On May 15, 2012, at 6:00 AM, Bruce Walker wrote: > On Tue, May 15, 2012 at 12:34 AM, Larry Colen <[email protected]> wrote: >> I was asking a serious question here. > > Larry, believe it or not, I was answering you seriously. My silly > scenario has been repeated at many studio workshop shoots I've > attended. And it contains the basic truth that you aren't yet > grokking. But I hope I may be close to a breakthrough. :) > > Larry, I'm taking the time to answer you at length because I know you > to be a serious student of photography, like myself. You dig > deep--deeper than most--for understanding. I totally get that because > I spend many hours experimenting with basic principals myself in order > to get better results.
Thank you. > > > Now, perspective: my use of incident metering is confined almost > entirely to the studio environment. I've used my meter outdoors too, > and gotten good results, but the payback isn't as obvious for me so I > don't do it. I'm also using an incident *flash* meter. When I bought a big pile of gear from a widow friend, whose husband had done photography, there was a flash meter in the pile. I will have to experiment with it. > > So, no, I don't advocate it for solving the sunlight-off-silver-car > issues that Doug has. That makes sens. > > > Thought experiment time. Your scene: > > - white seamless background > - black clothing on model with fair skin I meant black background black clothing, white background white clothing. > > Goal: light this scene so detail is reasonably retained in both skin > and black clothing. Some artistic latitude is allowed. > > How to set exposure? After initial light placement, here are the usual ways: > > 1. take a shot, study histo, adjust until highlights are not blown, > don't worry about too-dark areas. > 2. take a shot, study histo, adjust until dark areas are detailed, > don't worry about blown-out areas. > 3. use HDR: combine 1 & 2 and bracket. layer results later and tone-map. > > 1 and 2 are fine for art and many fashion shots. If that's what you're > after, great! We're done here and you and I can go our separate ways. > > 3 is a fine method too and works for landscapes and architectural interiors. > > But! #3 will not work for moving subjects, like models (or racing cars). > > Wait, there's a method #4 ... This is starting to get me what I was asking for. > > 4. roam the scene with an incident meter and measure the light falling > on it. Adjust the light sources by turning them up or down, moving > them, adding a reflector, a scrim, block too-bright sources with > gobos. Keep light ratios in mind: you may want to light one side of > subject twice as brightly as other for some shape definition. (This is > what the lighting guys in TV and film do. Ask them if histograms are > enough.) Now, we are getting to the meat of the matter. An incident light meter will measure the intensity of the light field in the whole space. That is what a light meter will get me that the histogram won't. There is some question as to whether a clever individual could use their camera to obtain similar results, by setting it up and checking the histogram of the whole scene, or photographing grey cards (or better yet, grey spheres) in any place in the scene. Hmm, that gives me an idea for a gadget. A small (8-10") 18% neutral grey umbrella mounted so that it is facing the camera a couple feet away from the lens to get histograms of the light intensity throughout a set. It probably wouldn't actually work better than an expodisc though. > > Basically, you reduce the overall contrast of the scene until all > camera-facing parts of the scene have some light falling on them. You > are attempting to contain the dynamic range to 6-7 stops or so to > match what your sensor can handle. Now, I understand what you are using the meter for, to adjust the light intensity, not to measure the light intensity. > > Then, and this the key thing: you measure the *brightest* light > falling on your scene, often the subject's face. By definition, all > points in your scene will return light the same or less than that. You > point your meter at the light source itself to do this, and you > measure this right at the subject. You cannot trust reflected light. > > If you set your camera manually to the aperture you get from the > meter, you will not have blow-outs (nothing returns light brighter > than the brightest light falling on it unless something lases). And > you will get detail in the darker objects because you took care to get > light on them too and they are within your sensor's range. > > > >>> Definitely, that's one of the big advantages. But there's more. >>> >>> The histogram is fooled entirely by the scene as it's showing you >>> what's reflecting from it. If the scene is a white dress against a >>> white backdrop, or a largely black business suit against a black >>> backdrop, I wish you good luck histogramming that. >> >> I dunno about you, but my K-5 isn't glued to the tripod that's nailed to >> the floor. If I've got a tricky situation I have to histogram, I'll just >> carry the camera over close enough that it pretty much fills the screen, >> take a photo and look at the histogram. > > That still won't help you. It's still the same old reflected light. > > >>> The meter OTOH tells you the correct exposure for the light actually >>> hitting the scene. Put your meter under the model's chin, pop your >>> lights, read off the exposure, set it and you are done. It doesn't >>> matter the clothing or skin colour, the textures, the backdrop, >>> nothin'. You may have issues with hotspots in the scene or areas that >>> are too dark, but that's lighting design. You need to add reflectors >>> to get fill into too dark areas or add gobos to solve hotspots, but >>> that doesn't alter the basic exposure. >> >> I'm confused here. Because if I set up the lights, and I have two scenes. >> One of which has a black backdrop, perhaps cut velvet so there is subtle >> details in the dark, and a model wearing dark clothes, and a hat, with their >> face in the shadow. The other has white cut velvet, and a blond, fair >> skinned model, wearing a white satin dress with white lace and embroidery, >> I'm going to need to expose the scene completely differently, even if the >> incident meter says the same thing with the same lights. > > Nope. One exposure. Once the lights are setup, no matter what falls > under them it's the same single exposure to get the full tonal range > within your sensor's range. > > > New scenario: like the one way above, but now it's a series of models, > one after another. Red dress. Yellow bikini. Black pin-striped suit. > White sundress. Model in well polished chainmail, followed by the black knight wearing black velvet. One with an albedo of .99 the other with an albedo of .1, both with perfect exposure. > > My way: I go click, click, click & click -- thanks ladies! Perfect > exposure for each model. I may tweak the sliders in Lightroom after > but that's all. > > >> I'll want to >> adjust the exposure to get as much detail, and as little noise, on the >> sensor/in the raw file, and then I'll process the the final image to be as >> light or dark as I want the final image to be. > > Are you saying that you're going to stop each model, then take two or > three shots adjusting the exposure? In the words of Trump: You're > fired!! > > In a recent creative shoot I did, we took two models, a MUA and a > hairstylist into a dark basement workroom, bare concrete walls. The > models were 3.5 hours in makeup, hair and wardrobe (custom clothing). > I spent an hour and a half setting up the lights. Between my shooting > partner and I we took about 500 shots. > > I re-measured the exposure after each lighting change. We took test > shots to get the light right, plenty of shadow detail, no out of scene > reflections. Then we shot dozens of shots, but what we're bracketing > is poses and expressions. Not light. > > > So how did I do? Convinced to try borrowing a flash meter yet? :-) Or at least to find mine and give it a try. With instant feedback with the camera, I had never really seen the need, or advantage. Working by intuition, aided by the histogram I have almost always gotten the exposure nailed within a couple of test shots, while setting up the lights. My SOP has pretty much been to take a Scientifically Calculated Wild Assed Guess, then adjust based on the histogram. I surprise myself more often than not when my guess turns out to be pretty much dead nuts on. One thing that I've found is that (especially with the K-5) sometimes I feel like I'm just getting the best that I can with the situation, and when I get the photos into lightroom, they look much better than I expected. The dynamic range of the K-5 covers an unbelievable range of sins. Thank you for the explanation, and answering my question in a manner that my feeble brain can digest. -- Larry Colen [email protected] sent from i4est -- PDML Pentax-Discuss Mail List [email protected] http://pdml.net/mailman/listinfo/pdml_pdml.net to UNSUBSCRIBE from the PDML, please visit the link directly above and follow the directions.
