There are numerous implementations of perceptual coders in use in commercial products. Sonys ATRAC, used in the MiniDisc system, and the famous MPEG-1 Layer I, II and III [1] are well known examples . MPEG-2 AAC (Advanced Audio Coding, also called NBC, non-backwards-compatible) [2], and Dolby's AC-3 are currently considered top-of-the-line. The differences between the different coders are in the human audio perceptual model, the type of subband coding, and becoming more important, in special tricks to handle certain artifacts. An example of such a trick is pre-echo detection, which is used to make transient sounds like castanets sound better.
Some examples of the current implementations will be presented here, ordered after the type of subband splitting they use. All perceptual audio coders use some kind of subband splitting (or almost all, see e.g [3]). Any linear subband splitting can be interpreted as linear transforms, though, the the difference is really in the name of the method.