Sound Files

Optimizing sound for the Web


Choosing the file format is pretty straight forward depending on what you're trying to achieve. MIDI(.mid) files are small and quick, but since MIDI is just a pure data stream, it can only sound like the MIDI board that your computer has, which is or mimics the General MIDI Module. Which means to say your music will sound like the cantina scene in Star Wars. Or, as Thomas Dolby put it, "like a squished bug".

WAVE(.wav) files, on the other hand , can sound like anything you want, from music to speech to sound effects. Unfortunately, they are also notoriously fat hungry beasts that gobble up both memory and processing speed. A file just a few seconds long can end up being 500K or more. Not Internet friendly, to say the least. But until the bandwidth issue is solved, it's something we must work with. The file size can be managed and minimized by the format the sound sample is before it is converted.

Trading off download time and sound quality

The format of the audio involves a number of things:

1) The number of channels (mono or stereo)
2) The sample rate
3) The data format (8 bit audio, 16 bit audio, or some compressed format)

For sounds that are to be localized in a VRML world, it's suggested that you start with good quality mono sounds that have been sampled at 16 bit resolution. For speech sounds, you can usually get by with 8 bit. The sample rate can vary depending on the type of sound. Generally, the higher the sample rate, the better the quality and the longer the download time. Music and sound effects sound decent at 22050 samples per second, and tolerable at 11025 samples per second.

Once you have the mono sounds recorded at 16 bit quality and the right sample rate, you need to decide if you should compress them. By selecting a good compression algorithm, you can almost always get better sound quality and a smaller file size than if you just drop the sample rate and use 8 bit audio. On the other hand, you need to pick a compression format that will be understood by the platform that you are playing back on. Of course, the major advantage of the Internet and VRML is their cross-platform capability and open standards/open protocols structure. Compression can limit your audience, so if you use compression, be sure it's suited to your particular project and goals.

The real key to being an Internet sound designer is just like polygon counting in VRML- keep it small! You really don't need that much time to create that little groove or effect you're looking for. Be creative!

< < < < _ > > > >