There is an immediate answer...

...to the weird lighting in the Vashon High School gym issue.

That issue was the apparent fact that at a shutter speed of 250th sec and up it's quite possible to capture the changing light output/light color as the metal halide gym lights cycle on-and-off-and-on at 120 hz or 120 times a second.

Consider the following, as-shot:



ISO 12,800, 400th second at f5.0, Auto White Balance.

Not bad, really, and maybe some snapshot-shooters would be satisfied, but to me it's way, way too cool and bluish.

(I've been researching the detailed EXIF data for Canon EOS 1D Mark IV RAW files, and I have yet to figure out just exactly how Canon encodes the white balance data for a photo shot on "Auto White Balance". It's got to be in there somewhere, but currently I'm thinking it's not expressed in degrees Kelvin but rather in an RGGB quadruplet, which I have yet to find/decypher).

Merely doing the RAW conversion with the White Balance set to Fluorescent (approximately 4000 degrees K) improves the photo quite a bit:



except that now it's too yellow-y and the skin tones are too flat.

So, clicking on "More settings" brings up the "Additional Settings" (!) dialog, where we can fiddle with all sort of stuff.

We've got a Blue B9 ... B3 ... 0 ... A3 ... A9 Amber slider, a Green G9 ... 0 ... M9 Magenta slider, and a "Color Tone" drop-down -- which is actually Reddish skin tone -4 ... 0/as-shot ... +4 Yellowish skin tone range -- to play with.

Remember since we're talking about light, we're talking Additive Color here.

Take a look at the Additive Color wheel and green is opposite magenta (which is the combination of red and blue) while blue is opposite yellow ("amber" here) -- yellow being red and green combined.

Which is why our pairings here are "blue" to "amber" and "green" to "magenta".

Why "skin tone" rather than a "red" to "cyan" pairing? I'd guess that a lot of real-world photography has a greater need of fiddling with skin tone rather than fiddling with colors ranging from red to cyan. Also, "skin tone" is a specific adjustment in the Canon EOS Picture Style system, which is what we're really doing here: making micro-adjustments to Canon RAW files before we convert them to tiffs.

Anyhoo, when I pull over the Blue to B5, and dropped down the Skin Tone to -4 Reddish, I get something that I like better, at least, than any other setting.

The floor particularly has lost a little yellow, the skin tones look a little more alive, the green of the uniforms seems to have a little more pop, and we're good to go.

(Note that despite all the sliders and pull-down dialogs this all ends up being really subjective, and also implies a calibrated monitor, which I have).



Here's the finished product:



So to correct the weird lighting issues from the volleyball match I shot in the Vashon High School gym I ended up fiddling with the Blue-Amber slider particularly, added in a touch of Magenta on a couple, and pulled up or down the Skin Tone a taste on a few.

And they came out OK, at least as OK as I was willing to spend time with them.

The real test will be photographing the next VHS volleyball match this coming Friday.

Well, that may have answered...

...that question.

The question being, why the hell do photos shot in the Vashon High School gym come out with such wildly different color renditions between two photos shot within fractions of a second from one another?

Example #1 (you'll need to click on the image to see the full details);


And Example #2:



Obviously these two photos were shot within a fraction of a second from one another.

And yes, (Spoiler Alert!) they were shot on Auto White Balance, which may turn out to be half the problem.

But only half.

So while the color cast in photo Number One looks pretty much OK, Number Two (particularly look at the color of the back wall) has a distinctly orangey-pinky cast to it.

W, as they say, TF?

I haven't moved; the gym hasn't moved; so why the hell is the color cast so different (and this is a relatively mild example) between the two photos?

Now, it's a standing joke that the halide lights in the Vashon High School gym are appallingly bad anyway: they're literally different colors, such that you can stand out on the gym floor and almost see no two light of the same visible color.

This is due to age and maintenance, I'd bet: when the individual bulbs have needed replacement I'm sure that no thought whatsoever was given to buying a metal-halide bulb identical to those already installed. Let's get the cheapest, and be done with it.

Add in age (halide lights change color as they get older) and you've got lighting that pretty much sucks.

But why would two photos shot within moments of one another, pointed at identical parts of the gym, have such different color casts?

Turns out the real culprit is alternating current, and the way halide (and fluourescent) lights work.

Halide (and fluorescent) lights are actually cycling on and off at 120 hertz or 120 times per second -- at least those with older non-electronic ballasts -- which I'm sure is what the VHS gym lights are.

Way too fast for the human optic system to see, but not too fast for a camera taking ten shots per second at an exposure of 250th second per.

So apparently what I'm catching here is the actual change in light output/color output of the lights as they cycle on and off at 120 hz.

OK: fine.

What to do about it?

Slowing the shutter speed down to, say, a 60th or a 30th of a second might do it, but that's not going to work for shooting sports.

So I think the strategy is to register a Custom White Balance in my EOS 1D Mark IV and 5D Mark II, and not have the camera think that it's doing it correctly when it's looking at light that's constantly flickering and thinking it's got a valid reading for Auto White Balance.

During RAW conversion and post-processing I just tell BreezeBrowser "This is what color the light is, period" instead of whatever the camera thought it was at the instant a particular photo was shot.

I think.

At least that's the Plan of the Moment(tm).