ETSI TR 102 082-1999 Speech Processing Transmission and Quality Aspect (STQ) Guidance on Writing Specifications and Tests for Non-Linear and Time Variant Telephony Terminals (V1 1 .pdf

1、STD-ETSI TR LO2 082-ENGL 1999 W 3400855 0372435 892 w TR 102 082 1.1.1 (iggg-oi) Technical Report Speech Processing, Transmission and Quality Aspect (STQ); Guidance on writing specifications and tests for non-linear and time variant telephony terminals STDoETSI TR LO2 082-ENGL 1999 3400855 037241b 7

2、29 TR 102 082 V1.l.l (199941) 2 Reference DTWSTQ-O0002 (ja000ics.PDF) Keywords quality, signal processing, speech, telephony, testing, transmission ETSI Postal address F-O6921 Sophia Antipolis Cedex - FRANCE Office address 650 Route des Lucioles - Sophia Antipolis Valbonne - FRANCE Siret No 348 623

3、562 00017 - NAF 742 C Association A but non lucratif enregistre la Cous-Prfecture de Grasse (06) No 780388 Tel.: +33 4 92 94 42 O0 Fax: i33 4 93 65 47 16 Internet secretariat Q etsi.fr Individual copies of this ETSI deliverable can be downloaded from http:/hww.etsi.org If you find errors in the pres

4、ent document, send your comment to: editorQetsi.fr Copyright Notification No pari may be reproduced except as authorized by written permission. The copyright and the foregoing restriction extend to reproduction in all media. O European Telecommunications Standards institute 1999. All rights reserved

5、. ETSI STD*ETSI TR LO2 082-ENGL 1999 3400855 037Z417 hh5 0 3 TR 102 082 Vl.l.l (1999-01) Contents Intellectual Property Rights 4 Foreword 4 Introduction 4 1 scope 5 2 References 5 3 Definitions and abbreviations 6 3.1 Definitions . 6 3.2 Abbreviations 6 4 4.1 General Principles 6 4.2 4.2.1 Purpose .

6、 7 4.2.2 Assumptions . 8 4.2.3 Relevant standards and recommendations 8 4.2.4 Test arrangement 8 4.2.5 Decision variable(s) . 8 4.2.6 Mapping between objective and subjective evaluation 8 4.2.7 4.3 4.3.1 Test vector 10 4.3.2 Result presentation . 10 4.3.3 Conditioning signal 10 4.3.4 Analysis princi

7、ple . 10 4.3.5 Measuring signal 11 4.3.6 Validation . 11 Derivation of requirement and selection of test method 6 Basis for Requirements 7 Expression of the Requirement 9 Guidelines for Test Procedure Specification 9 5 5.1 5.2 5.3 5.4 5.4.1 5.4.2 5.4.3 Descriptions of terminals considered as systems

8、 . 11 Classification and Coverage of terminals . 11 Description of a SPM as a subdivided system . 13 Black-box approach or use of prior knowledge . 14 Detector Types . 14 Level Detection 14 Speech Detection 14 Response Detection 14 6 6.1 Commented template . 15 Annex A (informative): Examples of SPM

9、 descriptions . 17 A.l Frequency Multiplication . 17 A.2 Centre Clipping 18 A.3 Automatic Gain Control (AGC) . 19 History 21 Generic Signal Processing Method 15 ETSI STD=ETSI TR LO2 082-ENGL 1999 U 3400855 0372418 5TL TR 102 082 V1.l.l (i9994i) 4 intellectual Property Rights IPRs essential or potent

10、ially essential to the present document may have been declared to ETSI. The information pertaining to these essential IPRs, if any, is publicly available for ETSI members and non-members, and can be found in SR O00 3 14: “lntellectual Property Rights (IPRs); Essential, or potentially Essential, IPRs

11、 notified to ETSI in respect of ETSI standards“, which is available free of charge from the ETSI Secretariat. Latest updates are available on the ETSI Web server (http:/www.etsi.org/ipr). Pursuant to the ETSI IPR Policy, no investigation, including IPR searches, has been carried out by ETSI. No guar

12、antee can be given as to the existence of other IPRs not referenced in SR o00 314 (or the updates on the ETSI Web server) which are, or may be, or may become, essential to the present document. Foreword This Technical Report (TR) has been produced by ETSI Technical Committee Speech Processing, Trans

13、mission and Quality Aspects (STQ). Introduction From being a relatively stable discipline for many years, telephonometry is presently facing a massive need for development and renewal. A major reason for this need is the current rapid development of communications technology in general. This has bee

14、n reflected in the establishment of several new telephone transmission networks and in an effective break up from the previously dominating handsevoffice user situation. Examples of important areas with unresolved measurement problems are terminals with hands-free mode of operation, conference syste

15、ms, and the often difficult conditions associated with mobile telephony. Equally important, however, is the evolution from a national protectionist environment via a more objective European regulatory environment to tomorrows fully open competition on quality as perceived by the individual user. In

16、this environment, it is most important for the credibility of objective testing that results are much less dependent on specific test procedures than in the past. Testing should be based on well-defined test procedures. The results should be reproducible and comparable, by using the same procedures

17、so that results are independent of test system implementations. The need for documented test system performance, necessitates a significant increase in the current specification level of standards and recommendations. A methodology needs to be developed that is applicable to a much wider variety of

18、telephony applications while ensuring the development of more subjectively related though objectively more well defined tests. It is essential for the future to eliminate the perceived difficulties that have previously been associated with telephonometry in the ordinary telephone engineers mind. On

19、the basis of presently accepted objective criteria for speech transmission quality, the present document deals with the various considerations that have to be made to avoid pitfalls in the selection and description of practical instrumentation and test procedures. There is, however, a clearly recogn

20、ized need to derive a range of new or improved objective parameters describing various aspects of the subjectively perceived transmission quality. Definition of new objective parameters is outside the scope of the present document, where the intention is to form the general basis for the standards e

21、ngineer in the work of describing test methods for existing objective parameters. However the considerations are expected also to be useful when defining new or improved objective parameters for terminals containing various types of signal processing. The considerations should also be useful when sp

22、ecifying test methods for end-to-end measurements (“mouth-to-ear measurements including the telephony terminal and the acoustical environment). ETSI STD-ETSI TR 102 082-ENGL L999 3400855 0372419 438 W 5 TR 102 082 V1.l.l (199941) 1 Scope The present document provides guidance on writing objective sp

23、ecifications and tests for non- linear and time variant telephony terminals but excluding the speech coding functions. The present document provides general guidance on the formulation of test specifications including emphasizing the need for the requirement and test specification to be prepared tog

24、ether so that they are correctly matched to each other. This general guidance has wider application than the main scope of the present document. The present document provides specific guidance on the treatment of non-linearity and time variance in order to enable existing requirement and test method

25、ologies to be applied to these terminals. The method used is to treat the terminal as a system that can be subdivided into different parts so that the non-linearity or time variant element can be characterized separately. A checklist is given for the description of any specific non-linearity or time

26、 variant element so that the necessary decisions can be taken on the most appropriate way to formulate a test method including the selection of the test stimulus. 2 Refe re nces The following documents contain provisions which, through reference in this text, constitute provisions of the present doc

27、ument. 0 References are either specific (identified by date of publication, edition number, version number, etc.) or non-specific. For a specific reference, subsequent revisions do not apply. For a non-specific reference, the latest version applies. A non-specific reference to an ETS shall also be t

28、aken to refer to later versions published as an EN with the same number. il CCITT Recommendation P.10 (1988) (Blue Book): “Vocabulary of terms on telephone transmission quality and telephone sets“. ITU-T Recommendation P.501: “Test signals for use in telephonometry“. ITU-T Recommendation 4.7: “Signa

29、lling systems to be used for international automatic and semi-automatic telephone working“. 121 31 NOTE 1: Relevant to ITU-T Recommendation P.501 is ITU-T Q.71ITU-T 4.12 NOTE 2: In general a large number of the ITU-T P series of recommendations will be relevant to use together with the present docum

30、ent when writing specific test specifications. Question 20/12 provides important information on activities in the ITU-T SG 12. STD.ETS1 TR LO2 082-ENGL 1999 3400855 0372420 15T 6 TR 102 082 V1.l.l (199SOl) 3 Definitions and abbreviations 3.1 Definitions For the purposes of the present document, the

31、following terms and definitions apply. non-linearity: a property whereby the magnitude of the instantaneous output signal is intended not to have a direct linear relationship to the applied stimulus. Examples of non linearity are pi-law / A-law companding and centre clipping. time Variant: a propert

32、y whereby the output signal of the device is intended to be dependent not only on the instantaneous value of the applied stimulus, but on earlier values of stimulus. Examples are echo cancellers which take time to optimize and automatic level systems which may have long (attack +) decay times. NOTE:

33、 Devices may exhibit both non-linear and Time dependent characteristics. An example is VAD (Voice Activated Detection) used in DCME. which are inherently non-linear but the switching threshold may be time dependent. 3.2 Abbreviations For the purposes of the present document, the following abbreviati

34、ons apply: AGC DCME FFr IUT LR RMS SLR SPM TCL VAD Automatic Gain Control Digital Circuit Multiplication Equipment Fast Fourier Transform Implementation Under Test Loudness Rating Root Mean Square (value) Sending Loudness Rating Signal Processing Method Terminal Coupling Loss Voice Activity Detectio

35、n 4 Derivation of requirement and selection of test method 4.1 General Principles When defining new test standards or practices it is essential to separately consider the two following processes: - the derivation of a well-defined objective requirement which is linked as closely as possible to a bas

36、ic quality (auditorylsubjective perception); selection and description of the most appropriate method(s) to assess the relevant decision variable. - This separation may not seem necessary when testing products based on conventional technology, but it is crucial for proper evaluation of products with

37、 properties and characteristics not foreseen in the standard and for the possible later acceptance of alternative, typically more advanced, measurement methods. NOTE 1: In some cases a telephone terminal containing non-linear or time-varying circuitry may be tested against the standard requirement b

38、y modifying the test method slightly e.g. to ensure that the IUT is in a well- defined state. Such a modification can essentially be justified by technical arguments. In other cases it may be necessary to re-evaluate the requirement as such in view of the acclaimed technology. This will typically re

39、quire psycho-acoustic considerations as to whether one quality parameter can be traded in for another, e.g. lower echo in return for higher distortion. ETSI STDaETSI TR 102 O82-ENGL 1999 3400855 0372421 09b W 7 TR 102 082 V1.l.l (199941) It is necessary to consider carefully the range of conditions

40、under which the requirement has to be fulfilled to satisfy the basic purpose. In most cases practical testing has to be restricted to a scarce sampling of these conditions. Nevertheless, it is essential for the obligations of the equipment manufacturer whether such a limitation of conditions is spec

41、ified as part of the requirement or as part the test specification. NOTE 2: The underlying assumption for the sampling: that the result sampled at 1 kHz would be representative for the whole range of frequencies specified by the requirement, would clearly not be fulfilled. If a requirement for third

42、 harmonic distortion is only applicable at 1 kHz, the equipment manufacturer may ensure the necessary performance of his product by integrating a 3 kHz notch filter (possibly triggered by the 1 kHz signai only) in his product. If alternatively, the requirement is stated to cover a wider range of fre

43、quencies, and the actual frequency is specified in the test specification, such a solution would not be acceptable. The present document is meant to help “concerned parties“ (manufacturers, test laboratories, standards body etc) in enhancing standards and to ensure and/or improve the speech quality

44、of systems using complex signal processing techniques. 4.2 Basis for Requirements The items mentioned below summon all the application specific knowledge, including the psycho-acoustics aspects, necessary to derive a specific requirement, which can be objectively evaluated. A clear understanding of

45、these items is essential for both the standards engineer and the designers ability to implement improvements to the product. - Purpose; - assumptions; - relevant Standards and Recommendations: - test Arrangement; - decision Variable(s); - - expression of the Requirement. mapping between objective an

46、d subjective evaluation; An explanation and discussion of the individual items is given in subclause 4.2.1 to subclause 4.2.7. 4.2.1 Purpose Each specific requirement shall be justified in terms of a test purpose, i.e. a description of the nature of the supposed impact on the speech transmission as

47、well as the documentation for the relationship between this impact and the subjectively perceived quality of the object to be tested. EXAMPLE: The need to optimize the use of the limited dynamic range of the network and the need to have a somewhat uniform listening levei for receivers not having a v

48、olume control justifies the requirement for Sending Loudness Rating. ETSI STD-ETSI TR LO2 082-ENGL 1999 = 3400855 0372422 T22 = 8 TR 102 082 V1.l.l (1999-01) 4.2.2 Assumptions The assumptions shall contain information specific to the particular implementation of the functionality of the test object

49、or its application. Often such assumptions can lead to dramatic simplifications of the test procedures and the requirements for specific instrumentation. When a requirement and the consequent test specification are based on such assumptions, however, a means for a test engineer to check the validity of the assumptions should be annexed to the test specification. EXAMPLE 1: If a test object is required to have (and separately tested for) an amplitude linear performance, the requirements for frequency sensitivity and loudness ratings may be limited to be applic