Bit Depth and Image Quality

Digital images are made of pixels, solid blocks of colour or tone that are defined by a numeric value. This numeric value speaks a universal language that all digital devices understand, although now more complex than in the past but simple in nature. The first digital images produced were 1bit images. The pixels were either black or white, on or off, represented in the binary language of 1 or 0. As digital technology evolved we had 2bit images where the information was two bits long with the possibility of creating four shades from black to white; Black, Dark Grey, Light Grey and White.

Greyscale gradient of 1bit, 2bit, 4bit and 8bit

At presents a basic digital image will have 8bits of colour information giving 256 shades from black to white. As digital systems are generally colour and use three colour channels, Red Green and Blue, each with 256 tones from for example; from the blackest to the whitest red. Multiplying 3 8bit colour channels produces a 24bit colour image also known as true colour. When we talk about the bit depth of a digital image we are talking about how many tones are created in a greyscale from light to dark. We would say a photo is an 8bit image not a 24bit colour image.
To create a continuous tone you need a minimum of 250 shades, 8 Bit sounds enough then for photography?

This is where the question of bit depth becomes important to photography. 8bit is sufficient to create a continuous tone without its gradient being broken up and looking artificial. Also if we add the fact that standard printers also print 8bit images (there are 16bit printers on the market), stock and publishing companies will only accept 8bit images. 8bit sounds brilliant. However the rub comes when we edit our files and in doing so we chop up the information contained within reducing the number of tones and degrading the quality of the image. This degradation of quality can be seen in the histogram at the end of the editing process with an effect called combing.

Combing occurs because as you edit an image you are effectively stretching the information inside. An image is best represented by its histogram, which shows all 256 tones in a neat little graph from dark to light. If you were to increase the contrast of an image using levels for example; moving the point of black from 0 to 49, all the 49 shades are now black slicing away the information leaving the different tones with the same tonal value. The same happens when you move the white point from 255 to 199, all 56 tones are white. As you stretch the image information gaps appear creating combing in the histogram and posterization also known as banding in the image on the screen.

The main outcome of the editing process should be to create the best and highest quality image that you can. In the same way it is best to have the highest resolution image possible (you can read more about resolution and image quality here), it is best to start with the highest bit image you can and then convert at the point of output. Most cameras will shoot at 8, 12,14 and 16 bit. If you were to shoot at 12 bit where each colour channel will have 4096 tones with a possible 68 billion colours, leaving room for editing.
Shooting higher bit images adds more information to the file meaning a larger file size yet this is another trade-off in creating the best image possible.

Left 16 bit raw image Right 8bit jpeg.
The 16bit image has more definition in the shadows and stronger colours than its jpeg counterpart.Both zoomed in at 100%

Bit depth also contributes to the, now in my opinion “dead”, debate of Jepg vs Raw. A Jpeg image is an 8 bit image yet the camera sensor can capture a higher bit image. When taking a photo processes the larger bit information from the sensor and then in saving slices away the information to create the Jpeg file. This has wasted potential tones and colours that could otherwise be captured. With a Raw file you get the information unadulterated from the sensor, with all the bit depth that it can offer. You can keep the image as a high bit image all the way through editing up until the point of output. It is even possible to create HDR images that are 32bit and edit them in Lightroom (4.2) and Adobe Camera Raw (7.2) meaning you have even more room to create amazing images (this is different from tone mapping, which gives you a simulation of what the image would look like but at 16 bit).

ISO 400 50mm f/8.0 1/1000 Edited in Lightroom

Of course I must note that not every image will benefit from being shot at a higher bit depth, as the output would show no added benefit; for example a passport photo. This being said when cooking tastiest dish you can, you need to start with highest quality produce. The same is with digital photography. To create the highest quality image you need the best resolution and the highest bit depth that your camera can capture. Keeping the image at a high bit depth until you send the file for output is the best practice.

Remember if you liked this post to; like, share and subscribe.

If you wish to get notifications when I post on my blog, you can follow me on Twitter@apertureF64, on Facebook.com/aperturesixtyfour or alternatively be emailed by subscribing below. All images are the Copyright of Benjamin Rowe , ALL RIGHTS Reserved unless credited to another photographer. For more information please read my Copyright Statement