Convergence Channel: H.264 Helps Megapixel Go Mainstream

February 28, 2009 SSI Staff

Normally, things being “in” and then “out” faster than the blink of an eye is an idea found in the fashion industry or Hollywood. Yet we have just witnessed it happen in our own video security industry.

Yes, the standard resolution IP camera, after making a huge splash in our industry the past couple of years, is now “out,” having been replaced by an increasingly amazing new technology: megapixel. Alas 4SIF, we hardly knew ye.

OK, technically megapixel is just an upgraded version of an IP camera, but it is also much more. The benefits are enormous: A single camera covering the same area as several standard resolution cameras or the ability to pan and zoom within a scene after the video has been recorded to reveal information a standard resolution camera would never be able to uncover.

We have discussed megapixel cameras before (see May 2008). In that column, we looked at what megapixel was, as well as some of the hardware. Here, however, I want to focus on the compression technologies making megapixel not only possible, but also a more powerful and affordable solution.

How Compression Works

Understanding that raw video needs to be compressed before it can be sent across a network and recorded is not new to us. However, the way video is compressed is constantly changing and improving, a progression vital to megapixel video.

Until now, MPEG-4 and MJPEG compression algorithms (codecs) were the norm for standard resolution IP cameras. For resolutions varying between SIF (320 X 240) and 4SIF (720 X 480), these methods provided more than sufficient bitrates for network transmission and storage.

Now, however, with resolutions approaching 1,280 X 720 (720p) and 1,920 X 1,080 (1080p) or even 2,584 X 1,936 (5MP, 4:3 ratio), new compression technologies are needed to help squeeze the increasing number and size of megapixel video sources into today’s networks.

GOP in the context of video stands for group of pictures. When each frame of analog video is captured from the CCD or CMOS chip, a resulting digital image is formed (for the sake of this column, the use of frame and image are interchangeable).

Each of these images, or frames, of video has a designation. The first one to be captured is called an I-frame, or intra frame. This I-frame becomes the reference for the next few frames of video that comes in.

Here’s how compression works: Our eyes can only see so much information like colors, shades, movement, etc. A computer can “see” much more than we can. Knowing this, a compression algorithm will toss out redundant info that we can’t see. Getting rid of this excess data is one way files are made smaller. On top of this, any information that hasn’t changed between frames is also removed, as it can be copied from the previous frame.

So again, the I-frame is used as the reference for several frames that are captured next. In MPEG-4 video, these are called P-frames, or predicted. When a P-frame comes into the compression circuit, it is compared to the previous I-frame, and anything in the image that hasn’t changed gets removed, resulting in a smaller, compressed frame of video.

For standard resolution video, this method was fine. But the world changed rapidly.

Meeting Megapixel’s Metrics

When higher megapixel resolutions created much larger video files, this compression method became less able to squeeze the huge frames down to usable size. A new, more efficient compression method was needed. H.264 came to the rescue.

MPEG-4 and other, earlier video codecs, such as H.263, were developed by two different standards groups. MPEG-4 came out of the ISO Motion Picture Experts Group and was designed for, what else, motion picture compression and transport. The H.26x series came from the ITU telecommunications groups and was primarily for the emerging video teleconferencing market. In 2003, the two standards bodies decided to team up to develop the next generation of video compression, as video quality and network requirements were growing ever-more similar for the two groups.

Out of this partnership, a new, extremely flexible video codec was born. What it was called depended on which group was talking about it. For the MPEG group, it was called MPEG-4 AVC (Advanced Video Codec) Part 10. ITU had a shorter and catchier title, H.264.

If you enjoyed this article and want to receive more valuable industry content like this, click here to sign up for our FREE digital newsletters!

Pages 1 of 2 Next