GardenersHelper
In Memoriam
- Messages
- 6,343
- Name
- Nick
- Edit My Images
- Yes
I am new to using raw. Processing my first set of raw test images proved to me that raw white balance manipulation is a powerful tool, and that Lightroom (as no doubt other raw editors too) makes it easy to change the white balance. However, it very quickly became obvious that although it was easy to change the white balance, getting it right was another matter altogether. So, I decided to do the sensible thing and use a grey card.
The procedure, as I read it, was straightforward. Include a grey card in a “calibration” image, using the same lighting as for the normal captures, and then in Lightroom, Photoshop or whatever take the colour values from the grey area of the calibration image to set up the white balance on the normal captures. And the colours would then be correct. Easy.
What I did not read, but after some confusion and head-scratching discovered, was that if you take multiple readings from the grey area of a single calibration image you will get different values of temperature and tint. This and what follows is for my G3, but I would be surprised if it was not true for other cameras.
The variations from one measurement to another (on the same image) are quite small at low ISOs, and at base ISO only make a subtle difference to the appearance of an image. However, as the ISO increases the differences get much larger and can have a huge impact.
As an experiment I captured four images, at (base) ISO 160, ISO 400, ISO 1600 and ISO 6400, of a grey card on a table with light coming through a roof light window just above the table. As there was 100% cloud cover the light almost certainly did not change during the few seconds while the four shots were taken.
The following graph shows the temperature and tint for 25 measurements taken from each of the four images. As you can see, the spread of temperature and tint values increases significantly as the ISO increases.

Spread of White Balance measurements for four ISOs by gardenersassistant, on Flickr
For those who prefer numbers, here are some statistics for the measurements.

White balance sample measurements by gardenersassistant, on Flickr
When I applied samples of the settings to several images it seemed to me that although the base ISO values did vary, the impact on the look of the image was probably less than would matter to most people. At the other end of the spectrum, using settings from the ISO 6400 calibration image, there were huge colour shifts. This composite image superimposes two of the white balance settings measured from the ISO 6400 calibration image.

Image with two White Balance settings taken from the same calibration image by gardenersassistant, on Flickr
Here is what the grey card on the table looked like with the same two settings.

Grey card with two white balance adjustments by gardenersassistant, on Flickr
------------------------------------------
The other thing I discovered from this exercise was that even for base ISO, where there is not much variability about the measured white balance, using this white balance does not guarantee that the colours look like what you saw when capturing the image. For example, this pair of images is of the same flower. (Yes, the left hand one is underexposed, but I have deliberately only changed the white balance to avoid unwanted complications. As it happens though, correcting the exposure does not remove the redness, which is the issue of concern here.)

Same flower, different time of day, both grey card corrected by gardenersassistant, on Flickr
When I took both pictures the colour I saw was similar to the right hand version. I pass by that flower very often, and it always has an “orangy” aspect, and never red, in any light. However, both those images have been adjusted to the white balance taken from a calibration image taken at the time the image was captured.
Interestingly enough, using JPEG with my current camera, and with my two previous cameras, this flower tended to come out red like in the left hand image, so this is a familiar issue. It was not cured by the white-balance calibration of the left hand image. It was corrected for the right hand image. Very curiously, the more appropriately coloured version is ISO 3200, with its undependable white balance calibration, while the wrongly red one is base ISO, 160, with its much more consistent white balance calibration.
Another colour that is sometimes not brought into the realm of “reality as I saw it” by white balance correction is green, sometimes. With JPEG greens sometimes took on a totally unnatural bright emerald green character, and this is still happening with some white balance corrected raw images.
The good news is that it turns out that Lightroom has some very powerful, subtle and easy to use colour adjustment facilities. Much as I would have liked grey card white balance adjustment to handle it all, it looks like I am going to be grateful for those colour adjustment facilities.
The procedure, as I read it, was straightforward. Include a grey card in a “calibration” image, using the same lighting as for the normal captures, and then in Lightroom, Photoshop or whatever take the colour values from the grey area of the calibration image to set up the white balance on the normal captures. And the colours would then be correct. Easy.
What I did not read, but after some confusion and head-scratching discovered, was that if you take multiple readings from the grey area of a single calibration image you will get different values of temperature and tint. This and what follows is for my G3, but I would be surprised if it was not true for other cameras.
The variations from one measurement to another (on the same image) are quite small at low ISOs, and at base ISO only make a subtle difference to the appearance of an image. However, as the ISO increases the differences get much larger and can have a huge impact.
As an experiment I captured four images, at (base) ISO 160, ISO 400, ISO 1600 and ISO 6400, of a grey card on a table with light coming through a roof light window just above the table. As there was 100% cloud cover the light almost certainly did not change during the few seconds while the four shots were taken.
The following graph shows the temperature and tint for 25 measurements taken from each of the four images. As you can see, the spread of temperature and tint values increases significantly as the ISO increases.

Spread of White Balance measurements for four ISOs by gardenersassistant, on Flickr
For those who prefer numbers, here are some statistics for the measurements.

White balance sample measurements by gardenersassistant, on Flickr
When I applied samples of the settings to several images it seemed to me that although the base ISO values did vary, the impact on the look of the image was probably less than would matter to most people. At the other end of the spectrum, using settings from the ISO 6400 calibration image, there were huge colour shifts. This composite image superimposes two of the white balance settings measured from the ISO 6400 calibration image.

Image with two White Balance settings taken from the same calibration image by gardenersassistant, on Flickr
Here is what the grey card on the table looked like with the same two settings.

Grey card with two white balance adjustments by gardenersassistant, on Flickr
------------------------------------------
The other thing I discovered from this exercise was that even for base ISO, where there is not much variability about the measured white balance, using this white balance does not guarantee that the colours look like what you saw when capturing the image. For example, this pair of images is of the same flower. (Yes, the left hand one is underexposed, but I have deliberately only changed the white balance to avoid unwanted complications. As it happens though, correcting the exposure does not remove the redness, which is the issue of concern here.)

Same flower, different time of day, both grey card corrected by gardenersassistant, on Flickr
When I took both pictures the colour I saw was similar to the right hand version. I pass by that flower very often, and it always has an “orangy” aspect, and never red, in any light. However, both those images have been adjusted to the white balance taken from a calibration image taken at the time the image was captured.
Interestingly enough, using JPEG with my current camera, and with my two previous cameras, this flower tended to come out red like in the left hand image, so this is a familiar issue. It was not cured by the white-balance calibration of the left hand image. It was corrected for the right hand image. Very curiously, the more appropriately coloured version is ISO 3200, with its undependable white balance calibration, while the wrongly red one is base ISO, 160, with its much more consistent white balance calibration.
Another colour that is sometimes not brought into the realm of “reality as I saw it” by white balance correction is green, sometimes. With JPEG greens sometimes took on a totally unnatural bright emerald green character, and this is still happening with some white balance corrected raw images.
The good news is that it turns out that Lightroom has some very powerful, subtle and easy to use colour adjustment facilities. Much as I would have liked grey card white balance adjustment to handle it all, it looks like I am going to be grateful for those colour adjustment facilities.
Last edited: