Nuts & Volts - September 2004

Let’s Get Technical

330 KB of storage was converted into a 220 KB MP3 file. The MP3 file sounds as good as — if not better than — the original WAV file, with the absence of 110 KB of sound information.

Even sampling the analog sound information can be approached from different directions. Using an analog-to-digital converter, you may sample the converter 8,000 times per second, reading an eight-bit sample each time. Thus, a one second audio waveform would require 64,000 bits of storage.

Now, suppose you change the sampling hardware to one that uses a delta modulation method, where the sample can only increase or decrease by a small value each time. (This is the delta.) Then, each sample only requires one bit of storage.

Now, one second of audio only requires 8,000 bits of information. If the delta modulation method does not have significant slope overload losses (signal level changing faster than one delta between samples), it provides a nice, built-in compression over the eight-bit A/D method.

Once the MP3 compressor has thrown away the WAV file’s redundant and unnecessary signal information, there is no way to get it back. In this case, it does not matter, since the MP3 format has high quality.

What if we need to have every bit of the original data back when it is

Graphics Interchange Format. JPEG stands for Joint Photographic Experts Group. GIF files were created by CompuServe as a method of exchanging graphical information. The features of a .GIF file are as follows:

• Maximum of 256 simultaneous colors

Figure 2. Sample image containing 24,320 pixels.

• Lossless compression using LZW ( Lempel-Ziv-Welch) algorithm

uncompressed? If we compress a program file and then uncompress it, every byte in the file has to be correct or the program will not function correctly when the program is executed.

In this case, we need a compression algorithm that is “lossless.” The technique used to convert the WAV file to an MP3 file is a “lossy” compression technique, since some of the original data is thrown away.

Okay, so the MP3 audio file is one place where lossy compression may be used to advantage, but where else can we get away with it? We cannot do so in program files, for sure, or even files that contain important information, such as databases, but what about image files?

Browsing the Internet is a widespread activity, awash with images of many shapes and sizes. Web browsers that display graphical images accept two different types of image files: GIF images and JPG (or JPEG) images. GIF stands for

• Support for animation and transparency built-in

• Interlaced and non-interlaced formats

The quality of the GIF image was kept low so that its file size could be managed. People with dial-up modems for their Internet connection do not want to wait forever for an image to load.

When the GIF image data is decompressed for display purposes, an exact copy of the original data is reproduced, due to the lossless compression used on the original image data.