SHARE

Video Compression And The Future

Mobile phones have screens, speakers and microphones. They can take pictures and record video. And they can connect to the Internet. So why you can’t make a video-conference call from your mobile phone? For the same reason on-demand video doesn’t really work on mobile devices. And for the same reason you can only watch eight […]

Written By

Nicholas Carlson

Apr 18, 2006

Datamation content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More

Mobile phones have screens, speakers and microphones. They can take pictures and record video. And they can connect to the Internet. So why you can’t make a video-conference call from your mobile phone?

For the same reason on-demand video doesn’t really work on mobile devices. And for the same reason you can only watch eight HDTV channels at home.

Bandwidth.

Most mobile devices connect to the Internet at about 10KB per second. Video requires a bandwidth of at least 50KB per second.

That’s where video-compression technology currently stands, according to Deepak Turaga, video-compression researcher at IBM’s T.J. Watson Research Center.

The distance between 10KB and 50KB is far, but it’s a distance that Euclid Discoveries claims to have closed.

Euclid CEO Richard Wingard said that, after six years of research, his company can compress video-conferencing streams down to 1000 bits per second.

Object Versus Blocks

The advancement in Euclid’s video-compression technology, Wingard said, is its integration of object-based compression. To achieve this, the company had to overcome block-based video compression.

Block-based compression works by segmenting the video into rows and columns that form blocks, which are the large pixels you sometimes see in streaming video over the Internet. Data is sent over the network only for the blocks that need updating in each frame.

Think about a television screen when the local news is on.

Because the camera is fixed on a talking head, most of the screen is still. If the screen were divided into blocks, the blocks in the areas that are stationary would not need to change very often, because sending that data would be redundant. By sending less data, video streams can be compressed using block-based compression.

But those few blocks in a newscast that do change with every frame can cause some trouble. This is because each block has to be re-transmitted in every frame.

Imagine, then, if there were some way for the compression software to recognize the talking head as a single object, rather than a collection of changing blocks.

This is where object-based compression enters the picture. Now it can transmit only what’s changed about an object from frame to frame, rather than having to transmit new data for a host of blocks. Less data would be transmitted and higher compression would be achieved.

This article was first published on InternetNews.com. To read the full article, click here.