Digital Video Compression: An Overview of the Technology Behind the Pixels

From YouTube to webcams to online TV shows, having fun and passing time online is now synonymous with video content. Watching YouTube has become a new national pastime. Click here and there and everywhere, and there's always another related video. All these videos are continuously being watched and downloaded. Digital video content online has increased extremely quickly in the last few years. So, just how many videos are online? YouTube has changed its search query to return "video results 1 - 20 of millions" when you use the wildcard("*") search, but the last time I was able to check(about 9 months ago) YouTube had over 82 million videos. Today the number must be around 130-180 million, and nobody can get an exact number as its constantly changing. That means a video for every other American. Furthermore, live streaming videos (such as through webcams, Yahoo Messenger, and Skype) and online television shows are now very popular and diverse.

A major part of the reason for the huge increase in videos can be attributed to video compression. There are many algorithms and digital video formats that have made storing all these videos more efficient and affordable. Efficient bandwidth utilization saves both money and time, and thus compression has been a topic of much interest in computer science.

There are two categories of compression, termed lossless and lossy. The compression used for all popular online video is lossy - by necessity, you need to lose some of the original data to get it all through the electronic pipelines. You upload a 30 megabyte video to YouTube and they manage to chop and shave it down to a nice little compact file of 3-5 megabytes. Lossless compression is useful in specific situations where you need exactly the same data, like in programs.

Even before delving into video compression, the original fundamental topic boils down to image compression. This topic is a precursor to video compression, as images are the discrete units that comprise a video. Many of the techniques used in the Joint Photographic Experts Group(JPEG) format were applied in the Moving Picture Experts Group(MPEG) standard. Thus, to be able to build an effective video compression method we need to know about how images are converted to data, and about the characteristics of images. JPEG uses advanced algorithms and functions to compress pictures. The different techniques used for image and video compression include run-length and entropy encoding. Run-length encoding is intuitive - instead of storing 10 blue pixels in a row, store 10 * blue pixel. Entropy encoding is a technique loosely based on the Morse code idea of using short symbols for common data values - in Morse code, 'e' is the most common letter and 'z' is the least common, so you use a short value for 'e' and a long value for 'z'.

Today, there are many standards that have made it easier to distribute video, such as Flash Video, MPEG-1, MPEG-2, and MPEG-4. These standards all utilize many of the basic compression techniques used for JPEG. In addition there are human-characteristics such as motion compensation, where you only need to store the moving objects(and your eyes blur moving objects and so there is repetition of data), that are employed to minimize storage space for digital video and are vital for compression.

Another one of our traits that makes compression easier is that our eyes cannot tell different colors apart very accurately. We can only notice the hue differences among about 1000 colors, and so color information can be compressed more. We do however have a sharper sensitivity to luminance(at night we can perceive far more shades of darkness than any digital camera) and so that demands less compression.

And as we all know, television is now going digital too. In only a few weeks, everyone who watches public network television will be watching digital TV that utilizes the above technologies. This is a very welcome upgrade, and so the days of fiddling with rabbit ears and wrapping up aluminum foil should soon be history. Now you'll just have to patiently hope the signal strengthens if you start seeing funny green Lego shapes dance around on your TV.