SMPTE RP 2073-2-2015 VC-5 Video Essence - Part 2 Conformance Specification (Includes Access to Additional Content).pdf

1、 Access to Content for (SMPTE RP 2073-2:2015) (Click here to view the publication) This Page is not part of the original publication: This page has been created by IHS as a convenience to the user in order to provide access to the content as authorized by the Copyright holder of this document. Click

2、 the link(s) below to access the content and use normal procedures for downloading or opening the files. SMPTE_Additional Data Information contained in the above is the property of the Copyright holder and all Notice of Disclaimer or that a certain course of action is preferred but not necessarily r

3、equired; 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 defined at this tim

4、e, 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 (“shall“) and, i

5、f implemented, all recommended provisions (“should“) as described. A conformant implementation need not implement optional provisions (“may“) and need not implement them as described. SMPTE RP 2073-2:2015 Page 5 of 27 pages Unless otherwise specified, the order of precedence of the types of normativ

6、e information in this document shall be as follows: Normative prose shall be the authoritative definition; Tables shall be next; then formal languages; then figures; and then any other language forms. 3 Normative References The following standards contain provisions which, through reference in this

7、text, constitute provisions of this engineering document. At the time of publication, the editions indicated were valid. All standards are subject to revision, and parties to agreements based on this engineering document are encouraged to investigate the possibility of applying the most recent editi

8、on of the standards indicated below. SMPTE ST 2073-1:2013, VC-5 Video Essence Part 1: Elementary Bitstream SMPTE ST 2073-3:2015, VC-5 Video Essence Part 3: Image Formats SMPTE ST 2073-4:2015, VC-5 Video Essence Part 4: Subsampled Color Difference Components SMPTE VC-5 Test Materials, Version 2.0 or

9、later ISO/IEC 9899:2011 Programming Language C (C11) 4 Terms and Definitions 4.1 little-endian Order of bytes that comprise a number with less significant bytes before more significant bytes. 4.2 VC-5 essence standard SMPTE standards designated SMPTE ST 2073-1, ST 2073-3, and ST 2073-4. 5 Overview (

10、Informative) The test materials provided as part of this conformance specification include: 1. Source code for a reference decoder; 2. Source code for a sample encoder; 3. Utility programs and scripts useful for testing VC-5 decoder and encoder implementations; 4. Reference bitstreams for testing co

11、nformance of a decoder implementation to the VC-5 essence standard; 5. Images used to encode the reference bitstreams. The bitstream defined in the VC-5 essence standard consists of an ordered set of component arrays. In typical practice, each component array corresponds to an image plane containing

12、 color values of the same type. An image unpacking process unpacks an input image into component arrays for input to the encoding process. The image unpacking process is not defined by the VC-5 essence standard. Likewise, a decoding process outputs an ordered set of component arrays that can be pack

13、ed into an image by an image repacking process that is not defined by the VC-5 essence standard. SMPTE RP 2073-2:2015 Page 6 of 27 pages The sample encoder and reference decoder provided with the test materials include image unpacking and repacking code to enable testing with common image formats. T

14、he test materials are described in Section 6. The reference decoder for testing bitstreams created by implementations of a VC-5 encoder for compliance with the VC-5 essence standard is described in Section 6.2. The test materials include a sample encoder (Section 6.3), a program for converting betwe

15、en image formats (Section 6.4), a program for comparing image files (Section 6.5), and the images used by the sample encoder to create the reference bitstreams used for conformance testing (Section 6.6). The image file formats used by the sample encoder, reference decoder, and utility programs are d

16、escribed in Section 7. Procedures and scripts that can be used for conformance testing are described in Section 8. The criteria for conformance to the VC-5 essence standard are specified in Section 9. Instructions for installing the test materials are provided in section 10 and in a text file provid

17、ed with the test materials distribution. Instructions for installing and building the programs and contact information for submitting bug reports are provided in a separate document that is included in the test materials distribution. A VC-5 bitstream can represent common image formats as well as Co

18、lor Filter Array (CFA) images such as Bayer. The component values represented in a VC-5 bitstream can have up to 24 bits of precision. To accommodate the variety of images that can be encoded into a VC-5 bitstream, the software provided with the test materials can use images stored as unformatted bi

19、nary files (Section 7.2) or an ordered set of component arrays stored as one array per file with the channel number encoded into the pathname (Section 7.3). For convenience, the software also supports the DPX file format (Section 7.4). Reference bitstreams for testing VC-5 decoder implementations fo

20、r compliance with the VC-5 essence standard and the corresponding reference images decoded from the reference bitstreams by the reference decoder are listed in Annex A. 6 Description of Materials 6.1 Source Code The source code for the reference decoder and sample encoder is written in the C program

21、ming language. 6.2 Reference Decoder The reference decoder reads a bitstream that is compliant with the VC-5 essence standard and outputs a single image that is either an unformatted file (Section 7.2), an ordered set of component array files (Section 7.3), or a DPX file with 10-bit packed RGB color

22、 values (Section 7.4). The reference decoder includes an implementation of an image repacking process that packs the component arrays output by the decoder into an output image. To invoke the reference decoder from the command line: decoder options bitstream_file image_file Command-line options for

23、the reference decoder are listed in Table 1. The reference decoder provided with the test materials includes an implementation of an image repacking process to allow the decoder to output packed images. The packed image file formats that are supported by the reference decoder are listed in Table 6.

24、The code for the image repacking process can be modified to support other image formats. SMPTE RP 2073-2:2015 Page 7 of 27 pages Table 1 Reference decoder command-line options Option (Short or Long Format) Description -p FileFormat -pixel FileFormat File format of the image input to the sample encod

25、er to enable the decoder to output an image in the same format as input to the sample encoder (default is the file format of the output image) -o FileFormat -output FileFormat File format of the output image (default is the file format corresponding to the filename extension of the output image file

26、) -w ImageWidth -width ImageWidth Width of the input image in samples (default is the width obtained from the bitstream) -h ImageHeight -height ImageHeight Height of the input image in sample rows (default is the height obtained from the bitstream) -v -verbose Enable verbose output during decoding (

27、default is no verbose output) -h -help Print the program usage and command-line options (default is not to print the program usage text) -P -parts List of part numbers from Table 5, separated by commas without spaces, to enable the corresponding parts of the decoder (Section 6.7) The value of FileFo

28、rmat is case insensitive and can be any of the file formats listed in Table 6. The value of ImageWidth or ImageHeight can be any unsigned integer that corresponds to valid dimensions of the component arrays produced by the image unpacking process. For example, the minimum width or height of a compon

29、ent array is 48, which implies that the minimum width and height of an image in BYR4 or NV12 format is 96. 6.3 Sample Encoder The sample encoder takes as input a single image that is either an unformatted file (Section 7.2) or a DPX file with 10-bit packed RGB color values (Section 7.4) and outputs

30、a VC-5 compliant bitstream. The sample encoder includes an implementation of an image unpacking process for unpacking the input image into component arrays. SMPTE RP 2073-2:2015 Page 8 of 27 pages To invoke the sample encoder from the command line: encoder options image_file bitstream_file Command-l

31、ine options for the sample encoder are listed in Table 2. The image file can be one of the packed image file formats listed in Table 6. The encoded bitstream is placed in a binary file with the filename extension “vc5”. Table 2 Sample encoder command-line options Option (Short or Long Format) Descri

32、ption -w ImageWidth -width ImageWidth Width of the image in samples (required if the file format does not provide the width) -h ImageHeight -height ImageHeight Height of the image in sample rows (required if the file format does not provide the height) -p FileFormat -pixel FileFormat File format of

33、the image (default is the file format corresponding to the filename extension of the image file) -q Q1,Q2,Q3,Q4,Q5,Q6,Q7,Q8,Q9 Quantization divisors for the highpass subbands (default values are compiled into the encoder) -v -verbose Enable verbose output during encoding (default is no verbose outpu

34、t) -h -help Print the program usage and command-line options (default is not to print the program usage text) -P -parts List of part numbers from Table 5, separated by commas without spaces, to enable the corresponding parts of the encoder (Section 6.7) The unformatted image file formats (Section 7.

35、2) do not provide the image dimensions so the width and height must be provided as command-line options. The file format can be inferred from the filename extension (Table 5) or passed as a command-line option. The sample encoder does not support using an ordered set of component arrays as an input

36、format (Section 7.3). A file format provided explicitly as a command-line option takes precedence over a file format inferred from the filename extension. SMPTE RP 2073-2:2015 Page 9 of 27 pages 6.4 Image Converter The image converter is a utility program that is provided as a convenience. The progr

37、am can convert between several of the image formats used by the test materials. To invoke the converter from the command line: converter options input_image_file output_image_file The converter command-line options are described in Table 3. The file formats of the input and output images are inferre

38、d from the filename extensions, but can be overridden using commandline options. Table 3 Image converter command-line options Option (Short or Long Format) Description -w ImageWidth -width ImageWidth Width of the input image in samples (required if the file format does not provide the width) -h Imag

39、eHeight -height ImageHeight Height of the input image in sample rows (required if the file format does not provide the height) -p FileFormat -pixel FileFormat File format of the input image (default is the file format inferred from the input filename extension) -o FileFormat -output FileFormat File

40、format of the output image (default is the file format inferred from the output filename extension) -v -verbose Enable verbose output during conversion (default is no verbose output) -h -help Print the program usage and command-line options (default is not to print the program usage text) SMPTE RP 2

41、073-2:2015 Page 10 of 27 pages 6.5 Image Comparer The image comparer is a C language utility program that is provided as a convenience. The program computes the PSNR between two images. To invoke the comparer from the command line: comparer options image_file_1 image_file_2 The comparer command-line

42、 options are described in Table 4. Table 4 Image comparer command-line options Option (Short or Long Format) Description -w ImageWidth -width ImageWidth Width of both images in samples (required if the file format does not provide the width) -h ImageHeight -height ImageHeight Height of both images i

43、n sample rows (required if the file format does not provide the height) -p FileFormat -pixel FileFormat File format of both images (default is the filename extension) -v -verbose Enable verbose output during conversion (default is no verbose output) -h -help Print the program usage and command-line

44、options (default is not to print the program usage text) Both image files must have the same format due to limitations in the comparer program. SMPTE RP 2073-2:2015 Page 11 of 27 pages 6.6 Test Images The test materials include the following sample images organized into the directory structure speci

45、fied in Section 8. The name of the test case is the prefix of the filename preceding the image dimensions. boxes-1280x720-0000.byr4 boxes-1920x1080-0000.byr4 boxes-640x480-0000.byr4 boxes-1280x720-0000.rg48 boxes-1920x1080-0000.rg48 boxes-640x480-0000.rg48 boxes-1280x720-0000.nv12 boxes-1920x1080-00

46、00.nv12 boxes-640x480-0000.nv12 boxes-96x96-0000.nv12 gradient-640x480-0000.b64a gradient-1280x720-0000.byr4 gradient-1280x720-0000.nv12 gradient-96x96-0000.nv12 solid-48x48-0000.b64a solid-1280x720-0000.byr4 solid-96x96-0000.byr4 solid-96x97-0000.byr4 solid-97x96-0000.byr4 solid-97x97-0000.byr4 sol

47、id-48x48-0000.rg48 solid-48x49-0000.rg48 solid-49x48-0000.rg48 solid-49x49-0000.rg48 solid-640x480-0000.rg48 solid-1280x720-0000.nv12 solid-96x96-0000.nv12 The boxes test case consists of synthetic images of randomly placed, overlapping rectangles with randomly chosen colors. The solid test case con

48、sists of synthetic images that are a single randomly chosen color. The gradient test case consists of synthetic images that are a single gray wedge from black to white. The test images that have width or height equal to 48 (or equal to 96 in the case of BYR4 and NV12) were chosen so that the compone

49、nt arrays output by the image unpacking process have the minimum width or height allowed by the VC-5 essence standard. The test images that have width or height equal to 49 (or equal to 97 in the case of BYR4 and NV12) were chosen so that the component arrays output by the image unpacking process have width or height that is an odd number. Wavelet transforms normally require an even input dimension, but SMPTE ST 2073-1 specifies a mechanism for handling an odd width or height. The mechanism adds one to an odd dimension. In the case where the inp