Video Compression Basics

Raw digital video uses on the order of 3 GB of storage space per minute. This size makes storing and transmitting uncompressed digital video data prohibitive. Thus, video needs to be compressed before being stored or transmitted, and restored to its original format for editing and viewing.

A codec, short for “coder-decoder,” is a mechanism for compressing and restoring digital video. A codec can be hardware- or software-based. An encoder analyzes digital video data and looks for certain patterns that it can express via mathematical algorithms. For example, an encoder looks for areas of pixels that stay the same over time, or for predictable movement patterns among pixels. Then, the codec generates an “encoded” description of the digital video, saving a considerable amount of space by avoiding redundancies. When it’s time to view the video, a decoder quickly recreates the original data, usually with small losses, based on its own encoded description.

As video is compressed, a certain amount of data loss is incurred, depending on how much compression is required. The data loss is manifested in a visible drop in video quality, once the video is decoded. The more compressed a video stream, the less space it takes up, but the greater the drop in video quality. Hence, video codecs strike a balance between size and quality. This is often referred to as the rate-distortion ratio.

At first, most video codecs were hardware-based. Computer CPUs were simply not fast enough to allow software to compress and decompress digital video data on the fly. However, as computer technology evolved, a variety of software-based codecs have emerged, using several different video encoding standards.

Furthermore, the codecs themselves have evolved to be increasingly more complex in order to yield ever better rate-distortion ratios. A powerful codec can use complicated algorithms to reduce the size of the encoded video, while allowing it to retain good quality. However, the more complicated the algorithms it uses, the more the codec is a strain on a computer’s resources. Compression with a complex codec in real-time might not be possible on a slower computer. The amount of resources a codec requires to operate is the third main consideration in compression technology, along with compression ratio and video quality.
The most popular video compression standards today include:

• MPEG-1, developed in 1992, for video of VHS quality at a rate of about 1.5 Mbps (megabits per second).
• MPEG-2, developed in 1993, for a broad range of high-quality video applications, including TV broadcasting and DVDs; standard definition (D1) at a rate of 2-8 Mbps.
• MPEG-4, developed in 1998, for low bit-rate applications such as video conferencing; high-quality compression achieves a similar bitrate to MPEG-2.
• H.264/MPEG-4 Part 10 or AVC, developed in 2003, for the full spectrum of quality and bitrate requirements; it is the most advanced codec, achieving full D1 resolution with DVD quality in under 1 Mbps.