29
\$\begingroup\$

I have read the existing question - Does the camera white balance setting affect the raw image at all?

After reading the above post, I am still not clear on the part around exposure. Is it possible that selecting one white balance over the other can cause the exposure to be different? This answer from that question in particular brings that point up exactly: https://photo.stackexchange.com/a/3598/4892

As far as I have heard, the WB does not affect the RAW data, but it does affect the exposure.

So in difficult lighting situations the camera auto exposure might react differently depending on the WB.

For an example, what if I always shot in RAW and always shot in Auto WB, or always Incandescent WB. If I am WB correcting in post regardless and ignore any "preview", am I losing any information or altering the exposure by leaving the WB in one of these settings?

I found additional discussion of this topic but no real answer here. I also found someone who recommended instead of Auto WB, to always leave the camera in WB 5000k to keep the most information.

Improve this question
\$\endgroup\$
7
  • 1
    \$\begingroup\$ This is what you are looking for: photo.stackexchange.com/questions/664/… \$\endgroup\$
    – mattdm
    Commented Nov 14, 2012 at 3:14
  • 1
    \$\begingroup\$ White balance affects the histogram, and in ONE respect, that MIGHT affect exposure: Clipping a single color channel due to how the camera meters. If you are photographing a scene with very wide dynamic range that is pushing the limitations of your gear, using an incorrect white balance could do one of two things: force you to underexpose to avoid clipping one of the color channels, or simply result in a clipped channel if you expose for the shadows. Using a white balance setting that actually balances the histogram can give you more exposure latitude...in that case, and that case only... \$\endgroup\$
    – jrista
    Commented Nov 14, 2012 at 4:06
  • 1
    \$\begingroup\$ ...YES, white balance can affect exposure in high dynamic range scenes. \$\endgroup\$
    – jrista
    Commented Nov 14, 2012 at 4:06
  • 2
    \$\begingroup\$ White balance does not directly affect exposure, as exposure is the amount of light coming through the lens (the photons don't sense the WB setting and steer out of the way). White balance may affect metering in some cases. \$\endgroup\$
    – Matt Grum
    Commented Nov 14, 2012 at 10:59
  • 2
    \$\begingroup\$ I guess the better way to term it would be WB can affect exposure settings, thus changing the final exposure when you actually press the shutter button. \$\endgroup\$
    – jrista
    Commented Nov 17, 2012 at 16:26

6 Answers 6

Reset to default
27
+300
\$\begingroup\$

This is actually fairly easy to test, and so I did. I used my Pentax K-7, so this doesn't speak to all cameras, but I think at least many work the same way.

I worked in a dark room, lit only by an iPad app which simply turns the whole screen a certain color. I put the camera close enough to the screen that the color filled the entire frame, and, although I don't think it matters, manually focused. I put the camera in aperture priority mode with the lens at F/2.8 and ISO fixed at 1600. And I set capture to RAW.

With a red light and daylight white balance, the camera selected shutter speed 800. When I set white balance to tungsten (without changing anything else), it instead chose 640. I went back and forth several times to make sure nothing else was influencing the result.

Then I changed the light to blue. Here, with tungsten, the camera picked 1250, but with daylight, it chose 1000.

So, clearly the white balance selected does influence metering decisions made by the camera even in RAW mode.

However, it's also worth noting that even in this contrived, extreme case, the difference only one third of a stop in either case. Therefore, I think you're pretty safe with using Auto WB in RAW. (Which, by the way, happened to give the same exposure as Daylight in my tests.)

Improve this answer
\$\endgroup\$
7
  • 1
    \$\begingroup\$ +1 Thanks for providing your test procedure and result here... \$\endgroup\$
    – woliveirajr
    Commented Nov 16, 2012 at 18:02
  • 5
    \$\begingroup\$ +1 simple test is worth a thousand hours of internet speculation. I'd bet any camera with an RGB metering sensor will do this (otherwise why bother with full colour metering?). Older cameras without colour sensitive metering will not be able to exhibit this behaviour. \$\endgroup\$
    – Matt Grum
    Commented Nov 16, 2012 at 20:04
  • \$\begingroup\$ @MattGrum: If you are indicating that I was only speculating when I offered up my answer, I assure you I was not. I offered my answer based on actual experience (in my case, photographing birds lit by direct sunlight that was sometimes obscured by clouds). I also noted in a comment to my own answer that this does only apply to modern color metering sensors. \$\endgroup\$
    – jrista
    Commented Nov 17, 2012 at 20:28
  • 1
    \$\begingroup\$ @jrista I was talking in the general sense, not referring to any specific answer on this site. I would have upvoted it if there was a reference to said experience with a specific model in the answer. I have not read all of the comments. \$\endgroup\$
    – Matt Grum
    Commented Nov 18, 2012 at 9:24
  • 1
    \$\begingroup\$ @MattGrum You are assuming that the monochromatic light meters are equally sensitive at all wavelengths of light. That is seldom the case. In fact, it turns out that RGB light meters are less affected by color differences than monochromatic light meters are. The mono meters tend to be more sensitive at one end of the spectrum or the other. \$\endgroup\$
    – Michael C
    Commented Jan 17, 2017 at 22:34
6
\$\begingroup\$

Possibly:

White balance affects the histogram, and in ONE respect, that MIGHT affect exposure: Clipping a single color channel due to how the camera meters. If you are photographing a scene with very wide dynamic range that is pushing the limitations of your gear, using an incorrect white balance could do one of two things: force you to underexpose to avoid clipping one of the color channels, or simply result in a clipped channel if you expose for the shadows. Using a white balance setting that actually balances the histogram can give you more exposure latitude...in that case, and that case only...

YES, white balance can affect exposure by resulting in a change to the exposure settings (aperture, shutter, and/or ISO) in high dynamic range scenes, or if you Expose to the Right (ETTR).

Improve this answer
\$\endgroup\$
8
  • \$\begingroup\$ So if this is the case, I would think Auto WB which can do fairly decent these days- would be adequate to at least get within the general range of a correct WB. At least close enough to prevent improper exposure I would think. Do you agree? One person recommended to me to shoot at 5000K every time, as this was somehow "safer" than Auto WB. \$\endgroup\$
    – dpollitt
    Commented Nov 14, 2012 at 4:51
  • \$\begingroup\$ It depends on what your goal is. Auto WB is more likely to maximize your exposure latitude, however it may not produce the best results. AWB can cause slight differences in balance from shot to shot, and if you need white balance consistency, AWB sometimes leads to a nightmare in post. I wouldn't say using 5000K "every time" is right either...you need to use the right white balance for the lighting in your scene. That may be daylight, or shadow, or tungsten, or perhaps a custom WB setting would be best. If you need consistency, though, using a fixed WB setting rather than AWB is better. \$\endgroup\$
    – jrista
    Commented Nov 14, 2012 at 18:03
  • 1
    \$\begingroup\$ I recently ran into this problem about a week ago. I was photographing some bucks around sunset. Before sunset, the sun was sliding behind patchy clouds, and I was using daylight WB. I was at ISO 3200, and I was pretty sure none of the shots would be anything I could use outside of at 750x500 web size...no printing. As the sun actually set, the sky lit up like red fire, but I forgot to change my WB setting. At a 5200K WB, the scene ended up VERY red and slightly magenta, and the meter caused my shots to underexpose enough that none of the post-sunset shots were usable AT ALL. \$\endgroup\$
    – jrista
    Commented Nov 14, 2012 at 18:07
  • \$\begingroup\$ Now, I could have used AWB. That would have taken care of the before- to after- sunset WB problem. But the light was changing fairly frequently before sunset, and in the past when I've used AWB in that situation, trying to adjust my photos so they appeared consistent in post was an extremely tedious job, and one that usually results in failure. Using a fixed "daylight" setting before sunset, and a "shade" setting or custom setting after, usually results in far more consistent results when processing in post. The catch is you have to have the presence of mind to change WB as the light changes. \$\endgroup\$
    – jrista
    Commented Nov 14, 2012 at 18:10
  • \$\begingroup\$ I think people are conflating spot metering with exposure. Use an incident meter and the equation changes, it measures the amount of light falling and not the reflecting capability of the subject under certain temperatures of light. \$\endgroup\$
    – Joanne C
    Commented Nov 16, 2012 at 19:47
5
\$\begingroup\$

It may close as a dupe, but I don't think it is... I don't think the idea that the white balance effects exposure was established at all. I don't really see how it can be, white balance is about the colour temperature of light, not how much of it there is in the scene. In other words, we have an exposure triangle, not a rectangle, so the temperature is not a factor.

At any rate, the original answer is still correct, the raw sensor data is subject to interpretation in post processing without loss of information. The camera white balance setting is information in that exposure, sure, but it's a "hint" to the software and will not result in loss should it be changed.

Improve this answer
\$\endgroup\$
2
  • 1
    \$\begingroup\$ Thank you for the answer. I guess the "theory" that I was reading about subscribes to an idea that the exposure may be adjusted at the time of capture depending upon what WB is selected. I agree about the triangle, but what if the camera decided to push the shutter speed a bit faster because the WB caused the image to be bright? Something like that is kind of the argument I am seeing. \$\endgroup\$
    – dpollitt
    Commented Nov 14, 2012 at 2:48
  • \$\begingroup\$ That's just a bad meter reading I think, more likely a result of the subject than the light, as it just misread the grey because of it. Use an incident meter, which doesn't involve the subject, and the result is very likely a correct exposure that has no consideration for colour temperature. \$\endgroup\$
    – Joanne C
    Commented Nov 14, 2012 at 2:53
2
\$\begingroup\$

In my experience, the Canon bodies that I have used show some variance in automatic metering of incandescent light when the white balance setting is changed. Generally, I leave the white balance to "automatic" and I deal with setting consistent and pleasing WB values after exposure. That said, when set on automatic with my body, I've learned to increase the exposure compensation when shooting where the primary light source is incandescent bulbs.

This is easy enough to find out for yourself. Shoot some photos under varying light sources. Change the white balance. Meter with the camera. Do you notice any difference? I haven't done this particular experiment. I've just learned from my past experience with my cameras and my preferences.

Improve this answer
\$\endgroup\$
1
\$\begingroup\$

I think that this is POTENTIALLY more complex than some are giving credit.

Whether it is ACTUALLY more complex depends on the camera manufacturer.

If colour balance is altered you change the relative energy in each colour channel.
If you meter and make exposure decisions on the processed data then WB setting can affect the result. If all your decision making is carried out on raw sensor data then WB does not affect exposure.

Note that such post WB adjustment metering would not mean that the RAW data had been altered after capture because of the WB adjustment. BUT it would mean that the actual RAW data that was presented to be captured had been altered somewhat as a result of WB processing.

It's a philosophical decision in the hands of the camera maker and could vary between brands and models., although there is probably industry best-practice based on experience of the result. Trying it on various extreme scenes may help to answer the question for a given camera.

Improve this answer
\$\endgroup\$
3
  • \$\begingroup\$ See my comment to the question above. Once the exposure is taken and piped into a RAW file, you're done, and you can tweak WB to taste in post as much as you like. But because metering affects white balance, it is indeed possible for it to affect the "exposure" as you configure it in-camera, especially if you ETTR, or encounter a high dynamic range scene and are forced to reduce exposure to avoid clipping a single channel. \$\endgroup\$
    – jrista
    Commented Nov 14, 2012 at 4:39
  • \$\begingroup\$ @jrista Yes and ... :-). That overlaps with what I said but is not identical. Even if a scene is not of high dynamic range, if the colours are not equally saturated (even if there is lots of "headroom" at the black end) then if the WB is adjusted the system may move the white point up and down. That's where my comment about what the manufacturer does comes in. If they look at the raw sensor data before WB adjusment then WB changes do not affect exposure. If they make exposure decision after WB adjustment then the white level may shift - even if dynamic range is low. . \$\endgroup\$
    – Russell McMahon
    Commented Nov 14, 2012 at 7:57
  • \$\begingroup\$ With DSLRs the exposure decision is made before the sensor collects any raw data. The raw data collected is of course affected by selected exposure. The WB decision in terms of metering is based on how the camera uses the WB setting selected by the user to calibrate/correct the reading from the exposure meter. Before RGB metering became more common, if the WB was set to "Auto" the camera would meter based on a setting of 5200 K. If any other WB setting was selected the camera would adjust the metering based on the known response of the meter to light centered on that particular temperature. \$\endgroup\$
    – Michael C
    Commented Jan 17, 2017 at 22:48
0
\$\begingroup\$

Apart from minor differences in exposure that have been well covered by the other answers, there is at least one other possibility where trying to adjust the raw data captured using a specific color temperature and white balance might be more limited in post-processing than by using a "Custom WB" in-camera at the time the image is taken.

At least with the Canon cameras I use, taking a photo of a white object and using that photo as a basis for Custom WB allows a wider latitude of adjustment than the raw processing tools I typically use (Canon's DPP 4 and LR).

If I need a WB correction of more than "10" (which equates to about 50 mireds of color correction), I'm out of luck from within DPP 4 or LR. The same is true for a color temperature lower than 2000K. But if the light I am shooting under requires such extreme correction (say a night street scene illuminated by several different types of high pressure vapor lights), using the in-camera Custom WB setting at the time I take the photo allows that more extreme correction to be applied when the raw file is later opened in DPP 4.

All of the other post-processing options are still on the table as well, but I can't go back later and take my white object shot and apply that custom WB to an image unless the reference image was already loaded into the camera and Custom WB using that image was selected at the time the raw data was collected.

Improve this answer
\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.