Boston University Medical Campus
Internet and Technology Help

Audio Files

In the right situation, sounds and audio can greatly enhance a Web page, however, when dealing with audio on the Internet, several consessions must be made. The first problem stems from the fairly ‘chaotic’ nature of audio, which is made up of overlapping pressure waves that our brain interprets as sound. When sound is digitized by an audio card, the computer is sampling, or measuring, the input thousands of timeseach second, and recording the data. For example, audio CD’s are created by sampling over 44 thousand times per second, allowing for reproduction of the frequencies audible by the human ear. Unfortunately, the data collected is fairly “random”, and thus, hard to compress yet retain complete fidelity.

Besides finding a method other than digitizing the audio (such as Fourier Transformation, which is very slow to perform even on fast computers, or wavelet compression which is in development), one solution to the compression hurdle is to sample the audio at lower rates, such as 8 thousand times per second (8KHz), opposed to 44KHz. This will cut the file size down to a fraction of the original, and yields data which is more conducive to compression. When saving the data to a file, you can store the numbers in a highly precice 16-bit format, or a smaller and less precise 8-bit format. Unfortunately, if you use 8 (or less) KHz 8-bit recording, fidelity will be lost (since you are recording less data and saving it less precisely), which can result in poor quality recordings of music and other “rich” sounds. Human voice, however, is usually well reproduced at this sampling rate (The U.S. telephone system, which has been digital since the late 1960′s, is based around a dynamic 8KHz compression scheme).

There are many audio compression formats available which can provide differing levels of fidelty based upon their minimum sampling and bit rate. Modern browsers and operating systems can support playing, and other programs recording, the most popular formats which include:

For most of these formats, you would create a link to the URL of the audio file, such as <a href="hello.au">Click to say 'Hello'</a>. The client’s browser would then be responsible for transferring the file and launching the appropriate helper application in order to play the audio. Some formats, however, would require the installation of the proper plug-in extension in your browser, and then provide playback within the Web page (Check with your plug-in manufacturer on how to embed the audio file within the page).

Progressive Audio

While lower sampling rates may improve the level of compression achieved, digitizing even at 8Khz can still result in file sizes approaching one megabyte per minute. This presents a funtional limit on the amount of audio which users will be patient enough to download, since the above formats must be fully transferred before they can be played.

New progressive, or streaming, audio formats provide improved compression and allow for the client browser to begin playing the file while it is being transferred, even over slow 14.4Kbps modems. However, you will probably need to first download the necessary proprietary plug-in or player software, assuming it exists for your computer platform and browser. Creating, or encoding, your own audio files may require purchase of additional software from the respective company. Current formats include: