If you have ever downloaded anything off the internet, you know that sometimes the downloaded files are in the .zip, .gz or .tar format. These are examples of compression systems, a very handy invention, especially for internet users, because they reduce the overall size of the file to allow transfer over slow internet connections, or take up lesser space on disk. Once you download the file, you can retrieve the original uncompressed file using utilities like WinZip (for Windows) or gzip or tar for Unix. If everything works correctly, the compressed file is identical to the original uncompressed file, so that when you uncompress the .zip file, you can retrieve all the information in the original file.

Okay, alright, so you guys have already heard about Winzip and .zip files, and are starting to yawn. But now, what if I told you that almost all the files that you download, including .mpeg, .mp3, .avi, .rm, .jpg, .gif files, aren't actually uncompressed files? And that the very television shows that you love and trust to make you laugh consist of compressed information?!?! If you aren't impressed, then you probably know most of what these pages are about - compression of audio, video and video stills.
But:

if you just sat up straight in your chair and lost that yawn, welcome!

Some more facts, just to prove the point:

The storage of uncompressed graphics, audio and video not is not feasible even with DVDs (which can store about 20 GB of information). Uncompressed graphics also require high bandwidth for transmission, meaning that the number of bits to be transferred per second is very high. This you have probably experienced before, especially when you are streaming a movie online and are trying to load another page with lots of graphics on another window at the same time - you notice that the graphics don't load as fast as they would if you weren't streaming the movie, because your internet connection probably allows you to transfer (upload or download) only a fixed number of bits per second, and the movie is taking up most of this alloted bandwidth. This also happens if you are using a P2P software like Kazaa to download an mp3 file while someone else is uploading a movie file from you - if there are too many users trying to access your files, most of your bandwidth is expended in them uploading your files and your download becomes slow.

Without compression:

• A text file (let's say it has 1 byte alloted to each character, and there are 55 lines per screen with 80 characters per line) which is about ten pages will be about 44 KB in size
• A still image (even low resolution images, 640x480 pixels, with more than 1 byte required per pixel) would be about 30 KB (the size would be a lot larger if it was a true colour image, where you use 8 bytes for each of three colours per pixel!)
• Each second of an audio signal (sampled at 44.1 kHz and quantized at 16 bits per sample, 2 channel stereo) would require about 180 KB, and Falco's Rock Me Amadeus (length = 3:15) would be about 35 MB in size - this, in fact, is about the size of the .wav or .cda files that exist on your music CDs. Compare this to the mp3 file, which is about 3 MB!
• NTSC (the format used in the United States for TV broadcasting) TV signals have bandwidths of about 10 MBps. This means that an entire episode of The Simpsons (about 25 minutes) will need 15 GB! Once again, did you just go:

• HDTV (High Definition TV) doubles the number of lines from NTSC and uses an aspect ratio of 16:9 compared to the NTSC value of 4:3, a factor of 1.33 higher, and has a bandwidth of about 50 MBps (it can be as high as 1 GBps) - this only makes matters worse, even this bandwidth is too high to actually achieve for regular transmission.
• Digital video (let's say its low resolution again, 640x480 pixels), running at 29.9 frames per second (fps), encoded with True Colour, would require a bandwidth of about 900 MBps! A typical Britney Spears music video (length about 3.5 minutes) needs about 190 TB!!

So, there are three things to think about here -

• How is it that we can reduce the size of a file and still retrieve all of the original information on uncompression?
• Why is there a need for compressing audio, video and images? I heard that TV broadcasts also consist of compressed data! Is this true? How does that work?
• Okay, I understand how this can work for text files, but what do you do in case you have images, audio and/or video?

Go to the next page, or use one of the following quick links:

Main Page
Motivation For Compression - Some True Stories
A Brief History of Compression (heheh - brief, compress... get it?)
Requirements From Any Compression Algorithm
Data Compression Fundamentals
Some Compression Techniques
Video Compression Techniques: The MPEG-1 Standard
The Future: MPEG-4 And MPEG-7
Related Links
 
 
 

Contact Me: sundar@pha.jhu.edu
Copyright © Sundar Srinivasan 2002