ISO/IEC 23003-3:2012/Amd 1:2014

INTERNATIONAL ISO/IEC
STANDARD 23003-3
First edition
2012-04-01
AMENDMENT 1
2014-03-15
Information technology — MPEG
audio technologies —
Part 3:
Unified speech and audio coding
AMENDMENT 1: Conformance
Technologies de l’information — Technologies audio MPEG —
Partie 3: Discours unifié et codage audio
AMENDEMENT 1: Conformité
Reference number
ISO/IEC 23003-3:2012/Amd.1:2014(E)
©
ISO/IEC 2014
ISO/IEC 23003-3:2012/Amd.1:2014(E)

© ISO/IEC 2014
All rights reserved. Unless otherwise specified, no part of this publication may be reproduced or utilized otherwise in any form
or by any means, electronic or mechanical, including photocopying, or posting on the internet or an intranet, without prior
written permission. Permission can be requested 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 2014 – All rights reserved

ISO/IEC 23003-3:2012/Amd.1:2014(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 1 to ISO/IEC 23003-3:2012 was prepared by Joint Technical Committee ISO/IEC JTC 1,
Information technology, Subcommittee SC 29, Coding of audio, picture, multimedia and hypermedia
information.
© ISO/IEC 2014 – All rights reserved iii

ISO/IEC 23003-3:2012/Amd.1:2014(E)
Information technology — MPEG audio technologies —
Part 3:
Unified speech and audio coding
AMENDMENT 1: Conformance
In Clause 2, “Normative References”, add the following entry:
ISO/IEC 14496-26:2010, Information technology — Coding of audio-visual objects — Part 26: Audio
conformance
In 4.5.4 replace:
Four different hierarchical levels are defined with increasing number of audio channels and increasing
complexity. All four levels include Level 2 of the Baseline USAC profile. The definition of the four levels
of the Extended HE AAC profile is given in Table 3. All notes in Table 3 and all restrictions listed in the
columns 2, 3, 4, and 5 (“Max. channels/object”, “Max. AAC sampling rate, SBR not present [kHz]”, “Max.
AAC sampling rate, SBR present [kHz]”, “Max. SBR sampling rate [kHz] (in/out)”) of Table 3 apply only
when decoding HE AAC v2 profile compliant bit streams.
Table 3 — Levels for the Extended HE AAC profile
Max. AAC Max. AAC
Max. SBR Max. PCU Max. RCU
Level Max. sampling sampling
sampling HQ / LP HQ / LP
(NOTE channels / rate, SBR rate, SBR Max. PCU Max. RCU
rate [kHz] SBR SBR
1) object not present present
(in/out) (NOTE 5) (NOTE 5)
[kHz] [kHz]
1 NA NA NA NA NA NA NA NA
2 2 48 24 24/48 12 11 12 11
3 2 48 24/48 48/48 15 11 15 11
(NOTE 3) (NOTE 2)
4 5 48 24/48 48/48 25 28 20 23
(NOTE 4) (NOTE 2)
5 5 96 48 48/96 49 28 39 23
NOTE 1: Level 2, 3, and 4 Extended HE AAC profile decoders implement the baseline version of the parametric stereo tool. A
level 5 decoder 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 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 are present the maximum AAC sampling rate is 24kHz, if Parametric Stereo data are 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.
with:
A number of hierarchical levels are defined with increasing number of audio channels and increasing
complexity. All levels include Level 2 of the Baseline USAC profile. The definition of the levels of the
Extended HE AAC profile is given in Table 3. All notes in Table 3 and all restrictions listed in the columns
2, 3, 4, and 5 (“Max. channels/object”, “Max. AAC sampling rate, SBR not present [kHz]”, “Max. AAC
© ISO/IEC 2014 – All rights reserved 1

ISO/IEC 23003-3:2012/Amd.1:2014(E)

sampling rate, SBR present [kHz]”, “Max. SBR sampling rate [kHz] (in/out)”) of Table 3 apply only when
decoding HE AAC v2 profile compliant bit streams.
Table 3 — Levels for the Extended HE AAC profile
Max. AAC Max. AAC
Max. SBR Max. PCU Max. RCU
Level Max. sampling sampling
sampling HQ / LP HQ / LP
(NOTE channels / rate, SBR rate, SBR Max. PCU Max. RCU
rate [kHz] SBR SBR
1) object not present present
(in/out) (NOTE 5) (NOTE 5)
[kHz] [kHz]
1 NA NA NA NA NA NA NA NA
2 2 48 24 24/48 12 11 12 11
3 2 48 24/48 48/48 15 11 15 11
(NOTE 3) (NOTE 2)
4 5 48 24/48 48/48 25 28 20 23
(NOTE 4) (NOTE 2)
5 5 96 48 48/96 49 28 39 23
6 7 48 24/48 48/48 34 37 27 30
(NOTE 4)
7 7 96 48 48/96 67 37 53 30
NOTE 1: Level 2, 3, 4, 6 and 7 Extended HE AAC profile decoders implement the baseline version of the parametric stereo
tool. A level 5 decoder shall not be limited to the baseline version of the parametric stereo tool.
NOTE 2: For level 3, 4 and 6 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 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 are present the maximum AAC sampling rate is 24kHz, if Parametric Stereo data are 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.
NOTE 6: A Level 6 or 7 decoder is not required to decode a Level 5 stream.
In 5.3.2 amend Table 36 as follows:
2 © ISO/IEC 2014 – All rights reserved

ISO/IEC 23003-3:2012/Amd.1:2014(E)

Table 36 — Syntax of acelp_coding()
Syntax No. of bits Mnemonic
acelp_coding(acelp_core_mode)
{
[…]
switch (acelp_core_mode) {
case 0
icb_index[sfr]; 20 uimsbf
break;
case 1
icb_index[sfr]; 28 uimsbf
break;
case 2
icb_index[sfr]; 36 uimsbf
break;
case 3
icb_index[sfr]; 44 uimsbf
break;
case 4
icb_index[sfr]; 52 uimsbf
break;
case 5
icb_index[sfr]; 64 uimsbf
break;
case 6
icb_index[sfr]; 12 uimsbf
break;
case 7
icb_index[sfr]; 16 uimsbf
break;
}
gains[sfr]; 7 uimsbf
}
NOTE: coreCoderFrameLength designates the core frame length in samples and is equal to either 1024 or 768. See also
6.1.1.2.
In 7.14.5.2.1, replace:
Depending on the coding mode, the following codebooks are used:
with:
Depending on the coding mode, the following codebooks are used:
— 12-bit codebook with 2 pulses i and i . Pulse i can be selected from either track 0 or 2, pulse i
0 1 0 1
can be selected from either track 1 or 3 (5×2+2)
© ISO/IEC 2014 – All rights reserved 3

ISO/IEC 23003-3:2012/Amd.1:2014(E)

— 16-bit codebook with 3 pulses on three tracks. One pulse on track 0, one pulse on track 2 and one
pulse on either track 1 or 3 (selected track signalled by a 1 bit field), which amounts to (5×3+1) = 16
bits.
Add a new Clause 8, “Conformance testing”, as shown below:
8 Conformance testing
8.1 Introduction
The present Clause 8 specifies conformance criteria for both bitstreams and decoders compliant with
the USAC standard as defined in this document. This is done to assist implementers and to ensure
interoperability.
8.2 Terms and definitions
bitstream
encoded audio data
conformance data
conformance test sequences and conformance tools
conformance tool
tool to check certain conformance criteria
conformance test sequence
generic term for conformance test bitstreams and corresponding reference waveforms
conformance test bitstream
USAC bitstream used for testing the conformance of a USAC decoder
conformance test condition
condition which applies to properties of a conformance test bitstream in order to test a certain func-
tionality of the USAC decoder
conformance test case
combination of one or more conformance test conditions for which a set of conformance test sequences
is provided
main audio channel
audio channel conveyed by means of a UsacSingleChannelElement or UsacChannelPairElement
reference waveform
decoded counterpart of a bitstream
USAC bitstream
data encoded according to the USAC standard
UsacCPE
UsacChannelPairElement
UsacEXT
UsacExtElement
UsacLFE
UsacLfeElement
UsacSCE
UsacSingleChannelElement
4 © ISO/IEC 2014 – All rights reserved

ISO/IEC 23003-3:2012/Amd.1:2014(E)

8.3 USAC conformance testing
8.3.1 Profiles
Profiles are defined in 4.5. Some conformance criteria apply to USAC in general, while others are specific
to certain profiles and their respective levels. Conformance shall be tested for the level of the profile
with which a given bitstream or decoder claims to comply.
In addition to the conformance requirements described in this clause, a decoder which claims to comply
with the Extended HE AAC Profile shall fulfill conformance for the HE AAC v2 profile according to
ISO/IEC 14496-26:2010.
8.3.2 Conformance tools and test procedure
To test USAC compliant audio decoders, ISO/IEC JTC 1/SC 29/WG 11 supplies a number of conformance
test sequences. Supplied sequences cover all profiles as defined in 4.5. For a supplied test sequence,
testing can be done by comparing the output of a decoder under test with a reference waveform also
supplied by ISO/IEC JTC 1/SC 29/WG 11. In cases where the decoder under test is followed by additional
operations (e.g. quantizing a signal to a 16 bit output signal) the conformance point is prior to such
additional operations, i.e. it is permitted to use the actual decoder output (e.g. with more than 16 bit) for
conformance testing.
Measurements are carried out relative to full scale where the output signals of the decoders are
normalized to be in the range between −1.0 and +1.0.
In ISO/IEC 14496-26:2010 a set of test methods is defined to test the output of the decoder under test
against the reference output. RMS/LSB Measurement, Segmental SNR and PNS conformance criteria are
used for the comparison. A particular test method for a certain test sequence is specified in 8.5.
For elements producing output that cannot be tested with the methods described in ISO/IEC 14496-
26:2010, specific conformance testing procedures are described in 8.5.
8.3.2.1 Conformance data
All test sequences are provided in the shape of a zip archive as an electronic attachment. Furthermore,
an MS Excel worksheet (“Usac_Conformance_Tables.xlsx”) is provided as an electronic attachment that
lists all test sequences for each module.
For all conformance test sequences, the file names are composed of several parts which convey
information about:
— which module of the decoder is tested
— which channelConfigurationIndex is employed
— which test conditions apply to the test sequence
— which coreSbrFrameLengthIndex applies to the test sequence
— which sampling frequency is signalled in the test sequence
The file naming convention given in Table 149 is used. Values in box brackets are optional.
© ISO/IEC 2014 – All rights reserved 5

ISO/IEC 23003-3:2012/Amd.1:2014(E)

Table 149 — File name conventions
Module File Name (compressed) File Name (uncompressed)
Frequency domain cod- Fd__c__.mp4 FD__c__.wav
ing (FD mode), 8.4.4
Linear predictive Lpd__c__[].mp4 Lpd__c__[].wav
domain coding (LPD
mode), 8.4.5
Common core coding Cct__c__[].mp4 Cct__c__[].wav
tools, 8.4.6
Enhanced spectral eSbr__.mp4 eSbr__.wav
band replication
(eSBR), 8.4.7
MPEG Surround 2-1-2, Mps__Sc_.mp4 Mps__Sc_.wav
8.4.10
channelConfigurationIndex as described in Table 68.
Setup string. May consist of a concatenation of one or more abbrevia-
tions as listed in Table 150. If no setup string is specified the basic test
conditions apply
coreSbrFrameLengthIndex as described in Table 70.
usacSamplingFrequencyIndex as described in Table 67. If the escape
value is specified the used sampling frequency is appended, e.g.
“xx_1f_42000.mp4” for a sampling frequency of 42 kHz.
bsFreqRes as described in ISO/IEC 23003-1:2007, Table 39
stereoConfigIndex as described in Table 72
Table 150 — Test conditions and abbreviations
FD core mode
Test Condition Abbrev.
FD window switching test condition Win
Noise filling test condition Nf
Tns test condition Tns
Varying max_sfb test condition Sfb
Handling of extensions condition Ex
Arithmetic coder test condition Ac
Non-meaningful FD window switching test Nmf
condtion
M/S stereo test condition Ms
Complex prediction stereo test condition Cp
6 © ISO/IEC 2014 – All rights reserved

ISO/IEC 23003-3:2012/Amd.1:2014(E)

LPD core mode
Test Condition Abbrev.
LPC coding test condition Lpc
ACELP core mode test condition Ace
TCX and noise filling test condition Tcx
LPD mode coverage and FAC test condition Lpd
Bass-post filter test condition Bpf
AVQ test condition Avq
Combined core coding
Test Condition Abbrev.
FD-LPD transition and FAC test condition Flt
FD/TCX noise filling test condition Cnf
Bass-post filter test condition Cbf
synchr. FD-LPD transition and FAC test Flts
condition
asynchr. FD-LPD transition and FAC test Flta
condition
Arithmetic coder test condition CAc

eSbr
Test Condition Abbrev.
QMF accuracy test condition Qma
Envelope adjuster accuracy and SBR pre- Eaa
processing test condition
Header and grid control test condition test Hgt
condition
Inverse filtering test condition Ift
Additional sine test (missing harmonics) Ast
test condition
Sampling rate test condition Sr
Channel mode test condition Cm
interTes test condition Tes
PVC test condition Pvc
Harmonic transposition (QMF) test condi- Htq
tion
Harmonic transposition (crossproducts) Xp
test condition
Transposer toggle test condition Ttt
Envelope shaping toggle (PVC on/off) test Est
condition
Varying crossover frequency test condi- Xo
tion
stereoConfigIndex test condition Mps
© ISO/IEC 2014 – All rights reserved 7

ISO/IEC 23003-3:2012/Amd.1:2014(E)

Mpeg surround 212
Test Condition Abbrev.
TSD test condition Tsd
Rate mode test condition Rm
Phase coding test condition Pc
Decorrelator configuration. test condition Dc
DMX gain test condition Dm
Bands phase test condition Bp
Pseudo lr test condition Plr
Residual bands test condition Rb
8.4 USAC Bitstreams
8.4.1 General
8.4.1.1 Characteristics
Characteristics of bitstreams specify the constraints that are applied by the encoder in generating the
bitstreams. These syntactic and semantic constraints may, for example, restrict the range or the values
of parameters that are encoded directly or indirectly in the bitstreams. The constraints applied to a
given bitstreams may or may not be known a priori.
8.4.1.2 Test procedure
Each USAC bitstream shall meet the syntactic and semantic requirements specified in this document.
The present subclause defines the conformance criteria that shall be fulfilled by a compliant bitstream.
These criteria are specified for the syntactic elements of the bitstream and for some parameters decoded
from the USAC bitstream payload.
For each tool a set of semantic tests to be performed on the bitstreams is described. To verify whether
the syntax is correct is straightforward and therefore not defined herein after. In the description of
the semantic tests it is assumed that the tested bitstreams contains no errors due to transmission or
other causes. For each test the condition or conditions that must be satisfied are given, as well as the
prerequisites or conditions in which the test can be applied.
8.4.2 USAC Configuration
8.4.2.1 Characteristics
Encoders may apply restrictions to the following parameters of the bitstream:
a) usacSamplingFrequencyIndex
b) usacSamplingFrequency
c) coreSbrFrameLengthIndex
d) channelConfigurationIndex
e) presence of configuration extensions
f) numOutChannels
g) bsOutputChannelPos
8 © ISO/IEC 2014 – All rights reserved

ISO/IEC 23003-3:2012/Amd.1:2014(E)

h) numElements
i) stereoConfigIndex
j) use of time warped MDCT
k) use of noise filling in FD mode
l) use of the eSBR harmonic transposer
m) use of the eSBR inter-TES tool
n) use of the eSBR PVC tool
o) SBR default header, for details see 8.4.7.
p) MPS config, for details see 8.4.10.
8.4.2.2 Test procedure
8.4.2.2.1 UsacConfig()
usacSamplingFrequencyIndex Shall be encoded with a non-reserved value specified in Table 67. For
further profile and level dependent restrictions see 8.4.11.
usacSamplingFrequency No restrictions apply. For profile and level dependent restrictions see
8.4.11.
coreSbrFrameLengthIndex no restrictions apply
channelConfigurationIndex Shall be encoded with a non-reserved value specified in Table 68.
For further profile and level dependent restrictions see 8.4.11. In the
case of channelConfigurationIndex==0 further restrictions apply as
described in 8.4.2.2.2.
usacConfigExtensionPresent no restrictions apply
8.4.2.2.2 UsacChannelConfig()
numOutChannels no restrictions apply. For profile and level dependent restrictions see
8.4.11.
bsOutputChannelPos A bsOutputChannelPos of value 3 or 26 (LFE speaker positions) shall
be associated with an LFE channel. Any other value shall be associ-
ated with a main audio channel.
8.4.2.2.3 UsacDecoderConfig()
numElements the value of this data element shall be such that the accumulated sum
of all channels contained in the bitstream complies with the restric-
tions outlined in 8.4.2.2.1.
usacElementType no restrictions apply. For profile and level dependent restrictions see
8.4.11.
© ISO/IEC 2014 – All rights reserved 9

ISO/IEC 23003-3:2012/Amd.1:2014(E)

8.4.2.2.4 UsacSingleChannelElementConfig()
No restrictions are applicable to this bitstream element.
8.4.2.2.5 UsacChannelPairElementConfig()
NOTE The UsacChannelPairElementConfig() element and all included elements may only be present when
coding more than one output channel (see restrictions applying to UsacConfig() in 8.4.2.2.1).
stereoConfigIndex no restrictions apply
8.4.2.2.6 UsacLfeElementConfig()
No restrictions are applicable to this bitstream element.
8.4.2.2.7 UsacCoreConfig()
tw_mdct no restrictions apply. For profile and level dependent restrictions see
8.4.11.
noiseFilling no restrictions apply
8.4.2.2.8 SbrConfig()
harmonicSBR no restrictions apply
bs_interTes no restrictions apply
bs_pvc no restrictions apply
8.4.2.2.9 SbrDfltHeader()
dflt_start_freq no restrictions apply
dflt_stop_freq no restrictions apply
dflt_header_extra1 no restrictions apply
dflt_header_extra2 no restrictions apply
dflt_freq_scale no restrictions apply
dflt_alter_scale no restrictions apply
dftl_nose_bands no restrictions apply
dflt_limiter_bands no restrictions apply
dflt_limiter_gains no restrictions apply
dflt_interpol_freq no restrictions apply
dflt_smoothing_mode no restrictions apply
10 © ISO/IEC 2014 – All rights reserved

ISO/IEC 23003-3:2012/Amd.1:2014(E)

8.4.2.2.10 Mps212Config()
bsFreqRes shall not be encoded with a value of 0
bsFixedGainDMX no restrictions apply
bsTempShapeConfig no restrictions apply
bsDecorrConfig shall not be encoded with a value of 3
bsHighRateMode no restrictions apply
bsPhaseCoding no restrictions apply
bsOttBandsPhasePresent no restrictions apply
bsOttBandsPhase shall not be encoded with a value larger than the value of numBands
as given by ISO/IEC 23003-1:2007, 5.2, Table 39 and depends on
bsFreqRes.
bsResidualBands shall not be encoded with a value larger than the value of numBands
as given by ISO/IEC 23003-1:2007, 5.2, Table 39 and depends on
bsFreqRes.
bsPseudoLr no restrictions apply
bsEnvQuantMode shall be 0
8.4.2.2.11 UsacExtElementConfig()
usacExtElementType no restrictions apply
usacExtElementConfigLength no restrictions apply
usacExtElementDefaultLengthPresent no restrictions apply
usacExtElementDefaultLength no restrictions apply
usacExtElementPayloadFrag no restrictions apply
8.4.2.2.12 UsacConfigExtension()
numConfigExtensions no restrictions apply
usacConfigExtType[] no restrictions apply
usacConfigExtLength[] no restrictions apply
fill_byte should be ‘10100101’
8.4.3 Framework
8.4.3.1 Characteristics
Encoders may apply restrictions to the following parameters of the bitstream:
a) signalling of independently decodable frames
© ISO/IEC 2014 – All rights reserved 11

ISO/IEC 23003-3:2012/Amd.1:2014(E)

b) presence of extension elements
c) core_mode
d) presence of TNS
8.4.3.2 Test procedure
8.4.3.2.1 UsacFrame()
usacIndependencyFlag no restrictions apply
8.4.3.2.2 UsacSingleChannelElement
No restrictions are applicable to this bitstream element.
8.4.3.2.3 UsacChannelPairElement
No restrictions are applicable to this bitstream element.
8.4.3.2.4 UsacLfeElement
No restrictions are applicable to this bitstream element.
8.4.3.2.5 UsacExtElement
usacExtElementPresent no restrictions apply
usacExtElementUseDefaultLength no restrictions apply
usacExtElementPayloadLength no restrictions apply
usacExtElementStart no restrictions apply
usacExtElementStop no restrictions apply
usacExtElementSegmentData no restrictions apply
8.4.3.2.6 UsacCoreCoderData
core_mode no restrictions apply.
tns_data_present no restrictions apply
8.4.4 Frequency domain coding (FD mode)
8.4.4.1 Characteristics
Encoders may apply restrictions to the following parameters of the bitstream:
a) use of noise filling
b) window_shape
c) M/S Stereo
12 © ISO/IEC 2014 – All rights reserved

ISO/IEC 23003-3:2012/Amd.1:2014(E)

d) use of TNS
e) Complex prediction stereo coding
f) max_sfb
g) use of time warped MDCT
h) use of long blocks
i) use of short blocks
8.4.4.2 Test procedure
8.4.4.2.1 fd_channel_stream
global_gain no restrictions apply.
noise_level no restrictions apply
noise_offset no restrictions apply
fac_data_present shall be 0, if the core_mode of the preceding frame of the same chan-
nel was 0 or if mod[3] of the preceding frame of the same channel
was > 0.
8.4.4.2.2 ics_info
window_sequence A conformant bitstream shall consist of only meaningful window_
sequence transitions. However, decoders are required to handle non-
meaningful window_sequence transitions as well. The meaningful
window_sequence transitions are shown in Table 133.
window_shape no restrictions apply
max_sfb shall be <= num_swb_long or num_swb_short as appropriate for win-
dow_sequence and sampling frequency and core coder frame length.
scale_factor_grouping no restrictions apply
8.4.4.2.3 tw_data
tw_data_present no restrictions apply
tw_ratio no restrictions apply
8.4.4.2.4 scale_factor_data
hcod_sf Shall only be encoded with the values listed in the scalefactor Huff-
man table. Shall be encoded such that the decoded scalefactors sf[g]
[sfb] are within the range of zero to 255, both inclusive.
© ISO/IEC 2014 – All rights reserved 13

ISO/IEC 23003-3:2012/Amd.1:2014(E)

8.4.4.2.5 tns_data
n_filt no restrictions apply
coef_res no restrictions apply
length shall be small enough such that the lower bound of the filtered
region, does not exceed the start of the array containing the spectral
coefficients.
order shall not exceed the values listed in in Table 130.
direction no restrictions apply
coef_compress no restrictions apply
coef no restrictions apply
8.4.4.2.6 ac_spectral_data
arith_reset_flag no restrictions apply
8.4.4.2.7 StereoCoreToolInfo
tns_active no restrictions apply
common_window no restrictions apply
common_max_sfb no restrictions apply
max_sfb1 shall be <= num_swb_long or num_swb_short as appropriate for win-
dow_sequence and sampling frequency and core coder frame length.
ms_mask_present no restrictions apply
ms_used no restrictions apply
common_tw no restrictions apply
common_tns no restrictions apply
tns_on_lr no restrictions apply
tns_present_both no restrictions apply
tns_data_present no restrictions apply
14 © ISO/IEC 2014 – All rights reserved

ISO/IEC 23003-3:2012/Amd.1:2014(E)

8.4.4.2.8 cplx_pred_data
cplx_pred_all no restrictions apply
cplx_pred_used no restrictions apply
pred_dir no restrictions apply
complex_coef no restrictions apply
use_prev_frame shall be 0 if the core transform length of previous frame is different
from the core transform length of the current frame or if the core_
mode of the previous frame is 1.
delta_code_time no restrictions apply
hcod_sf no restrictions apply
8.4.5 Linear predictive domain coding (LPD mode)
8.4.5.1 Characteristics
Encoders may apply restrictions to the following parameters of the bitstream:
a) acelp_core_mode
b) lpd_mode (use of ACELP, short TCX, medium TCX, and long TCX)
c) activation of bass-post filter
8.4.5.2 Test procedure
8.4.5.2.1 lpd_channel_stream
acelp_core_mode shall be encoded with a value in the range of 0 to 5, both inclusive.
lpd_mode shall be encoded with a non-reserved value listed in Table 89.
bpf_control_info no restrictions apply
core_mode_last shall be encoded with the value of data element core_mode of the pre-
vious frame
fac_data_present shall be 0, if the core_mode of the preceding frame of the same chan-
nel was 0 and mod[0] of the current frame is > 0, or if mod[0] of the
current frame is > 0 and mod[3] of the preceding frame of the same
channel was > 0.
short_fac_flag shall be encoded with a value of 1 if the window_sequence of the
previous frame was 2 (EIGHT_SHORT_SEQUENCE). Otherwise short_
fac_flag shall be encoded with a value of 0.
8.4.5.2.2 lpc_data
lpc_first_approximation_index no restrictions apply
© ISO/IEC 2014 – All rights reserved 15

ISO/IEC 23003-3:2012/Amd.1:2014(E)

8.4.5.2.3 qn_data
qn the codebook number shall be encoded as described in 7.13.7.2.
qn_base no restrictions apply
qn_ext no restrictions apply
8.4.5.2.4 get_mode_lpc
binary_code shall be encoded with the values listed in Table 143 in the column
Binary Code
8.4.5.2.5 code_book_indices
code_book_index no restrictions apply
kv no restrictions apply
8.4.5.2.6 acelp_coding
mean_energy no restrictions apply
acb_index the adaptive codebook index shall be encoded as described in
7.14.5.1.
ltp_filtering_flag no restrictions apply
icb_index the innovation codebook excitation shall be encoded as described in
7.14.5.2.
gains no restrictions apply
8.4.5.2.7 tcx_coding
noise_factor no restrictions apply
global_gain no restrictions apply
arith_reset_flag no restrictions apply
8.4.6 Common core coding tools
8.4.6.1 Characteristics
Encoders may apply restrictions to the following parameters of the bitstream:
a) use of arithmetic coder reset
8.4.6.2 Test procedure
16 © ISO/IEC 2014 – All rights reserved

ISO/IEC 23003-3:2012/Amd.1:2014(E)

8.4.6.2.1 arith_data
acod_m shall be encoded as described in 7.4.3
acod_r shall be encoded as described in 7.4.3
s no restrictions apply
8.4.6.2.2 fac_data
fac_gain no restrictions apply
8.4.7 Enhanced spectral band replication (eSBR)
8.4.7.1 Characteristics
Encoders may apply restrictions to the following parameters of the bitstream:
a) use of the eSBR harmonic transposer
b) use of Crossproducts in eSBR harmonic transposer
c) use of the eSBR inter-TES tool
d) choice of SBR ratio
e) choice of amplitude resolution
f) choice of SBR crossover band
g) use of SBR preprocessing (prewhitening)
h) use of the eSBR PVC tool
8.4.7.2 Test procedure
The present subclause defines the conformance criteria that shall be fulfilled by a compliant bitstream
that utilize the Enhanced SBR tool.
8.4.7.2.1 UsacSbrData
sbrInfoPresent no restrictions apply
sbrHeaderPresent no restrictions apply
sbrUseDfltHeader no restrictions apply
© ISO/IEC 2014 – All rights reserved 17

ISO/IEC 23003-3:2012/Amd.1:2014(E)

8.4.7.2.2 SbrInfo
bs_amp_res no restrictions apply
bs_xover_band shall define a value that does not exceed the limits defined in ISO/
IEC 14496-3:2009, 4.6.18.3.6.
bs_sbr_preprocessing no restrictions apply
bs_pvc_mode shall be encoded with a non-reserved value specified in Table 96
8.4.7.2.3 SbrHeader
bs_start_freq shall define a frequency band that does not exceed the limits defined
in 7.5.5 and ISO/IEC 14496-3:2009, 4.6.18.3.6.
bs_stop_freq shall define a frequency band that does not exceed the limits defined
in 7.5.5 and ISO/IEC 14496-3:2009, 4.6.18.3.6.
bs_header_extra1 no restrictions apply
bs_header_extra2 no restrictions apply
bs_freq_scale no restrictions apply
bs_alter_scale no restrictions apply
bs_noise_bands shall define a value that does not exceed the limits defined in ISO/
IEC 14496-3:2009, 4.6.18.3.6.
bs_limiter_bands no restrictions apply
bs_limiter_gains no restrictions apply
bs_interpol_freq no restrictions apply
bs_smoothing_mode no restrictions apply
8.4.7.2.4 sbr_single_channel_element
sbrPatchingMode no restrictions apply
sbrOversamplingFlag no restrictions apply
sbrPitchInBinsFlag no restrictions apply
sbrPitchInBins no restrictions apply
bs_add_harmonic_flag no restrictions apply
18 © ISO/IEC 2014 – All rights reserved

ISO/IEC 23003-3:2012/Amd.1:2014(E)

8.4.7.2.5 sbr_channel_pair_element
bs_coupling no restrictions apply
sbrPatchingMode no restrictions apply
sbrOversamplingFlag no restrictions apply
sbrPitchInBinsFlag no restrictions apply
sbrPitchInBins no restrictions apply
bs_add_harmonic_flag no restrictions apply
8.4.7.2.6 sbr_grid
bs_frame_class shall define a value that does not exceed the limits defined in 7.5.1.3
and ISO/IEC 14496-3:2009, 4.6.18.3.6.
tmp (determines bs_num_env), no restrictions apply
bs_freq_res no restrictions apply
bs_pointer shall be encoded with a value listed in ISO/IEC 14496-3:2009,
Table 4.174.
The restrictions defined in ISO/IEC 14496-26:2010, 7.17.1.2.1.3 sbr_grid() shall be applied to the
following corresponding bitstream elements:
bs_var_bord_0
bs_var_bord_1
bs_num_rel_0
bs_num_rel_1
bs_noise_position shall be chosen so that the time slot borders for noise floors fall
within the leading and trailing SBR frame borders (i.e. the SBR frame
boundaries)
bs_var_len_hf shall be encoded with a non-reserved value specified in Table 97
8.4.7.2.7 sbr_envelope
bs_env_start_value_balance no restrictions apply
bs_env_start_value_level no restrictions apply
bs_codeword shall be encoded as defined in sbr_huff_dec() in ISO/IEC 14496-
3:2009, 4.A.6.1.
Additionally, the restrictions defined in ISO/IEC 14496-26:2010, 7.17.1.2.1.5 sbr_envelope() apply.
© ISO/IEC 2014 – All rights reserved 19

ISO/IEC 23003-3:2012/Amd.1:2014(E)

8.4.7.2.8 dtdf
bs_df_env no restrictions apply
bs_df_noise no restrictions apply
8.4.7.2.9 sbr_sinusoidal_coding
bs_add_harmonic no restrictions apply
bs_sinusoidal_position_flag no restrictions apply
bs_sinusoidal_position shall be chosen so that the position of the starting time slot for sinu-
soidals fall within the SBR frame boundaries
8.4.7.2.10 sbr_invf
No restrictions are applicable to this bitstream element.
8.4.7.2.11 sbr_noise
The restrictions defined in ISO/IEC 14496-26:2010, 7.17.1.2.1.6 sbr_noise() apply.
8.4.8 eSBR – Predictive vector coding (PVC)
8.4.8.1 Characteristics
Encoders may apply restrictions to the following parameters of the bitstream:
a) activation of PVC
b) use of IDs from the previous frame
c) length
8.4.8.2 Test procedures
8.4.8.2.1 pvc_envelope
divMode no restrictions apply
nsMode no restrictions apply
reuse_pvcID shall be 0 if the bs_pvc_mode of the preceding SBR frame was 0
pvcID no restrictions apply
length shall be chosen so that the time slot borders for pvcID fall within the
SBR frame boundaries
grid_info the first grid_info (grid_info[0]) shall be 1 if the bs_pvc_mode of the
preceding SBR frame was 0
8.4.9 eSBR – Inter temporal envelope shaping (inter-TES)
20 © ISO/IEC 2014 – All rights reserved

ISO/IEC 23003-3:2012/Amd.1:2014(E)

8.4.9.1 Characteristics
Encoders may apply restrictions to the following parameters of the bitstream:
a) activation of inter-TES
8.4.9.2 Test procedure
8.4.9.2.1 sbr_envelope
bs_temp_shape no restrictions apply
bs_inter_temp_shape_mode no restrictions apply
8.4.10 MPEG Surround 2-1-2
8.4.10.1 Characteristics
Encoders may apply restrictions to the following parameters of the bitstream:
a) use of phase coding
b) use of residual coding
c) use of pseudo LR
d) use of Transient Steering Decorrelator
8.4.10.2 Test procedure
8.4.10.2.1 Mps212Data
bsIndependencyFlag no restrictions apply
8.4.10.2.2 FramingInfo
bsFramingType no restrictions apply
bsNumParamSets shall have a value not larger than (numSlots-1)/4, where the division
shall be interpreted as an ANSI C integer division
bsParamSlot shall be in the range 0.numSlots-1
8.4.10.2.3 OttData
bsPhaseMode no restrictions apply
bsOPDSmoothingMode no restrictions apply
© ISO/IEC 2014 – All rights reserved 21

ISO/IEC 23003-3:2012/Amd.1:2014(E)

8.4.10.2.4 SmgData
bsSmoothMode no restrictions apply
bsSmoothTime no restrictions apply
bsFreqResStrideSmg no restrictions apply
bsSmgData no restrictions apply
8.4.10.2.5 TempShapeData
bsTsdEnable no restrictions apply
bsTempShapeEnable no restrictions apply
bsTempShapeEnableChannel no restrictions apply
8.4.10.2.6 TsdData
bsTsdNumTrSlots shall be encoded with 4 or 5 bits depending on numSlots
bsTsdCodedPos no restrictions apply
bsTsdTrPhaseData no restrictions apply
8.4.10.2.7 EcData
bsXXXdataMode shall fulfill the requirements outlined in ISO/IEC 23003-1:2007,
6.1.13. Shall not be encoded with a value of 2 if residual coding is
applied. Shall have the value 0 or 3 if ps==0 and bsIndependency-
Flag is set to 1
bsDataPairXXX shall have the value 0 if setIdx == datasets-1. No further restrictions
apply
bsQuantCoarseXXX no restrictions apply
bsFreqResStrideXXX no restrictions apply
8.4.10.2.8 EcDataPair
bsPcmCodingXXX no restrictions apply
8.4.10.2.9 GroupedPcmData
bsPcmWord no restrictions apply
22 © ISO/IEC 2014 – All rights reserved

ISO/IEC 23003-3:2012/Amd.1:2014(E)

8.4.10.2.10 DiffHuffData
bsDiffType no restrictions apply
bsCodingScheme no restrictions apply
bsPairing no restrictions apply
bsDiffTimeDirection no restrictions apply
8.4.10.2.11 HuffData1D
hcodFirstband_XXX bsCodeW shall have a value out of a set of values as defined by
column ‘codeword’ in ISO/IEC 23003-1:2007, Tables A.2 and A.3, for
CLD and ICC respectively. For IPD, in Table A.2. Shall have a length as
defined by the corresponding entry in column ‘length’
hcod1D_XXX_YY bsCodeW shall have a value out of a set of values as defined by
column ‘codeword’ in ISO/IEC 23003-1:2007, Tables A.5 and A.6, for
CLD and ICC respectively. For IPD, in Table A.3. Shall have a length as
defined by the corresponding entry in column ‘length’
bsSign do not apply to the encoding of IPD parameters. No further restric-
tions apply
8.4.10.2.12 HuffData2DFreqPair, HuffData2DTimePair
hcodLavIdx bsCodeW shall have a value out of a set of values as defined by col-
umn ‘codeword’ in ISO/IEC 23003-1:2007, Tables A.24, and shall have
a length as defined by the corresponding entry in column ‘length’
hcod2D_XXX_YY_ZZ_LL_escape bsCodeW shall have a value out of a set of values as defined by
column ‘codeword’ in ISO/IEC 23003-1:2007, Tables A.8 and A.9, for
CLD and ICC respectively. For IPD, in Table A.4. Shall have a length as
defined by the corresponding entry in column ‘length’
hcod2D_XXX_YY_ZZ_LL bsCodeW shall have a value out of a set of values as defined by
column ‘codeword’ of the applicable table in ISO/IEC 23003-1:2007,
Tables A.11 to A.18, for CLD and ICC. For IPD, in Tables A.5 to A.8.
Shall have a length as defined by the corresponding entry in column
‘length’
8.4.10.2.13 SymmetryData
bsSymBit no restrictions apply
8.4.10.2.14 LsbData
bsLsb no restrictions apply
8.4.11 Restrictions depending on profiles and levels
© ISO/IEC 2014 – All rights reserved 23

ISO/IEC 23003-3:2012/Amd.1:2014(E)

8.4.11.1 Introduction
Depending on the profile and level associated with the USAC bitstream, further restrictions may apply.
8.4.11.2 Baseline USAC profile
8.4.11.2.1 usacSamplingFrequencyIndex
For Baseline USAC Profile usacSamplingFrequencyIndex shall be encoded with a value specified in
Table 151.
Table 151 — Specification of usacSamplingFrequencyIndex
and usacSamplingFrequency in Baseline USAC Profile
Level
1 2 3 4 5
0x03…0x0c, 0x03…0x0c, 0x03…0x0c, 0x00…0x0c,
0x11…0x1b 0x11…0x1b 0x11…0x1b 0x0f…0x1b
usacSamplingFrequencyIndex/
N / A
usacSamplingFrequency
0x1f / 0x1f / 0x1f / 0x1f /
≤ 48000 ≤ 48000 ≤ 48000 ≤ 96000
Furthermore, for the Baseline USAC Profile the employed sampling rates shall be one out of those listed
in Table 3.
8.4.11.2.2 channelConfigurationIndex
For Baseline USAC Profile channelConfigurationIndex shall be encoded with a value specified in
Table 152.
Table 152 — Specification of channelConfigurationIndex
in Baseline USAC Profile
Level
1 2 3 4 5
channelConfigurationIndex 0, 1 0, 1, 2, 8 0.6, 8.10 0.6, 8.10 N / A
8.4.11.2.3 numOutChannels
For Baseline USAC Profile numOutChannels shall be encoded with a value specified in Table 153. Further
restrictions apply to the number of main audio channels (channels conveyed in UsacSCEs and UsacCPEs)
and LFE channels (conveyed in UsacLFEs) as shown in Table 153.
Table 153 — Specification of numOutChannels
for Baseline USAC Profile
Level
1 2 3 4 5
numOutChannels ≤ 1 ≤ 2 ≤ 6 ≤ 6 N / A
number of main audio chan-
≤ 1 ≤ 2 ≤ 5 ≤ 5 N / A
nels
number of LFE channels 0 0 ≤ 1 ≤ 1 N / A
24 © ISO/IEC 2014 – All rights reserved

ISO/IEC 23003-3:2012/Amd.1:2014(E)

8.4.11.2.4 usacElementType
For the Baseline USAC Profile usacElementType shall take values such that the number of main audio
channels and LFE channels comply with the restrictions outlined in 8.4.11.2.3.
8.4.11.2.5 tw_mdct
For Baseline USAC Profile tw_mdct shall be encoded with 0.
8.4.11.2.6 tw_data
tw_data should not be present in Baseline USAC Profile complying bitstreams, due to restrictions of
bitstream element tw_mdct.
8.4.11.3 Extended HE AAC profile
8.4.11.3.1 usacSamplingFrequencyIndex
For Extended HE AAC Profile usacSamplingFrequencyIndex shall be encoded with a value specified in
Table 154.
Table 154 — Specification of usacSamplingFrequencyIndex
and usacSamplingFrequency in Extended HE AAC Profile
Level
1 2 3 4 5
0x03…0x0c, 0x03…0x0c, 0x03…0x0c, 0x03…0x0c,
0x11…0x1b 0x11…0x1b 0x11…0x1b 0x11…0x1b
usacSamplingFrequencyIndex/
N / A
usacSamplingFrequency
0x1f / 0x1f / 0x1f / 0x1f /
≤ 48000 ≤ 48000 ≤ 48000 ≤ 48000
8.4.11.3.2 channelConfigurationIndex
For Extended HE AAC Profile channelConfigurationIndex shall be encoded with a value specified in
Table 155.
Table 155 — Specification of channelConfigurationIndex
in Extended HE AAC Profile
Level
1 2 3 4 5
channelConfigurationIndex N / A 0, 1, 2, 8 0, 1, 2, 8 0, 1, 2, 8 0, 1, 2, 8
8.4.11.3.3 numOutChannels
For Extended HE AAC Profile numOutChannels shall be encoded with a value specified in Table 156.
Table 156 — Specification of numOutChannels
for Extended HE AAC Profile
Level
1 2 3 4 5
numOutChannels N / A ≤ 2 ≤ 2 ≤ 2 ≤ 2
© ISO/IEC 2014 – All rights reserved 25

ISO/IEC 23003-3:2012/Amd.1:2014(E)

8.4.11.3.4 tw_mdct
For Extended HE AAC Profile tw_mdct shall be encoded with 0.
8.4.11.3.5 tw_data
The bitstream element tw_data should not be present in Extended HE AAC Profile complying bitstreams,
due to restrictions of bitstream element tw_mdct.
8.5 USAC Decoders
8.5.1 General
This document describes a set of test conditions that shall be applied to verify that a given USAC decoder
implementation complies with this standard. Test conditions are designed such that each tool can be
tested isolated, thus setting the constraints for the corresponding conformance test sequences.
However, some tools show interactions and dependencies. To cover that fact, test cases are defined that
can be composed of one or more test conditions.
Every line in the electronic attachment “Usac_Conformance_Tables.xlsx” represents a test case. For each
test case in the worksheet a set of conformance test sequences are provided as an electronic attachment
to this document. Which tool or tool combination is tested by a given test sequence can be deduced
from its filename, as it follows the nomenclature defined in Table 149. In most cases a conformance test
sequence consists of an USAC encoded bitstream wrapped in the MP4 file format and the corresponding
decoded wave file. Decoded wave files are always supplied with 24 bit resolution (RIFF (little-endian)
data, WAVE audio, Microsoft PCM, 24 bit).
To claim conformance, every test sequence mandatory for a certain profile / level combination has to
meet the conformance criteria specified for the given test. Bitstream restrictions depending on profile
and level are described in 8.4.11.
For each test case varying conformance criteria may apply. The output of the implementation under test
has to be tested against the reference by applying the appropriate test procedure. Test procedures as
well as constraints for each test case are listed in the electronic attachment “Usac_Conformance_Tables.
xlsx”. All test procedures are defined in 8.3.2.
8.5.2 FD core mode tests
This Subclause describes test conditions to test the transform based (FD: frequency domain) part of the
decoder.
A full list of all FD core related
...