1、Rec. ITU-R BT.802-1 1RECOMMENDATION ITU-R BT.802-1*TEST PICTURES AND SEQUENCES FOR SUBJECTIVE ASSESSMENTSOF DIGITAL CODECS CONVEYING SIGNALS PRODUCEDACCORDING TO RECOMMENDATION ITU-R BT.601(Question ITU-R 82/11)(1992-1994)Rec. ITU-R BT.802-1The ITU Radiocommunication Assembly,consideringa) that digi
2、tal signals defined according to Recommendation ITU-R BT.601 are required to be transmittedthrough point-to-point transmission links;b) that coding algorithms have been devised and standards established or proposed to enable such transmission tobe effected using bit rate reduction techniques;c) that
3、 prototype codec equipment using these algorithms is being developed and needs to be assessed;d) that Recommendation ITU-R BT.800 defines the relevant user requirements;e) that bit rate reduction techniques are also being considered for other applications, e.g. recording;f) that the evaluation of su
4、ch codecs can best be performed on a subjective basis, due to the complex andadaptive natures of the multi-dimensioned impairments introduced by bit rate reduction;g) that subjective testing methods are defined in Recommendations ITU-R BT.500 and ITU-R BT.800;h) that the selection of appropriate tes
5、t material is a key step in the planning of subjective assessment and thatwhen results are to be correlated between a number of testing sites or groups the same pictures should be used andshould be available at the highest level of quality;j) that the availability of digital recording according to R
6、ecommendation ITU-R BT.657 enables a consistentlevel of high quality to be achieved in the compilation of subjective test sequences,recommends1. that the still pictures and moving sequences specified in Recommendation ITU-R BT.800 and used in thesubjective evaluation of codecs conveying signals orig
7、inated according to Recommendation ITU-R BT.601 be selectedfrom the library tape defined in Annex 1.Note 1 A number of methods for subjective tests are under consideration, each requiring a somewhat differentarrangement of material. The arrangement defined in Annex 2 is the one currently preferred._
8、*This Recommendation should be brought to the attention of Telecommunication Standardization Study Group 9.2 Rec. ITU-R BT.802-1ANNEX 1Content of library tapeTABLE 1Scene Title Characteristic Motion Source AvailabilityNo.525 62511 Formal pond Luminance resolution Still Slide o o12 Boats Luminance an
9、d colourresolutionStill Slide o o13 Clown Horizontal resolution Still Slide o o14 Boy with toys Skin and colour edges Still Slide o o15 Girl with toys Skin and colour edges Still Slide o o16 Young couple Luminance and fine detail Still Slide o o17 Blackboard Colour, vertical resolution Still Slide o
10、 o18 Tree Luminance patterns Still Slide o o19 Old master Chroma key FG Still Video o10 Old master Chroma key BG Still Video o11 Still life Chroma key FG Still Video o o12 Still life Chroma key BG Still Video o o13 Kiel Harbour-1 High resolution Still 8 10 slide o o14 Sailboat Luminance resolution S
11、low Video o o15 Flower garden Colour details Slow pan Video o o16 Susie Skin tones Slow Video o o17 Diva with noise Rapid entropy changes Prod. wipe Video o o18 Dinner party Chroma key BG Video o o19 Boy with toys Skin and colour edges Pan (H, V) Slide o20 Old master Chroma key FG Slow pan Video o21
12、 Old master Chroma key BG Slow pan Video o22 Clown Luminance and colourhorizontal resolutionPan (H, V) Slide o23 BBC disc Random movement Circular Video o24 Kiel Harbour-2 Cycle motion (narrow filter) Rapid rocking Component o o25 Kiel Harbour-3 Cycle motion (wide filter) Rapid rocking Component o o
13、26 Kiel Harbour-4 High resolution in H, V, TdimensionsSlow pan/zoom Component o o27 Balls of wool Moving colours Medium Video o o28 Popple Moving colours Pan/rotate Video o o29 Table tennis Multiple rapid motions Pan/zoom/cut Video o o30 Mobile and calendar Random motion of objects Slow Video o o31
14、Autumn leaves Colour details Slow pan/zoom Camera o 32 Summer flowers Saturated colours, texture Slow pan Camera o Rec. ITU-R BT.802-1 3TABLE 1 (continued)Scene Title Characteristic Motion Source AvailabilityNo.525 62533 Birches Luminance details, sky Slow tilt up Camera o 34 Horse riding Landscape
15、Zoom Camera o 35 Bicycles Bicycle wheels Complex, fast Camera o 36 Ferris wheel Luminance and colour details Fast, complex Camera o 37 Shinjuku Horizontal and vertical detail Slow pan Camera o 38 Football Sports Rapid motion Camera o 39 Cheerleaders Fast, complex Zoom Camera o 40 Ciao! CK, FG, Lumin
16、ance, colourdetailsSlow pan/zoom Camera o o41 Ciao! CK, BG, Luminance, colourdetailsSlow pan/zoom Camera o o42 Portrait de famille Progressive tilisation Wipe Camera/SE o 43 Diva Cuts on titles/busy scene Cuts Camera/SE o 44 Tempte H, V, Luminance, colour details Random motion Camera o o45 Tempte wi
17、th noise H, V, Luminance, colour details Random motion Camera o o46 TV trip 3D graph, H, V luminance,colour detailsZoom/rotate Graphics o 47 Cruising Animated freeze frames 2-10 freezes Camera o o48 Decoded NTSC Cross luminance colour Slow pan/zoom Camera o 49 Decoded PAL Cross luminance colour Slow
18、 pan/zoom Camera o50 Un gnrique Rolling and crawling titles Crawl/roll Camera/CG o o51 Error recovery Frame and calibration Slow Camera/Key o o52 Text for 625 diva Cuts on titles Cuts SE o53 Basketball Sports, multiple motion,detailed backgroundZoom, H pan CCDcameraoNote 1 The content of the library
19、 tape will be updated from time to time; information about the criteria governing the selection ofitems for the library tape is given in Annex 3; the library tape is available in the D1 digital format specified in Recommen-dation ITU-R BR.657; all segments of the library tape are in the public domai
20、n and may be used freely for evaluations anddemonstrations.Note 2 Scene No. 17, “diva with noise”, is known to contain pixels with Y, CR, CBcombinations that result in out-of-range valueswhen converted to R, G, B or composite signals.Copies of the test material can be made available; requests should
21、 be addressed to the ITU Radiocommunication Bureau.4 Rec. ITU-R BT.802-1ANNEX 2A possible arrangement for use in the assembly of a subjectivetest tape using sequences selected from Annex 1TABLE 2BlockNo.Duration(s)Time code(start)Video Audio11(a) 60 00:54:15:00 Colour bars Full level Tone 440 Hz(All
22、 channels)12(a) 30 00:55:15:00 Flat field 1 Tone (all channels)400 Hz13(a) 30 00:55:45:00 Flat field 2 Tone (all channels)400 Hz14(a) 30 00:56:15:00 Ramp Tone (all channels)400 Hz15(a) 20 00:56:45:00 Zone plate Y only StillCB= CR= 128Tone (all channels)400 Hz16(a) 20 00:57:05:00 Zone plate CBonly St
23、illY = 16, CR= 128Tone (all channels)400 Hz17(a) 20 00:57:25:00 Zone plate CRonly StillY = 16, CB = 128Tone (all channels)400 Hz18(a) 20 00:57:45:00 Zone plate Y only MobileCR= CB= 128Tone (all channels)400 Hz19(a) 20 00:58:25:00 Zone plate CBonly MovingY = 16, CR = 128Tone (all channels)400 Hz10(a)
24、 20 00:58:05:00 Zone plate CRonly MovingY = 16, CB = 128Tone 400 Hz(All channels)11(a) 10 00:58:45:00 Title caption Test tape according toRecommendation ITU-R BT.800, YY/MM/DDVoice announcement12(a) 10 00:58:55:00 Credits Voice over13(a) 40 00:59:05:00 Rolling index Silence14(a) 10 00:59:45:00 Title
25、: “Formal pond” “Formal pond” (Ch. 1)14(b) 05 00:59:55:00 Grey (Y = 64, CB= CR= 128) Silence14(c) 30 01:00:00:00 Test scene 1 Formal pond Tone seq. (Ch.1)14(d) 15 01:00:30:00 Grey Silence15(a) 10 01:00:45:00 Title: “Boats” “Boats” (Ch. 1)Etc. Etc.NN 60 Colour bars End of video Tone (all channels)400
26、 HzNote 1 Further information about the compilation of subjective test tapes from material contained in the library tape is given inAnnex 4.Rec. ITU-R BT.802-1 5ANNEX 3Criteria governing the choice of itemsincluded in the library tapeIn the preparation of subjective assessments, the assessor must ha
27、ve available a variety of scenes than can beconsidered critical but not unduly so, and which are representative in quality and production values with the best oftelevision broadcasting. The level of difficulty may be checked by objective measurements, such as the determination ofthe worst-case diffe
28、rential entropy.Subjective assessments may also consider a range of possible applications and situations making necessary theinclusion of scenes suitable for the evaluation of basic quality, processibility (chroma key, slow-motion, special effects,etc.), error performance and sound/vision synchroniz
29、ation. Tests for consistency of results as between the 525- and 625-line systems would require also that near-identical scenes for both systems be included.A library of test scenes must then be based on the essential following criteria and should include: images at a number of levels of difficulty a
30、nd with different rates and modes of motion including stills; material suitable for contribution and distribution uses; scene pairs suitable for colour matte evaluation; highest quality images with the lowest possible noise levels, but noting the need for some scenes to haveknown amounts of noise ad
31、ded to evaluate certain facets of codec performance; scenes covering a range of programme types; scenes suitable for testing codec performance under normal conditions and in the presence of errors, bothconcealed and unconcealed.It should be noted that because Y, CR, CBsignals can represent a substan
32、tially greater gamut of signal valuesthan can be supported by the corresponding range of R, G, B signals, care should be taken to ensure that sequences forsubjective assessment do not contain pixels with combinations of Y, CR, CBwhich would result in out-of-range R, G, Bvalues.Table 1 describes a li
33、brary of scenes meeting these criteria, which has been assembled in both the 525-line andthe 625-line versions. The scenes were developed by a number of organizations engaged in codec development andevaluation including the Administrations of Canada, France, Germany, Italy, Japan and the United King
34、dom.ANNEX 4Compilation of subjective test tapesThe need for a large selection of scenes and for their presentation in a random order for different presentationsprecludes the preparation of a tape suitable for direct use in sessions for subjective evaluation. A better arrangement is toproduce a libra
35、ry tape of the scenes in a format that will allow the assembly of tapes for specific sessions with aminimum of inconvenience by editing between the play back of the library tape and the specific play back medium forthat session. The quality loss in this process can be very small, if the interface be
36、tween machines is in the digital formspecified in Recommendation ITU-R BT.601 and Recommendation ITU-R BT.656.The library tape consists of a leader for technical evaluation, time-code for edit control, an index in video formand a number of scene blocks to be used for evaluation.Scene identification
37、is carried in a caption title located in the video and audio ahead of the actual scene. Scenesmay thus be located visually or by reference to time-code based on addresses from the Index Page at the head of thetape.Scene identification can be carried in time-code and in the longitudinal or digital audio areas of the D1-DTTRformat specified in Recommendation ITU-R BT.657._
