NLE Color correction that is similar to parametric equalizer in audio?

[Warren Kawamoto]

In another thread, there is discussion about how LED light sources have spikes and valleys in their spectral output, thus making accurate color correction virtually impossible. That got me wondering. Is there such a thing as a color correction tool or third party program that acts like a parametric equalizer, where you can select only the range of frequency(s) you want to manipulate without affecting everything else? Something like this would be extremely useful for color correction and grading!

[Gary Nattrass]

I find that an interesting analogy but it shows an acceptance of using kit that is not designed for the job, would it also be acceptable to record audio with microphones or recorders with known problems only to try and sort those problems out in post with a graphic equaliser which would still affect the quality of the audio and remove other elements that could be still valid on the recording?

[Mark Britton]

Yes, color correction tools have tools to allow you to target a specific hue and manipulate it.

The the point being made in the thread you are referring to was you can't amplify what isn't there to begin with.

Extreme example to just conceptualize: imagine a red apple under a pure green light source. The apple is going to look black. You can't amplify the red in that shot, because there isn't any. Could you key the apple in a compositing tool and 'recolor' the apple? Possibly, but it won't look nearly as natural as if you had used "white" light. And there's going to be fun had trying to seperate the 'black' of the apple from the black you'd find in the scene elsewhere.

The moral of the story is: Be aware that your tools are not perfect. Don't go creating problems for yourself by knowingly using the wrong tool at the wrong time. If you're in a situation where you're stuck with the wrong tool for the job, at least you have an idea of what you might be in for, and can work to minimize or avoid as much trouble as possible.

[Dean Sensui]

The question isn't whether you can get perfect fidelity, but if you can get something that looks good.

In reality, color perception is relative.

There are some problems that are tough to solve, such as black turning brown under certain circumstances with the first version of the Sony EX1. Otherwise, can you get a "look" that works?

[Tom Morrow]

Photoshop's default color correction tool has always had this "parametric" feel, trying to replace particular colors only. But in most cases a more broad spectrum approach, adjusting all colors equally, gives a more natural look.

I remember when this feature of being able to replace particular colors came out. My roomate started a catalog business on the theory that one could now just take a picture of an item in one color, and use photoshop to change it to any other color. But it's just not the same; real color is so much more than an RGB value. Real color is speckled, spacial, complex. Each pixel of a solid fabric has a different RGB value.

[Evan Donn]

Apple's Color has this functionality in the secondaries, not sure what the official term for it is. You get a rainbow line that can be manipulated as a b-spline via points to raise or lower specific ranges of color, and you can add as many points as you like to control the width and falloff of your adjustments, so it's very much like working with a parametric EQ for color. There's actually 3 per secondary, one each for hue, saturation and luminance. Sure - they won't let you fix extreme problems like the 'green apple under a red light source' example, but when we're talking about things like correcting the uneven spectrum of lights (that do in fact cover the whole spectrum to some degree) they are useful tools.

Originally Posted by Gary Nattrass

would it also be acceptable to record audio with microphones or recorders with known problems only to try and sort those problems out in post with a graphic equaliser which would still affect the quality of the audio and remove other elements that could be still valid on the recording?

Sure, it would be acceptable if you've tested the workflow and know you can get the results you need out of it. There's a big difference between just doing something at random on set and hoping it can be fixed in post later, and using a known combination of production and post production tools to achieve a desired result.