ISO/IEC 18477-8:2020

INTERNATIONAL ISO/IEC
STANDARD 18477-8
Second edition
2020-05
Information technology — Scalable
compression and coding of
continuous-tone still images —
Part 8:
Lossless and near-lossless coding
Technologies de l'information — Compression échelonnable et codage
d'images plates en ton continu —
Partie 8: Codage sans perte et quasi sans perte
Reference number
©
ISO/IEC 2020
© ISO/IEC 2020
All rights reserved. Unless otherwise specified, or required in the context of its implementation, 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
CP 401 • Ch. de Blandonnet 8
CH-1214 Vernier, Geneva
Phone: +41 22 749 01 11
Fax: +41 22 749 09 47
Email: copyright@iso.org
Website: www.iso.org
Published in Switzerland
ii © ISO/IEC 2020 – All rights reserved

Contents Page
Foreword .iv
Introduction .v
1 Scope . 1
2 Normative references . 1
3 Terms, definitions, symbols and abbreviated terms . 1
3.1 Terms and definitions . 1
3.2 Symbols . 5
3.3 Abbreviated terms . 6
4 Conventions . 6
4.1 Conformance language . 6
4.2 Operators . 6
4.2.1 Arithmetic operators . 6
4.2.2 Logical operators . 7
4.2.3 Relational operators . 7
4.2.4 Precedence order of operators . 7
4.2.5 Mathematical functions . 7
5 General . 8
5.1 General definitions . 8
5.2 Overview of this document . 8
5.3 Profiles .10
5.4 Encoder requirements .10
5.5 Decoder requirements.10
Annex A (normative) Encoding and decoding process .12
Annex B (normative) Boxes .17
Annex C (normative) Multi-component decorrelation transformation .25
Annex D (normative) Entropy coding of residual data in the DCT-bypass and large range mode .28
Annex E (normative) Discrete cosine transformation .39
Annex F (normative) Component upsampling .52
Annex G (normative) Quantization and noise shaping for the DCT-bypass process .54
Annex H (normative) Profiles .57
Bibliography .58
© ISO/IEC 2020 – All rights reserved iii

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.
The procedures used to develop this document and those intended for its further maintenance are
described in the ISO/IEC Directives, Part 1. In particular, the different approval criteria needed for
the different types of document should be noted. This document was drafted in accordance with the
editorial rules of the ISO/IEC Directives, Part 2 (see www .iso .org/ directives).
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. Details of any patent rights identified during the development of the document will be in the
Introduction and/or on the ISO list of patent declarations received (see www .iso .org/ patents) or the IEC
list of patent declarations received (see http:// patents .iec .ch).
Any trade name used in this document is information given for the convenience of users and does not
constitute an endorsement.
For an explanation of the voluntary nature of standards, the meaning of ISO specific terms and
expressions related to conformity assessment, as well as information about ISO's adherence to the
World Trade Organization (WTO) principles in the Technical Barriers to Trade (TBT) see www .iso .org/
iso/ foreword .html.
This document was prepared by Joint Technical Committee ISO/IEC JTC 1, Information technology,
Subcommittee SC 29, Coding of audio, picture, multimedia and hypermedia information.
This second edition cancels and replaces the first edition (ISO/IEC 18477-8:2016), which has been
technically revised.
The main changes compared to the previous edition are as follows:
— Annex F.2 has been revised to adopt centred upsampling by default;
— minor editorial changes throughout.
A list of all parts in the ISO/IEC 18477 series can be found on the ISO website.
Any feedback or questions on this document should be directed to the user’s national standards body. A
complete listing of these bodies can be found at www .iso .org/ members .html.
iv © ISO/IEC 2020 – All rights reserved

Introduction
This document specifies a coded codestream format for storage of continuous-tone high and low
dynamic range photographic content. This is a scalable lossy to lossless image coding system supporting
multiple component images consisting of integer samples between 8- and 16-bit resolution, or floating
point samples of 16-bit resolution. It is by itself an extension of ISO/IEC 18477-6 and ISO/IEC 18477-7,
which specify intermediate range and high-dynamic range image decoding algorithms. Both of these
are based on the box-based file format specified in ISO/IEC 18477-3, which is again an extension of
ISO/IEC 18477-1; the codestream is composed in such a way that legacy applications conforming to
Rec. ITU-T T.81 | ISO/IEC 10918-1 are able to reconstruct a lossy, low dynamic range, 8 bits per sample
version of the image.
Today, the most widely used digital photography format, a minimal implementation of JPEG (specified
in Rec. ITU-T T.81 | ISO/IEC 10918-1), uses a bit depth of 8; each of the three channels that together
compose an image pixel is represented by 8 bits, providing 256 representable values per channel.
For more demanding applications, it is not uncommon to use a bit depth of 16, providing 65 536
representable values to describe each channel within a pixel, resulting in over 2.8×10 representable
colour values. In some less common scenarios, even greater bit depths are used, requiring a floating-
point sample representation.
Most common photo and image formats use an 8-bit or 16-bit unsigned integer value to represent some
function of the intensity of each colour channel. While it might be theoretically possible to agree on
one method for assigning specific numerical values to real world colours, doing so is not practical.
Since any specific device has its own limited range for colour reproduction, the device’s range may be a
small portion of the agreed-upon universal colour range. As a result, such an approach is an extremely
inefficient use of the available numerical values, especially when using only 8 bits (or 256 unique
values) per channel. To represent pixel values as efficiently as possible, devices use a numeric encoding
optimized for their own range of possible colours or gamut.
This document is primarily designed to encode intermediate or high dynamic image sample values
without loss, or with a precisely controllable bounded loss using the tools defined in ISO/IEC 18477-1
and some minimal extensions of those tools. The goal is to provide a backwards-compatible coding
specification that allows legacy applications and existing toolchains to continue to operate on
codestreams conforming to this document.
JPEG XT has been designed to be backwards compatible to legacy applications while at the same time
having a small coding complexity; JPEG XT uses, whenever possible, functional blocks of Rec. ITU-T T.81
| ISO/IEC 10918-1 to extend the functionality of the legacy JPEG coding system. It is optimized for
storage and transmission of intermediate and high dynamic range and wide colour gamut 8- to 16-
bit integer or 16-bit floating point images while also enabling low-complexity encoder and decoder
implementations.
This document is an extension of ISO/IEC 18477-1, a compression system for continuous tone digital
still images which is backwards compatible with Rec. ITU-T T.81 | ISO/IEC 10918-1. That is, legacy
applications conforming to Rec. ITU-T T.81 | ISO/IEC 10918-1 will be able to reconstruct streams
generated by an encoder conforming to this document, though will possibly not be able to reconstruct
such streams in full dynamic range, full quality or without loss.
This document is itself based on ISO/IEC 18477-3 that defines a box-based file format similar to
other JPEG standards. It also contains elements of ISO/IEC 18477-6 and ISO/IEC 18477-7. The aim
of this document is to provide a migration path for legacy applications to support lossless coding of
intermediate and high dynamic range images, that is images that are either represented by sample
values requiring 8- to 16-bit precision, or even using 16-bit floating point sample resolution. While Rec.
ITU-T T.81 | ISO/IEC 10918-1 already defines a lossless mode for integer samples, images encoded in
this mode cannot be decoded by applications only supporting the lossy 8-bit-mode; the coding engine
for lossless coding in Rec. ITU-T T.81 | ISO/IEC 10918-1 is completely different from the lossy coding
mode. Unlike the legacy standard, this document defines a lossless scalable coding engine supporting
all bit depths between 8 and 16 bits per sample, including 16-bit floating point samples, while also
staying compatible with legacy applications. Such applications will continue to work, but will only able
© ISO/IEC 2020 – All rights reserved v

to reconstruct a lossy 8-bit standard low dynamic range (LDR) version of the full image contained in
the codestream. The ISO/IEC 18477 series specifies a coded file format, referred to as JPEG XT, which is
designed primarily for storage and interchange of continuous-tone photographic content.
vi © ISO/IEC 2020 – All rights reserved

INTERNATIONAL STANDARD ISO/IEC 18477-8:2020(E)
Information technology — Scalable compression and
coding of continuous-tone still images —
Part 8:
Lossless and near-lossless coding
1 Scope
This document specifies a coding format, referred to as JPEG XT, which is designed primarily for
continuous-tone photographic content. This document defines extensions that allow lossless coding of
such content while staying compatible with the core coding system specified in ISO/IEC 18477-1.
2 Normative references
The following documents are referred to in the text in such a way that some or all of their content
constitutes requirements of this document. For dated references, only the edition cited applies. For
undated references, the latest edition of the referenced document (including any amendments) applies.
ISO/IEC 18477-1:2020, Information technology — Scalable compression and coding of continuous-tone still
images — Part 1: Scalable compression and coding of continuous-tone still images
ISO/IEC 18477-3:2015, Information technology — Scalable compression and coding of continuous-tone still
images — Part 3: Box file format
ISO/IEC 18477-6:2016, Information technology — Scalable compression and coding of continuous-tone
still images — Part 6: IDR Integer Coding
ISO/IEC 18477-7:2017, Information technology — Scalable compression and coding of continuous-tone still
images — Part 7: HDR Floating-Point Coding
ITU-T T.81 | ISO/IEC 10918-1, Information technology — Digital compression and coding of continuous
tone still images — Requirements and guidelines
ITU-T BT.601, Studio encoding parameters of digital television for standard 4:3 and wide screen 16:9
aspect ratios
3 Terms, definitions, symbols and abbreviated terms
3.1 Terms and definitions
For the purposes of this document, the following terms and definitions apply.
ISO and IEC maintain terminological databases for use in standardization at the following addresses:
— IEC Electropedia: available at http:// www .electropedia .org/
— ISO Online browsing platform: available at http:// www .iso .org/ obp
3.1.1
AC coefficient
any DCT coefficient for which the frequency is not zero in at least one dimension
© ISO/IEC 2020 – All rights reserved 1

3.1.2
ASCII encoding
encoding of text characters and text strings
Note 1 to entry: In accordance with ISO/IEC 10646.
3.1.3
base decoding path
process of decoding legacy codestream and refinement data to the base image, jointly with all further
steps until residual data is added to the values obtained from the residual codestream
3.1.4
base image
collection of sample values obtained by entropy decoding the DCT coefficients of the legacy codestream
and the refinement codestream, and inversely DCT transforming them jointly
3.1.5
box
structured collection of data describing the image or the image decoding process embedded into one or
multiple APP marker segments
Note 1 to entry: See ISO/IEC 18477-3:2015, Annex B for the definition of boxes.
3.1.6
coding process
encoding process, a decoding process, or both
3.1.7
DC coefficient
DCT coefficient for which the frequency is zero in both dimensions
3.1.8
DCT coefficient
amplitude of a specific cosine basis function
Note 1 to entry: May refer to an original DCT coefficient, to a quantized DCT coefficient, or to a dequantized DCT
coefficient.
3.1.9
decoder
embodiment of a decoding process
3.1.10
decoding process
process which takes as its input compressed image data and outputs a continuous-tone image
3.1.11
dequantization
inverse procedure to quantization by which the decoder recovers a representation of the DCT
coefficients
3.1.12
encoder
embodiment of an encoding process
3.1.13
encoding process
process which takes as its input a continuous-tone image and outputs compressed image data
2 © ISO/IEC 2020 – All rights reserved

3.1.14
extension image
residual image
sample values as reconstructed by inverse quantization and inverse DCT transformation applied to the
entropy-decoded coefficients described by the residual scan and residual refinement scans
[SOURCE: ISO/IEC 18477-6:2016, 3.1.54]
3.1.15
fixed point discrete cosine transformation
implementation of the discrete cosine transformation based on fixed point arithmetic
Note 1 to entry: As specified in Annex E.
3.1.16
forward DCT bypass
transformation that takes an 8×8 sample block and prepares it for entropy coding without applying a
discrete cosine transformation
3.1.17
forward fixed point DCT
transformation of an 8×8 sample block from the spatial domain to the frequency domain using the fixed
point arithmetic
Note 1 to entry: As specified in Annex E.
3.1.18
forward integer DCT
transformation of an 8×8 sample block from the spatial domain to the frequency domain using the
integer approximation of the discrete cosine transformation
Note 1 to entry: As specified in Annex E.
3.1.19
inverse DCT bypass
transformation that takes an 8×8 sample block as generated by entropy decoding and level-shifts it
without applying a discrete cosine transformation
3.1.20
inverse fixed point DCT
transformation of an 8×8 sample block from the frequency domain to the spatial domain using the fixed
point arithmetic
Note 1 to entry: As specified in Annex E.
3.1.21
inverse integer DCT
the transformation of an 8×8 sample block from the frequency domain to the spatial domain using the
integer approximation of the discrete cosine transformation
Note 1 to entry: As specified in Annex E.
3.1.22
frequency
two-dimensional index into the two-dimensional array of DCT coefficients
[SOURCE: ISO/IEC 10918-1:1994, 3.1.61]
3.1.23
high dynamic range
HDR
image or image data comprised of more than eight bits per sample
© ISO/IEC 2020 – All rights reserved 3

3.1.24
Huffman encoding
entropy encoding procedure which assigns a variable length code to each input symbol
3.1.25
intermediate dynamic range
image or image data comprised of more than eight bits per sample
3.1.26
legacy codestream
collection of markers and syntax elements
Note 1 to entry: The legacy codestream, as defined by Rec. ITU-T T.81 | ISO/IEC 10918-1 and any syntax elements
defined by the ISO/IEC 18477 series, consists of the collection of all markers except those APP markers that
describe JPEG XT boxes by the syntax defined in ISO/IEC 18477-3:2015, Annex A.
3.1.27
lossless
encoding and decoding processes and procedures in which the output of the decoding procedure(s) is
identical to the input to the encoding procedure(s)
3.1.28
lossless coding
mode of operation which refers to any one of the coding processes in which all of the procedures are
lossless
Note 1 to entry: Coding processes defined in ISO/IEC 18477-8.
3.1.29
lossy
encoding and decoding processes which are not lossless
3.1.30
low-dynamic range
LDR
image or image data comprised of data with no more than eight bits per sample
3.1.31
noise shaping
signal processing technique that removes quantization noise from the low frequency components and
injects it into the high frequency domain where it can be removed by filtering
3.1.32
point transformation
application of a location independent global function to reconstructed sample values in the spatial domain
3.1.33
residual decoding path
collection of operations applied to the entropy coded data contained in the residual data box and
residual refinement scan boxes up to the point where this data is merged with the legacy data to form
the final output image
3.1.34
residual image
sample values as reconstructed by inverse quantization and inverse DCT transformation applied to the
entropy-decoded coefficients described by the residual scan and residual refinement scans
4 © ISO/IEC 2020 – All rights reserved

3.1.35
residual scan
additional pass over the image data invisible to legacy decoders which provides additive and/or
multiplicative correction data of the legacy scans to allow reproduction of high-dynamic range or wide
colour gamut data
3.1.36
refinement scan
additional pass over the image data invisible to legacy decoders which provides additional least
significant bits to extend the precision of the DCT transformed coefficients
Note 1 to entry: Refinement scans can be either applied in the legacy or residual decoding path.
3.1.37
superbox
box that carries other boxes as payload data
3.1.38
sub box
box that is contained as payload data within a superbox
3.1.39
uniform quantization
procedure by which DCT coefficients are linearly scaled in order to achieve compression
3.1.40
upsampling
procedure by which the spatial resolution of a component is increased
3.2 Symbols
X width of the sample grid in positions
Y height of the sample grid in positions
Nf number of components in an image
s subsampling factor of component i in horizontal direction
i,x
s subsampling factor of component i in vertical direction
i,y
H subsampling indicator of component i in the frame header
i
V subsampling indicator of component i in the frame header
i
v sample value at the sample grid position x,y
x,y
R additional number of DCT coefficients bits represented by refinement scans in the base image,
h
8+R is the number of non-fractional bits (i.e. bits in front of the "binary dot") of the output of
h
the inverse DCT process in the base image
R additional number of DCT coefficients bits represented by refinement scans in the residual,
r
P+R is the number of non-fractional bits (i.e. bits in front of the "binary dot") of the output of
h
the inverse DCT process in the residual image where P is the bit depth indicated in the frame
header of the residual codestream
R additional bits in the HDR image. 8+R is the sample precision of the reconstructed HDR image
b b
© ISO/IEC 2020 – All rights reserved 5

3.3 Abbreviated terms
ASCII American Standard Code for Information Interchange
LSB least significant bit
MSB most significant bit
TMO tone mapping operator
DCT discrete cosine transformation
FCT fixed point multi-component transformation
ICT irreversible multi-component transformation
RCT reversible multi-component transformation
JPEG joint photographic experts group
4 Conventions
4.1 Conformance language
The keyword "reserved" indicates a provision that is not specified at this time, shall not be used, and
may be specified in the future. The keyword "forbidden" indicates "reserved" and in addition indicates
that the provision will never be specified in the future.
4.2 Operators
NOTE Many of the operators used in this document are similar to those used in the C programming language.
4.2.1 Arithmetic operators
+ addition
− subtraction (as a binary operator) or negation (as a unary prefix operator)
× multiplication
/ division without truncation or rounding
smod x smod a is the unique value y between −⎾(a−1)/2⏋ and ⎿(a−1)/2⏌
for which y+N×a = x with a suitable integer N.
umod x umod a is the unique value y between 0 and a−1 for which
y+N×a = x with a suitable integer N
6 © ISO/IEC 2020 – All rights reserved

4.2.2 Logical operators
|| logical OR
&& logical AND
! logical NOT
∈ x ∈ {A, B} is defined as (x == A || x == B)
∉ x ∉ {A, B} is defined as (x != A && x != B)
4.2.3 Relational operators
> greater than
>= greater than or equal to
< less than
<= less than or equal to
== equal to
!= not equal to
4.2.4 Precedence order of operators
Operators are listed in descending order of precedence. If several operators appear in the same line,
they have equal precedence. When several operators of equal precedence appear at the same level in an
expression, evaluation proceeds according to the associativity of the operator either from right to left
or from left to right.
Operators Type of operation Associativity
(), [ ], . expression left to right
− unary negation
×, / multiplication left to right
umod, smod modulo (remainder) left to right
+, − addition and subtraction left to right
< , >, <=, >= relational left to right
4.2.5 Mathematical functions
⎾x⏋ ceil of x: returns the smallest integer that is greater than or equal to x
⎿x⏌ floor of x: returns the largest integer that is lesser than or equal to x
|x| absolute value, is –x for x < 0, otherwise x
sign(x) sign of x, 0 if x is zero, +1 if x is positive, −1 if x is negative
clamp(x,min,max) clamps x to the range [min,max]: returns min if x < min, max if x > max or oth-
erwise x
a
x raises the value of x to the power of a: x is a non-negative real number, a is a real
a
number; x is equal to exp(a×log(x)) where exp is the exponential function and
a
log() the natural logarithm; if x is 0 and a is positive, x is defined to be 0
© ISO/IEC 2020 – All rights reserved 7

5 General
5.1 General definitions
This clause gives an informative overview of the elements specified in this document. It also introduces
many of the terms which are defined in Clause 3. These terms are printed in italics upon first usage in
this clause.
There are three elements specified in this document:
a) An encoder is an embodiment of an encoding process. An encoder takes as input digital source image
data and encoder specifications and, by means of a specified set of procedures, generates as output
codestream.
b) A decoder is an embodiment of a decoding process. A decoder takes as input a codestream and, by
means of a specified set of procedures, generates as output digital reconstructed image data.
c) The codestream is a compressed image data representation, which includes all necessary data to
allow a (full or approximate) reconstruction of the sample values of a digital image. Additional data
might be required that define the interpretation of the sample data, such as colour space or the
spatial dimensions of the samples.
5.2 Overview of this document
This document allows near-lossless and lossless coding of high and intermediate dynamic range
of photographic images in a way that is backwards compatible to Rec. ITU-T T.81 | ISO/IEC 10918-1.
Decoders compliant to Rec. ITU-T T.81 | ISO/IEC 10918-1 will be able to parse codestreams conforming
to this document correctly, albeit in less precision, with a limited dynamic range, and potential loss in
sample bit precision.
This document includes multiple tools to reach the above functionality, defined in Annex B and on. A
short overview of these coding tools is given in this clause.
The syntax of an ISO/IEC 18477-8 compliant codestream is specified in ISO/IEC 18477-3, that is, this
document uses a syntax element denoted as "box" to annotate its syntactical elements. The definition of
the box syntax element is not repeated here (refer to ISO/IEC 18477-3). Additional boxes besides those
already specified in ISO/IEC 18477-3 are defined in Annex B. In addition, this document also reuses
boxes defined in ISO/IEC 18477-6:2016, Annex B and ISO/IEC 18477-7:2017, Annex B.
8 © ISO/IEC 2020 – All rights reserved

Figure 1 — Overview of the decoding process
To allow lossless and near-lossless coding, this document provides a stricter definition of two elements
of the reconstruction process defined in Rec. ITU-T T.81 | ISO/IEC 10918-1 and ISO/IEC 18477-1. The DCT
process, only loosely defined in an implementation-agnostic way in Rec. ITU-T T.81 | ISO/IEC 10918-1
is replaced by strictly defined algorithms (specified in Annex E) that a conforming decoder shall
follow. Not following these steps will compromise lossless reconstruction. Annex C replaces the ICT
transformation by a precise fixed-point implementation denoted as FCT operating entirely on integer
samples and thus allowing a fully reproducible transformation conforming decoders shall follow as
well. Again, deriving from the specifications of the FCT specified in Annex C will compromise lossless
coding. The DCT operations in Annex E and the FCT in Annex C are fully backwards-compatible to the
DCT in Rec. ITU-T T.81 | ISO/IEC 10918-1 and the ICT in ISO/IEC 18477-1 and approximate them within
the error bounds of Rec. ITU-T T.83 | ISO/IEC 10918-2. Thus, a possible implementation choice for the
ISO/IEC 18477 series is to always use the DCT and/or FCT as specified here, and not to provide a second
implementation based on floating point or other technology.
Lossless coding can be achieved by two alternative mechanisms. First, by applying the integer DCT
specified in Annex E, and replacing the base transformation by an identity transformation. This coding
mode can only be applied to bit precisions of 8 bits per sample, or up to 12 bits per sample in the
presence of refinement scans already defined in ISO/IEC 18477-6. Residual scans are not required if
the integer DCT is deployed.
Second, by replacing the DCT by the fixed point DCT and by selecting the FCT, the scaled identity
transformation or an integer-based free-form transformation as base transformation. The fixed
point DCT is specified in Annex E. The FCT and modifications to free-form integer transformations
are defined in Annex C. In this case, FCT and the fixed point DCT create an additional coding error
that is precisely defined by the coding procedure. This coding error is then corrected by an additive
residual scan. While residual scans were already defined in ISO/IEC 18477-6, the Residual DCT
Specification box defined in Annex B allows users to bypass the DCT in the residual image completely
© ISO/IEC 2020 – All rights reserved 9

and thus avoids additional complexity not required for lossless coding. Residual data using the DCT
bypass mode are entropy coded in a new scan type denoted as residual scan defined in Annex D. It
is closely related to the regular (baseline, extended or progressive) Huffman scan types specified in
Rec. ITU-T T.81 | ISO/IEC 10918-1. Since all coefficients are now error residuals in the spatial domain,
the natural distinction between DC and AC coefficients no longer applies. This means that the special
role the DC coefficient has in the Rec. ITU-T T.81 | ISO/IEC 10918-1 decoding procedure is no longer
justified. The residual scan type thus extends the AC decoding procedure to the top-left coefficient of
an 8×8 block while keeping everything else unchanged. It is thus only a very minor modification of the
decoding process specified in Rec. ITU-T T.81 | ISO/IEC 10918-1 that improves the coding efficiency of
DCT-bypassed coded error residuals.
If the DCT is bypassed, quantization in the residual domain may cause ringing and stair-casing artefacts.
Such artefacts can be eliminated by the noise shaping algorithm specified in Annex G.
To decorrelate the components of the additive error residuals, Annex C specifies an additional lossless
component decorrelation transformation denoted as the RCT. It is related, but not identical, to the RCT
of Rec. ITU-T T.801 | ISO/IEC 15444-1.
The decoding procedure of this document is otherwise closely related to that of ISO/IEC 18477-6.
Legacy codestream and refinement scans in the Refinement box form the DCT coefficients of the base
image. The image is dequantized and then processed by either the fixed point DCT or the integer DCT. A
multi-component decorrelation transformation, the identity, the scaled identity, the FCT or a free-form
integer transformation follows. The output of this process is optionally processed by a non-linear point
transformation selected by the base Non-linear Point Transformation box. The output of this process is
called the precursor image.
Image reconstruction proceeds with the entropy decoding of the residual image, if present,
encapsulated by the Residual Data box and the Residual Refinement box, both of which also specified
in ISO/IEC 18477-6:2016, Annex B. Data is dequantized and either processed through the integer DCT,
or the DCT bypass process specified in Annex E. The DCT bypass process requires residual data to be
encoded in the residual scan type of Annex D. Output of this data is then linearly scaled to range, if
required, and inversely decorrelated either by the RCT or an identity transformation. The error residuals
are finally added to the precursor image, and modulo arithmetic is used in the addition to ensure a
final reconstruction value in range. If the encoded data is floating point, the lossless conversion from
floating point to integer specified in ISO/IEC 18477-7:2017, Annex D completes decoding, otherwise the
sum of residual and precursor image is already the final output.
For the detailed specification of the decoding and merging process, see Annex A.
5.3 Profiles
The profiles define the implementation of a particular technology within the functional blocks of
Figure 1. Profiles are defined in Annex H.
5.4 Encoder requirements
There is no requirement in this document that any encoder shall support all profiles. An encoder is
only required to meet the compliance tests and to generate the codestream according to the syntax and
to limit the coding parameters to those valid within the profile it conforms to. Profiles are defined in
Annex H.
5.5 Decoder requirements
A decoding process converts compressed image data to reconstructed image data. It may follow the
decoding operation specified in this document and ISO/IEC 18477-1 to generate an LDR image from
the legacy codestream, and it shall follow the operations in this document to decode an IDR or an
HDR image from the data in the full file. The Decoder shall parse the codestream syntax to extract the
parameters, the residual image and the base image. The parameters shall be used to merge the residual
image with the base image into the reconstructed IDR image.
10 © ISO/IEC 2020 – All rights reserved

In order to comply with this document, a decoder:
a) may convert a codestream conforming to this document without considering the information
in any box into to a low dynamic range image;
b) shall convert a conforming codestream within the profile it claims to be conforming to into an
intermediate dynamic range image to exactly the same sample values as the reference decoder
specified in ISO/IEC 18477-5. Additional details on reference testing and allowable error bounds
are specified in ISO/IEC 18477-4.
© ISO/IEC 2020 – All rights reserved 11

Annex A
(normative)
Encoding and decoding process
A.1 Decoding process
The decoding process relies on a layered approach to extend the capabilities of the Rec.
ITU-T T.81 | ISO/IEC 10918-1 process. The encoder decomposes an IDR or HDR image into a base layer,
which consists of a tone-mapped version of the IDR/HDR image, and a residual layer. In addition to
the residual layer, the codestream includes a description of an approximate inverse tone mapping
operation that allows the decoder to reconstruct from the LDR image an approximate IDR/HDR image
denoted as precursor image. The errors of this approximation process, if any, are corrected by the
residual codestream included in the residual data box and residual refinement box (see Annex B). Both
the description of the tone mapping and the residual image are included in boxes invisible to legacy
decoders. Such decoders will thus only see the tone mapped LDR image. While the base image complies
to ISO/IEC 18477-1 and thus supports only the 8-bit extended or baseline, extended or progressive
Huffman modes, the residual image may optionally be encoded in the 12-bit Huffman or progressive
modes and may optionally use the residual scan type of Annex D bypassing the inverse DCT.
Figure A.1 illustrates the functionality of a compliant decoder.
Figure A.1 — High-level overview of the decoding process of a compliant decoder
Bold lines carry three components (or one for greyscale). Round boxes implement point-transformations,
square boxes (except B1, B1a, B5, B5a) multiplications by 3×3 matrices. Letters denote signal names.
12 © ISO/IEC 2020 – All rights reserved

The reconstruction process of an intermediate dynamic range image from a LDR image and a residual
image decoders shall follow the following steps (see Figure A.1).
— In steps B1 and B1a, decode the base DCT coefficients from legacy codestream and the refinement
codestream if a Refinement Data box is present. Refinement coding is specified in Annex D.
Otherwise, the decoding process of Rec. ITU-T T.81 | ISO/IEC 10918-1 is used unaltered.
— In step B1b, apply inverse quantization as specified in Rec. ITU-T T.81 | ISO/IEC 10918-1.
— In step B1c, apply inverse discrete cosine transformation. The inverse DCT process is either
the integer DCT or the fixed point DCT of Annex E, and the DCT process is selected by the DCT
Specification box specified in Annex B.
— In step B2, the upsampling process specified in Annex F shall be followed to generate samples for all
positions on the sample grid.
— In step B3 the linear transformation, as selected by the Base Transformation box defined in
Annex B, shall be applied to inversely decorrelate the image components. Table B.1 defines which
transformation to pick. The output of this block consists of either one or three samples per grid
point O , depending on the number of components in the legacy codestream. The output of this
i
transformation is rounded to integers and clipped to [0,255].
— In step B4, a non-linear point transformation shall be applied to each of the output components
O . This process is selected according to the Base non-linear Point Transformation subbox of
i
the Merging Specification box, implementing L of Annex C and following the specifications of
i
ISO/IEC 18477-3:2015, Annex C. The outputs of this process are the predicted high dynamic range
samples J . As above, i=1.Nf.
i
— Also in step B4, an integer colour transformation is applied to the input values J resulting in the
i
output pixel values H . The transformation is selected by the Colour Transformation subbox of the
i
Merging Specification box, which selects one of the transformations defined in Annex C. If Nf equals
1, no transformation is performed.
— In steps B5 and B5a, the residual image shall be reconstructed from the data contained in the
Residual Codestream box and the Residual Refinement box. The codestream contained in this box
either follows the specifications defined in Rec. ITU-T T.81 | ISO/IEC 10918-1 or is encoded in the
residual scan specified in Annex D. If a Residual Refinement box is present, the precision of the
samples of the residual codestream shall be extended by refinement coding as specified in Annex D
of this document and ISO/IEC 18477-6:2016, Annex D. The number of components of the residual
image shall be equal to the number of components signalled in the base image.
— In step B5b, inverse quantization is performed. If the DCT bypass operation is selected, an optional
Noise Shaping process is applied to the reconstructed coefficients. Noise shaping is specified in
Annex G. Otherwise, the inverse quantization procedure of Rec. ITU-T T.81 | ISO/IEC 10918-1 applies
unaltered.
— In step B5c apply either the Integer DCT or the DCT bypass as specified in Annex E. The DCT
operation is selected by the Residual DCT Specification box specified in Annex B and shall fit to
the scan type of the residual image.
— In step B6, residual data is upsampled to the common sample grid following the specification of
Annex F.
— In step B7, apply a non-linear point transformation that is defined by the Residual non-linear Point
Transformation subbox of the Merging Specification box. The outputs of this operation are one or
three residual sample values per pixel denoted by P .
i
— The outputs of the residual decoding process are one or three integer sample values P per sample
i
grid point.
© ISO/IEC 2020 – All rights reserved 13

— In step B8, an inverse colour decorrelation shall be applied to the P data. This is either the identity
i
transformation or the Reversible Colour Transformation (RCT), selected by the Residual
Transformation box of Annex B. The RCT is specified in Annex C. The outputs of this process are the
inversely decorrelated prediction errors Q .
i
— In step B9, the intermediate dynamic range output F , i.e. the output of the decoding process, is
i
reconstructed from H , the predicted high dynamic range signal, and the inversely decorrelation
i
prediction errors Q . The output bit precision R is taken from the Output Conversion subbox of the
i b
Merging Specification box defined in Annex B. For this, compute
Rb+8−1 Rb+8
F = H + Q − 2   umod   2
i i i
— If the Oc flag of the Output Conversion box is set, the val
...