a file, but no fading. Some audio editing programs do edit MPEG files, but they actually convert the file to PCM16 first, and then convert it back again after editing. This has two problems. First, it is slow, and second, each time the conversion is done, the sound quality is degraded. MPEG is only valid for 32, 44.1 and 48 KHz. MPEG requires a great deal of DSP processing power. As a result, the
ADPCM is typically 4 to 1 compression, however, lower sample rates are possible. Therefore, an ADPCM file recorded at 8 KHz will use half the disk space of an MPEG file recorded at 32 KHz with 8 to 1 compression. The sound quality will be much worse, but if the application is to record a conversation for archive pur- poses, the sound quality may be adequate. ADPCM is also easier to edit.
In a network environment, where all sound files are stored on a server and sent to the individual computers via a network, the data rate is an issue of concern. If the network cannot send the data from the file on the server to the computer fast enough, there will be objectionable “dropouts” or “pops and clicks” in the audio. Using a compressed file format will allow for more network traffic before this problem occurs. Also, compressed file formats are advantageous if the files must be sent by modem or emailed somewhere.
ISO/MPEG-1 BITRATES
ISO/MPEG supports several compression bitrates. (The bitrates in the Antex Demo program are selected by
17