White balance pre-set

<p>Ok guys, Im using a Nikon D700, i usually shoot in RAW using the "standard" or "neutral" pic control.<br>

I have read that WB setting doesnt really matter as it can be corrected in PP.<br>

Im no expert by all means, but surely WB must be of some importance! <br>

So im interested to hear other photographers opinion on how to set your WB ie: card, white wall, another photo etc....<br>

thx in advance,<br>

steven</p>

<p>Nikon D50 -<br>

I shoot raw, and I choose one of the presets - Full Sun, Incandescent, Cloudy, etc. This gets me "close" to what the actual condition is. Then in Adobe Camera Raw, WB can be changed completely, or just fine tuned. It's true that in a RAW process, WB can be easily changed. I shot some pics with WB set to "flash" and then got outside a minute later to take more pictures, but I forgot to reset the WB. Well, the outside pics were "ruined" - until I got home and processed them in ACR, where I simply changed the WB to "full sun" and the pictures were fine. I could be wrong, but the "data" collected in RAW has not been WB encoded. In your LCD preview screen, it is showing a processed JPG, so it might look wrong, but the data is the same no matter what WB is selected.</p>

<p>I set white balance according to the source of light.</p>

<p>The white balance scale we use today is based on a set of observable temperatures. It runs from 1,000 to 10,000K (as a practical matter). When something is on fire at ___Kelvin, what kind of white does it emit?</p>

<ul>

<li>Candles come in at about 2500 Kelvin</li>

<li>Tungsten (regular incandescent lightbulbs) come in at 3200K</li>

<li>Sunset/sunrise twilights (Golden Hour) are about 4500K</li>

<li>Common daylight will be between 5000K and 5600K (color films)</li>

<li>Fluorescent lighting and computer monitors will be 6000K</li>

</ul>

<p>Those would be the white definitions, based on comparable observed temperatures in a laboratory. When we're referring to this problem, we almost always see people referring to the <em>balancing temperature </em>by its apparent observed temperature definition.</p>

<p>[it's just easier that way. There are two halves of the problem, and we have a tendency to pick the half that contains the practical answer to photography. Logically, this is a problem that tackles the question of defining an identity over a process that has some aspects which are proportional and holistic, not just linear and totalitarian.]</p>

<p>The balance of those temperatures is what counters that, so that as the recording is made and reproduced, the white definition (just like its opposite, black point or black definition) comes out white no matter what the color of the original light source was.</p>

<p>This is how we can proceed through the process and consistently reproduce the answer, or edit it in a controlled way.</p>

<p>The original light source is a set of limiting parameters. Any white light can be split into the colors of the rainbow, but the starting point for what's defined as white sets the limits on what those colors of the rainbow look like.</p>

<p><em>Some of what that is and why:</em></p>

<p><em>White definition and white balance are about reproducing white so that it looks white, regardless of the original light source; the goal is to define something that is a logical limit: black and white.</em></p>

<p><em>In color, black and white have logical values similar to 0 or SUM, depending on what aspect of the energy we're concentrating our minds on. Meanwhile, the subordinate colors, (red, yellow and so on) have values that are both observable by empirical wavelength and also proportional to whole encompassing problem of recording and reproducing the colors as they were first seen. </em></p>

<p><em>That is, in nature, what's white and black is set by the actual flow of the energy. When you're making a picture, your capture of that energy and recording and reproducing of that energy will always be mechanically imperfect. One of the harder aspects of those imperfections is in determining what the 0 and the SUM should be. For example, if you record an unknown quantity at an unknown efficiency, then how are you going to hypothetically set a target for what the SUM should be? Things won't add up because of the inefficiency of recording, and also, you would have no way of confirming whether or not you were right because the experiment was unrepeatable. </em></p>

<p><em>There's a way to resolve an unrepeatable instance, but often that requires a heckuva lotta data. That big pile of data would generate some commonalities which could be used as a form of data themselves. </em></p>

<p>So, how to set the definition of white and its balance?</p>

<p>As a practical matter, the easiest way is to know what the apparent white definition of the light source was. Do you know that the light is a candle flame? A light bulb, daylight, fluorescent, and so on? Well, we don't have to prove in a lab what color the energy was. We just want to reproduce or produce a picture as a simple (and sometimes unscientific) matter.</p>

<p>So, use what you know about the apparent thermal definitions of white (3200K, 5600K, 6000 or 6500K) of the light sources to set your white balance.</p>

<p>The truth will be a subtle variation of that, but those guideline temperatures will get you a lot closer. With some basic practice, you ought to be able to look at a photo and get within 500K just by sight.</p>

<p>There's more to white balance than all that, but as a keep it simple, practical method, I use those guidepost temperatures to get the job done. Lightbulbs are 3200K, sunsets are about 4500K, daylight is 5600K at high noon.</p>

<p>FWIW (this may blow your mind, but here it goes), <br /> those tiny sensors in the DSLR are a matrix of very little sensors cut into a chip.</p>

<p>Each one of those tiny sensors <strong>is</strong> <em>a very sophisticated, refined, small, and limited</em><strong> thermometer.</strong></p>

<p>This tiny, limited thermometer answers a very basic question: did energy of _____ frequency get intercepted at this point, yes or no, pass or fail, true or false?</p>

<p>1 or 0?</p>

<p>The thermometers we usually hold in our hands and take something's temperature with are actually much more sophisticated instruments capable of telling us a range of answers. In the case of DSLR sensors and the like, there is actually an advantage to the asking of a very basic question: it can be codified in terms of yes or no much easier, lending itself to digital logics. Instead of telling us a range of applicable energies and describing it or labeling it, the sensor tells us only yes or no this did or did not happen to meet the criteria the very basic, limited, thermometer could tell us.</p>

<p>The answers to those questions are rolled up in some matrix math. The first layer of the matrix math is for grouping four thermometers together (black/white, red, green, blue). The next layers of matrix math have to do with grouping those small matrices together so that they can be handled as a string of numbers. Those numbers in a list become the computer file that's your digital picture as it is made by the camera.</p>

<p>When you are setting changes in post-process, you are actually telling the computer to limit the minimums and maximums (or balance of proportions within a function) on how to do that math.</p>

<p>So, trust me on this, the easiest thing to do is to look at the light source and ask yourself: about which ones of those is it?</p>

<p>Is it like a lightbulb (3200K), a sunset (4500K), daylight (5600K), or fluorescent lighting & computer monitors (6000+ Kelvin)? Use that answer to set your white balances.</p>

<p>Steven,</p>

 

<p>If the goal is accurate color and if the light is consistent across the scene, a gray balance target

will give you the best results. A $15 cardboard gray card such as what Kodak used to make and what

others still make is surprisingly good, and certainly in the “good enough” category for

most people — and it is also very good for setting proper exposure. If you want something a bit

better and cheaper, get a quart of Laura Ashley “Chimney Sweep” LA1616 paint from

Lowe’s and roller it on some cardboard.</p>

 

<p>The WhiBal is a better gray balance target, but not $50 better. And it can’t be used to set

exposure. If you want something better that you can also use for exposure, you’re looking at the

X-Rite ColorChecker Mini Gray Balance for about as much as the WhiBal, or you’re looking at

$1500 for a Spectralon target.</p>

 

<p>If you don’t need to set exposure, the brighter the target the better since the camera renders

it with less noise. The BabelColor Watch Your White target is the gold standard for photographers in

this regards, with 99.5%+ reflectance across the entire visible spectrum. Just make sure you don’t overexpose it! (There are scientific targets

significantly better including ones made from Spectralon. They’re all much more

expensive. They’re also serious mega-overkill for photography. What would almost make the Spectralon worth the price for a photographic gray card is its near-perfect Lambertian reflectance, meaning it’s impervious to glare.)</p>

 

<p>Two dirt-cheap alternatives to the BabelColor target are Tyvek (used in many tear-proof envelopes)

and PTFE (Teflon) thread seal tape. Both have very similar spectral characteristics to the BabelColor

target.</p>

 

<p>But…perhaps better still is a styrofoam coffee cup. It’s not quite as bright, but

it’s much brighter than the WhiBal — and, more importantly, it’s a much, much,

much more neutral color. But what makes it probably your best bet is that it’s <em>not

flat.</em>Put it upside down in your scene, and you get a nice gradated sampling of the color of all the light

sources. If you want to balance for the light coming through the window on the left, click on the left part

of the cup. If you’d rather go for the warmer incandescents on the right, click on the right part of

the cup. Click inbetween to target a blend of the two.</p>

 

<p>Or, if the cup is big enough, stick it over the end of your lens and use that for a custom white

balance both in-camera and in post processing. You’ll get an average of all the light sources

blended together in their actual proportions, resulting in a white balance essentially identical to what

your brain would have seen. (Be sure to increase exposure to get a nice, bright sample.)</p>

 

<p>Some of the worst choices are “white” paper and clothes. Paper is typically yellow

with added fluorescent brighteners to give it a blue tint to compensate. The result is decidedly not-white.</p>

 

<p>Laundry detergents have similar agents to whiten and brighten clothes — and that’s

assuming that the fabric itself is white, which it rarely is. And skin and undergarments often give a tint

to clothing. If you want to make sure the bride’s dress is white, start with a good white balance

from a neutral target; that’ll get her skin and flowers (basically) correct. If the dress isn’t

neutral, selectively desaturate just the dress…assuming that the dress isn’s

<em>supposed</em> to be subtly off-white.</p>

 

<p>Other found objects are also problematic. That “white” wall in the background

probably isn’t, and it’s certainly getting illuminated differently than your subject. The

dinner plate is throwing specular reflections and is completely unreliable. The wall clock is a bad

combination of the wall and the dinner plate — and so on.</p>

 

<p>Lastly, if you don’t use a target (for whatever reason), you can still get excellent white balance in post-processing pretty easily. Just crank up the saturation all the way. If the white balance is off, it’ll look hideous. Adjust the white balance until it just looks over-saturated, not hideous, and then return the saturation to normal. Hey-presto! Perfect white balance.</p>

 

<p>Cheers,</p>

 

<p>b&</p>

<p><strong><a href="http://www.flickr.com/photos/mamk/2854988029/">If your brain feels full,</a></strong> I'll tell you, while I cannot completely escape this problem of colored light energy and its headaches this way, I prefer to relax by using black and white films. With those, I only have to think about spectral energy in bulk for editing shapes, and the volume I record it with is a mass of random reactive particles. This is why I prefer black and white film.</p>

<p>If you are an ultra-left brained multivariable calculus hog, you're gonna love DSLRs. Personally, I am more of a right brained thinker, and therefore prefer films. [in some ways, it's same-same, but the path of the logics to yield similar styles of answers is a little different.]</p>

 

<p>By the way, if you have multiple or competing sources of light, there are two good ways to go: you can either pick the one that's dominant and use that, or you can average them up and pick a spot in between. I usually go with the dominant one.</p>

<p>If the sun is involved, it will almost always be the dominant light source. </p>

<p>John’s last sentence reminds me of a super-important point: outdoors on sunny days,

you’ve got two radically different white balances in your scene.</p>

 

<p>The sun is about 5000K (or thereabouts).</p>

 

<p>The sky, which is illuminating the shady parts, has a color temperature of about 15,000K. (The

actual sky isn’t anywhere near that hot, of course; the same Rayleigh scattering that makes it

blue causes it to have a spectral distribution very similar to an object that temperature.)</p>

 

<p>The two mixed together, in a typical outdoor scene, result in a continuous range of color

temperatures ranging from about 5000K for objects directly lit by the sun to about 5500K for most

things you see to about 7000K for objects completely shaded from the sun and entirely lit by the

sky. (Remember that the sky is very, very dim compared to the sun, so it’s not really the sky that’s lighting your subject so much as the sun bouncing off everything around you.)</p>

 

<p>So, for general outdoor scenes you’d mostly go for 5500K, since that’s a close

approximation of the sense your brain makes of it all. With the sun right at your back, you’d

generally go for a lower temperature. But, if it’s a mostly shady back-lit scene, you’d

probably go for a much higher temperature.</p>

 

<p>Cheers,</p>

 

<p>b&</p>

<p>Hey Guys, wow! thanks very much for taking the time to help out in great detail.<br>

This info will be extremely helpful when shooting from now on.<br>

i never realized there was soo much to it. once again, i cant thank you guys enough, its people like yourselves that make this informative site so worthwhile.<br>

Steven</p>

<p>Steven,<br>

The Xrite Passport ColorChecker is a great investment at $98 from Amazon.com. Visit their website, <a href="http://www.xritephoto.com">www.xritephoto.com</a>, and review their demonstration videos.</p>

<p> </p>

When I started using an expodisc or gray card, whatever is handy at the time, it changed my life. There is, in my opinion,

no substitute for custom white balance and getting it right, or as close to right as you can in camera. I also shoot raw and

adjust if needed in lightroom.