ISO/IEC 13818-2:1996/Cor 2:1997
(Corrigendum)Information technology — Generic coding of moving pictures and associated audio information: Video — Technical Corrigendum 2
Information technology — Generic coding of moving pictures and associated audio information: Video — Technical Corrigendum 2
Technologies de l'information — Codage générique des images animées et du son associé: Données vidéo — Rectificatif technique 2
General Information
Relations
Buy Standard
Standards Content (Sample)
INTERNATIONAL STANDARD ISO/IEC 13818-2:1996
TECHNICAL CORRIGENDUM 2
Published 1997-l 2-15
INTERNATIONAL ORGANIZATION FOR STANDARDIZATION l MEmYHAPOflHAR OPfAHM3ALW-4 f-l0 CTAHJJAPTM3ALWl . ORGANISATION INTERNATIONALE DE NORMALISATION
l MEJKJjVHAPOjJHAfi 3JlEKTPOTEXHM~ECKAR KOMMCCMR l coMMlssl0N ~LECTROTECHNIOUE ~NTERNATIONALE
INTERNATIONAL ELECTROTECHNICAL COMMISSION
eu
Information technology - Generic coding of moving
pictures and associated audio information: Video
TECHNICAL CORRIGENDUM 2
Codage des images animkes et du son associk Vidko
Technologies de I’informa tion -
RECTIFICA T/F TECHNIQUE 2
Technical Corrigendum 2 to International Standard ISOAEC 13818-2:1995 was prepared by Joint Technical
Committee ISOAEC JTC 1, Information technology, Subcommittee SC 29, Coding of audio, picture, multimedia and
hypermedia information.
Ref. No. lSO/IEC 13818-2:1996/Cor.2:1997(E)
KS 35.040
Descriptors: data processing, moving pictures, image processing, video recording, video data, data converting, coding (data conversion).
0 lSO/lEC 1997
Printed in Switzerland
---------------------- Page: 1 ----------------------
ISO/IEC 13818-2 : 1996Kor.2 : 1997 (E)
INTERNATIONAL STANDARD
ITU-T RECOMMENDATION
INFORMATION TECHNOLOGY - GENERIC CODING OF MOVING PICTURES
AND A4SSOCIATED AUDIO INFORMATION: VIDEO
TECHNICAL CORRIGENDUM 2
Introduction
Some corrigenda items are followed by an informative background.
Subclause 3.44
1)
Replace “in SI units” by “in spatial measurement units such as centimetre”
Background
SI (Systeme International) units are used by engineers to measure lots of different things. In this case, we are talking
only of spatial measurement units such as centimetre. Using the term SI is not necessary and, furthermore, the
abbreviation SI is not defined in this Recommendation 1 International Standard.
Subclause 3.78
2)
Add the following definition immediately after 3.78.
3.79 Inverse DCT, IDCT: Inverse discrete cosine transform, as defined in Annex A.
and renumber the subsequent items.
Subclause 4.6
3)
At the end of this subclause, after the dfinition of “vlclbf’, add the following text that clariJies the use of those
abbreviations in the syntax clause:
In the clauses describing the syntax, any syntactic element that can only take positive or unsigned values (such as a flag
that can be equal to 0 or 1) is described with the mnemonic ‘uimsbf ‘. If the syntactic element can have a negative value,
it is described with the mnemonic ‘simsbf ‘. If the syntactic element has a constant value (e.g. marker bit) then it is
described with the mnemonic ‘bslbf ‘. If the syntactic element represents a variable-length code, it is de&bed with the
mnemonic ‘vlclbf ‘.
Subclause 6.1.1.4
4)
At the beginning of 6. I. 1.4, insert the following text:
A coded picture is made of a picture header, the optional extensions immediately following it and the following picture
data. A coded picture may be a coded frame or a coded field.
ITU-T Rec. H.262 (1995)/Cor.2 (1996 E)
1
---------------------- Page: 2 ----------------------
ISO/IEC 13818-2 : 1996Kor.2 : 1997 (E)
An I-frame picture or a pair of field pictures, where the first field picture is an I-picture and the second field picture is an
I-picture or a P-picture, is called a coded I-frame.
A P-frame picture or a pair of P-field pictures is called a coded P-frame.
A B-frame picture or a pair of B-field pictures is called a coded B-frame.
A coded I-frame, a coded P-frame or a coded B-frame is called a coded frame.
Background
Currently coded I-frame, coded P-frame and coded B-frame are not defined by any text except in clause 3.
Subclause 6.2.1
5)
In the fifth paragraph, change “Table 6- 1 defines the slice start code values . .” to “Table 6-l defines the start code
values . . .“.
Subclause 6.2.2.6
6)
codefiom “bslbf’ to “uimsbf ‘.
Change the mnemonic for time
-
Subclause 6.2.3.6
7)
or copyfiom “bslbf’ to “uimsbf ‘.
Change the mnemonic for original
- -
“bslbf’ to “uimsbf ‘.
Change the mnemonic for copyright flag@om
-
Change the mnemonic for reservedporn “uimsbf’ to “bslbf ‘.
Background
today, only one possible value is
As of for those reserved bits.
Subclause 6.2.4
8)
Change the mnemonic for reserved bitsjkom “uimsbf’ to “bslbf ‘.
-
Background
As of today, only one possible value is permitted for those reserved bits.
Subclause 6.2.6
9)
Define the mnemonic for “First DCT coefflicient” and “Subsequent DCT coefficients” to be “vlclbf’.
Subclause 6.3.9
10)
Replace the de$nition of temporal reference with the following:
-
temporal reference -
The temporal reference is a lo-bit unsigned integer associated with each coded picture.
- -
The following simple specification applies only when low delay is equal to zero.
-
When a coded frame is in the form of two field pictures, the temporal reference associated with each picture shall be the
same (it is called the temporal-reference of the coded frame). The
temporal reference of each coded frame shall
-
increment by one modulo 1024 when examined in display order at the output of the decoding process, except when a
group of pictures header occurs. Among the frames coded after a group of pictures header, the temporal reference of the
-
coded frame that is displayed first, shall be set to zero.
2 ITU-T Rec. H.262 (1995)/Cor.2 (1996 E)
---------------------- Page: 3 ----------------------
ISO/IEC 13818-2 : 1996Kor.2 : 1997 (E)
The following more general specification applies when low delay is equal to zero or one.
-
If picture A is not a big picture, i.e. the VBV buffer is only examined once before the coded picture A is removed from
the VBV buffer and if N is the temporal-reference of picture A, then the temporal-reference of picture B immediately
following picture A in display order is equal to:
0
0 if there is a group of pictures header present between picture A and picture B (in coded order).
l (N + 1) % 1024 if picture B is a frame picture or is the first field of a pair of field pictures.
0
N if picture B is the second field of a pair of field pictures.
When low delay is equal to one, there may be situations where the VBV buffer shall be re-examined several times
before removing a coded picture (referred to as a big picture) from the VBV buffer.
If picture A is a big picture and if K is the number of times that the VBV buffer is re-examined as defined in C.7 (K > 0),
if N is the temporal-reference of picture A, then the temporal-reference of picture B immediately following picture A in
display order is equal to:
0
K % 1024 if there is a group of pictures header present between picture A and picture B (in coded order).
l (N + K + 1) % 1024 if picture B is a frame picture or is the first field of a pair of field pictures.
l (N + K) % 1024 if picture B is the second field of a pair of field pictures.
NOTE 1 - If the big picture is the first field of a frame coded with field pictures, then the temporal reference of the two
-
field pictures of that coded frame are not identical.
Background
A minor technical error was found in 6.3.9, definition of temporal reference in the presence of big pictures. The current
specification is incorrect when the big picture happens to be the first field of a pair of field pictures, like in the following
example:
The sequence consists of six field pictures.
Picture Tl (first field of the second coded frame) is big.
TO BO -- --
decode Tl B2 T3 B3
display TO TO Tl T3 top field
BO BO B2 B3 bottom field
Clearly, in this case, the temporal reference of the bottom field following Tl (Tl is the big field-picture) should be
2 (B2), but according to the current text, it would have to be 3, which is incorrect. The definition of temporal
reference
-
should be changed to handle the case of the first field of a pair of field pictures being a big picture.
In Table 6-l 2, add a missing space between “EC” and “11172”.
Subclause 6.3.10
11)
Right before the paragraph for top field first, add the following two bullets:
- -
0
vbv delay;
-
0
temporal reference.
-
Subclause 6.3.11
12)
Iden@ by “u” the horizontal axis and by “v” the vertical axis, as in Figure 7-2.
Background
Axis of the default quantiser matrices are not identified. They should match the convention of 7.3.
by “The first value shall always be 8
Under intra quantiser matrix, replace “The first value shall always be 8.”
(values 1 to land 9 to 235 are reserved).”
Under chroma intra quantiser matrix, replclce “The first value shall always be 8.” by “The first value shall always
be 8 (values 1 to 7 and9 to 255 ai reserved).”
ITU-T Rec. H.262 (1995)/Cor.2 (1996 E) 3
---------------------- Page: 4 ----------------------
ISO/IEC 13818-2 : 1996Kor.2 : 1997 (E)
Subclause 6.3.17
13)
Thefirst line in the NOTE, add a missing space between IEC and 11172.
Remove the second bullet that is misleading:
l
The first and last macroblock of a slice shall not be skipped.
Add the following text at the end of 6.3. I7 just before 6.3. I7. I.
It should be noted that the syntax does not allow the first and last macroblock of a slice to be skipped.
Subclause 6.3.17.3
14)
Under dmvector[t], change “as described in 7.6.3.1” to “as described in 7.6.3.6”.
Subclause 6.3.17.4
15)
‘4:4:4’)“, change “for (i = 8; i < 12; i++)” to “for (i = 6; i < 12; i++)“.
Under “if (chroma format = =
-
Two paragraphs before Table 6-20, replace “. defined in Figures 6-8, 6-9 and 6-l 0” by “. defined in Figures 6-l 0,
6-l 1 and 6-12”.
Clause 7
16)
Remove the secondparagraph “With the exception of the Inverse Discrete Transform . . . this Specification.”
Background
The above paragraph belongs to ISO/IEC 138 18-4 (Conformance).
This corrigendum provides a solution to a genuine problem identified in ITU-T Rec. H.262 1 ISO/IEC 138 18-2 (MPEG-2
Video) known as the IDCT internal register overflow problem. The proposed solution is a compromise between various
possible solutions discussed by email by members of the Video group.
Without this correction, it is not possible to define unambiguously what is a conformant video bitstream and what is a
conformant video decoder in ISO/IEC 13 8 18-4 (Conformance).
The proposed corrigendum clarifies the definiti on and requirements on the IDCT used by a conforming decoding
process.
It was demonstrated that, in some cases, compliant bitstreams cannot be decoded properly by hardware-based decoders
that have an IDCT that complies with the requirements currently in ITU-T Rec. H.262 1 ISO/IEC 138 18-2 (Annex A).
Because it is required that a compliant decoder decode properly (i.e. with the arithmetic accuracy requirements of
part-2), a compliant bitstream (that has the profile and level of the decoder), this situation is causing a contradiction. The
solution to this problem is to change the specification and requirements on the IDCT used by the decoding process.
The correction does not put unreasonable burden on decoder designs to cover cases that are very infrequent, where some
internal registers used in some hardware IDCT implementations could overflow.
The correction does not call for non-conformance of bitstreams that would cause such IDCT implementations to
overflow. Instead, it just specifies that the output of the decoding process is undefined in this case (i.e. IDCT internal
register overflow may happen).
However, the correction introduces a new requirement that IDCT b
...
Questions, Comments and Discussion
Ask us and Technical Secretary will try to provide an answer. You can facilitate discussion about the standard in here.