Bit Depth

To illustrate bit depth further, imagine you want to draw a picture of a sunset, but all you have is the basic box of 16 crayons. Real life sunsets have a huge variety of colors, from blinding yellows and oranges, to faint reds and purples. If all you have are those 16 crayons, you won’t really be able to draw all those different shades. You’ll still be able to draw a picture, but it won’t look that great.

But what if you had the 32-crayon box? Well, you’d be able to use twice as many colors, but your picture still wouldn’t look like the real deal. Now what if you had a box of 1024 crayons? At this point, you might actually be able to draw a pretty decent picture. You won’t have all the colors you need to draw a photorealistic sunset, but you will do much better than you did with just 16 or 32 crayons. And if you keep adding more crayons to your box, you’ll be able to draw a better and better picture.

This example might seem a bit simplistic, but it helps to understand how bit depth works. A codec’s bit depth is basically the box of crayons it can use to draw images. The deeper the bit depth, the bigger the box of crayons (number of colors), and the more realistic image it can produce.

How Bit Depth is Measured

To fully understand bit depth, we need to explore how digital imagery works in greater detail.

You probably already know that digital still and motion images are made up of pixels, and that each pixel mixes together three primary color channels (red, green, and blue) to make many different colors in the image.

What you may not know is how these colors are defined by a computer. Each of the three color channels has a range of possible values it can be assigned, and this range is stored as a single number. What determines how big or small that number can be is the number of bits the computer uses to store it.

A bit is just a binary unit of information, digitally displayed as a 0 or a 1. To store increasingly complex information, computers need to use more bits per integer. So, a 1-bit integer can only have 2 values (0 or 1), but a 2-bit integer can have 4 values (00, 01, 10, and 11), and a 3-bit integer can have 8 values (000, 001, 010, 011, 100, 101, 110, and 111). For every extra bit in the integer, a digital value can store twice as much information.

By increasing the number of bits per primary color channel, computers can store more complex color information. If each primary color channel uses an 8-bit binary integer, that means there are 256 possible shades of red, green, and blue that can be used for every pixel. Mixing every single one of these shades together would make 16,777,216 different colors (256x256x256), from pure black (a value of 00000000 for each channel) to pure white (11111111 for each channel), and every shade in between.

8-bit vs 10-bit

The most common bit depth for video is 8-bit. It has been the digital video standard for decades, both in television and film production. DVDs use 8 bits of color depth, as do Blu-rays, and most streamed content (for now). Even many cutting-edge cameras still default to 8-bit color when recording video. In fact, almost every digital video you’ve ever watched was 8-bit. We want to emphasize these points to demonstrate that video with 8 bits of color depth is technically capable, and can be used to tell incredible stories.

However, despite this, you should consider codecs with a bit depth higher than 8 bits if your workflow can handle it. Why? Because 8-bit codecs have some significant shortcomings that are important to understand.

The fundamental problem with 8-bit color depth is a lack of available colors. While 16.7 million colors might seem like a lot, remember that each primary color channel only has 256 unique shades to itself. That can make it difficult to reproduce subtle gradients between the same or similar colors.

This issue is called banding. It occurs because there is a relatively small number of steps between the brightest and the darkest shade of any given color. The lines in the image are artifacts left over from a computer trying to stretch too few available colors over too wide an information gradient. This problem is amplified even more in low contrast scenes, such as those that take place in dark rooms or under shadows.

The limitations of 8-bit codecs span the entire length of post-production processes, but are especially problematic for color correction and VFX. Strong color grading in an 8-bit codec often leaves noticeable artifacts in footage, simply because there aren’t enough shades to make the fine-tuned adjustments you want. And chroma-keying (green screening) with an 8-bit codec can leave jagged or undefined edges, which makes VFX look bad.

Again, as the last 30+ years of digital video production has demonstrated, 8-bit color depth is usable for many projects. But, there are higher-quality options that you should consider for your workflow.

10-bit color depth is exactly like 8-bit color depth, except that each primary color channel has two extra bits for storing the color value. That means that instead of just 256 possible values per RGB channel, 10-bit yields 1024 possible values per RGB channel. That is a fourfold increase to 1,073,741,824 colors (1024x1024x1024).

10-bit has existed for high-end productions for decades, but has only become usable for everyday workflows recently as new cameras and software began to support it. It is now possible for even mid range DSLRs to shoot in 10-bit color depth, and to edit 10-bit codecs on standard editing workstations.

The advantages of 10-bit color depth almost always outweigh the extra storage and processing power you need to handle it, if you are doing any color correction or VFX. Your original camera files will be much higher quality, your color correction will be much more precise, and your VFX work will look much more realistic. Even if you will be delivering the video in an 8-bit codec, you should still consider mastering the project at 10 bits because the intermediate steps will benefit.