Файл: Audio and Radio Engineering.pdf

7-1

Section

7

Compression Technologies for Audio

A number of existing and proposed audio compression systems employ a combination of pro-
cessing techniques. Any scheme that becomes widely adopted can enjoy economies of scale and
reduced market confusion. Timing, however, is critical to market acceptance of any standard. If a
standard is selected well ahead of market demand, more cost-effective or higher-performance
approaches may become available before the market takes off. On the other hand, a standard may
be merely academic if it is established after alternative schemes already have become well
entrenched in the marketplace.

These factors have placed a great deal of importance on the standards-setting activities of

leading organizations around the world. It has been through hard work, inspiration, and even a
little compromise that the various standards have developed and evolved to the levels of utility
and acceptance that they enjoy today.

With these important benchmarks in place, audio industry manufacturers have been able to

focus on implementation of the technology and offering specific user-centered features. Fortu-
nately, the days of the video (and audio) tape “format wars” appear to have passed as the stan-
dards-setting bodies take the lead in product direction and interface.

The function of any audio compression device or system is to provide for efficient storage

and/or transmission of information from one location or device to another. The encoding pro-
cess, naturally, is the beginning point of this chain. Like any chain, audio encoding represents not
just a single link but many interconnected and interdependent links. The bottom line in audio
encoding is to ensure that the compressed signal or data stream represents the information
required for recording and/or transmission, and only that information. If there is additional infor-
mation of any nature remaining in the data stream, it will take bits to store and/or transmit, which
will result in fewer bits being available for the required data. Surplus information is irrelevant
because the intended recipient(s) do not require it and can make no use of it.

Surplus information can take many forms. For example, it can be information in the original

signal or data stream that exceeds the capabilities of the receiving device to process and repro-
duce. There is little point in transmitting more information than the receiving device can use.

Noise is another form of surplus information. Noise is—by nature—random or nearly so, and

this makes it essentially incompressible. Many other types of artifacts exist, ranging from filter
ringing to disc scratches. Some may seem trivial, but in the field of compression they can be very
important. Compression relies on order and consistency for best performance, and such artifacts

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)

Copyright © 2004 The McGraw-Hill Companies. All rights reserved.

Any use is subject to the Terms of Use as given at the website.

Source: Standard Handbook of Audio and Radio Engineering

7-2 Section Seven

can compromise the final reproduction or at least lower the achievable bit rate reduction. Gener-
ally speaking, compression systems are designed for particular tasks, and make use of certain
basic assumptions about the nature of the data being compressed.

Such requirements have brought about a true “systems approach” to compression. From the

algorithm to the input audio, every step must be taken with care and precision for the overall
product to be of high quality.

These forces are shaping the audio and video technologies of tomorrow. Any number of sce-

narios have been postulated as to the hardware and software that will drive the digital facility of
the future. One thing is certain, however: It will revolve around compressed audio and video sig-
nals.

In This Section:

Chapter 7.1: Audio Compression Systems

7-7

Introduction

7-7

PCM Versus Compression

7-7

Audio Bit Rate Reduction

7-8

Redundancy and Irrelevancy

7-8

Human Auditory System

7-8

Quantization

7-9

Sampling Frequency and Bit Rate

7-10

Prediction and Transform Algorithms

7-11

Subband Coding

7-11

Subband Gain

7-11

APCM Coding

7-12

Processing and Propagation Delay

7-13

Bit Rate and Compression Ratio

7-14

Bit Rate Mismatch

7-14

Editing Compressed Data

7-14

Common Audio Compression Techniques

7-14

apt-X100

7-15

ISO/MPEG-1 Layer 2

7-16

MPEG-2 AAC

7-20

MPEG-4

7-21

Dolby E Coding System

7-23

Architectural Overview

7-25

Coded Frame Format

7-25

Objective Quality Measurements

7-26

Perspective on Audio Compression

7-28

References

7-29

Bibliography

7-30

Chapter 7.2: ATSC DTV System Compression Issues

7-31

Introduction

7-31

MPEG-2 Layer Structure

7-31

Slices

7-32

Pictures, Groups of Pictures, and Sequences

7-33

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)

Copyright © 2004 The McGraw-Hill Companies. All rights reserved.

Any use is subject to the Terms of Use as given at the website.

Compression Technologies for Audio

Compression Technologies for Audio 7-3

I-Frames

7-34

P-Frames

7-34

B-Frames

7-34

Motion Estimation

7-35

Vector Search Algorithm

7-35

Motion-Vector Precision

7-35

Motion-Vector Coding

7-35

Encoder Prediction Loop

7-35

Spatial Transform Block—DCT

7-37

Quantizer

7-37

Entropy Coder

7-38

Inverse Quantizer

7-39

Inverse Spatial Transform Block—IDCT

7-39

Motion Compensator

7-39

Dual Prime Prediction Mode

7-39

Adaptive Field/Frame Prediction Mode

7-40

Image Refresh

7-40

Periodic Transmission of I-Frames

7-40

Progressive Refresh

7-41

Discrete Cosine Transform

7-41

Blocks of 8 

×

 8 Pixels

7-41

Adaptive Field/Frame DCT

7-42

Adaptive Quantization

7-42

Perceptual Weighting

7-42

Entropy Coding of Video Data

7-43

Huffman Coding

7-43

Run Length Coding

7-44

Channel Buffer

7-45

Decoder Block Diagram

7-45

Spatial and S/N Scalability

7-45

References

7-46

Chapter 7.3: DTV Audio Encoding and Decoding

7-47

Introduction

7-47

AES Audio

7-47

AES3 Data Format

7-49

SMPTE 324M

7-50

Audio Compression

7-50

Encoding

7-51

Decoding

7-53

Implementation of the AC-3 System

7-53

Audio-Encoder Interface

7-54

Sampling Parameters

7-55

Output Signal Specification

7-55

Operational Details of the AC-3 Standard

7-56

Transform Filterbank

7-57

Window Function

7-57

Time-Division Aliasing Cancellation Transform

7-57

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)

Copyright © 2004 The McGraw-Hill Companies. All rights reserved.

Any use is subject to the Terms of Use as given at the website.

Compression Technologies for Audio

7-4 Section Seven

Transient Handling

7-57

Coded Audio Representation

7-57

Exponent Coding

7-58

Mantissas

7-58

Bit Allocation

7-58

Forward Adaptive

7-59

Rematrixing

7-59

Coupling

7-59

Bit Stream Elements and Syntax

7-60

Splicing and Insertion

7-60

Error-Detection Codes

7-61

Loudness and Dynamic Range

7-61

Dynamic Range Compression

7-61

Encoding the AC-3 Bit Stream

7-61

Input Word Length/Sample Rate

7-62

AC-3/MPEG Bit Stream

7-62

Decoding the AC-3 Bit Stream

7-64

Continuous or Burst Input

7-64

Synchronization and Error Detection

7-64

Decoding Components

7-66

Algorithmic Details

7-67

Bit Allocation

7-67

Audio System Level Control

7-68

Dialogue Normalization

7-68

Example Situation

7-69

Dynamic Range Compression

7-69

Example Situation

7-70

Heavy Compression

7-71

Audio System Features

7-72

Complete Main Audio Service (CM)

7-72

Main Audio Service, Music and Effects (ME)

7-72

Visually Impaired (VI)

7-73

Hearing Impaired (HI)

7-73

Dialogue (D)

7-73

Commentary (C)

7-74

Emergency (E)

7-74

Voice-Over (V0)

7-74

Multilingual Services

7-75

Channel Assignments and Levels

7-75

References

7-77

Bibliography

7-78

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)

Copyright © 2004 The McGraw-Hill Companies. All rights reserved.

Any use is subject to the Terms of Use as given at the website.

Compression Technologies for Audio

Compression Technologies for Audio 7-5

On the CD-ROM:

•

ATSC, “Digital Audio Compression Standard (AC-3),” Advanced Television Systems Com-
mittee, Washington, D.C., Doc. A/52, Dec. 20, 1995.

•

ATSC, “Digital Television Standard,” Advanced Television Systems Committee, Washington,
D.C., Doc. A/53, Sept.16, 1995.

•

ATSC, “Guide to the Use of the Digital Television Standard,” Advanced Television Systems
Committee, Washington, D.C., Doc. A/54, Oct. 4, 1995.

•

The Tektronix publication A Guide to MPEG Fundamentals and Protocol Analysis as an
Acrobat (PDF) file. This document, copyright 1997 by Tektronix, provides a detailed discus-
sion of MPEG as applied to DTV and DVB, and quality analysis requirements and measure-
ment techniques for MPEG-based systems.

Reference Documents for this Section

Brandenburg, K., and Gerhard Stoll: “ISO-MPEG-1 Audio: A Generic Standard for Coding of

High Quality Digital Audio,” 92nd AES Convention Proceedings, Audio Engineering Soci-
ety, New York, N.Y., 1992, revised 1994.

Ehmer, R. H.: “Masking Patterns of Tones,” J. Acoust. Soc. Am., vol. 31, pp. 1115–1120, August

1959.

Fibush, David K.: “Testing MPEG-Compressed Signals,” Broadcast Engineering, Overland

Park, Kan., pp. 76–86, February 1996.

Herre, J., and B. Grill: “MPEG-4 Audio—Overview and Perspectives for the Broadcaster,” IBC

2000 Proceedings, International Broadcast Convention, Amsterdam, September 2000.

IEEE Standard Dictionary of Electrical and Electronics Terms, ANSI/IEEE Standard 100-1984,

Institute of Electrical and Electronics Engineers, New York, 1984.

ITU-R Recommendation BS-775, “Multi-channel Stereophonic Sound System with and Without

Accompanying Picture.”

Lyman, Stephen, “A Multichannel Audio Infrastructure Based on Dolby E Coding,” Proceedings

of the NAB Broadcast Engineering Conference, National Association of Broadcasters,
Washington, D.C., 1999.

Moore, B. C. J., and B. R. Glasberg: “Formulae Describing Frequency Selectivity as a Function

of Frequency and Level, and Their Use in Calculating Excitation Patterns,” Hearing
Research, vol. 28, pp. 209–225, 1987.

Robin, Michael, and Michel Poulin: Digital Television Fundamentals, McGraw-Hill, New York,

N.Y., 1998.

SMPTE Standard for Television: “12-Channel Serial Interface for Digital Audio and Auxiliary

Data,” SMPTE 324M (Proposed), SMPTE, White Plains, N.Y., 1999.

SMPTE Standard for Television: “Channel Assignments and Levels on Multichannel Audio

Media,” SMPTE 320M-1999, SMPTE, White Plains, N.Y., 1999.

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)

Copyright © 2004 The McGraw-Hill Companies. All rights reserved.

Any use is subject to the Terms of Use as given at the website.

Compression Technologies for Audio