Using a Custom White Balance
White balancing means adjusting the color balance in an image to compensate for the color temperature of the illumination source. This can be done automatically in the camera, or manually in subsequent image processing.
Human Visual Perception and Color Balance
Human visual perception is very good at working over a tremendous brightness range. It is also excellent at adapting to the color of the illumination in a scene so that colors appear natural to us.
Millions of years of evolution have produced human vision that is based on the light from the Sun being perceived as white. But the color of sunlight changes throughout the day. It can be very red when the Sun is close to the horizon. It can be very blue when sunlight is scattered in our atmosphere, and the sky is the only source of illumination when the sun is behind a cloud or when we are in open shade. Our eyes can adapt to this and other changes in the color of illumination from the dominant light source so that we see the brightest tones in a scene as white. This means that we will see the same piece of typewriter paper as white under direct sun, or open shade, whose illumination is very blue from the blue sky. Our eyes will also see this piece of typewriter paper as white under indoor tungsten illumination, which is very red.
The color of illumination is usually described as “color temperature”. This is an explanation based on the actual physical temperature of a “black body” which, when heated, gives off light. At low temperatures, the light is very red. At about 5,000 – 7,000 degrees K, the light is seen as white. As the color temperature goes higher, the light gets more blue.
Here are the color temperatures of a variety of light sources:
Digital Cameras and Color Balance
Camera manufacturers go to great effort to make sure that cameras record colors accurately. However, to a camera set for daylight white balance, a scene illuminated by tungsten light at 3,200K will accurately record as very red. Likewise, a scene illuminated by the sky in open shade will accurately record as being very blue when the camera is set to daylight white balance. Although the camera records these scenes “accurately”, this is not how we perceive them. We don’t perceive them as being very red or very blue, we perceive them as being normal, with neutral colors. To overcome this problem, digital cameras come with the ability to automatically white balance a scene, or to manually select the correct color temperature for a scene so that the camera’s representation of the scene is close to the way we perceive it.
In the examples below, all images were shot with the camera set to daylight white balance. The color temperatures were read with a Minolta color temperature meter.
Digital SLR cameras usually come with an auto white balance setting that tries to correct for whatever type of illumination is present in the scene. This usually works pretty well with daytime subjects, and less well with warm, tungsten illuminated subjects. These cameras usually also offer several standard white balance settings depending on the type of illumination, such as tungsten, daylight, open shade, and cloudy. Some DLSR cameras offer the ability to pick a specific color temperature in degrees Kelvin for white balance. Some even offer the ability to set a custom white balance on a gray card to specifically match the dominant source of illumination in the scene.
These cameras can do this because the data is digital. Numbers that represent colors in a scene can be adjusted so that white and gray tones are neutral from a wide variety of different light sources. Digital cameras usually produce fairly accurate colors with these standard light sources.
Deep-sky astronomical images are different than normal daytime photographic images however. Daytime images usually have a normal scene with objects in it that reflect and absorb light from the dominant source of illumination such as the Sun, a tungsten light, the open sky, or even a flash. Deep-sky images usually include stars that are self luminous, emission nebulae which are also self luminous, and reflection nebulae which shine by reflected starlight.
Color Balance in Astronomical Images
So what color balance setting should we use for astronomical images? If we want to record accurate color, we should use daylight color balance. This is what our eyes evolved to see as a standard. If we use daylight white balance for our night astronomical photos, any stars that are similar in color to our own Sun will record correctly as white, and other star’s colors will also be accurate.
There are, however, several problems that we can encounter in trying to do this.
First, to shoot good pictures of red emission nebulae, we need a modified camera whose original long-wavelength filter has been replaced or modified. Once we do this, the daylight white balance setting is no longer accurate.
The other main problem is that some astronomical image processing programs do not support standard or custom white balances, even if that data tag is embedded in the image file when it is recorded in the camera.
Some astronomical image processing programs, like Images Plus v2.8 do support these white balance settings.
Depending on which program you prefer, you may have to use a different method to achieve correct white balance in an astronomical image.
White Balancing a Modified DSLR Camera
In a stock DSLR camera, the sensor, Bayer filters, and processing algorithms have all been finely tuned by the manufacturer’s engineers and color scientists to produce accurate color with the original long-wavelength filter in place. Modifying the filter by removing it, or substituting another filter with different passband characteristics changes this situation. The automatic color balancing algorithms, and the built-in preset color balances like daylight, won’t work correctly with a modified camera.
For astrophotography, white balancing a DSLR camera that has had the manufacturer’s low-pass long-wavelength filter removed or modified can be accomplished in one of three ways:
Note that we should use only one of these methods at a time.