Perceptual audio coder

From Wikipedia, the free encyclopedia
Jump to: navigation, search

Perceptual Audio Coder (PAC) is an algorithm, like MPEG's MP3 standard, used to compress digital audio by removing extraneous information not perceived by most people. It is used by Sirius Satellite Radio for their DARS service.

The PAC originally developed by James Johnston and Anibal Ferriera [1] at AT&T's Bell Labs, has a flexible format and bitrate. It provides efficient compression of high-quality audio over a variety of formats from 16 kbit/s for a monophonic channel to 1024 kbit/s for a 5.1 format with four or six auxiliary audio channels, and provisions for an ancillary (fixed rate) and auxiliary (variable rate) side data channel. For stereo audio signals, it is claimed that it provides near compact disc (CD) quality at about 56-64 kbit/s, with transparent coding at bit rates approaching 128 kbit/s[citation needed]. In the 1993, ISO-MPEG-2 5-channel test, PAC demonstrated the best decoded audio signal quality available from any algorithm at 320 kbit/s, far outperforming all algorithms, including the layer-II and layer-III backward compatible algorithms.[citation needed]

Over the years PAC has evolved considerably. A known software implementation of this codec is CelestialTech's AudioLib. Later, it was considerably improved and renamed to ePAC (enhanced Perceptual Audio Coder) by Lucent. It's available in the AudioVeda music library manager.

iBiquity initially tested PAC for the HD-Radio IBOC digital radio upgrade for FM and AM, but chose an MPEG4-derived codec, HE-AAC, instead. It is interesting to note that MPEG-2 AAC is substantially similar to the original AT&T PAC algorithm written by Johnston and Ferreira, including the specifics of stereo pair coding, bitstream sectioning, handling of 1 or 2 channels at a time, multiple codebooks responding to the same Largest Absolute Value, and block switching triggers. The version of PAC tested for the MPEG-NBC (later to become AAC) trials used 1024/128 sample block lengths, rather than 512/128 sample block lengths.

The non-capitalized term is used to describe lossy compression codecs that utilize perceptual coding techniques for audio.

  1. ^ Johnston, J. D. and Ferreira, A. J., (March 1992). Sum-difference stereo transform coding, from ICASSP 1992 Transactions.