white balance

[Pallycow]

so...I'm trying to learn to do custom white balance and also edit/check in acr7. I'm not fairing very well using auto and presets on camera and I know the right thing to do is set it myself, so I am practicing it.

I seem to have trouble "seeing" it, so I am relying on numbers and histograms to train my eye to "see" it.

I notice when I set custom white balance using a gray card, it seems to give it about two points extra on the blue side. When I pull up raw image...I want to know how to check it in there, and/or fix it.

I tried adjusting various things, I think I get it, look away, come back..it seems off. I also use the wb tool in acr and I think it gets it close.

Is there a way to check it? a numeric way?

anyone have any good reads on wb and how to set and adjust effictively?

thanks.

 

[Pallycow]

oh, I'm shooting with a 7d

 

[bratkinson]

This past March, I finally figured out I needed to learn about WB rather than doing tons of highly inconsistent color correcting in post.

So I started out by getting a grey card. It's easy as pie to use. Take a 'fairly tight' picture of the grey card being held or sitting in 'the light' you'll be shooting in. Then, in the Menu for your camera, go to Set White Balance, select 'custom' and then scan backwards through the pix to the shot of the grey card, and select that. Done. All subsequent shots will have that WB set. But if your camera is faced in more than one direction, the lighting to the left will likely be different than the lighting to the right, or behind you, etc. So it's necessary to get a grey card shot in each of the situations, and set custom WB to the correct one before taking (more) pictures in that lighting.

ExpoDisk works the same way. Hold it up to, or click it into the front of your lens, turn off the AF, take a shot while aimed at the location where the subject will be, set the custom WB to that 'grey' imageless picture, and shoot away. I found this to be very effective where I was taking pictures at a small party in a restaurant banquet room with solid windows along one wall, dark walls on 2 sides, and windows to the adjoining banquet room, in use by another group. By taking 4 Expodisk shots 'up front', I simply selected the correct WB before shooting in each direction.

Those two methods work quite well for in-camera WB. As you, me, and countless others have discovered, the pre-determined WB settings in the camera are somewhere between barely close enough and downright useless. There's nothing preset that handles an incandescent lamp here, and flourescent lamp (many temperatures!) over there, etc. But for sunny or cloudy day shots, and maybe even night shots, the presets are fairly good.

A better solution that I and the bulk of the experienced photographers use is to set custom WB in post processing. To get the best results, the RAW images must be used, as JPGs have already had much of the color information altered/reduced/compressed/lost. Using grey card shots for each situation as indicated above, using the WB tool (I typically use Lightroom), click on the grey card shot desired, and the grey card itself with the pointer, and set the WB to that. Then highlight all the pix taken that will use that WB setting, and with one click (I forget what the button is labelled), it will set ALL those pictures to that WB. I'll usually start with only the first picture in that series, maybe tweek the color temperature a smidge and other settings as appropriate, THEN select all the pictures to be handled the same, and do it with a single click. In a jam, if you forgot to take a grey card shot of some angle, find someone in a white shirt, a white table cloth, or something else white and select that with the WB pointer (I've had to do this too often...). You'll end up pretty close to the desired WB. Then simply weak the temperature slider until you get the 'right' WB.

Method number three is an X-Rite Color Checker. Shooting everything in RAW, and taking a shot of the Color Checker 'up front' (or later on, if you wish) and processing all the desired pictures through their custom add-on to Lightroom (and other products) gets ALL the colors even better, if not absolutely dead-on accurate. It even allows photos shot by different cameras, even mixed brands, to have the identical color balancing. It's a bit of screwing around to get it initially installed for each camera, but I think it produces the best results.

I have and use all 3 tools for setting WB. Depending on the situation, my 'hurriedness', and my remembering to take at least ONE of them along with me (getting old really sucks!), I'm quite satisfied with all the results.

And as far as setting a custom WB to some Kelvin temperature...unless you are in a laboratory or in your own studio, it's near impossible to get right. And for inquiring minds that want to know, 1 degree Farenheit = -457.87 degrees Kelvin. Then the degrees go up and down in parallel with each other.

 

[Pallycow]

Thank you for taking the time to write all of that. Very helpful.

I don't think I was zooming in enough on my gray card, which was accounting for the extra blue. I'm done practicing tonight, so tomorrow I will def do that and see the better results.

I practiced self portraits tonight, and while I got closer than I usually do, out of the dozen or so keepers...I'm pretty sure I have a dozen different wb's from doing each one individually. lol

In lightroom, I am still learning it as I mostly used cs6, I noticed the wb tool you spoke of. Stupid me deleted the one with the gray card in it, and it didn't all click until you wrote it there. When I used the tool my thought was "now hell the hell am I supposed to find 18% gray in this image somewhere. lolol. Had I imported the image with the card...then it would have been quite easy. lol.

I do love lightroom for workflow...so I am glad I am finally forcing myself to use it.

Thanks again for the post and I look forward to practicing again tomorrow given new knowledge.

 

[KmH]

In lightroom, I am still learning it as I mostly used cs6, I noticed the wb tool you spoke of. Stupid me deleted the one with the gray card in it, and it didn't all click until you wrote it there. When I used the tool my thought was "now hell the hell am I supposed to find 18% gray in this image somewhere.

Using a release of Photoshop Elements or CS one can use layers, blending modes, and a Levels adjustment layer to identify parts of a photo that are neutral gray, if there are any, so they can be used to set white balance. There is now an after market layers plug-in available for Lightroom. I have never used it so I don't know how capable it is.
It doesn't have to be 18% gray, it just has to have = values in each of the RGB color channels. - http://www.bythom.com/graycards.htm

Lightroom has a ton of benefits for photographers, but it also has some serious limitations for photographers too.
Adobe designed Lightroom to be a supplement to Elements or CS, not a replacement for them. Lightroom's main reason for being is image database management. Raw conversion and editing is intended as a secondary or tertiary function. while Lightroom can do a lot of the image editing photographers need to do, it can't do all the editing photographers need to do.

 

[Helen B]

And as far as setting a custom WB to some Kelvin temperature...unless you are in a laboratory or in your own studio, it's near impossible to get right. And for inquiring minds that want to know, 1 degree Farenheit = -457.87 degrees Kelvin. Then the degrees go up and down in parallel with each other.

Very minor point on an excellent and helpful post: the unit is called a kelvin, not a Kelvin or a degree Kelvin. You have the conversion the wrong way round (0 K = - 459.67 degF and a change of 1 K is a change of 1.8 degF) - not least because negative kelvin temperatures have a meaning that is way beyond the beginner level, and they are purely theoretical. Nobody has reached zero kelvin yet. I offer this on the grounds that if you think something is worth mentioning, you may as well get it right.

There are plenty of good reasons why using a CT or CCT value doesn't work well. Either or both of the camera and measured values may be wrong, CT doesn't take magenta-green balance into account ... and more.

 

[Ysarex]

In lightroom, I am still learning it as I mostly used cs6, I noticed the wb tool you spoke of. Stupid me deleted the one with the gray card in it, and it didn't all click until you wrote it there. When I used the tool my thought was "now hell the hell am I supposed to find 18% gray in this image somewhere.

Using a release of Photoshop Elements or CS one can use layers, blending modes, and a Levels adjustment layer to identify parts of a photo that are neutral gray, if there are any, so they can be used to set white balance....

I don't think so. If you're referring to this method: An Easy Way To Find Neutral Gray In A Photo With Photoshop -- that doesn't work, but it is one of the bigger misdirection fallacies available on the Internet.

Joe

 

[unpopular]

... going back to the OP. When you set the grey point using a known reference (greycard, etc) are the values of the greycard actually grey - i.e. 128,128,128? It is more likely that your monitor is pitched blue than your image.

And as far as setting a custom WB to some Kelvin temperature...unless you are in a laboratory or in your own studio, it's near impossible to get right. And for inquiring minds that want to know, 1 degree Farenheit = -457.87 degrees Kelvin. Then the degrees go up and down in parallel with each other.

This is why we have color meters, and it would be extremely easy to modify mySpectral Arduino Spectrophotometer: Spectruino to be a precise, handheld meter by implementing Wiens Displacement Law.

 

[KmH]

In lightroom, I am still learning it as I mostly used cs6, I noticed the wb tool you spoke of. Stupid me deleted the one with the gray card in it, and it didn't all click until you wrote it there. When I used the tool my thought was "now hell the hell am I supposed to find 18% gray in this image somewhere.

Using a release of Photoshop Elements or CS one can use layers, blending modes, and a Levels adjustment layer to identify parts of a photo that are neutral gray, if there are any, so they can be used to set white balance....

I don't think so. If you're referring to this method: An Easy Way To Find Neutral Gray In A Photo With Photoshop -- that doesn't work, but it is one of the bigger misdirection fallacies available on the Internet.

Joe

I hadn't seen that link. I based my comment on books written by acknowledged Photoshop experts, like Scott Kelby and Dave Cross.

 

[Ysarex]

Using a release of Photoshop Elements or CS one can use layers, blending modes, and a Levels adjustment layer to identify parts of a photo that are neutral gray, if there are any, so they can be used to set white balance....

I don't think so. If you're referring to this method: An Easy Way To Find Neutral Gray In A Photo With Photoshop -- that doesn't work, but it is one of the bigger misdirection fallacies available on the Internet.

Joe

I hadn't seen that link. I based my comment on books written by acknowledged Photoshop experts, like Scott Kelby and Dave Cross.

Yep, Kelby I believe has published that same basic method using a Difference blend with a 50% gray layer but, I'm afraid that's not enough to make wrong into right.

Joe

P.S. As for that link, you can find a hundred of the same all over the internet -- some of them do acknowledge Kelby as a source for the error.

 

[unpopular]

even if it did work, this method seems like a lot of work when all you'd need to do is select by color using 50% grey, and adjust feather. I can think of a zillion methods that would determine degrees of middle grey within an image, but none as easy as this.

 

[bratkinson]

Thank you, Helen, for the correction. I was trying to remember my college physics stuff from 45 years ago, and simply snapped up the number from a quick Google search. I should have done the 'does this make sense' test, knowing that zero on the Kelvin scale as well as the Rankin scale is absolute zero, at which temperature all molecular activity theoretically stops. But then, if that could be reached, the thermometer wouldn't be working, either.

 

[Buckster]

I don't think so. If you're referring to this method: An Easy Way To Find Neutral Gray In A Photo With Photoshop -- that doesn't work, but it is one of the bigger misdirection fallacies available on the Internet.

Joe

I hadn't seen that link. I based my comment on books written by acknowledged Photoshop experts, like Scott Kelby and Dave Cross.

Yep, Kelby I believe has published that same basic method using a Difference blend with a 50% gray layer but, I'm afraid that's not enough to make wrong into right.

Joe

P.S. As for that link, you can find a hundred of the same all over the internet -- some of them do acknowledge Kelby as a source for the error.

Just reiterating the point - it's a confirmed kill. The method can be used to find the middle gray value of any image, but that middle gray value it finds may have supposed to have been warmer or cooler originally, not the 50% gray it is in the image that came out of the camera, which is the whole problem in a nutshell. So it's worthless to set white balance by it.

 

[Ysarex]

I hadn't seen that link. I based my comment on books written by acknowledged Photoshop experts, like Scott Kelby and Dave Cross.

Yep, Kelby I believe has published that same basic method using a Difference blend with a 50% gray layer but, I'm afraid that's not enough to make wrong into right.

Joe

P.S. As for that link, you can find a hundred of the same all over the internet -- some of them do acknowledge Kelby as a source for the error.

Just reiterating the point - it's a confirmed kill. The method can be used to find the middle gray value of any image, but that middle gray value it finds may have supposed to have been warmer or cooler originally, not the 50% gray it is in the image that came out of the camera, which is the whole problem in a nutshell. So it's worthless to set white balance by it.

Thank you Buckster, that is precisely correct. You only have to think about it for a minute. How could any software or software method help you determine that a color you have that isn't neutral gray should have been?

Unfortunately we're not going to win this one. It's like the 1/3 -- 2/3 DOF distribution and so many other photo myths that can't be killed. In Keith's defense Kelby is generally a pretty good source. The problem is for every minor slip up by a Scott Kelby there's a couple dozen Brian Petersens out there repeating the mistakes as in a game of wack-a-mole. Next thing you know they end up on You-tube a bazillion times. You keep wackin' but you never win.

Joe

 

[unpopular]

Thank you Buckster, that is precisely correct. You only have to think about it for a minute. How could any software or software method help you determine that a color you have that isn't neutral gray should have been?

Using color select feather you could determine a color within a given threshold of middle grey. This could help determine what region should be middle grey by finding the region closest to middle grey.

However, in a good number fo scenes, there is no middle grey reference to find.