ISO/IEC 14496-3:2005/Amd 2:2006
(Amendment)Information technology - Coding of audio-visual objects - Part 3: Audio - Amendment 2: Audio Lossless Coding (ALS), new audio profiles and BSAC extensions
Information technology - Coding of audio-visual objects - Part 3: Audio - Amendment 2: Audio Lossless Coding (ALS), new audio profiles and BSAC extensions
Technologies de l'information — Codage des objets audiovisuels — Partie 3: Codage audio — Amendement 2: Codage audio sans perte (ALS), nouveaux profils audio et extensions BSAC
General Information
Relations
Frequently Asked Questions
ISO/IEC 14496-3:2005/Amd 2:2006 is a standard published by the International Organization for Standardization (ISO). Its full title is "Information technology - Coding of audio-visual objects - Part 3: Audio - Amendment 2: Audio Lossless Coding (ALS), new audio profiles and BSAC extensions". This standard covers: Information technology - Coding of audio-visual objects - Part 3: Audio - Amendment 2: Audio Lossless Coding (ALS), new audio profiles and BSAC extensions
Information technology - Coding of audio-visual objects - Part 3: Audio - Amendment 2: Audio Lossless Coding (ALS), new audio profiles and BSAC extensions
ISO/IEC 14496-3:2005/Amd 2:2006 is classified under the following ICS (International Classification for Standards) categories: 35.040 - Information coding; 35.040.40 - Coding of audio, video, multimedia and hypermedia information. The ICS classification helps identify the subject area and facilitates finding related standards.
ISO/IEC 14496-3:2005/Amd 2:2006 has the following relationships with other standards: It is inter standard links to ISO/IEC 14496-3:2005, ISO/IEC 14496-3:2009; is excused to ISO/IEC 14496-3:2005. Understanding these relationships helps ensure you are using the most current and applicable version of the standard.
You can purchase ISO/IEC 14496-3:2005/Amd 2:2006 directly from iTeh Standards. The document is available in PDF format and is delivered instantly after payment. Add the standard to your cart and complete the secure checkout process. iTeh Standards is an authorized distributor of ISO standards.
Standards Content (Sample)
INTERNATIONAL ISO/IEC
STANDARD 14496-3
Third edition
2005-12-01
AMENDMENT 2
2006-03-15
Information technology — Coding of
audio-visual objects —
Part 3:
Audio
AMENDMENT 2: Audio Lossless Coding
(ALS), new audio profiles and BSAC
extensions
Technologies de l'information — Codage des objets audiovisuels —
Partie 3: Codage audio
AMENDEMENT 2: Codage audio sans perte (ALS), nouveaux profils
audio et extensions BSAC
Reference number
ISO/IEC 14496-3:2005/Amd.2:2006(E)
©
ISO/IEC 2006
ISO/IEC 14496-3:2005/Amd.2:2006(E)
PDF disclaimer
This PDF file may contain embedded typefaces. In accordance with Adobe's licensing policy, this file may be printed or viewed but
shall not be edited unless the typefaces which are embedded are licensed to and installed on the computer performing the editing. In
downloading this file, parties accept therein the responsibility of not infringing Adobe's licensing policy. The ISO Central Secretariat
accepts no liability in this area.
Adobe is a trademark of Adobe Systems Incorporated.
Details of the software products used to create this PDF file can be found in the General Info relative to the file; the PDF-creation
parameters were optimized for printing. Every care has been taken to ensure that the file is suitable for use by ISO member bodies. In
the unlikely event that a problem relating to it is found, please inform the Central Secretariat at the address given below.
© ISO/IEC 2006
All rights reserved. Unless otherwise specified, no part of this publication may be reproduced or utilized in any form or by any means,
electronic or mechanical, including photocopying and microfilm, without permission in writing from either ISO at the address below or
ISO's member body in the country of the requester.
ISO copyright office
Case postale 56 • CH-1211 Geneva 20
Tel. + 41 22 749 01 11
Fax + 41 22 749 09 47
E-mail copyright@iso.org
Web www.iso.org
Published in Switzerland
ii © ISO/IEC 2006 – All rights reserved
ISO/IEC 14496-3:2005/Amd.2:2006(E)
Foreword
ISO (the International Organization for Standardization) and IEC (the International Electrotechnical
Commission) form the specialized system for worldwide standardization. National bodies that are members of
ISO or IEC participate in the development of International Standards through technical committees
established by the respective organization to deal with particular fields of technical activity. ISO and IEC
technical committees collaborate in fields of mutual interest. Other international organizations, governmental
and non-governmental, in liaison with ISO and IEC, also take part in the work. In the field of information
technology, ISO and IEC have established a joint technical committee, ISO/IEC JTC 1.
International Standards are drafted in accordance with the rules given in the ISO/IEC Directives, Part 2.
The main task of the joint technical committee is to prepare International Standards. Draft International
Standards adopted by the joint technical committee are circulated to national bodies for voting. Publication as
an International Standard requires approval by at least 75 % of the national bodies casting a vote.
Attention is drawn to the possibility that some of the elements of this document may be the subject of patent
rights. ISO and IEC shall not be held responsible for identifying any or all such patent rights.
Amendment 2 to ISO/IEC 14496-3:2005 was prepared by Joint Technical Committee ISO/IEC JTC 1,
Information technology, Subcommittee SC 29, Coding of audio, picture, multimedia and hypermedia
information.
This amendment specifies the Audio Lossless Coding (ALS) scheme. The amendment further defines a new
profile, the High Efficiency AAC v2 Profile, that incorporates all the features of the High Efficiency AAC Profile
and in addition the Parametric Stereo tool. The amendment also specifies the way in which the audio object
type ER BSAC is extended to support multi-channel format, providing backward compatibility.
© ISO/IEC 2006 – All rights reserved iii
ISO/IEC 14496-3:2005/Amd.2:2006(E)
Information technology — Coding of audio-visual objects —
Part 3:
Audio
AMENDMENT 2: Audio Lossless Coding (ALS), new audio profiles
and BSAC extensions
In the Introduction, at the end of subclause "Lossless Audio Coding Tools", add:
MPEG-4 ALS (Audio Lossless Coding) provides lossless coding of digital audio signals. Input signals can be
integer PCM data with 8 to 32-bit word length or 32-bit IEEE floating-point data. Up to 65536 channels are
supported.
In Part 3: Audio, Subpart 1, in subclause 1.3 Terms and Definitions, add:
ALS: Audio Lossless Coding
and increase the index-number of subsequent entries.
© ISO/IEC 2006 – All rights reserved 1
ISO/IEC 14496-3:2005/Amd.2:2006(E)
In Part 3: Audio, Subpart 1, in subclause 1.5.1.1 Audio object type definition, replace table 1.1 with the table
below:
Table 1.1 — Audio Object Type definition based on Tools/Modules
Audio Object
Type
0 Null
1 AAC main X X X X X X X X X X 2)
2 AAC LC X X X X X X X X X
3 AAC SSR X X X X X X X X X X
4 AAC LTP X X X X X X X X X X 2)
5 SBR X
6 AAC Scalable X X X X X X X X X X X X 6)
7 TwinVQ X X X X X X X
8 CELP X
9 HVXC X
10 (reserved)
11 (reserved)
12 TTSI X
13 Main X X X 3)
synthetic
14 Wavetable X X 4)
synthesis
15 General MIDI X
16 Algorithmic X
Synthesis and
Audio FX
17 ER AAC LC X X X X X X X X X X X
18 (reserved)
19 ER AAC LTP X X X X X X X X X X X X 5)
20 ER AAC X X X X X X X X X X X X X X 6)
scalable
21 ER TwinVQ X X X X X X X X
22 ER BSAC X X X X X X X X X X
23 ER AAC LD X X X X X X X X X X X
24 ER CELP X X X X
25 ER HVXC X X X X
26 ER HILN X X X
27 ER X X X X X
Parametric
28 SSC X X
29 PS X X
30 (reserved)
31 (escape)
32 Layer-1 X
33 Layer-2 X
34 Layer-3 X
35 DST X
36 ALS X
37 - (reserved)
In Part 3: Audio, Subpart 1, in subclause 1.5.1.2 Description, add:
1.5.1.2.30 ALS object type
The ALS object type is the counterpart of the Audio Lossless Coding (ALS) scheme and contains the
corresponding ALS tools.
2 © ISO/IEC 2006 – All rights reserved
Object Type ID
gain control
block switching
window shapes - standard
window shapes – AAC LD
filterbank - standard
filterbank - SSR
TNS
LTP
intensity
coupling
frequency deomain prediction
PNS
MS
SIAQ
FSS
upsampling filter tool
quantisation&coding - AAC
quantisation&coding – TwinVQ
quantisation&coding - BSAC
AAC ER Tools
ER payload syntax
EP Tool 1)
CELP
Silence Compression
HVXC
HVXC 4kbit/s VR
SA tools
SASBF
MIDI
HILN
TTSI
SBR
Layer-1
Layer-2
Layer-3
SSC (Transient, Sinusoid, Noise)
Parametric stereo
DST
ALS
Remark
ISO/IEC 14496-3:2005/Amd.2:2006(E)
In Part 3: Audio, Subpart 1, replace Table 1.3 (Audio Profiles definition) with the following table:
Table 1.3 – Audio Profiles definition
Mobile
High High
High Low
AAC
Main Scalable Speech Syntheti Natural Audio
Object Audio Object Quality Delay Efficiency Efficiency
Profile
Audio Audio Audio c Audio Audio Internet-
Type ID Type Audio Audio AAC AAC v2
Profile Profile Profile Profile Profile working
Profile Profile Profile Profile
Profile
0 Null
1 AAC main X X
2 AAC LC X X X X X X X
3 AAC SSR X X
4 AAC LTP X X X X
5 SBR X X
6 AAC Scalable X X X X
7 TwinVQ X X X
8 CELP X X X X X X
9 HVXC X X X X X
10 (reserved)
11 (reserved)
12 TTSI X X X X X X
13 Main X X
synthetic
14 Wavetable
synthesis
15 General MIDI
16 Algorithmic
Synthesis and
Audio FX
17 ER AAC LC X X X
18 (reserved)
19 ER AAC LTP X X
20 ER AAC X X X
Scalable
21 ER TwinVQ X X
22 ER BSAC X X
23 ER AAC LD X X X
24 ER CELP X X X
25 ER HVXC X X
26 ER HILN X
27 ER X
Parametric
28 SSC
29 PS X
30 (reserved)
31 (escape)
32 Layer-1
33 Layer-2
34 Layer-3
35 DST
36 ALS
In Part 3: Audio, Subpart 1, subclause 1.5.2.3 (Levels within the profiles), add at the end:
• Levels for the High Efficiency AAC v2 Profile
© ISO/IEC 2006 – All rights reserved 3
ISO/IEC 14496-3:2005/Amd.2:2006(E)
Table 1.11A - Levels for the High Efficiency AAC v2 Profile
Level Max. Max. AAC
Max. AAC Max. SBR Max. PCU Max. RCU Max. PCU Max. RCU
channels/ sampling sampling sampling rate HQ / LP HQ / LP
object rate, SBR not rate, SBR [kHz] (in/out) SBR SBR
present [kHz] present [kHz] (Note 5) (Note 5)
1 NA NA NA NA NA NA NA NA
2 2 48 24 24/48 (Note 9 10 9 10
1)
3 2 48 24/48 (Note 48/48 (Note 15 10 15 10
3) 2)
4 5 48 24/48 (Note 48/48 (Note 25 28 20 23
4) 2)
5 5 96 48 48/96 49 28 39 23
Note 1: A level 2 HE AAC v2 Profile decoder implements the baseline version of the parametric stereo tool.
Higher level decoders shall not be limited to the baseline version of the parametric stereo tool.
Note 2: For level 3 and level 4 decoders, it is mandatory to operate the SBR tool in downsampled mode if the
sampling rate of the AAC core is higher than 24kHz. Hence, if the SBR tool operates on a 48kHz AAC signal,
the internal sampling rate of the SBR tool will be 96kHz, however, the output signal will be downsampled by
the SBR tool to 48kHz.
Note 3: If Parametric Stereo data is present the maximum AAC sampling rate is 24kHz, if Parametric Stereo
data is not present the maximum AAC sampling rate is 48kHz.
Note 4: For one or two channels the maximum AAC sampling rate, with SBR present, is 48kHz. For more
than two channels the maximum AAC sampling rate, with SBR present, is 24kHz.
Note 5: The PCU/RCU number are given for a decoder operating the LP SBR tool whenever applicable.
A HE AAC v2 Profile decoder of a certain level shall operate the HQ SBR tool for streams containing
Parametric Stereo data. For streams not containing Parametric Stereo data, the HE AAC v2 Profile decoder
may operate the HQ SBR tool, or the LP SBR tool.
In Part 3: Audio, Subpart 1, subclause 1.5.2.4 (Table 1.12 - audioProfileLevelIndication Values), replace the
row:
0x30-0x7F reserved for ISO use -
with:
0x28 AAC Profile L1
0x29 AAC Profile L2
0x2A AAC Profile L4
0x2B AAC Profile L5
0x2C High Efficiency AAC Profile L2
0x2D High Efficiency AAC Profile L3
0x2E High Efficiency AAC Profile L4
0x2F High Efficiency AAC Profile L5
0x30 High Efficiency AAC v2 Profile L2
0x31 High Efficiency AAC v2 Profile L3
0x32 High Efficiency AAC v2 Profile L4
0x33 High Efficiency AAC v2 Profile L5
0x34-0x7F reserved for ISO use -
4 © ISO/IEC 2006 – All rights reserved
ISO/IEC 14496-3:2005/Amd.2:2006(E)
In Part 3: Audio, Subpart 1, in subclause 1.6.2.1 AudioSpecificConfig, replace table 1.13 with the table below:
Table 1.13 — Syntax of AudioSpecificConfig()
Syntax No. of bits Mnemonic
AudioSpecificConfig ()
{
audioObjectType = GetAudioObjectType();
samplingFrequencyIndex; 4 bslbf
if ( samplingFrequencyIndex == 0xf ) {
samplingFrequency; 24 uimsbf
}
channelConfiguration; 4 bslbf
sbrPresentFlag = -1;
psPresentFlag = -1;
if ( audioObjectType == 5 ||
audioObjectType == 29) {
extensionAudioObjectType = 5;
sbrPresentFlag = 1;
if ( audioObjectType == 29 ) {
psPresentFlag = 1;
}
extensionSamplingFrequencyIndex; 4 uimsbf
if ( extensionSamplingFrequencyIndex == 0xf ) {
extensionSamplingFrequency; 24 uimsbf
}
audioObjectType = GetAudioObjectType();
}
else {
extensionAudioObjectType = 0;
}
switch (audioObjectType) {
case 1:
case 2:
case 3:
case 4:
case 6:
case 7:
case 17:
case 19:
case 20:
case 21:
case 22:
case 23:
GASpecificConfig();
break:
case 8:
CelpSpecificConfig();
break;
case 9:
HvxcSpecificConfig();
break:
case 12:
TTSSpecificConfig();
break;
© ISO/IEC 2006 – All rights reserved 5
ISO/IEC 14496-3:2005/Amd.2:2006(E)
case 13:
case 14:
case 15:
case 16:
StructuredAudioSpecificConfig();
break;
case 24:
ErrorResilientCelpSpecificConfig();
break;
case 25:
ErrorResilientHvxcSpecificConfig();
break;
case 26:
case 27:
ParametricSpecificConfig();
break;
case 28:
SSCSpecificConfig();
break;
case 32:
case 33:
case 34:
MPEG_1_2_SpecificConfig();
break;
case 35:
DSTSpecificConfig();
break;
case 36:
ALSSpecificConfig();
break;
default:
/* reserved */
}
switch (audioObjectType) {
case 17:
case 19:
case 20:
case 21:
case 22:
case 23:
case 24:
case 25:
case 26:
case 27:
epConfig; 2 bslbf
if ( epConfig == 2 || epConfig == 3 ) {
ErrorProtectionSpecificConfig();
}
if ( epConfig == 3 ) {
directMapping; 1 bslbf
if ( ! directMapping ) {
/* tbd */
}
}
}
if ( extensionAudioObjectType != 5 && bits_to_decode() >= 16 ) {
syncExtensionType; 11 bslbf
if (syncExtensionType == 0x2b7) {
extensionAudioObjectType = GetAudioObjectType();
6 © ISO/IEC 2006 – All rights reserved
ISO/IEC 14496-3:2005/Amd.2:2006(E)
if ( extensionAudioObjectType == 5 ) {
sbrPresentFlag; 1 uimsbf
if (sbrPresentFlag == 1) {
extensionSamplingFrequencyIndex; 4 uimsbf
if ( extensionSamplingFrequencyIndex == 0xf ) {
extensionSamplingFrequency; 24 uimsbf
}
if ( bits_to_decode() >= 12 ) {
syncExtensionType; 11 bslbf
if (syncExtensionType == 0x548) {
psPresentFlag; 1 uimsbf
}
}
}
}
}
}
}
In Part 3: Audio, Subpart 1, in subclause 1.6.2.1 AudioSpecificConfig, add:
1.6.2.1.12 ALSSpecificConfig
Defined in ISO/IEC 14496-3 subpart 11.
In Part 3: Audio, Subpart 1, in subclause 1.6.2.2.1 Overview, replace table 1.15 by the following table:
Table 1.15 – Audio Object Types
Audio Object Type Object definition of elementary stream Mapping of audio payloads to
Type ID payloads and detailed syntax access units and elementary
streams
AAC MAIN 1 ISO/IEC 14496-3 subpart 4 see subclause 1.6.2.2.2.1.2
AAC LC 2 ISO/IEC 14496-3 subpart 4 see subclause 1.6.2.2.2.1.2
AAC SSR 3 ISO/IEC 14496-3 subpart 4 see subclause 1.6.2.2.2.1.2
AAC LTP 4 ISO/IEC 14496-3 subpart 4 see subclause 1.6.2.2.2.1.2
SBR 5 ISO/IEC 14496-3 subpart 4
AAC scalable 6 ISO/IEC 14496-3 subpart 4 see subclause 1.6.2.2.2.1.3
TwinVQ 7 ISO/IEC 14496-3 subpart 4
CELP 8 ISO/IEC 14496-3 subpart 3
HVXC 9 ISO/IEC 14496-3 subpart 2
TTSI 12 ISO/IEC 14496-3 subpart 6
Main synthetic 13 ISO/IEC 14496-3 subpart 5
Wavetable synthesis 14 ISO/IEC 14496-3 subpart 5
General MIDI 15 ISO/IEC 14496-3 subpart 5
Algorithmic Synthesis 16 ISO/IEC 14496-3 subpart 5
and Audio FX
ER AAC LC 17 ISO/IEC 14496-3 subpart 4 see subclause 1.6.2.2.2.1.4
ER AAC LTP 19 ISO/IEC 14496-3 subpart 4 see subclause 1.6.2.2.2.1.4
ER AAC scalable 20 ISO/IEC 14496-3 subpart 4 see subclause 1.6.2.2.2.1.4
ER Twin VQ 21 ISO/IEC 14496-3 subpart 4
ER BSAC 22 ISO/IEC 14496-3 subpart 4
© ISO/IEC 2006 – All rights reserved 7
ISO/IEC 14496-3:2005/Amd.2:2006(E)
ER AAC LD 23 ISO/IEC 14496-3 subpart 4 see subclause 1.6.2.2.2.1.4
ER CELP 24 ISO/IEC 14496-3 subpart 3
ER HVXC 25 ISO/IEC 14496-3 subpart 2
ER HILN 26 ISO/IEC 14496-3 subpart 7
ER Parametric 27 ISO/IEC 14496-3 subpart 2 and 7
SSC 28 ISO/IEC 14496-3 subpart 8
PS 29 ISO/IEC 14496-3 subpart 8
(reserved) 30
(escape) 31
Layer-1 32 ISO/IEC 14496-3 subpart 9
Layer-2 33 ISO/IEC 14496-3 subpart 9
Layer-3 34 ISO/IEC 14496-3 subpart 9
DST 35 ISO/IEC 14496-3 subpart 10
ALS 36 ISO/IEC 14496-3 subpart 11
In Part 3: Audio, Subpart 1, under 1.6.3 Semantics, after 1.6.3.13 extensionAudioObjectType add:
1.6.3.14 psPresentFlag
A one bit field indicating the presence or absence of Parametric Stereo data. The value –1 indicates that the
psPresentFlag was not conveyed in the AudioSpecificConfig(). In this case, a High Efficiency AAC v2 Profile
decoder shall support implicit signaling (see subclause 1.6.6).
In Part 3: Audio, Subpart 1, after 1.6.5 Signaling of SBR, add the following subclause:
1.6.6 Signaling of Parametric Stereo (PS)
1.6.6.1 Generating and Signaling HE AAC + PS Content
The PS tool in combination with the HE AAC coder enables good stereo quality at very low bitrates. At the
same time it allows for compatibility with existing HE AAC-only decoders. However, the output from a HE AAC
decoder will only be mono for a HE AAC v2 stream carrying PS data.
Therefore, depending on the application, a content provider or content creator may want to choose between
the two alternatives given below. In general, the PS data is always embedded in the HE AAC stream in a HE
AAC compatible way (in the sbr_extension element), and PS is a pure post processing step in the decoder.
Therefore, compatibility can be achieved. However, by means of different signaling the content creator can
select between the full-quality mode and the backward compatibility mode as outlined in 1.6.6.1.1 and
1.6.6.1.2.
For the hierarchical profiles, a profile higher in the profile hierarchy is of course able to decode the content of a
profile lower in the profile hierarchy. In Figure 1.0A the hierarchical structure of the AAC, HE AAC and HE
AAC v2 Profile is displayed. The figure shows that a HE AAC Profile decoder is fully capable of decoding any
AAC-Profile stream, given that the HE AAC Profile decoder is of the same or a higher level as indicated in the
AAC Profile stream. Similarly the HE AAC v2 decoder can handle all HE AAC Profile streams as well as all
AAC Profile streams.
8 © ISO/IEC 2006 – All rights reserved
ISO/IEC 14496-3:2005/Amd.2:2006(E)
AAC SBR PS
AAC Profile
High Efficiency AAC Profile
High Efficiency AAC v2 Profile
Figure 1.0A – Hierarchical structure of AAC, HE AAC and HE AAC v2 Profile,
and compatibility between them.
1.6.6.1.1 Ensuring Full Audio Quality of AAC+SBR+PS for the Listener
To ensure that listeners get the full audio quality of AAC+SBR+PS, the stream should indicate the HE AAC v2
Profile and use the explicit, hierarchical signaling (signaling 2.A. as described below), so that it is played by
HE AAC v2 Profile decoders, i.e., PS capable decoders. With regard to HE AAC-only streams or AAC-only
streams, an HE AAC v2 Profile decoder will decode all HE AAC Profile streams and AAC Profile streams of
the appropriate level, as the HE AAC v2 Profile is a superset of the HE AAC Profile and the AAC Profile.
1.6.6.1.2 Achieving Backward Compatibility with Existing HE AAC and AAC Decoders
The aim of this mode is to get all AAC-based and HE AAC-based decoders to play the stream, even if they do
not support the PS tool. Compatible streams can be created using the following two signaling methods:
a) indicate a profile containing SBR (e.g. the HE AAC Profile), but not the HE AAC v2 Profile, and use
the explicit backward compatible signalling (2.B. as described below). This method is recommended
for all MPEG-4 based systems in which the length of the AudioSpecificConfig() is known in the
decoder. As this is not the case for LATM with audioMuxVersion==0 (see clause 1.7), this method
cannot be used for LATM with audioMuxVersion==0. In explicit backward compatible signaling, PS-
specific configuration data is added at the end of the AudioSpecificConfig(). Decoders that do not
know about PS will ignore these parts, while HE AAC v2 Profile decoders will detect its presence and
configure the decoder accordingly.
b) indicate a profile containing SBR (e.g. the HE AAC Profile), but not the HE AAC v2 Profile, and use
implicit signalling. In this mode, there is no explicit indication of the presence of PS data. Instead, HE
AAC v2 Profile decoders shall open two output channels for a stream containing SBR data with
channelConfiguration==1, i.e., a mono stream using a single channel element, and check the
presence of PS data while decoding the stream and use the PS tool if PS data is found. This is
possible because PS can be decoded without PS-specific configuration data if a certain way of
handling decoder number of output channels is obeyed, as described below for HE AAC v2 Profile
decoders.
Both methods lead to the result that, provided that the profile indication indicates a profile supported by the
decoder, the AAC+SBR part of an AAC+SBR+PS streams will be decoded by HE AAC-only decoders, and the
AAC part of an AAC+SBR+PS stream will be decoded by AAC-only decoders. HE AAC v2 decoders will
detect the presence of PS and decode the full quality AAC+SBR+PS stream.
© ISO/IEC 2006 – All rights reserved 9
ISO/IEC 14496-3:2005/Amd.2:2006(E)
1.6.6.2 Implicit and Explicit Signaling of Parametric Stereo
This subclause outlines the different signaling methods of PS, and the decoder behavior for different types of
signaling.
There are several ways to signal the presence of PS data:
1. implicit signaling: If bs_extension_id equals EXTENSION_ID_PS, PS data is present in the
sbr_extension element, and this implicitly signals the presence of PS data. The ability to detect and
decode implicitly signaled PS is mandatory for all High Efficiency AAC v2 Profile (HE AAC v2 Profile)
decoders.
2. explicit signaling: The presence of PS data is signaled explicitly by means of the PS Audio Object
Type and the psPresentFlag in the AudioSpecificConfig(). When explicit signaling of PS is used,
implicit signaling of PS shall not occur. Two different types of explicit signaling are available:
2.A. hierarchical signaling: If the first audioObjectType (AOT) signaled is the PS AOT, the
extensionAudioObjectType is set to SBR, and a second audio object type is signaled which indicates
the underlying audio object type. This signaling method is not backward compatible. This method
may be needed in systems that do not convey the length of the AudioSpecificConfig(), such as LATM
with audioMuxVersion==0, and content authors are encouraged to use it only when thus needed.
2.B. backward compatible signaling: If the extensionAudioObjectType SBR is signaled at the end of
the AudioSpecificConfig(), a psPresentFlag is transmitted at the end of the backward compatible
explicit SBR signaling, indicating the presence or absence of PS data. This method shall only be
used in systems that convey the length of the AudioSpecificConfig(). Hence, it shall not be used for
LATM with audioMuxVersion==0.
For all types of parametric stereo signaling, the channelConfiguration in the audioSpecifcConfig indicates the
number of channels of the underlying AAC coded stream. Hence, if parametric stereo data is available, the
channelConfiguration will be one, indicating a single channel element, while the parametric stereo tool will
produce two output channels based on the single channel element and the parametric stereo data.
Table 1.22A shows the decoder behavior depending on profile and audio object type indication when implicit
or explicit signaling is used.
10 © ISO/IEC 2006 – All rights reserved
ISO/IEC 14496-3:2005/Amd.2:2006(E)
Table 1.22A – PS Signaling and Corresponding Decoder Behavior
Bitstream characteristics Decoder behavior
Profile PS signaling psPresent raw_data_block HE AAC HE AAC v2
indication Flag Profile Profile
Decoders Decoders
High signaling 1, implicit -1 AAC+SBR Play AAC+SBR Play AAC+SBR
Efficiency signaling (Note 1)
AAC Profile (first AOT != PS)
AAC+SBR+PS Play AAC+SBR Play at least
AAC+SBR,
should play
AAC+SBR+PS
(Note 1)
signaling 2.B, 0 AAC+SBR Play AAC+SBR Play AAC+SBR
backwards (Note 2)
compatible explicit
1 AAC+SBR+PS Play AAC+SBR Play at least
signaling
AAC+SBR,
(second AOT ==
should play
SBR)
AAC+SBR+PS
(Note 3)
High signaling 2.A, non- 1 AAC+SBR+PS Undefined Play
Efficiency backwards AAC+SBR+PS
AAC v2 compatible (Note 3)
Profile signaling
(first AOT == PS)
signaling 2.B, 1 AAC+SBR+PS Undefined Play
backwards AAC+SBR+PS
compatible signling (Note 3)
(second AOT ==
SBR)
Note 1: Implicit signaling, assume the presence of PS data in the payload, giving two output channels
for a single channel element.
Note 2: Explicitly signals that there is no PS data, hence no implicit signaling is present.
Note 3: Number of output channels is two for a single channel element containing AAC+SBR+PS
data.
The upper part of Table 1.22A displays bitstream characteristics and decoder behavior if the profile indication
is the High Efficiency AAC Profile. The lower part displays bitstream characteristics and decoder behavior if
the profile indication is the High Efficiency AAC v2 Profile.
1.6.6.3 HE AAC v2 Profile Decoder Behavior in Case of Implicit Signaling
If the presence of PS data is backward compatible implicitly signaled (signaling 1, in the list above) the first
AudioObjectType signaled is not the PS AOT, and the psPresentFlag is not read from the
AudioSpecificConfig(). Hence, the psPresentFlag is set to –1, indicating that implicit signaling of parametric
stereo may occur.
Since a received mono stream will result in a stereo output if Parametric Stereo data is present in the stream,
the HE AAC v2 Profile decoder shall assume that PS data is available and decide the number of output
channels to be two for a single channel element containing SBR data, and thus also possibly PS data. If no
PS data is found the mono output shall be mapped to the two opened channels for every single channel
element.
© ISO/IEC 2006 – All rights reserved 11
ISO/IEC 14496-3:2005/Amd.2:2006(E)
1.6.6.4 HE AAC v2 Profile Decoder Behavior in Case of Explicit Signaling
If the presence of PS data is explicitly signaled (signaling 2, in the list above) the presence of PS data is
backward compatible explicitly signaled (signaling 2.B) or non-backward compatible explicitly signaled
(signaling 2.A).
For the backward compatible explicit signaled (signaling 2.B) the extensionAudioObjectType signaled is the
SBR AOT. The explicit signaling of PS is done by means of the psPresentFlag that can be either zero or one.
If the psPresentFlag is zero, this indicates that PS data is not present, and hence the HE AAC v2 Profile
decoder should not make assumptions on the number of output channels in anticipation of PS data (as in case
of implicit signaling of PS) and instead employ the original channelConfiguration. If the psPresentFlag is one,
PS data is present and the HE AAC v2 Profile decoder shall operate the PS Tool.
For the non-backward compatible explicit signaling of PS (signaling 2.A) the first AudioObjectType signaled is
the PS AOT. The extensionAudioObjectType is assigned the SBR AOT. For this hierarchical explicit signaling,
the psPresentFlag is set to one if the first signaled AOT is the PS AOT. The psPresentFlag is not transmitted
and hence it is not possible to explicitly signal the absence of implicit signaling. Hence, for the hierarchical
explicit signaling of parametric stereo, PS data is always present and the HE AAC v2 Profile decoder shall
operate the PS Tool.
In Part 3: Audio, Subpart 4, in subclause 4.4.2.6 Payloads for the audio object type ER BSAC, replace table
4.33 bsac_raw_data_block with the following table:
Table 4.33 – Syntax of bsac_raw_data_block()
No. of bits Mnemonic
• Syntax
bsac_raw_data_block()
{
bsac_base_element();
layer=slayer_size;
while(data_available() && layer<(top_layer+slayer_size)) {
bsac_layer_element(layer);
layer++;
}
byte_alignment();
if (data_available()) {
zero_code 32 bslbf
syncword 8 bslbf
while( data_available() )
extended_bsac_raw_data_block();
}
}
12 © ISO/IEC 2006 – All rights reserved
ISO/IEC 14496-3:2005/Amd.2:2006(E)
In Part 3: Audio, Subpart 4, in subclause 4.4.2.6 Payloads for the audio object type ER BSAC, after Table
4.43 Syntax of bsac_spectral_data, add the following two tables:
Table 4.35 – Syntax of extended_bsac_raw_data_block()
Syntax No. of bits Mnemonic
extended_bsac_raw_data_block()
{
extended_bsac_base_element();
layer=slayer_size;
while(data_available() && layer<(top_layer+slayer_size)) {
bsac_layer_element(layer);
layer++;
}
byte_alignment();
}
Table 4.36 – Syntax of extended_bsac_base_element()
Syntax No. of bits Mnemonic
extended_bsac_base_element()
{
element_length 11 uimbf
channel_configuration_index 3 uimbf
reserved_bit 1 uimbf
bsac_header();
general_header();
byte_alignment();
for (slayer = 0; slayer < slayer_size; slayer++)
bsac_layer_element(slayer);
}
In Part 3: Audio, Subpart 4, under Bitstream elements in subclause 4.5.2.6.2.1 Definitions, replace
bsac_raw_data_block with the following:
bsac_raw_data_block() block of raw data that contains coded audio data, related information
and other data. A bsac_raw_data_block() basically consists of
bsac_base_element() and several bsac_layer_element(). There exists
a module that determines whether the BSAC bitstream has an
extended part.
In Part 3: Audio, Subpart 4, under Bitstream elements in subclause 4.5.2.6.2.1 Definitions, after
bsac_raw_data_block, add the following:
zero_code 32-bit zero values in order to terminate the arithmetic decoding for the
stereo part.
syncword a eight bit code that identifies the start of the extended part. The bit
string ‘1111 1111’.
© ISO/IEC 2006 – All rights reserved 13
ISO/IEC 14496-3:2005/Amd.2:2006(E)
In Part 3: Audio, Subpart 4, under Bitstream elements in subclause 4.5.2.6.2.1 Definitions, replace
header_length with the following:
header_length the length of the headers including frame_length, bsac_header() and
general_header() in bytes. The actual length is (header_length+7)
bytes. However if header_length is 0, it represents that the actual
length is smaller than or equal to 7 bytes. And if header_length is 15,
it represents that the actual length is larger than or equal to (15+7)
bytes and should be calculated through the decoding of the headers.
In case of extended_bsac_base_element(), header_length includes
element_length, channel_configuration_index, reserved_bit,
bsac_header and general_header().
In Part 3: Audio, Subpart 4 under Bitstream elements in subclause 4.5.2.6.2.1 Definitions, after
bsac_spectral_data, add the following:
extended_bsac_raw_data_block() block of raw data that contains coded audio data, related information
and other data for the extended part. A extended_bsac_ raw_data
_block() basically consists of extended_ bsac _base_ element() and
several bsac_layer_element().
extended_bsac_base_element() syntactic element of the base layer bitstream containing coded audio
data, related information and other data for the extended part of
BSAC.
element_length the length of the extended_bsac_raw_data_block() in bytes. This is
used for proper arithmetic decoding.
channel_configuration_index a three bit field that indicates the audio output channel configuration
in the extended part. Each index specifies the number of channels
given the channel to speaker mapping.
Table 4.68 – channel_configuration_index
Index channel to speaker mapping number of channels (nch)
0 center front speaker 1
1 left, right front speakers 2
2 rear surround speakers 1
3 left surround, right surround rear speakers 2
4 front low frequency effects speaker 1
5 left, right outside front speakers 2
6-7 reserved -
reserved_bit bit reserved for future use
In Part 3: Audio, Subpart 4, after subclause 4.5.2.6.2.2.13 Reconstruction of the decoded sample from bit-
sliced data, add the subclause below:
4.5.2.6.2.2.14 Decoding the extended part
The structure of the extended part of BSAC is a simple replica of mono or stereo BSAC bitstream. New
functions called extended_bsac_raw_data_block and extended_bsac_base_element are added for the
extended BSAC.
14 © ISO/IEC 2006 – All rights reserved
ISO/IEC 14496-3:2005/Amd.2:2006(E)
4.5.2.6.2.2.14.1 extended_bsac_raw_data_block
An extended_bsac_raw_data_block also has the layered structure as bsac_raw_data_block. In case where
data is still available after decoding the stereo part, zero_code and syncword are parsed. zero_code is used
for the arithmetic termination of stereo part, and syncword is for the proper decoding of extended part.
4.5.2.6.2.2.14.2 extended_bsac_base_element
An extended bsac_base_element consists of element_length, channel_configuration_index, reserved_bit,
bsac_header, general_header and bsac_layer_element. For the stereo part, the value of nch is obtained from
channelConfiguration in Table 1.8 (Syntax of AudioSpecificConfig) and it is limited to either 1 or 2 (left and
right front speakers). For the extended part, the parameter, nch, is concerned with the rest of speakers, and
the exact value is determined by channel_configuration_index specified in Table 4.68. Each index indicates
the number of channels given the channel to speaker mapping.
In Part 3: Audio, Subpart 4, at the end of subclause 4.B.17.8 Payload transmitted over Elementary Steam bit-
sliced data, add the following subclause:
4.B.17.8.1 The functionality of fine-grain scalability in extended or multi-channel data
When the BSAC data extends to multi-channel data, each ES consists of large-step layers for a certain
channel element. To provide the functionality of fine-grain scalability in the multi-channel data, one might use
streamPriority specified in the ES descriptor in ISO/IEC 14496-1:2004. The values of streamPriority are
assigned to elementary streams according to the priority of channel elements. Different numbers of layers per
channel element can be truncated, because the extended BSAC bitstream consists of separate channel
elements. The values of streamPriority and the number of layers to be truncated per channel element depend
on application scenarios.
In Part 3: Audio, Subpart 8, in clause 8.A.1, replace:
The usage of this parametric stereo extension to HE AAC is signalled implicitly in the bitstream. Hence, if
with:
The usage of this parametric stereo extension to HE AAC is signalled either implicitly by the presence of
parametric stereo data in the bitstream, or explicitly by signalling the corresponding AudioObjectType in the
audioSpecificConfig. Hence, implicit signalling requires that, if
© ISO/IEC 2006 – All rights reserved 15
ISO/IEC 14496-3:2005/Amd.2:2006(E)
Create Part 3: Audio, Subpart 11:
Subpart 11: Technical description of Audio Lossless Coding for
lossless coding of audio signals
11.1 Scope
This subpart of ISO/IEC 14496-3 describes the MPEG-4 Audio Lossless Coding (ALS) algorithm for lossless
coding of audio signals.
MPEG-4 ALS is a lossless compression scheme for digital audio data, i.e. the decoded data is a bit-identical
reconstruction of the original input data. Input signals can be integer PCM data with 8 to 32-bit word length or
32-bit IEEE floating-point data. MPEG-4 ALS provides a wide range of flexibility in terms of compression-
complexity trade-off, since the combination of several tools allows for the definition of compression levels with
different complexities.
11.2 Technical Overview
11.2.1 Encoder and Decoder Structure
The basic structure of the ALS encoder and decoder is shown in Figure 11.1.
Data
Encoder Decoder
Control
Input
Frame / Block
Partition
Entropy
Decoding
(Short-Term)
Prediction
Joint Channel
Decoding
Compressed
Long-Term
Bitstream
Prediction
Long-Term
Prediction
Joint Channel
Coding
(Short-Term)
Prediction
Entropy
Coding
Block / Frame
Assembly
Output
Figure 11.1 – Block diagram of the ALS encoder and decoder
The input audio data is partitioned into frames. Within a frame, each channel can be further subdivided into
blocks of audio samples for further processing (block switching, see subclause 11.6.2). For each block, a
prediction residual is calculated using short-term prediction (see subclauses 11.6.3 and 11.6.5) and optionally
long-term prediction (LTP, see sublause 11.6.4). Inter-channel redundancy can be removed by joint channel
coding, using either difference coding of channel pairs (see subclause 11.6.7) or multi-channel coding (MCC,
see subclause 11.6.8). The remaining prediction residual is finally entropy coded (see subclause 11.6.6).
The encoder generates bitstream information allowing for random access at intervals of several frames. The
encoder can also provide a CRC checksum, which the decoder may use to verify the decoded data.
16 © ISO/IEC 2006 – All rights reserved
Multiplexing
Demultiplexing
ISO/IEC 14496-3:2005/Amd.2:2006(E)
11.2.2 Floating-Point Extensions
In addition to integer audio signals, MPEG-4 ALS also supports lossless compression of audio signals in the
IEEE 32-bit floating-point format. The floating-point sequence is modeled by the sum of an integer sequence
multiplied by a constant (ACF: Approximate Common Factor) and a residual sequence. The integer sequence
is compressed using the basic ALS tools for integer data, while the residual sequence is separately
compressed by the masked Lempel-Ziv tool. A detailed description of the floating-point extensions can be
found in subclause 11.6.9.
11.3 Terms and Definitions
11.3.1 Definitions
The following definitions and abbreviations are used in this document.
Frame Segment of the audio signal (containing all channels).
Block Segment of one audio channel.
Sub-block Subpart of a block that uses the same entropy coding parameters.
Random Access Frame Frame that can be decoded without decoding previous frames.
Residual Prediction error, i.e. original minus predicted signal.
Predictor/Prediction Filter Linear FIR filter which computes an estimate of the input signal using previous
samples.
Prediction order Order of the prediction filter (number of predictor coefficients).
LPC coefficients Coefficients of the direct form prediction filter.
Parcor coefficients Parcor representation of the predictor coefficients.
Quantized coefficients Quantized parcor coefficients.
LTP Long-term prediction.
Rice code Also known as Golomb-Rice code. In this document the short form is used.
BGMC Block Gilbert-Moore Code (also known as Elias-Shannon-Fano code).
CRC Cyclic Redundancy Check.
LPC Linear Predictive Coding.
PCM Pulse Code Modulation.
Mantissa Fractional part of floating-point data
Exponent Exponential part of floating-point data
ACFC Approximate Common Factor Coding
Masked-LZ Masked Lempel-Ziv Coding
MCC Multi-Channel Coding
MSB Most significant bit
LSB Least significant bit
© ISO/IEC 2006 – All rights reserved 17
ISO/IEC 14496-3:2005/Amd.2:2006(E)
11.3.2 Mnemonics
uimsbf Unsigned integer, most significant bit first
simsbf Signed integer, most significant bit first
bslbf Bit string, left bit first, where “left” is the order in which bits are written
IEEE32 IEEE 32-bit floating-point data (4 bytes), most significant bit first
The mnemonics Rice code and BGMC indicate that variable length codewords are used, which are described
in subclause 11.6.6.
11.3.3 Data Types
The following data types are used in the pseudo code sections:
INT64 64-bit signed integer (two's complement)
long 32-bit signed integer (two's complement)
short 16-bit signed integer (two's complement)
If "unsigned" is added in front of the data type, then the type is unsigned instead of signed.
11.4 Syntax
11.4.1 Decoder Configuration
Table 11.1 – Syntax of ALSSpecificConfig
Syntax No. of bits Mnemonic
ALSSpecificConfig()
{
samp_freq; 32 uimsbf
samples; 32 uimsbf
channels; 16 uimsbf
file_type; 3 uimsbf
resolution; 3 uimsbf
floating; 1 uimsbf
msb_first; 1 uimsbf
frame_length; 16 uimsbf
random_access; 8 uimsbf
ra_flag; 2 uimsbf
adapt_order; 1 uimsbf
coef_table; 2 uimsbf
long_term_prediction; 1 uimsbf
max_order; 10 uimsbf
block_switching; 2 uimsbf
bgmc_mode; 1 uimsbf
sb_part; 1 uimsbf
joint_stereo; 1 uimsbf
mc_coding; 1 uimsbf
chan_config; 1 uimsbf
chan_sort; 1 uimsbf
crc_enabled; 1 uimsbf
RLSLMS 1 uimsbf
18 © ISO/IEC 2006 – All rights reserved
ISO/IEC 14496-3:2005/Amd.2:2006(E)
(reserved) 5
aux_data_enabled; 1 uimsbf
if (chan_config) {
chan_config_info; 16 uimsbf
}
if (chan_sort) {
for (c = 0; c < channels; c++)
chan_pos[c]; 1.16 uimsbf
}
byte_align;
header_size; 16 uimsbf
trailer_size; 16 uimsbf
orig_header[]; header_size * 8 bslbf
orig_trailer[]; trailer_size * 8 bslbf
if (crc_enabled) {
crc; 32 uimsbf
}
if ((ra_flag == 2) && (random_access > 0)) {
for (f = 0; f < ((samples-1) / (frame_length+1)) + 1; f++) {
ra_unit_size[f] 32 uimsbf
}
}
if (aux_data_enabled) {
aux_size; 16 uimsbf
aux_data[]; aux_size * 8 bslbf
}
}
11.4.2 Bitstream Payloads
Table 11.2 – Syntax of top level payload (frame_data)
Syntax No. of bits Mnemonic
frame_data()
{
if ((ra_flag == 1) && (frame_id % random_access == 0)) {
ra_unit_size 32 uimsbf
}
if (mc_coding && joint_stereo) {
js_switch; 1 uimsbf
byte_align;
}
if (!mc_coding || js_switch) {
for (c = 0; c < channels; c++) {
if (block_switching) {
bs_info; 8,16,32 uimsbf
}
if (independent_bs) {
for (b = 0; b < blocks; b++) {
block_data(c);
}
}
else{
for (b = 0; b < blocks; b++) {
block_data(c);
block_data(c+1);
}
c++;
© ISO/IEC 2006 – All rights reserved 19
ISO/IEC 14496-3:2005/Amd.2:2006(E)
}
}
else{
if (block_switching) {
bs_info; 8,16,32 uimsbf
}
for (b = 0; b < blocks; b++) {
for (c = 0; c < channels; c++) {
block_data(c);
channel_data(c);
}
}
}
if (floating)
{
num_bytes_diff_float; 32 uimsbf
diff_float_data();
}
}
Note: If joint_stereo is off, or if c is the last channel, independent_bs is true by default. If joint_stereo is on,
ind
...
Questions, Comments and Discussion
Ask us and Technical Secretary will try to provide an answer. You can facilitate discussion about the standard in here.
Loading comments...