SMPTE RP 2047-1-2009 VC-2 Mezzanine Level Compression of 1080P High Definition Video Sources.pdf

1、 Copyright 2009 by THE SOCIETY OF MOTION PICTURE AND TELEVISION ENGINEERS 3 Barker Avenue., White Plains, NY 10601 (914) 761-11003 Barker Avenue., White Plains, NY 10601 Approved November 9, 2009 Table of Contents Page Foreword . 3 Intellectual Property 3 Introduction 3 1 Scope . 4 2 Conformance Not

2、ation . 4 3 Normative References . 4 4 General Description . 5 4.1 Acronyms and Terms 5 5 Coding Constraints . 5 5.1 Video Source Parameters . 5 5.2 Coding Parameters . 5 5.2.1 Sequence Header 6 5.2.1.1 Parse parameters . 6 5.2.1.2 Base video format . 6 5.2.1.3 Source parameters . 6 5.2.1.4 Picture

3、Coding Mode 6 5.2.2 Picture Syntax . 6 5.2.3 Transform Parameters 7 5.2.3.1 Wavelet filter . 7 5.2.3.2 Transform depth 7 5.2.3.3 Slice Coding Parameters 7 5.3 Entropy Coding . 7 6 Syntax Constraints . 7 Page 1 of 13 pages SMPTE RP 2047-1:2009 SMPTE RECOMMENDED PRACTICE VC-2 Mezzanine Level Compressi

4、on of 1080P High Definition Video Sources SMPTE RP 2047-1:2009 Page 2 of 13 pages Annex A Byte Values of Stream Components (Informative) . 9 A.1 Parse Info 1 9 A.2 Sequence Header 9 A.2.1 Parse Parameters . 9 A.2.2 Base Video Format . 10 A.2.3 Source Parameters . 10 A.2.4 Picture Coding Mode . 10 A.

5、2.5 Sequence Header Bytes . 10 A.3 Parse Info 2 11 A.4 Picture Number 11 A.5 Transform Parameters . 11 A.6 Transform Data 12 A.7 End of Sequence 12 Annex B Bibliography (Informative) 13 SMPTE RP 2047-1:2009 Page 3 of 13 pages Foreword SMPTE (the Society of Motion Picture and Television Engineers) is

6、 an internationally-recognized standards developing organization. Headquartered and incorporated in the United States of America, SMPTE has members in over 80 countries on six continents. SMPTEs Engineering Documents, including Standards, Recommended Practices, and Engineering Guidelines, are prepar

7、ed by SMPTEs Technology Committees. Participation in these Committees is open to all with a bona fide interest in their work. SMPTE cooperates closely with other standards-developing organizations, including ISO, IEC and ITU. SMPTE Engineering Documents are drafted in accordance with the rules given

8、 in Part XIII of its Administrative Practices. SMPTE RP 2047-1 was prepared by Technology Committee 10E. Intellectual Property At the time of publication no notice had been received by SMPTE claiming patent rights essential to the implementation of this Recommended Practice. However, attention is dr

9、awn to the possibility that some of the elements of this document may be the subject of patent rights. SMPTE shall not be held responsible for identifying any or all such patent rights. Introduction This recommended practice defines constraints on the VC-2 specification (SMPTE ST 2042-1) such that s

10、ource video images of 1920 x 1080 pixels using 4:2:2 sampling at 50, 60 and 60/1.001 frames per second will be compressed by a factor of approximately 2.5 to 1. The bit rate of the resulting VC-2 stream is such than it can be carried over a single-link serial digital interface (SDI) operating at 1.4

11、85Gbps (1.485/1.001 in the case of the 60/1.001 frame rate). This documents defines constraints on the coding and syntax, it does not define a mapping to any specific transport layer. SMPTE RP 2047-1:2009 Page 4 of 13 pages 1 Scope This recommended practice defines constraints on the VC-2 video comp

12、ression specification to provide a mezzanine level compression coding. It defines coding constraints such that source video of 1920 x 1080 pixels using 4:2:2 sampling at 50, 60 and 60/1.001 frames per second can be compressed by a factor of approximately 2.5 to 1 resulting in a compressed bit-rate t

13、hat permits carriage over a single-link SDI operating at 1.5 Gb/s. This document defines a VC-2 bit-stream for video compressed according to this practice. It also defines an application specific level for VC-2 compression. 2 Conformance Notation Normative text is text that describes elements of the

14、 design that are indispensable or contains the conformance language keywords: “shall“, “should“, or “may“. Informative text is text that is potentially helpful to the user, but not indispensable, and can be removed, changed, or added editorially without affecting interoperability. Informative text d

15、oes not contain any conformance keywords. All text in this document is, by default, normative, except: the Introduction, any section explicitly labeled as “Informative“ or individual paragraphs that start with “Note:” The keywords “shall“ and “shall not“ indicate requirements strictly to be followed

16、 in order to conform to the document and from which no deviation is permitted. The keywords, “should“ and “should not“ indicate that, among several possibilities, one is recommended as particularly suitable, without mentioning or excluding others; or that a certain course of action is preferred but

17、not necessarily required; or that (in the negative form) a certain possibility or course of action is deprecated but not prohibited. The keywords “may“ and “need not“ indicate courses of action permissible within the limits of the document. The keyword “reserved” indicates a provision that is not de

18、fined at this time, shall not be used, and may be defined in the future. The keyword “forbidden” indicates “reserved” and in addition indicates that the provision will never be defined in the future. A conformant implementation according to this document is one that includes all mandatory provisions

19、 (“shall“) and, if implemented, all recommended provisions (“should“) as described. A conformant implementation need not implement optional provisions (“may“) and need not implement them as described. Unless otherwise specified, the order of precedence of the types of normative information in this d

20、ocument shall be as follows: Normative prose shall be the authoritative definition; Tables shall be next; followed by formal languages; then figures; and then any other language forms. 3 Normative Reference The following standards contain provisions which, through reference in this text, constitute

21、provisions of this recommended practice. At the time of publication, the editions indicated were valid. All standards are subject to revision, and parties to agreements based on this recommended practice are encouraged to investigate the possibility of applying the most recent edition of the standar

22、ds indicated below. SMPTE 274M-2008, Television 1920 1080 Image Sample Structure, Digital Representation and Digital Timing Reference Sequences for Multiple Picture Rates SMPTE ST 2042-1:2009, VC-2 Video Compression SMPTE ST 2042-2:2009, VC-2 Level Definitions SMPTE RP 2047-1:2009 Page 5 of 13 pages

23、 4 General Description The VC-2 compression standard defines a compression system that uses Wavelet coding as the basis for image de-correlation. The VC-2 compression standard defines a number of compression tools including: A choice of Wavelet transform filters A choice in the number of levels of W

24、avelet decimation A choice between standard and low-delay operation A choice of entropy coding The VC-2 compression standard defines various coding parameters that are needed by a decoder to successfully decode and present the output video signal. This practice defines coding constraints to ensure t

25、he interoperability of VC-2 streams coded with mezzanine compression. It also constrains the stream syntax to ensure that coded streams may be mapped in an interoperable fashion onto a transport layer. The constraints defined here are signalled in the stream, as the specialized VC-2 level number 64

26、as indicated below in Section 5.2.1.1. That is, VC-2 streams conformant to this practice shall ensure compliance by specifying specialized level 64 in the stream. This practice does not define how the VC-2 stream is mapped to any particular transport layer. Note: Such VC-2 stream transport specifica

27、tions may be defined in other SMPTE documents such as SMPTE ST 2047-2 for the carriage of VC-2 Compressed Video over HD-SDI. 4.1 Acronyms and of Terms Lists of acronyms and terms are defined in SMPTE ST 2042-1. They are not replicated here to avoid any divergence of meaning. 5 Coding Constraints The

28、 Mezzanine compression process has the following key coding constraints: a) The codec uses only the low delay syntax. b) The video sources are limited to 1920 x 1080p with frame rates of 50 fps, 59.94 fps and 60 fps only. c) The codec uses only the Haar filter for simple implementation and to ensure

29、 that transformed samples remain bounded within each slice. 5.1 Video Source Parameters The video sources shall be as defined by SMPTE 274M for the frame rates of 50 fps, 59.94 fps and 60 fps only. The bit depth shall be limited to the 10 most significant bits. The video sampling format is 4:2:2 as

30、defined by SMPTE 274M. Only the active picture area of 1920 x 1080 pixels for the luminance channels and 960 x 1080 pixels for each of the color difference channels shall be coded. 5.2 Coding Parameters The codec shall use only the low delay syntax as defined by SMPTE ST 2042-1 (VC-2 Video Compressi

31、on). The coding syntax structure shall be as defined in SMPTE ST 2042-1. SMPTE RP 2047-1:2009 Page 6 of 13 pages 5.2.1 Sequence Header The Sequence Header comprises the parse parameters, base video format, source parameters and picture coding mode. 5.2.1.1 Parse parameters The major version and mino

32、r version numbers shall be as defined in SMPTE ST 2042-1. The profile value shall be Low Delay Profile as defined in SMPTE ST 2042-1. The level value shall be a specialized level value with the value of 64. This level value shall be used only for the VC-2 coding constraints that comply with this rec

33、ommended practice. 5.2.1.2 Base video format The base video format shall be either format number 13 or 14 (HD 1080P-60 or HD 1080P-50) as defined in SMPTE ST 2042-1. Note: The format is thus defined as follows: The picture size is 1920 x 1080 The sampling format is 4:2:2 The scanning is progressive

34、The frame rate is 60/1.001 or 50 (see below for 60-Hz frame rate) The pixel aspect ratio is 1:1 The clean picture area is the full 1920 x 1080 pixels The video depth is 10 bits The color primaries and color matrix are as defined in ITU-R BT.709 The transfer function is as defined in ITU-R BT.709 5.2

35、.1.3 Source parameters The source parameters part of the VC-2 specification allows the video parameters pre-defined by the base video format to be overridden with custom values. To comply with this document, the pre-defined video parameters shall not be overridden with the single exception that foll

36、ows: For a video source with a 60-Hz frame rate (rather than 60/1.001 Hz), the base video format shall be 13 (HD 1080P-60) and the frame rate shall be overridden by the frame rate index value 8 (60/1) as defined in SMPTE ST 2042-1, Section 11.3.5. 5.2.1.4 Picture Coding Mode Picture coding mode shal

37、l be 0, which means that VC-2 pictures are frames and thus the wavelet transform operates on each whole frame. 5.2.2 Picture Syntax The base video format and source parameters shall be static values over the VC-2 sequence. Thus, in the initialization process, all the sequence and source parameters c

38、an be accessed for each picture in the sequence. As per SMPTE ST 2042-1, the picture header value shall be an incrementing number for each picture. SMPTE RP 2047-1:2009 Page 7 of 13 pages 5.2.3 Transform Parameters The transform parameters shall be as defined below. 5.2.3.1 Wavelet filter The wavele

39、t filter shall be “Haar with single shift per level” filter (Wavelet index value = 4) as defined in SMPTE ST 2042-1. 5.2.3.2 Transform depth The transform depth shall be 2 as defined in SMPTE ST 2042-1. Note: VC-2 requires that the number of decimation stages be equal in both horizontal and vertical

40、 axes. 5.2.3.3 Slice Coding Parameters The number of slices horizontally per frame (slices x) shall be 120. This means that each slice corresponds to 16 input pixels for luminance. The number of slices vertically per frame (slices y) shall be 270. This means that each slice corresponds to 4 input pi

41、cture lines. Note: Therefore each picture contains a total of 32,400 slices. The number of bytes allocated to each slice (slice bytes) shall be 64. That is “slice bytes numerator” divided by slice bytes denominator shall be 64, where slice bytes numerator and denominator are as defined in SMPTE ST 2

42、042-1. The quantisation matrix shall be the default quantisation matrix corresponding to the Haar filter with single shift as defined in SMPTE ST 2042-1, Annex E, Table E.5. 5.3 Entropy Coding All entropy coded data shall comply with requirements of the low-delay syntax. Thus the entropy coding meth

43、ods will be as defined in SMPTE ST 2042-1, 9.1.2.1 and Annex A.3. Note: Only interleaved Exp-Golomb coding is used in the low-delay syntax. 6 Syntax Constraints The syntax of the VC-2 bit-stream shall be constrained as defined in this section to ensure interoperability and facilitate mapping to tran

44、sport layers. A sequence header shall be included in the bit-stream prior to each compressed frame. The VC-2 stream shall not contain any auxiliary or padding data units. The number of bytes allocated to each slice shall be signaled with “slice bytes numerator” = 64 and “slice bytes denominator” = 1

45、 Bits which are otherwise undefined shall be assigned the value zero. Note: Such bits are used to pad partially filled bytes to occupy a whole byte. The format of the bit-stream with the above syntax constraints will therefore be as indicated below. SMPTE RP 2047-1:2009 Page 8 of 13 pages Picture Pa

46、rse Info 2 Parse Info 1 Sequence Header End of Sequence Figure 1 Constrained Stream Syntax The structure of the sequence header and picture are defined in SMPTE ST 2042-1. For clarity the structure of the picture is as indicated below. Transform Parameters Picture Number Transform Data Figure 2 Pict

47、ure Syntax The string of bytes defining, parse info, sequence header, transform parameters and end of sequence are constant for streams that conform to this practice. The actual values of the bytes, given the constraints defined in this practice, are defined in SMPTE ST 2042-1 and are enumerated in

48、Annex A of this practice. Only the picture number and the transform data can vary between pictures in a coded bit-stream (see Section 5.2.2). The picture number increments for each successive frame and the transform data depends on the picture content of each coded frame. Each component of the bit s

49、tream shown above shall occupy a whole number of bytes. Note: Thus the components of the stream can be assembled by juxtaposing the separate components (in the correct order) without the need to shift bits within a byte. SMPTE RP 2047-1:2009 Page 9 of 13 pages Annex A Byte Values of Stream Components (Informative) This annex enumerates the specific bytes used as components of a compliant bit-stream. The byte values enumerated here can be derived unambiguously from SMPTE ST 2042-1 plus the constraints defined in prec