IEEE 1521-2003 en Measurement of Video Jitter and Wander《视频抖动和偏移测量的试用标准》.pdf

1、IEEE Std 1521-2003IEEE Standards1521TMIEEE Trial-Use Standard forMeasurement of Video Jitterand WanderPublished by The Institute of Electrical and Electronics Engineers, Inc.3 Park Avenue, New York, NY 10016-5997, USA6 February 2004IEEE Broadcast Technology SocietySponsored by theG-2.1 Audio Video T

2、echniques CommitteeIEEE StandardsPrint: SH95176PDF: SS95176IEEE Std 1521-2003(R2010)IEEE Standard for Measurementof Video Jitter and WanderSponsorG-2.1 Audio Video Techniques Committeeof theIEEE Broadcast Technology SocietyReaffirmed 25 March 2010Approved 11 September 2003IEEE-SA Standards BoardThe

3、Institute of Electrical and Electronics Engineers, Inc.3 Park Avenue, New York, NY 10016-5997, USACopyright 2004 by the Institute of Electrical and Electronics Engineers, Inc.All rights reserved. Published 6 February 2004. Printed in the United States of America.IEEE is a registered trademark in the

4、 U.S. Patent (978) 750-8400. Permission to photocopy portions of any individual standard for educationalclassroom use can also be obtained through the Copyright Clearance Center.Note: Attention is called to the possibility that implementation of this standard may require use of subject mat-ter cover

5、ed by patent rights. By publication of this standard, no position is taken with respect to the existence orvalidity of any patent rights in connection therewith. The IEEE shall not be responsible for identifying patentsfor which a license may be required by an IEEE standard or for conducting inquiri

6、es into the legal validity orscope of those patents that are brought to its attention.iv Copyright 2004 IEEE. All rights reserved.Introduction(This introduction is not part of IEEE Std 1521-2003, IEEE Trial-Use Standard for Measurement of Video Jitter andWander.)Digital video signals are now being m

7、ultiplexed with other services over nontraditional terrestrial datanetworks at high clock rates. Within these networks, the data is typically buffered to maintain an average bitrate. This can induce unwanted timing impairments in the video delivery, including a new, very low-frequency characteristic

8、 called wander. Heretofore, the measurement and specification of video wander, itsdifferentiation from the more familiar term jitter, and the associated terminology has been poorly defined.This trial-use standard was developed in close cooperation with the Society of Motion Picture andTelevision Eng

9、ineers, which is setting limits for acceptable timing performance of video signals transportedover such networks.ParticipantsAt the time this standard was completed, the G-2.1.4 Subcommittee on Video Distribution Measurementshad the following membership:James R. Redford, ChairMichel Poulin, Vice Cha

10、irAidan P. Moore, Project LeaderDaniel G. Baker, Technical EditorAt the time this standard was completed, the G-2.1 Audio Video Techniques Committee had the followingmembership:Alan Godber, ChairThe following members of the balloting committee voted on this standard. Balloters may have voted forappr

11、oval, disapproval, or abstention. Publication of this trial-use standard for comment and criticism has been approved by theInstitute of Electrical and Electronics Engineers, Inc. Trial-use standards are effective for 24months from the date of publication. Comments for revision will be accepted for 1

12、8 monthsafter publication. Suggestions for revision should be directed to the Secretary, IEEE-SAStandards Board, 445 Hoes Lane, P.O. Box 1331, Piscataway, NJ 08855-1331, and shouldbe received no later than 31 January 2006. It is expected that following the 24-month period,this trial-use standard, re

13、vised as necessary, shall be submitted to the IEEE-SA StandardsBoard for approval as a full-use standard.David FibushAlan GodberJohn J. GriggH. Douglas LungWallace MurrayJames E. ONealRichard StreeterAlexander WoernerJay BallardPaul BergerTim CarrollJoseph KaneH. Douglas LungJames R. RedfordEdmund W

14、illiamsDaniel G. BakerAlan GodberJohn J. GriggH. Douglas LungAidan P. MooreJames E. ONealMichel PoulinJames R. RedfordAlexander WoernerCopyright 2004 IEEE. All rights reserved. vWhen the IEEE-SA Standards Board approved this standard on 11 September 2003, it had the followingmembership:Don Wright, C

15、hairHoward M. Frazier, Vice ChairJudith Gorman, Secretary*Member EmeritusAlso included are the following nonvoting IEEE-SA Standards Board liaisons:Alan Cookson, NIST RepresentativeSatish K. Aggarwal, NRC RepresentativeSavoula AmanatidisIEEE Standards Managing EditorH. Stephen BergerJoseph A. Bruder

16、Bob DavisRichard DeBlasioJulian Forster*Toshio FukudaArnold M. GreenspanRaymond HapemanDonald N. HeirmanLaura HitchcockRichard H. HulettAnant Kumar JainLowell G. JohnsonJoseph L. Koepfinger*Tom McGeanSteve M. MillsDaleep C. MohlaWilliam J. MoylanPaul NikolichGary S. RobinsonMalcolm V. ThadenGeoffrey

17、 O. ThompsonDoug ToppingHoward L. Wolfmanvi Copyright 2004 IEEE. All rights reserved.Contents1. Overview 11.1 Scope 11.2 Purpose. 12. Definitions . 13. Measurement of jitter and wander . 43.1 Jitter measurement . 43.2 Wander measurement 63.3 Optional combination (automated) measurement of jitter, FO

18、, and DR from IFO data 7Annex A (informative) Examples for measuring IFO data . 9Annex B (informative) Terminology hierarchy and demarcation frequency examples 12Annex C (informative) Bibliography. 14Copyright 2004 IEEE. All rights reserved. 1IEEE Trial-Use Standard for Measurement of Video Jitteran

19、d Wander1. Overview1.1 ScopeThis trial-use standard allows the creation of instrumentation technology for consistent measurements ofvideo-related time-interval errors (TIEs).1.2 PurposeThis standard defines a set of measurements to provide metrics to quantify the timing perturbations of avideo signa

20、ls synchronization information. The goal of this measurement standard is to provide consistentand meaningful timing measurements of both digital (discrete-time) and analog (continuous-time) video thatcan be correlated to video system performance. To achieve this goal, this standard makes use of thet

21、raditional engineering concept of analyzing the timing perturbations in terms of a sinusoidal frequencyspectrum, rather than applying statistical metrics to time-domain measurements. A partitioning (filtering) ofthat spectrum is introduced and metrics are assigned for quantifying the timing perturba

22、tions so thatperformance limits, which are directly related to limitations of video timing recovery, buffering andsynchronization, can be set and verified.2. DefinitionsFor purposes of this standard, the following terms and definitions apply. The Authoritative Dictionary ofIEEE Standards Terms, Seve

23、nth Edition B11, should be referenced for terms not defined in this clause.2.1 demarcation frequency: In the context of the sinusoidal spectral content of TIE, it is the frequency thatseparates the low-frequency spectral components that are best quantified in terms of wander from the high-frequency

24、components that are best quantified as jitter. This frequency may be explicitly specified by a jittermeasurement standard for a particular format, such as bit-serial digital video. If not explicitly specified, agood choice for this frequency can be derived from the particular video formats drift-rat

25、e (DR) and jitterperformance limits in the format specifications. However, since these specifications depend on the level of1The numbers in brackets correspond to those of the bibliography in Annex C.IEEEStd 1521-2003 IEEE TRIAL-USE STANDARD FOR2 Copyright 2004 IEEE. All rights reserved.performance

26、required for the video signal (i.e., video production quality versus video distribution quality),different demarcation frequencies may coexist for a single video format. In any case, the particularfrequency used for the measurement should be called out along with the measured wander and jitter datav

27、alues. NOTESee Figure B.1 in Annex, which illustrates how the demarcation frequency separates jitter and wander measure-ment responses.2.2 frequency drift rate (DR): The time-rate-of-change of the frequency offset (FO). The preferred units ofmeasure are parts per million/second (ppm/s) or parts per

28、billion/second (ppb/s).2.3 frequency offset (FO): A low-pass filtered or averaged value of instantaneous frequency offset (IFO) toreduce the effects of TIE spectral components above the demarcation frequency, preferred to be specified inppm and sampled at a rate well above the demarcation frequency.

29、2.4 instantaneous frequency offset (IFO): The negative of the time-rate-of-change of the TIE, or equiva-lently, the difference in rate of occurrence of the significant timing instances from the expected or ideal ratenormalized to the ideal rate and preferably converted to ppm.2.5 intrinsic jitter: T

30、he output jitter of a video source that is not processing a video input signal (test signalgenerator, video storage playback, camera, etc.) or the component of the output jitter of a video signalprocessing element added by that element (processing or distribution amplifiers, routers, etc.). It is of

31、tenmeasured as the output jitter with a relatively jitter-free input.2.6 intrinsic wander: The output wander of a video source that is not processing a video input signal (testsignal generator, video storage playback, camera, etc.) or the component of the output wander of a video sig-nal processing

32、element (CODEC decoder, MPEG decoders, etc.) contributed by that element.2.7 jitter: The high-frequency spectral components of the TIE that are generally outside of the phase-tracking or synchronization bandwidth of subsequent video processing equipment. For the purposes of thisstandard, TIE spectra

33、l components above a specified demarcation frequency are considered jitter. Othervideo standards and recommended practices, such as SMPTE RP 184-2003 B3, which defines jittermeasurements specifically related to serial digital video distribution, share the concept that timing jitter ismeasured above

34、some specified frequency, and timing variations below that frequency are termed wander.NOTESMPTE RP 184-2003 B3 defines several types of jitter for serial digital video signals. In this recommendedpractice, timing jitter is defined as timing variations with spectral components above a specified freq

35、uency, typically 10Hz or less. Timing variations with spectral components below the specified frequency are termed wander and are notaddressed by SMPTE RP 184-2003 B3.2.8 jitter transfer: The video output jitter as a result of applied video input signal with jitter. 2.9 jitter transfer function: Dim

36、ensionless ratio as a function of frequency of the video output jitter to avideo input signal with jitter.2.10 output jitter: The jitter on a video output based on the combined effects of jitter transfer and intrinsicjitter.2.11 output wander: The wander on the video output based on the combined eff

37、ects of wander transfer andintrinsic wander.2.12 significant instance: A timing reference point of a video signal for purposes of synchronization ofsubsequent processing or distribution of the video signal. Ideally, the significant instances occur at uniformtime intervals and are defined in the appl

38、icable video signal format specification. For example, on an analogvideo signal this may be the 50% point of the leading edge of horizontal sync and/or the average of the zero-IEEEMEASUREMENT OF VIDEO JITTER AND WANDER Std 1521-2003Copyright 2004 IEEE. All rights reserved. 3crossings of the color re

39、ference burst on a composite signal. Other examples are the zero-crossings of aserial digital data stream or the beginning of data packets in a constant bit-rate transport stream. 2.13 TIE specification mask: The specification of performance limits for FO, frequency DR, and jitter of avideo signal d

40、efines an implied spectral limit mask that bounds the maximum TIE spectral magnitude of thevideo signal.NOTES1A log-log spectral plot showing a generalized mask and a compliant jitter spectral density are shown in Figure 1.2Explanation of TIE specification mask: In terms of TIE, the wander region is

41、 bounded by the limitations imposed bytwo independent specifications. Firstly, it is bounded by a maximum peak or IFO specification at the very low spectralregion. This is indicated by a 6 dB/octave slope and assumes the best case of a long-term, mean FO of zero. Secondly,above some corner frequency

42、, the peak frequency DR specification becomes more restrictive on TIE and is indicated bya 12 dB/octave slope. This corner frequency is the breakpoint frequency between the 6 dB/octave and 12 dB/octaveslopes. Below the breakpoint frequency, the peak frequency DR specification limit cannot be attaine

43、d without exceedingthe peak FO limit, whereas the reverse is true above the breakpoint frequency.Above the demarcation frequency, compliance with the DR limit forces the peak jitter to be below the jitter specificationand, since the peak jitter is allowed to at least equal the jitter specification l

44、imit, wander is no longer defined above thedemarcation frequency. Conversely, the peak TIE limit imposed by the jitter specification should not limit the peakvalues of TIE spectral components below the demarcation frequency, since they should only be limited by wanderspecifications that are more dir

45、ectly related to video system synchronization limitations. If all the spectral components ofa given video signals TIE fall below the specification mask, it is compliant with the jitter, FO, and DR specifications.Conversely, a TIE specification mask is implicitly specified when performance limit spec

46、ifications on jitter, FO, and DRare given.2.14 time interval error (TIE): The variation of the significant instances from their ideal positions in time.Since even the best video signals have some nonzero frequency error, the TIE relative to a precision timingreference must always be either increasin

47、g or decreasing without bound. To clarify this point, consider theTIE of the 50% point of the leading edge of a sync pulse on a particular line of a particular frame of a video-signal-under-test relative to a time-coincident significant instance on another reference video signal actingas an asynchro

48、nous reference clock. Even for a very small frequency error of the test video signal, the TIEcontinues to increase in a positive or negative direction without bound. Clearly, TIE may be difficult to mea-sure over any significant time interval, especially when the frequency error is quite large. Fort

49、unately, forthe purposes of jitter, frequency-error, and DR measurements, the rate-of-change of TIE or instantaneousfrequency error is all that is needed, and the TIE need not be measured directly.Figure 1TIE specification mask and compliant TIE spectral densityIEEEStd 1521-2003 IEEE TRIAL-USE STANDARD FOR4 Copyright 2004 IEEE. All rights reserved.2.15 wander: The low-frequency spectral components of the TIE that are generally within the phase-tracking or synchronization bandwidth of subsequent video processing equipment such that the peak timevariation of t