When you offer sound files for download on your Web site, you need to make them small enough that your site users can download them. To do this, you need to compress them, but you want your users to experience high-quality sound.

The human ear is really, really good at distinguishing among different sounds. We can usually spot “foreigners” — whether from another country or maybe just from another neighborhood — by their accents in a second or two, or recognize a favorite tune in just a couple of notes.

This is because there’s a tremendous amount of information in the sounds we can hear. To capture this information accurately for computer storage and transmission takes a lot of cleverness not to mention bandwidth. A relatively uncompressed file format such as the Windows WAV format and the roughly comparable Apple AIFF format take about 170 KB for each second of sound, or about 10 MB per minute. Even a fast Internet connection can’t deliver “live” WAV sound; users must wait anywhere from several seconds to many minutes before they can hear the sound, depending on how fast the connection is and how long the sound file is.

The most popular compressed file format, the MP3 format, using tricks of psychology and physics to capture music in a compressed file, takes about 17 KB per second of sound, or about 1 MB per minute, one-tenth the bit rate of a WAV file. This is just within the capability of most Internet connections to play back an MP3 live, or nearly live, or to download the file within a few minutes for later playback.

Despite its cleverness and success, MP3 may not be as aesthetically successful as its pictorial equivalent, JPEG. Why not? After all, the slight coldness and tinniness of MP3 files is probably, objectively, no worse than the slight visual “noise” found in moderately compressed JPEG files. But because sound is so emotionally involving, MP3 files are perhaps more likely than JPEGS to be experienced as not-quite-good-enough. Two indicators of this dissatisfaction: Vinyl record albums seem to be experiencing a slight resurgence, and there’s some support for more precise audio-file formats, despite the cost in file size and transmission time.

Not every computer out there has all the necessary elements to play back sounds in a sound file — a program to decode the sound and play it, sound hardware to deliver sound to output(s), and speakers and/or a headphone jack to actually produce sound the user can hear. Where all the elements are in place, they can be turned off (or the headphones not plugged in), and users may be anywhere from slightly disinclined to absolutely unwilling to turn on the sound gear just to hear the sound you want them to listen to.