cgardner
Offline
• • • • •
Upload & Sell: Off
|
 p.1 #4 · Shooting question for theatrical/dance shows that use LEDs | |
While not photography specific, this link will provide some insight into working with LED stage lighting:
http://www.onstagelighting.co.uk/led-stage-lighting/led-lighting-and-colour/
Color and exposure, while related, need to be dealt with separately when shooting.
With regard to exposure your camera has a range of about 7-stops over which is can record detail. A sunny day exceeds that range and so does most stage lighting. The most important content in scenes usually falls in the middle of the tonal range — mid-tones— and correctly exposing them will require blowing highlights whenever scene range is greater than sensor range. Like death and taxes it is unavoidable in ambient (natural or stage) lighting.
One possible workaround to the problem is to arrange to shoot at dress rehearsals where flash is allowed, working with dual flash. I've done that with flashes gelled with 1/2 CTO filters to match the color balance of the base tungsten lighting, with flash allowing me light the foreground performers with "normal" skin-tones while at the same time balancing the background lighting and in some instances like the concert below, get lighting effects that are different and more effective for the photographs than the stage lighting alone wound have been.
http://super.nova.org/TP/PSurge4.jpg
http://super.nova.org/TP/PSurge5.jpg
http://super.nova.org/TP/FirepowerWide.jpg
http://super.nova.org/TP/FirePowerCU.jpg
For the last two shots the stage lighting for a teen talent show was rather bland so I put my OCF on the stage behind and to the left as a "kicker" shooting from in front with another flash on my bracket using my 70-200mm f/2.8 IS.
If you can't use flash such as during actual performances, you must control tonal range with cropping decisions. For example at any performance I'll move to the back of the room and shoot some wide establishing shots to show the crowd and establish the context for tighter cropped shots taken later. I'll shoot from a tripod and bracket exposure using the clipping warning in the playback as a guide for highlight exposure. Here for example is a SOOC establishing shot with my 50D:
http://super.nova.org/TP/MAG_WideSOOC.jpg
Here is is after a couple minutes of editing in CS5:
http://super.nova.org/TP/MAG_WideEdit.jpg
Here's a closer shot take later - ambient only where I concentrated on getting the foreground exposure correct...
http://super.nova.org/TP/MAG_Dancer.jpg
...again guided during shooting by the clipping warning in the camera on the highlights and knowing I'd be able to pull more detail out of the mid- and 3/4 tones of the RAW file in ACR and CS5. More work that way that trying to get it "right" SOOC, but necessary when SOOC doesn't work due to the limitations of the camera.
With respect to color accuracy....
The problem isn't the stage lighting which is within the range (gamut) of human vision, again the problem is the tools used to record the facsimile image of it. Not so much the camera, which just records relative amounts of RGB in the lighting, but the color gamut of your monitor and printer. I use this shot taken with stage lighting in the background to illustrate the gamut limitations...
http://super.nova.org/TP/FirePowerCU.jpg
... in this tutorial on color management... http://super.nova.org/DPR/ColorManagement/
http://super.nova.org/DPR/ColorManagement/OGW2.jpg
The purple in the background is out of gamut for my printer and the monitor likely wasn't reproducing it as seen by eye either because the monitor gamut is smaller than that of human vision. The color will change and you can't avoid that, but you can manage how the change is perceived in the final output.
As with exposure whenever the color gamut of the original scene exceeds the range of the camera, monitor and printer the colors will change in the reproduction and look different than the impression by eye, both in terms of saturation and purity, and in hue. Again its a photographic fact of life you need to understand and workaround. The way to work around it is to balance the color so the most important content in the scene seems "normal" vs "off-color". "Normal" is a perceptual impression based on overall context, not something you can quantify with the eye dropper tool.
For conventional studio portrait or outdoor shot which appears to be taken at mid-day there is a very narrow range of skin tones that will seem "normal" in a photo. Why? Because we see faces in lighting like that as being very close to R=G=B neutral. But in other situations like in late afternoon, or a warmly lit tungsten-lit room our brains when adjusting color perception expect the faces to looks slightly warmer than the R=G=B "neutral" baseline. The problem photographically is that the camera doesn't adapt color perception in the same way the brain does so what the camera records while accurately recording the light that is there color-wise doesn't wind up looking "normal" or natural on the faces and familiar objects.
In that regard RAW format is a HUGE BLESSING. If at capture the most important content is exposed with normal (i.e. expected by eye) detail the color can be "normalized" by eye based on the background context during PP, as I did with the wide shot of the performance above. Is that a lot of extra work? Yes, but necessary in some cases to to overcome the physical limitations of the recording and output devices .
In terms of priorities in a photo you want the skin tone and costume of the principal actors on the stage to seem "normal" given the context seen in the shot. In a wide shot with a lot of background context less neutral skin tones will seem in context with the surroundings. But the closer you crop the photo and the less background context is seen the more R=G=B normal the viewer will expect the faces to be.
When shooting I never use AWB because it changes the baseline for color shot-to-shot trying to guess and adjust color like the brain adapts, making the brightest highlights neutral. For stage work I keep the camera set to Tungsten WB. That renders actors in the white spotlights as "seen by eye" neutral and records all the other funky stage lighting effects — to the ability the camera sensor range and gamut can — relative to that baseline. That's as close as I know I can get at capture to how my brain is adapting my in-person perception.
Typically some content will look "normal" color-wise, some too warm and some too cool. Those are perceptual judgements from some "looks normal" baseline. Where does that baseline come from? The context seen in the background of the photo. So what I will do is: 1) balance the RAW file in ACR so the most important content looks similar to my impression in person, then; 2) in CS5 make two duplicate layers and apply warm and cool photo filiters to them, then; 3) using masking I will blend the the warm and cool adjusted layers into the "normal" base layer. That's what I did with the wide shot above. I do the adjustment of the exposure in ACR based on the overall tonal range of the flle, but them after adjusting the color to "normalize" it across the stage I used masked screen adjustment layers to pull more detail out of the shadows.
I realize this is a lot to process, so to recap:
1) In each shot decide what is most important and expose for detail in it, realizing detail will be lost in other areas. I shoot in M mode not auto and as mentioned I expose based on the clipping warning in the camera with prevents me from blowing the highlights on the face and clothing of the "star" of whatever shot I'm taking and at the same time tells me what, if anything in the background is clipping. When possible I will bracket exposure, repeating the same sequence of wide , medium, close-up shots of the performer at different exposures.
2) For color I capture from a consistent baseline of tungsten WB which renders spot lit actor skin tones more or less normally. I adjust the WB from that consistent capture baseline perceptually, shot-to-shot in editing, to match the warm/neutral/cool background context of the shot.
3) PP techniques such as blended exposure and color balance layers are normal part of my workflow. I understand the gamut limits of screen and printer and how they differ from human vision. I also understand that perceptually of the more important content in a photo looks "normal" given the overall context of the photo then less accurate color rendering of background and less important content really isn't that critical to the overall effectiveness of the shot. Simply put if the most important face(s) in the shot seem normal to the viewer of the photo not much else really matters. As noted close-ups require more neutral WB than wider shots to seem "normal" in photos so its a case-by-case decision process for me.
It's just a matter of understanding the limitations of the process and how human perception works and finding workarounds to the limitations that work to fool the brain of the viewer.
|
