Luminous Landscape Forum
Raw & Post Processing, Printing => Digital Image Processing => Topic started by: wmchauncey on July 05, 2007, 01:11:16 pm
-
Most of my work, with my rebel xti, is landscape and critters in the woods.
I just ordered a cheapo gray card after watching the expodisc videos on their website. To use the expodisc you must aim your camera back to where your taking the image to set a custom white balance within your camera.
Now that seems fine in a more controlled milieu than where I usually work. Did I make the correct decision? What do you guys use for color correction?
Thanks a heap!
-
IF you're shooting raw, you'd want to WB on white, not gray. The 2nd patch of the Macbeth Color Checker is in the ballpark. There are all kinds of white cards out there.
This product just came to my attention, looks interesting. Have one coming for testing:
http://www.colorbalancecoach.com/ (http://www.colorbalancecoach.com/)
-
Most of my work, with my rebel xti, is landscape and critters in the woods.
I just ordered a cheapo gray card after watching the expodisc videos on their website. To use the expodisc you must aim your camera back to where your taking the image to set a custom white balance within your camera.
Now that seems fine in a more controlled milieu than where I usually work. Did I make the correct decision? What do you guys use for color correction?
Thanks a heap!
[a href=\"index.php?act=findpost&pid=126619\"][{POST_SNAPBACK}][/a]
I just noticed that in the latest video journal, Michael is using an Expodisc. He puts it over the lens and shoots; toward the shooting direction. I use mine the same way and it works.
Steve
-
ExpoDisc for complex lighting. Sometimes WhiBal for its convenience.
In the studio I've been using PhotoVision's target for almost four years. It helps with WB and exposure.
Click here. (http://www.bhphotovideo.com/c/product/429983-REG/Photovision_DT24011_24_One_Shot_Digital.html)
-
digitaldog-forgive me for being a neophyte but if I read you correctly, I've been doing this wrong.
In ACR on RAW images, I put the white eyedropper on a gray card, as I was told by somebody. If I crop in on the gray card and set the dropper I will get a white spike in the middle of the histogram.
If I'm doing it wrong, what's the correct way?
-
Its not what you're doing, its probably the tone of the gray card you have. If you got something like the Kodak gray cards its too dark a tone for setting WB in a DSLR properly.
Take a look at this website and compare the shade of gray you have against this.
WhiBal (http://www.rawworkflow.com/products/whibal/index.html)
-
The purpose of setting your WB is to measure the light balance at that moment and to correct it to neutral. This can only be done by measuring neutral tones, since that’s the only ‘color’ with a fixed mix of Red, Green and Blue AND are equal. In theory, you can use any neutral ‘color’ to set your WB.
If you measure RGB 128, 128, 150, you can tell your software/RAW converter to correct the blue color cast.
Both the Expodisk and the gray-cards have their strengths and limitations. For me in the studio, I use gray-cards, since they are the easiest to use. In the field, I sometimes use a gray-card in the scenery to make it possible to correct the WB at home on my RAW file.
Andrew Rodney recommends to use a white-card, but I think he made a little mistake here. This is why:
Your camera is not the most accurate device to measure light. It’s sensitivity decreases when the light is lower. So, measuring a whitecard is the best solution isn’t it? No, when you use a whitecard, you can clip channels without you knowing.
For example, lets say the light has a mix of colors, reading RGB 250, 253, 249. These can be corrected easily. Now you measure RGB 250, 255, 249, well this can be corrected also. In the next example, you see the problem: RGB 250, 260, 249. There is more blue, but it’s measured the same as the previous, namely RGB 250, 255, 249. This will give you an incorrect correction, ignoring the larger amount of blue in the light.
When you want to set WB, you use a graycard (or Expodisk). The standard graycards are bright enough for accurate measuring. The blackcards and whitecards are used for setting your black and whitepoints if you want (and can) preserve detail in the extremes. With digital, normally you want to avoid clipping in the whites (use RGB histogram on camera for this).
-
Andrew Rodney recommends to use a white-card, but I think he made a little mistake here.
[a href=\"index.php?act=findpost&pid=128392\"][{POST_SNAPBACK}][/a]
Actually, he recommended the second patch of the Color Checker, which has an optical density of 0.23 or a reflectance of 59%. White paper has a reflectance of about 90%. White here is relative, but light gray would have been a better description.
When you want to set WB, you use a graycard (or Expodisk). The standard graycards are bright enough for accurate measuring. The blackcards and whitecards are used for setting your black and whitepoints if you want (and can) preserve detail in the extremes. With digital, normally you want to avoid clipping in the whites (use RGB histogram on camera for this).
[a href=\"index.php?act=findpost&pid=128392\"][{POST_SNAPBACK}][/a]
Cards expressly made for setting white balance, such as the WhiBal, are somewhat brighter than 18% reflectance (optical density 0.74), and are light gray. In the Lab color space, the WhiBal has an L* value of about 0.75, whereas the Kodak 18% card has an L* value of 0.5. The higher reflectance of the WhiBal will give a more accurate reading than a gray card and there should be no risk of blown channels with reasonably accurate exposure. The WhiBal is in between squares 2 and 3 in the color checker in terms of reflectance, and I would estimate the reflectance at about 50%. Adobe Camera Raw will not allow use the white patch (left most neutral patch) of a normally exposed Color Checker to be used for white balance.
The Kodak gray card was originally intended to aid in determining exposure, not white balance, and it may not be spectrally neutral. That said, I have found that it usually gives reasonable white balance. Some photographers report good results from the top of a Pringles can or even with Kleenex. However, for critical work it is best to use the proper target for white balance.
-
Ok, that makes sence.
I use the graycards from Dynatech. When accurately exposed I measure RGB values around 50% (in Lightroom).
-
I wouldn't use gray cards unless they're specifically made for white balance. Gray cards are typically used for determining exposure when using reflective metering - but the gray not be truly color neutral, or may shift under certain lights, or change as it ages...
the standard is a gretag color checker. the paint used is specially formulated and guaranteed to be the precisely the same from batch to batch, year to year, and the tones on it are used by every camera and software manufacturer as the definitive standard they use for design and testing.
-
So, measuring a whitecard is the best solution isn’t it? No, when you use a whitecard, you can clip channels without you knowing.
I agree with this, but still prefer to use a stepped white balance with white, gray & black. I'm currently using older WhiteBal kits with the four steps, and am looking to buy the ColorChecker Gray Scale Card from Xrite. The stepped scales give me a better idea of the overall color situation than just a single swatch of gray.
-
Expo Disc all the way. I've been using it for three years now and have never looked back. I shoot only RAW and have had more consistent color balance, especially in tricky conditions, that I ever did with a Macbeth Color Checker, a white card or a gray card. It's accurate in about 95% of the lighting situations in which I use it.
-
For me, it depends on the application. White card all the way in our studio for custom white balances, but handheld incident meters for exposure.
On location with multiple light sources, it's back to the incident meter for better decisions about light balance and adjustment.
Fast moving action? I'll take Nikon's onboard color matrix metering (D2X) over any other body brand or metering method. It's simply too reliable and consistent to ignore.
So..... I guess the answer to your question is "neither."
I disposed of my gray cards the moment I bought my first incident meter. I've been given three expodiscs over the years by friends and clients. I gave them each (I think they may be successive generations due to differences) an honest try, but for my needs they all are useless.
-
I have always been suspect about how well the expodisk could work considering how the color content of incedent light can change. So I did a simple test. In a large number of scenes from bright sunlight to full shade I shot exposures with an expodisk and then shot a frame with a color checker. Then I processed the color checker images using the expodiusk gray balance. And no surprise to me I found what I expected to find. None of the colorchecker gray patches were gray and they were off in different directions depending on the scene contents. I'm open to any suggestions about how I might have gone wrong here. But this brings up an interesting point. Apparently an accurate gray balance, one based on a known gray source, is not as important to many as we might think or at least it is not apparent in the imagery and therefore the absolute neutrality of some gray cards is less important than we might think. Or it's just a matter of whether you want your amplifier to go to 11 or not!
As for gray verses white balance the world would be a better place if the words "white balance" passed from usage. As mentioned above in variuous forms, light gray is by far the best. You can be sure you're not clipping, there is less noise in light grays to corrupt the measurement it is the place where you will notice a slight cast first and there are numerous products that fit the bill.
-
It’s very important to get a WB to the same temperature if you want to match different pictures shot under different color temperatures, for example a publication on the same page in a magazine. With a grey card or better ‘white balance card’, I get good matching pictures. Maybe it can be done with the Expodisk, but I never tried it, because it’s expensive and it’s very easy to click the grey card in the RAW converter..
-
personally, I have a big problem with the current WB practice: photograph a white card and use it as a reference when processing the RAW image.
this technique assumes that the white card appears white. in most settings, it won't appear white. its colour will be affected by everything around it (assuming out-of-studio settings).
my (theoretical) solution was to print a booklet of white in variations involving the four colour temperature dimensions. admittedly, a lot of variations and cards would be needed. in my photo setting, I would open the booklet and find the card which appeared to me to be as close to white as possible. I would then photograph this card in the setting and use it to colour balance my image. this method would preserve all the natural colour affects of the environment and simultaneously adjust for colour-temperature affects.
it'd be nice if someone has an electronic method of this technique.
regards,
Gregory
-
I'm no doubt preaching to at least some of the choir, but we use the eyedropper tool in PS to check white balance on a file, then use that info to adjust small variations between shots to be used in the same spread.
For a quick and dirty repair of WB in a shot go to Levels in PS, select the white sampler (eyedropper), and click on what should be white in your image. It will startle you how often your eyes lie to you about what is white and what isn't.
-
this technique assumes that the white card appears white. in most settings, it won't appear white. its colour will be affected by everything around it (assuming out-of-studio settings).
Well really good white balance cards are spectrally neutral like the BableColor white tile I'm using.
http://www.babelcolor.com/main_level/White_Target.htm (http://www.babelcolor.com/main_level/White_Target.htm)
I measured mine with an EyeOne Pro Spectrophotometer and the LAB values are pretty close to prefect/ideal. This pup is both neutral and has an incredibly high L star value (99.7)
-
Nearly identical L*a*b* or RGB coordinates are obtained for a wide array of white illuminants (A, C, D50, D55, D65, F3, F11, etc.); this target thus exhibits almost no metamerism effect, i.e. no change in perceived white (Note: the exact technical term for a change in a single perceived color with various illuminants is Color Inconstancy, which is very low in this case).
does this mean that the target's white appearance is not affected by the surrounding environment or local light? that would be incredibly useful.
-
does this mean that the target's white appearance is not affected by the surrounding environment or local light? that would be incredibly useful.
[a href=\"index.php?act=findpost&pid=131517\"][{POST_SNAPBACK}][/a]
It can be affected of course. Not the case when measuring the tile using a Spectrophotometer however.
-
White balance is actually a fairly complicated topic. The fact that accomplished photographers using extremely advanced digital single lens reflex cameras continue to have white balance issues speaks to the fact that white balance is not simple, and a quick and reliable solution has not yet been discovered.
White balance is a process of measuring color temperature, and applying a correction to the data that comes out of the image sensor to remove color casts. In order to do this, the camera needs to be provided with an external reference, an internal reference, or it needs to use an algorithm to search the actual image data for a neutral reference.
The automatic white balance AWB feature on simple and sophisticated cameras is a process where the camera uses an algorithm to process the actual image data and extract a neutral reference or extract the color temperature. There are a number of different implementations of the automatic white balance feature. The quality of the automatic white balance depends on the processing power of the circuitry in the camera and on the underlying assumptions that are built into the algorithm.
The Expo disc White balance tool works on what is called the "grey world" algorithm. This algorithm assumes that if you average all of the light in any given image, it will balance out to a neutral point. The Expo disc has an acrylic layer built in that scrambles, and diffuses the image, and in essence averages the light in the image. Provided that the scene you are pointing the Expo disc at does indeed have an overall neutral color temperature.
Obviously, there are going to be scenes that will not balance out to neutral, that do not conform to the " gray world" paradigm.
Using a gray target or digital gray card is a better method because it is a direct measurement of color temperature, rather than an assumption or approximation. The digital gray card (httP://www.digitalimageflow.com - my company's product) acts as a color temperature mirror. It simply reflects the incident or illuminant light back into the cameras sensor, where the cameras electronics or the postprocessing software's algorithms can use it as a reference. Since in using a gray card you are in fact performing an effective measurement of color temperature, you will get a more accurate white balance.
Naturally, if you use a gray target that is not spectrally neutral, it will not perform as a faithful color temperature mirror, but instead it will add or subtract or somehow alter the reflected light, giving you an abnormal color balance. In a related way, if you do not place the great card where it will be illuminated by the light illuminating your subject, you will not achieve correct color balance.
Sorry about the long winded post , as I mentioned, this is actually a fairly complicated topic.
-
And lets not forget that defining the color of white using correlated color temperature (the K scale) is mildly inaccurate even with the best targets. I'd love to see the day when I have a tiny embedded Spectrophotometer in the camera that would measure the illuminant and write the necessary information into EXIF data for the Raw.
Photographers have been taught for years that tungsten film has a color temperature of 3400K (Kelvin; a unit of temperature). Lower Kelvin values appear more red, and as the Kelvin values get higher, the color becomes more blue. In actuality, a color temperature is a range of colors correlated to the temperature of a theoretical object known as a blackbody radiator. The blackbody reflects no light and emits energy in shorter wavelengths as it is being heated. Imagine a black cast iron pan on a very hot stove. As it is heated, it begins to glow dark red. As the temperature increases, shorter wavelengths are emitted, causing the color of the light emitted by the skillet to appear orange, then yellow-white, then blue-white. The tungsten filament of a light bulb behaves similarly to the blackbody and radiates energy in the form of light because its temperature is so great (around 3200K).It's somewhat dangerous to use color temperature to define what you want because the reality is if all light sources were true blackbodies a particular color temperature would produce the same color of light. Because natural materials are not theoretical blackbodies, heating them to a specific temperature creates deviates from the theoretical color from magenta to green. It's really much safer to use the term correlated color temperature (CCT) because many colors of white may correlate to the same blackbody color temperature. Different illuminants can have the same correlated color temperature.
This is one reason why the CIE defined the Standard Illuminants. These illuminants are defined spectrally meaning a certain amount of energy at each wavelength across the spectrum. This is an exact and non-ambiguous description of color. D65 is an exact color, it is not a range of colors. If you have a color meter that reports color temperature of a light source many light sources that appear different could read the same, that's kind of a problem!
-
Dear hdegroot,
what happens to the White Balance adjustment if for example the grey card is placed in a scene lit by a yellowish light? the way I see most people applying the dropper, the yellow hue would be lost producing an image not like the original scene.
regards,
Gregory
-
Dear hdegroot,
what happens to the White Balance adjustment if for example the grey card is placed in a scene lit by a yellowish light? the way I see most people applying the dropper, the yellow hue would be lost producing an image not like the original scene.
regards,
Gregory
[a href=\"index.php?act=findpost&pid=132346\"][{POST_SNAPBACK}][/a]
Exactamundo !
To me it seems sensible to calibrate your camera either using Grey or Macbeth cards (dependant on your botheredness level) in neutral conditions - grey day car park - sunny day car park - studio with nothing much in it
And then paste on those settings in RAW to images as a starting point for working the raw
This is of course dependant on the look you are going for - I like colour casts they are part of the environment - grass is green and trees cast green shadows
If however you are trying to accurately photograph a white product under tungsten light in an orange room a different approach will of course be needed
S
-
What about the cheapest one?
The qpcard!!!
http://www.qpcard.se/BizPart.aspx?tabId=28 (http://www.qpcard.se/BizPart.aspx?tabId=28)
-
Dear hdegroot,
what happens to the White Balance adjustment if for example the grey card is placed in a scene lit by a yellowish light? the way I see most people applying the dropper, the yellow hue would be lost producing an image not like the original scene.
regards,
Gregory
Exactamundo !
To me it seems sensible to calibrate your camera either using Grey or Macbeth cards (dependant on your botheredness level) in neutral conditions - grey day car park - sunny day car park - studio with nothing much in it
And then paste on those settings in RAW to images as a starting point for working the raw
This is of course dependant on the look you are going for - I like colour casts they are part of the environment - grass is green and trees cast green shadows
If however you are trying to accurately photograph a white product under tungsten light in an orange room a different approach will of course be needed
S
The answer to the first question is: A grey card (http://www.digitalimageflow.com) placed in a scene illuminated by light with a yellowish color temperature will reflect that light faithfully (without any alteration) into the camera, so that the image of the grey card will effectively encode the color temperature of the scene. When the image is processed, setting the color temperature to the reading off the grey card will correct the scene so that it appears as it did when it was photographed.
For the second question: I'm not sure what is being recommended there. It seems that the writer recommends taking three calibration readings in various "neutral" scenes. I imagine that would be three readings from a grey card. He then recommends you set put of those into your RAW converter as a starting point. Maybe I'm missing the point. The way I see it, if you need highly accurate color, why not do a grey card shot in the scene you are shooting, and then use that? Shooting a grey card image is as easy as giving your model the grey card to hold at the beginning and the end of each shot series (if you change your lighting between series). Why bother to shoot "neutral" scenes?
The point about color casts is well put - sometimes color casts are desirable, and using a grey card may reduce or eliminate them. For example, indoor shots with tungsten lights shot on film and developed at the drugstore typically look a bit yellow. A perfect way to add an "indoor snapshot" mood to a creative piece would be to try for a bit of a yellow cast. Color encodes meaning, says John Paul Caponigro, so when I'm shooting a scene, I try to make sure the color that means something to me stays where it belongs. Trying to neutralize every color, "correct" all colors, may lead to loss of valuable information.