Understanding Chroma And Luminance In Digital Imaging

This photo is an Audi TT taken from a car convention. The first thing you will probably notice, aside from the car itself, is the bright vibrant red background, the metallic color of the car body and the various shades of black. This was tone mapped to give more color and luminance, the two topics that this article will discuss.

The reason an image stands out is due to many things, color being one of them. This is because our eyes are trained to detect chroma (color) and luminance (light) in an image. Perhaps it is better to use the term luminous intensity of light. The retina in our eyes have what are called “photo receptors” and these perform a function. These receptors are the “Cone Cells” which handle the chroma and the “Rod Cells” which handle the luminance. To explain further, we see luminance as different shades of light in grays while chroma are different hues of color. Colors have intensity while light has brightness.

We see color in images because of light. In the absence of light, in total darkness, we do not see any colors. When photons from a light source strike an object, it gives off a wavelength of light that our eyes perceive as color. The cone cells interpret the various wavelengths with three primary colors: R-Red, G-Green and B-Blue aka RGB. In low light conditions, the rod cells in the eyes function to perceive shadows and darkness in the absence of color.

The human eye consists of photo-receptors called the “Cone Cells” (color) and “Rod Cells” (luminance or brightness).

The different hues of color we see is called the “gamut”, which are a range of different blending of colors. Chromaticity diagrams show the many different levels in gamut that a color space can provide when it comes to digital images. The chromaticity specifies the hue and saturation, but not the lightness. Lightness is the property that luminance can specify. In the diagram, we see a graph that plots the wavelength of light the human eye sees with coordinates. The coordinates are values that are plotted on an X and Y axis to show the wavelength and uses a reference point called the “white light”.

These color spaces measure the wide range of colors available and the more there are the more details in an image. With luminance, we add different tones of light on a scale from 0 to 100 or total white to completely black. Luminance comes from the root word luminance, which is the measure of light an object gives off or reflects from its surface. With luminance we deal with tones in digital images as the varying shades of gray i.e. tonal information. The eye perceives the difference in visual properties in luminance of an image as contrast. This is how we see the blackest blacks and whitest whites.

The black triangle shows the extent of the BT.709 color gamut which is what most HD television sets support. The D65 point is white light.

Digital images are represented as picture elements, also called pixels when being displayed on a screen. The device illuminates each pixel with light so that each pixel gives off a different wavelength which we perceive as color. In luminance, each pixel is measured in bits, with 0 meaning total darkness and 1 meaning total white. The bit depth of an image has become synonymous with luminance. When capturing images with a camera, the color and light information in an image is best stored in a RAW file. RAW is the best format for storing all the information contained in photo.

When you have more information stored, you have more details and higher resolution images. Thus the file size of a typical RAW image is very large and it is also uncompressed data. This allows more details to be processed by an Image Signal Processor or photo editing software on the RAW files. Higher resolution images of great quality are processed from these files that contain the color and light properties of the image during capture.

The number of colors that can be displayed is dependent on the device. The display i.e. screen, must be built to a corresponding specification. The image already stores the information that the device needs to display. There are different types of screen color specifications supported that are called color spaces. Common types are sRGB, Adobe RGB, DCI-P3 and Rec.2020. The consumer electronics industry requirement for devices is 8 bits per channel (also called bit depth).

In theory that is about the most the human eye can see. Anything more than that is just not possible, though there are color spaces capable of displaying more. In the past VGA monitors were capable of displaying a total of 262,144 colors only. Color monitors today are capable of 16,777,216 different colors. Color spaces like Rec. 2020 are capable of 12 bit (4096 gradations) 68,719,476,736 colors. However, the human eye cannot really see that many colors nor is there a screen today that can show that many colors. So why were they developed? It is because the colors look more vibrant and rich because it has a wider gamut. This is the difference between normal and superior quality in colored images.

A chromaticity diagram showing the color gamut of some popular color spaces ProPhoto RGB, Adobe RGB, Colormatch RGB, sRGB. ProPhoto RGB has the widest color gamut as can be seen from the purple triangle which is the area it occupies. The blue numbers represent the wavelength of color at a plotted coordinate on the graph, that is visible to the human eye.

When adjusting brightness in displays, the actual function being controlled is the luminance. The luminance of a display is measured in the unit cd/m² (Candela per Meter Squared). A good display has a luminance of 300 cd/m², often achieved by OLED and back lit LED screens. How we see brightness is based on the luminance of the object reflecting the light. Tones are the best way to see the contrast since it shows the lighter and darker parts of the image.

When you process a black and white image, you often see certain details that are striking but when in color it may look less impressive, even more flat. Tones are also a way to bring out clarity in an image by darkening the shadows and blacks to give the colors some more intensity. Looking at a histogram from an image shot on a DSLR camera, we can see the tones shift from light to dark as it shifts to the left.

A histogram shows whether an exposure of an image is too bright (overexposed) or too dark (underexposed). It is also a way to see the tonal range of lightness in an image’s luminance.

In order to get the accuracy of an images chroma and luminance, the display must either be calibrated to a supported color space, otherwise color correction during post is needed. Most displays in retail are not calibrated, so professional color grading would require a properly calibrated monitor to represent the accuracy of colors. Contrast is also best with luminance, so a monitor that has a bright display shows the best quality on an image.

The tonal range of darkness and lightness can be seen in black and white images.

We use the chroma and luminance of an image not to measure quality, but the details. Instead we look at chroma as color and luminance as lightness to measure the details in an image. The more colorful and luminous an image is, the more details you see. It is best to work with more details from RAW images and edit it non-destructively and later converting it back to a lower gamut that is supported on most displays.

For detail oriented photography and video for brands, the smallest details are just as important to portray the image. It can mean the difference between appealing to new customers or turning them away. The way an image looks in ads or editorials in fashion and beauty need color accuracy and details in skin, hair, eyes and body. For television shows and music videos, the details are about vivid and realistic viewing experiences. In creating images, the details are a key to having the best results.