As mentioned earlier in this book, encoding is the process of converting a stream of uncompressed digital audio to a compressed format. The mathematical process used for encoding and decoding is referred to as a CODEC.
MPEG Audio uses perceptual encoding (a type of lossy compression) to remove parts of the signal that most people can’t hear. The encoder also applies standard lossless data-compression techniques to compress the audio even more. The amount of information discarded, and therefore the sound quality, is dependent on parameters (such as bit-rate and sampling rate) that are chosen by the creator. Chapter 12 covers the effects of these parameters in more detail.
Perceptual encoding does not work perfectly, because the sensitivity of each person’s hearing is different. But the sensitivity of human hearing does fall within a finite range, and thus researchers can determine a range that applies to the vast majority of people.
Figure 10-1 shows the process used to encode uncompressed audio into an MPEG format. First, an uncompressed PCM audio signal is converted to AAC or MP3 by filtering the signal into several sub-bands and applying a “psychoacoustic” algorithm. The encoded audio is then packaged into frames, and ancillary data such as ID3 tag information and graphics is added.
Figure 10-1. MPEG encoder diagram
A perceptual encoder divides the incoming ...