1、INTERNATIONAL STANDARD ISO/IEC 14496-2 Second edition 2001-12-01 AMENDMENT 2 2002-02-01 Information technology - Coding of audio-visual objects - Part 2: Visual AMENDMENT 2: Streaming video profile Technologies de linformation - Codage des objets audiovisuels - Partie 2: Codage visuel AMENDEMENT 2:
2、Cours du profil vido Adopted by INCITS (InterNational Committee for Information Technology Standards) as an American National Standard. Date of ANSI Approval: 8/29/02 Published by American National Standards Institute, 25 West 43rd Street, New York, New York 10036 Copyright 2002 by Information Techn
3、ology Industry Council (ITI). All rights reserved. These materials are subject to copyright claims of International Standardization Organization (ISO), International Electrotechnical Commission (IEC), American National Standards Institute (ANSI), and Information Technology Industry Council (ITI). No
4、t for resale. No part of this publication may be reproduced in any form, including an electronic retrieval system, without the prior written permission of ITI. All requests pertaining to this standard should be submitted to ITI, 1250 Eye Street NW, Washington, DC 20005. Printed in the United States
5、of America Reference number ISOIIEC 14496-2:2001 /Amd.2:2002(E) O ISOIIEC 2002 ISO/I EC 196-2:2001/Amd.2:2002(E) PDF disclaimer This PDF file may contain embedded typefaces. In accordance with Adobes licensing policy, this file may be printed or viewed but shall not be edited unless the typefaces wh
6、ich are embedded are licensed to and installed on the computer performing the editing. In downloading this file, parties accept therein the responsibility of not infringing Adobes licensing policy. The IS0 Central Secretariat accepts no liability in this area. Adobe is a trademark of Adobe Systems I
7、ncorporated. Details of the software products used to create this PDF file can be found in the General Info relative to the file; the PDFcreation parameters were optimized for printing. Every care has been taken to ensure that the file is suitable for use by IS0 member bodies. In the unlikely event
8、that a problem relating to it is found, please inform the Central Secretariat at the address given below. O ISO/IEC 2002 All rights reserved. Unless othetwise specified, no part of this publication may be reproduced or utilized in any form or by any means, electronic or mechanical, including photoco
9、pying and microfilm, without permission in writing from either IS0 at the address below or ISOs member body in the country of the requester. IS0 copyright office Case postale 56 CH-1 21 1 Geneva 20 Tel. + 41 22 749 O1 11 Fax +41 227490947 E-mail copyrightiso.ch Web www.iso.ch Printed in Switzerland
10、ii O ISO/IEC 2002 -All rights reserved ISO/IEC 196-2:2001/Amd.2:2002( E) Foreword IS0 (the International Organization for Standardization) and IEC (the International Electrotechnical Commission) form the specialized system for worldwide standardization. National bodies that are members of IS0 or IEC
11、 participate in the development of International Standards through technical committees established by the respective organization to deal with particular fields of technical activity. IS0 and IEC technical committees collaborate in fields of mutual interest. Other international organizations, gover
12、nmental and non-governmental, in liaison with IS0 and IEC, also take part in the work. In the field of information technology, IS0 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, Pa
13、rt 3. 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
14、 bodies casting a vote. Attention is drawn to the possibility that some of the elements of this Amendment may be the subject of patent rights. IS0 and IEC shall not be held responsible for identifying any or all such patent rights. Amendment 2 to International Standard ISO/IEC 14496-2:2001 was prepa
15、red by Joint Technical Committee ISO/IEC JTC 1, Information technology, Subcommittee SC 29, Coding of audio, picture, multimedia and hypermedia information. O ISOIIEC 2002 -All rights reserved iii ISO/IEC 14496-2:2001/Amd.2:2002(E) Information technology - Coding of audio-visual objects - Part 2: Vi
16、sual AMENDMENT 2: Streaming video profile 1) Add the following text to the end of Purpose of Introduction: I Two profiles are developed in response to the growing need for a video coding method for Streaming Video on Internet applications. It provides the definition and description of Advanced Simpl
17、e (AS) Profile and Fine Granularity Scalable (FGS) Profile. AS Profile provides the capability to distribute single-layer frame based video at a wide range of bit rates available for the distribution of video on Internet. FGS Profile uses AS Video Object in the base layer and provides the descriptio
18、n of two enhancement layer types - Fine Granularity Scalability (FGS) and FGS Temporal Scalability (FGST). FGS Profile allows the coverage of a wide range of bit rates for the distribution of video on Internet with the flexibility of using multiple layers, where there is a wide range of bandwidth va
19、riation. I 2) Add the following text into Introduction following Error Resilience: I Fine Granularity Scalability Fine Granularity Scalability (FGS) provides quality scalability for each VOP. Figure AMD2-1 shows a basic FGS decoder structure. FGS Enhancement Decoding (optional output) Figure AMD2-1
20、- A Basic FGS Decoder Structure To reconstruct the enhanced VOP, the enhancement bitstream is first decoded using bit-plane VLD. The decoded block-bps are used to reconstruct the DCT coefficients in the DCT domain which are then right-shifted based on the frequency weighting and selective enhancemen
21、t shifting factors. The output of bit-plane shift is the DCT coefficients of the image domain residues. After the IDCT, the image domain residues are reconstructed. They are added to the reconstructed clipped base-layer pixels to reconstruct the enhanced VOP. The reconstructed enhanced VOP pixels O
22、ISOIIEC 2002 -All rights reserved 1 ISO/I EC 196-2:2001/Amd.2:2002(E) FGS VOP are limited into the value range between O and 255 by the clipping unit in the enhancement layer to generate the final enhanced video. The reconstructed base layer video is available as an optional output since each base l
23、ayer reconstructed VOP needs to be stored in the frame buffer for motion compensation. j The basic FGS enhancement layer consists of FGS VOPs that enhance the quality of the base-layer VOPs as shown in Figure AMD2-2. Base VOP FGS Layer VOP VOP VOP VOP VOP VOP i I I I I I I FGS VOP I“ FGST FGS FGST F
24、GS FGST FGS VOP VOP VOP VOP VOP VOP / Base Layer / / Figure AMD2-2 - Basic FGS Enhancement Structure When FGS temporal scalability (FGST) is used, there are two possible enhancement structures. One structure is to have two separate enhancement layers for FGS and FGST as shown in Figure AMD2-3 and th
25、e other structure is to have one combined enhancement layer for FGS and FGST as shown in Figure AMD2-4. FGST Layer pJ FGS ri/ YI/ YI/ YI Layer VOP VOP VOP VOP Base el Layer Figure AMD2-3 - Two Separate Enhancement Layers for FGS and FGST FGS-FGST Layer Base Layer Figure AMD2-4 - One Combined Enhance
26、ment Layer for FGS and FGST 2 O ISOIIEC 2002 -All rights reserved ISO/IEC 14496-2:2001/Amd.2:2002(E) In either one of these two structures that include FGS temporal scalability, the prediction for the FGS temporal scalable VOPs can only be from the base layer. Each FGS temporal scalable VOP has two
27、separate parts. The first part contains motion vector data and the second part contains the DCT texture data. The syntax for the first part is similar to that in the temporal scalability described in subclause 6.2. The DCT texture data in the second part are coded using bit-plane coding in the same
28、way as that in FGS. To distinguish the temporal scalability in subclause 6.2 and FGS temporal scalability, the FGS temporal scalability layer in Figure AMD2-3 is called “FGST layer”. The combined FGS and FGST layer in Figure AMD2-4 is called “FGS-FGST layer”. The “FGS VOP” shown in Figure AMD2-3 and
29、 Figure AMD2-4 is an fgsc vop with fgs-vopcodingtype being I. The “FGST VOP” shown in Figure AMD2-3 and Figure AMD2-4 is an fgsc vop with fgs-vopcodingtype being P or B. The code value of profile-and-level-indication in VisualObjectSequence() has been extended to include the profile and level indica
30、tions for AS Profile and FGS Profile. The identifier for an enhancement layer is the syntax videoobject-type-indication in VideoObjectLayer(). A unique code is defined for FGS Object Type to indicate that this VOL contains fgsc vops. Another unique code is defined for AS Object Type to indicate that
31、 this VOL is the base-layer. There is a syntax fgs-layer-type in VideoObjectLayer() to indicate whether this VOL is an FGS layer as shown in Figure AMD2-2 and Figure AMD2-3, or an FGST layer as shown in Figure AMD2-3, or an FGS-FGST layer as shown in Figure AMD2-4. Similar to the syntax structure in
32、 subclause 6.2, under each VOL for FGS, there is a hierarchy of fgsc vop, fgsc macroblock, and fgsc block. An fgsc vop starts with a unique fgs-vop-start-code. Within each fgsc vop, there are multiple vop-bps. Each vop-bp in an fgsc vop starts with an fgs-bp-start-code whose last 5 bits indicate the
33、 ID of the vop-bp. In each fgsc macroblock, there are 4 block-bps for the luminance component (Y), 2 block-bps for the two chrominance components (U and V) for the 4:2:0 chrominance format. Each block-bp is coded by VLC. I 3) Add the following subclauses in clause 3 II 3.AMD2.1 3.AMD2.2 3.AMD2.3 3.A
34、MD2.4 3.AMD2.5 3.AMD2.6 3.AMD2.7 3.AMD2.8 block-bp: An array of 64 bits, one from each DCT coefficient at the same significant position of accuracy in a zigzag scan order. When frequency weighting is used, block-bps are formed after the weighting is applied to the DCT coefficients in an 8x8 block. e
35、nd of plane; eop: A symbol to indicate whether a 1 bit is the last 1 bit of a block-bp. fgs block: An 8-row by 8-column matrix of bits, each from one DCT coefficient at the same significant position of accuracy, or its coded representation. The usage is clear from the context. fgs macroblock: The fo
36、ur block-bps of luminance component (Y) and the two (for 4:2:0 chrominance format) corresponding block-bps of chrominance components (U and V) with the same accuracy significance coming from the DCT coefficients of a macroblock. It may also be used to refer to the coded representation of the six blo
37、ck-bps. The usage is clear from the context. fgs macroblock number: A number for an fgsc macroblock within a vop-bp. The fgs macroblock number of the top-left fgsc macroblock in each vop-bp shall be zero. The fgsc macroblock number increments from left to right and from top to bottom. fgs run: The n
38、umber of O bits preceding a 1 bit within a block-bp. fgs temporal scalability; FGST: A type of scalability where an enhancement layer uses predictions from sample data derived from the base layer using motion vectors. The VOP size in the enhancement layer is the same as that in the base layer. FGST
39、is a specific type of temporal scalability where all DCT coefficients are coded using bit-plane coding as in FGS. fgs vop: The pixel differences between the original VOP and the reconstructed VOP in the base layer. It may be used to refer to the DCT coefficients of the pixel differences or the origi
40、nal VOP. It may also be used to refer to the coded representation of the DCT coefficients. In the context of FGST, fgsc vop refers to the original temporal scalable VOP. The usage is clear from the context. O ISOIIEC 2002 -All rights reserved 3 ISO/I EC 196-2:2001/Amd.2:2002(E) name 3.AMD2.9 fine gr
41、anularity scalability; FGS: A type of scalability where an enhancement layer uses prediction from sample data of reconstructed VOP in the base layer. The encoded bitstream for each fgsc vop can be truncated into any number of bits. The truncated bitstream for each fgsc vop can be decoded to provide
42、quality enhancement proportional to the amount of bits in the truncated bitstream of the fgsc vop. The fgsc vop has the same size and VOP rate as those of the base layer. start code value (hexadecimal) 3.AMD2.10 vop-bp: An array of block-bps with the same accuracy significance in an fgsc vop. There
43、are three color components (Y, U, and V) in a vop-bp. Each color component in a vop-bp consists of all the block-bps of that color. video-object-start-code video-object-layer-start-code reserved I O0 through IF 20 through 2F 30 throuah 3F 4) Add the following subclause to subclause 5.2: fgs-bp-start
44、-code reserved I 40 through 5F 60 throuah AF Definition of start-of-bit-plane() function visual-object-sequence-start-code visual-object-sequence-end-code user data start code The function start-of-bit-plane() returns 1 if the next bit in the bitstream is the first bit of the codes associated with a
45、 vop-bp. Otherwise it returns O. BO BI B2 I groupof-vop-start-code video-sessionerror-code 5) Add the following text to the end of subclause 6.1 : B3 B4 I visualobject-start-code VOD start code In a typical application of FGS, the bitstream at the input of an FGS decoder is a truncated version of th
46、e bitstream at the output of an FGS encoder. It is likely that, at the end of each fgsc vop before the next fgs-vop-start-code, only partial bits of the fgsc vop are at the input of the decoder due to truncation of the fgsc vop bitstream. Decoding of the truncated bitstream is not normative. An exam
47、ple of dealing with the truncated bitstream is described in Annex S. The FGS syntax description in this clause is for a complete bitstream without truncation. B5 B6 I 6) Replace Table 6-3 in subclause 6.2.1 with the following table: I 4 O ISOIIEC 2002 -All rights reserved ISO/IEC 14496-2:2001/Amd.2:
48、2002(E) slice-start-code extension start code B7 B8 fgs-vop-start-code fba-object-start-code fba-object-plane-start-code meshobject-start-code meshobject-plane-start-code still-texture-o bject-start-code texture-spatial-layer-start-code texture-snr-layer-start-code texture-tile-start-code texture-sh
49、ape-layer-start-code reserved Svstem start codes (see note) NOTE System start codes are defined in ISOIIEC 14496-1 I B9 BA BB BC BD BE BF CO CI c2 c3-c5 C6 throunh FF I videoobject-layer-start-code randomaccessi ble-vol 7) Replace VideoObjectLayer() in subclause 6.2.3: I 32 bslbf 1 bslbf I VideoObjectLayer() I No. of bits I Mnemonic I videoobject-type-indication if (is-object-layer-identifier) is-object-layer-identifier videoobject-layer-verid videoobject-layer-priority if(next-bits() = video-object-layer-start-code) short video he
