. Where that exists between film emulsions/video sensors and tungsten and/or daylight sources it is possible to mix dyes in a gelatin materials to create desired effects (it has taken decades to hone this system.) To use the available color correction and effects gels on the discontinuous spectrum of LEDs is a misapplication of a finely tuned system designed for continuous spectrum light sources only. Use this link for examples of what happens when you put a color effects gel designed for tungsten source on the discontinuous color spectrum of LEDs.

 

Guy Holt, Gaffer, ScreenLight & Grip, Lighting Rental and Sales in Boston.

[Stuart Brereton]

I'm a little confused as to whether you are referring to Color Temperature or Color Rendering. CTB is Color Temp gel and have should have no effect on CRI.

[Adrian Sierkowski]

CRI effects CT gels as without the full spectrum the gels designed for you don't really get the proper correction off of something like CTO or CTB-- at least in my experience. Its almost like putting some party gels on an HMI (Like for example a strong saturated yellow) you get something.... off..

With a party pack, it's not too bad; but when you're trying to do corrections, the discontinuous spectrum of LEDs and LEDRPs can be problematic as I think Guy is trying to show.

[Albion Hockney]

Yea I understand that if you have a low CRI source that color gel will effect it differently.

 

the question is. Do color correction gels effect CRI. If you for example put Full CTB on a high CRI Daylight source (a new HMI lets say) effectivley moving the color temperature into a very high kelvin reading what is the CRI like.

[John E Clark]

Yea I understand that if you have a low CRI source that color gel will effect it differently.

 

the question is. Do color correction gels effect CRI. If you for example put Full CTB on a high CRI Daylight source (a new HMI lets say) effectivley moving the color temperature into a very high kelvin reading what is the CRI like.

 

Color Temperature relates to 'black body' radiation, and a 'smooth' spectrum. CRI relates some types of 'spiky' spectra to that 'smooth' spectra, that is if the 'spiky' light is not so spiky, it has a better CRI relative to the smooth spectra ideals.

 

If a light has a 'blue' spike, and a red portion of the spectrum that doesn't match 'tungsten' in terms of spectral energy across the 'reddish' portion of the spectrum, the filter will suppress the 'red' portion relative more than the blue, relative to 'tungsten'.

[Tom Morrow]

Yes gels will tend to lower the cri if used on non black body sources because they are introducing further nonlinearities into the spectral response. How much this is an issue will depend on subtle interactions between the spectral response of gel source and sensor so there is no formula.

 

If you are adding ctb onto a daylight hmi I doubt it would be an issue. But I have seen horrible results adding cto to cheap daylight LEDs that looked ok without gels as daylight. They totally didn't match tungsten at all. Also mediocre results for adding plus green to low cri daylight to match green spiky practical flos.

 

Even reducing green spikes in high cri practical Flo lights with minus green gels has given inferior results versus skipping the minus green and balancing to white in post.

 

This is why I avoid gels on daylight sources.

[Tom Morrow]

My rule of thumb is to put gels on the tungsten source instead of the daylight source when trying to achieve a particular balance.

[Guy Holt]

I have seen horrible results adding cto to cheap daylight LEDs that looked ok without gels as daylight. They totally didn't match tungsten at all.

 

That's because LEDs don’t put out much beyond 625nm, so there is not much for a CTO filter to pass to rebalance the light output to 2900K. The result is that the “corrected” light is too cool. Another undesirable consequence comes from the fact that Full CTO is designed to pass extra green (there is a bump in the spectral transmission curve of Full CTO in the green portion of the spectrum) and so it creates, given the amount of green inherent in Daylight LEDs to begin with, a disproportionate amount of green (creating an overall green bias) to the "corrected" light when used on Phosphor White LEDs. Use this link to test results demonstrating this with a Lightpanel 1x1 Daylight Spot. The gel pack that eventually made the Lightpanel 1x1 Daylight Spot in the test marginally similar to a tungsten light, was only able to do so at the expense of two stops – so much for the greater efficiency of LEDs.

 

Guy Holt, Gaffer, ScreenLight & Grip, Lighting Rentals & Sales in Boston

Edited March 29, 2015 by Guy Holt

[Jeremy Cavanagh]

Guy,

 

Interesting posts and thanks for the link. Question: Can phosphors be used to smooth out spikes emitted in a LED's spectrum? I.e. if you take a red Led and put a phosphor around it that will absorb a spike and re emit across a wider part of the spectrum for red light (albeit at a lower level)? is this something that is practical i.e. outside of the lab and is it done with any sort of lighting?

[Phil Rhodes]

Red-emitting phosphors do exist and are used in certain types of backlighting for LCD displays. I am not aware of them being used for lighting devices, but they may be. Of course some lights have been using red and amber emitting LEDs to beef up longwave output for some time, but that does lead to a rather spiky spectrum in that area.

 

P

[Adrian Sierkowski]

As Phil kinda eludes to it's hard to fight spikes with spikes-- and the real issue comes that the only real way to get a smooth spectral output is to have some kind of black-body like radiator (e.g. a tungsten filament which heats up) when it comes to LEDs and even Remote Phosphor (which is kinda what you're leaning towards in your question), you'll still have a spiky output-- though some much better than others.