Viability of in-camera correction for LED and fluorescent lighting?

Question

I am getting an Olympus MFT system. The bodies in the system have in-camera color correction, and I am wondering how effective it will be compared to using gels.

My concern is that for intense lights like LEDs, the blue is so strong that it will mess up the exposure/white balance and since the sensor is saturated it will not really be able to correctly apply the compensation. I am concerned that I need a resin to knock the blues down before the light even gets to the sensor.

Is this a legitimate fear, or is the in-camera color compensation immune to this kind of problem and I can just throw out the gels?

(BTW I have very limited time to do post-processing, so my shooting style emphasizes getting the shot right out of the camera and minimizing post-processing, hopefully to zero, so post processing solutions are not what I am looking for.)

Answer 1

I do not use Olympus but I can offer a general statement: most auto-white-balance systems are pretty good, but are inherently limited by three issues - (1) "white balance" adjustments are aimed primarily at a color temperature, which presumes a light spectrum which is fairly smooth; artificial light other than incandescents are not smooth but have very discontinuous spectrums, (2) many lighting systems flicker, and besides changing brightness during the cycle change color, so the camera may correct well but by the time the shutter opens the light is different, and (3) your scene can be confusing to a system; imagine a florescent lit scene but with daylight from a window lighting a part of it, so that there is no single right setting.

None of these are easily handled by any kind of filters. Mixed lighting can be somewhat mitigated, e.g. gels on strobes to cool or warm the flash to match natural lighting.

There are two approaches that have the most leverage, one you discounted (mostly) which is post processing. In post processing you not only can adjust white balance, but apply it selectively over areas of the image that may be lit differently. Shooting raw gives you a LOT more leverage in post processing for white balance adjustments.

The other (that is complementary not exclusive) is a custom camera profile, assuming your camera shoots raw. While this also requires shooting raw, it does not require human attention to apply. If you produce a custom camera profile for a specific lighting venue, it can be applied to all shots in that venue at once. This does NOT correct white balance errors, but it does help correct color distortions due to unusual spectra (e.g. lighting that may have too little green would see green emphasized to adjust).

Another complementary approach is to look for a camera which has "flicker control" (may have different names). This is mostly available in higher end cameras, but it tracks the actual flicker of the lights, and opens the shutter only to coincide with the brightest point. This is not so much about being bright, as being consistent - instead of pink/green/whiter cycling you get only the whiter, which then can have more uniform white balance adjustments applied. A couple minutes with google did not show me any Olympus MFT with flicker control but I did not look carefully, plus that changes over time. Sports oriented cameras in particular are adopting this feature, as flickering lights in arenas and stadiums are a real problem there.

My recommendation, if you can do any post processing at all, is to (1) shoot raw, (2) plan on a brief look and adjustment on a calibrated screen to correct where the camera misses, and (3) if you shoot the same venue frequently consider a camera profile (and software, like Adobe, that supports it). If you can't post process at all (as in human review), shoot raw anyway, run through a good ACR, and trust the auto-white-balance. That way if you get that really spectacular shot and want to improve the colors after the fact, you have the raw image to provide extra leverage. Some cameras will shoot raw and JPG at the same time.

I should also mention that a manual color balance is a possibility, but one I mostly discount as (1) flicker kills it in many places, (2) unless you are shooting in one spot in one venue it changes depending on subject and where you aim and background. It is useful if you can work it in to have a neutral grey card in view at the edge of the shots, provided you can get it in identical light. You can use this as a quick reference in post processing. But be aware it might not work perfectly, as you may be amazed at the difference just a few feet can make. I shoot basketball (as an example), and the white balance varies hugely by whether a player is down near the floor (e.g. scrambling for the ball) or up high shooting, primarily due to the warming effect of the yellow floor proximity. Many venues are like that, with a lot of variation over short distances.

Answer 2

I don't shoot Olympus, but I've found that when used properly the in-camera White Balance Correction works pretty well the Canon cameras I use. Please note that white balance is a much broader term than color temperature. Color temperature is but one axis across the color wheel. White balance involves the entire color wheel and can usually be adjusted by combining the blue/purple←→amber 'color temperature' axis with the green←→magenta 'tint/hue' axis.

With my Canon cameras, each unit of WB adjustment is approximately equal to a 5 mired color correction filter. So a G9 setting is equivalent a 45 mireds green filter (magenta reduction). That's a pretty wide range of adjustment. If the light in which you are shooting exceeds that, then actual gels placed on the lighting might be the next option. In practice, I've usually found that even if a +9 setting in whichever direction(s) I need doesn't get me all the way there, it gets close enough to use an HSL tool in post to pull back the remaining color cast.

Using a correctly colored gel on light sources or glass color filters in front of the lens can allow you to increase the exposure before the strongest channel blows out. This even works when one is shooting to produce monochrome images. Unfortunately most gels are oriented on the blue/purple←→amber axis when what you need with typical LED stage lighting is usually a strong green filter to attenuate magenta.

When shooting in such unbalanced light it is particularly important to pay attention to an RGB histogram, rather than a total luminance histogram. The single histogram can fool you into thinking nothing is being overexposed even when, in fact, one channel is being fully saturated. Using the separate red, green, and blue histograms will show very clearly when one channel is much brighter than the others. One channel at full saturation can also make correctly focused images look blurry. Keep in mind that if shooting raw you have a stop or two of headroom beyond what the in-camera histogram, based on the camera generated jpeg preview image, is showing you.

Be aware that when shooting raw the correction will be applied to the jpeg preview, but may or may not be applied by a raw conversion application that you use to open the raw file. Most manufacturer's in house software will apply the in camera setting by default. Most third party raw converters, such as Adobe products that use Adobe Camera Raw, do not.

The further the lights you are using are from the axis along which black body radiators emit (the color temperature axis), the more likely you'll need very fine and customized color corrections. In camera adjustments are fairly coarse compared to what can be done with a good raw conversion application. For instance, in camera WB correction with Canon cameras is limited to integer values (i.e. 1, 2, 3, etc.). In Canon's Digital Photo Professional 4 the gradations are in tenths (i.e. 1.0, 1.1, 1.2, 1.3, etc.). Physical filters are even more limited in terms of the steps between conventional choices..