This preview shows page 1. Sign up to view the full content.
Unformatted text preview: PEG audio standard (known as
perceptual coding) takes advantage of this perceptual weakness of human ears
(pshychoacoustic phenomena). In perceptual coding, the audio spectrum is divided
into a set of a narrow frequency bands, called critical bands, to reflect the
frequency selectivity of human hearing. Then a filtering mechanism is used to
filter coding noise sharply to force it to stay close to the frequency of the
frequency components of the audio signal b6ing coded thereby masking out the
noise. By reducing the coding noise when no audio is present and allowing strong
audio to mask the noise at other times, the sound fidelity of the original signals can
be perceptually perceived.
The basic steps of MPEG audio (MPEG-1 audio, to be more precise) encoding,
shown in Figure 19.14(a), are as follows:
1. The audio signal to be compressed is first transformed from the time domain
to the frequency domain.
The resultant audio spectrum is then passed through a filter bank, which
divides it into multiple sub-bands.
3. The input audio signal is simultaneously subjected to a psychoacoustic model,
which exploits auditory masking and determines the signal-to-noise ratio of each
4. Each sub-band is then quantized using a quantizer. A bit/noise allocation
algorithm is used to determine the number of code bits allocated to quantize each
sub-band so that the audibility of quantization noise is minimized.
5. The quantized samples are then formatted into a series of bit streams, which is
the encoded bit stream for the input audio signal.
As shown in Figure 19.14(b), the decoder performs the inverse operation of the
encoder for reconstructing the time-domain audio signal from the encoded bit
To provide different quality and different bit rates, the standard supports three
distinct layers for compression layer I is the most basic and simplest coding
algorithm, and layers II and III are enhancements of layer I. Recently, a new version of MPEG audio (MPEG-2 audio) has been developed for
low-bit-rate coding of multichannel audio. It has 5.1 channels of audio, including
three front channels (i.e., left, center, right) plus two surrounds, and a lowfrequency bass-effect channel called th...
View Full Document
This document was uploaded on 04/07/2014.
- Spring '14