1、INTERNATIONAL TELECOMMUNICATION UNION)45G134 % TELECOMMUNICATIONSTANDARDIZATION SECTOROF ITU4%,%0(/.%G0G0.%47/2+G0G0!.$G0G0)3$.15!,)49G0G0/Note Transmission interruptions with durations longer than or equal to 10 seconds in a conversation phase arenot tolerable by telephone users (Annex A). Such tra
2、nsmission interruptions are considered as a premature release ofthe connection as defined in Recommendation E.850.(b) that speech loss causing transmission interruptions are caused by a change beyond given limits for a givenperiod of time in one or more parameters, e.g. power level, noise level, sig
3、nal-to-noise ratio, bit error ratio, etc.;(c) that the objective should take into account the expectations of the users for quality of voicecommunications as well as the capabilities of current technologies;(d) that the objective should take into account the concerns of network planners and system p
4、lanners, provideuseful guidance to each and that it can be used by Administrations in a consistent way to measure transmissioninterruptions;(e) that the objective should be in conformity with other Recommendations;(f) the definition of interruption as given in Recommendation E.800,recommends1 Defini
5、tions1.1 connection integrity for telephone serviceThe degree to which an established telephone connection is offered without excessive transmissioninterruptions.1.2 mean time between interruptions (MTBI)The expectation of the time between interruptions. The time between interruptions is the time du
6、ration betweenthe end of one interruption and the beginning of the next.1.3 mean interruption duration (MID)The expectation of interruption duration.1.4 transmission interruptionTemporary inability of the user-to-user transmission path to be provided persisting for less than 10 seconds(maximum durat
7、ion) and more than another given time duration (or minimum duration) characterized by a reductionbelow a certain threshold in received signal power level. The minimum duration of transmission interruption and theminimum power threshold are for further study. Transmission interruptions caused by chan
8、ges beyond certainthresholds of other parameters essential to connection integrity e.g. noise level, signal-to-distortion ratio, are for furtherstudy.2 Fascicle II.3 - Rec. E.8552 A measure to quantify telephone connection integrity performanceThe measure to be used shall be the complement of connec
9、tion integrity, namely the probability of speech loss,Pi, which is tolerable to telephone users due to transmission interruptions with durations shorter than 10 s. Theestimator of the speech loss probability, Pie, is the ratio of accumulated transmission interruption duration to the totalobservation
10、 period of time.PTDTie iiN=1/where T is the observation time and TDiis the time duration of the ith transmission interruption of N transmissioninterruptions measured during T (see Annex B).Note There are two major parameters: time between interruptions (or frequency) and duration to specifycharacter
11、istics of transmission interruptions. Those parameters should be easy to observe from the practical point ofview. Actually it seems very difficult to measure very short duration of transmission interruptions in analogue networksand to separate interruptions from burst errors in digital networks.3 Ov
12、erall objective for speech loss probabilityThe provisional objective for Pishall be such that the performance is better than the value given below:Pi= x (to be defined with further study)Note A percentage of speech loss of less than 0.5% due to transmission interruptions with durations shorterthan 1
13、0 s (Annex C) in a conversation phase is assumed to be tolerable to telephone users.4 Allocation of the overall objectiveFrom a practical point of view, instead of Pi, the value PPii1should be allocated to various networkcomponents.The method to allocate this value is for further study.Fascicle II.3
14、 - Rec. E.855 3ANNEX A(to Recommendation E.855)Tolerability of telephone user to transmission interruptionslasting several seconds or moreA.1 MeasureThe time interval between the start of transmission interruption occurring in the middle of a conversation andthe abandoning of the disturbed call eith
15、er by a calling or a called party is used as a measure to assess or evaluate thetolerability of telephone users.A.2 Measuring methodFifty intra-office calls were selected at random and deliberately interrupted by callers soon after the calls hadbeen established, and time intervals between the start
16、of transmission interruption and the release of calls by the calledparty were measured.A.3 Results of testsThe distribution of the durations of interruptions which forced the telephone users to give up their establishedcalls is depicted in Figure A-1/E.855. The distribution curve is well approximate
17、d by an exponential distributionfunction with the mean value of 17.26 sec.The figure shows that 50% of users released the established calls when the interruption lasted longer than11.96 sec.4 Fascicle II.3 - Rec. E.855 Fascicle II.3 - Rec. E.855 5ANNEX B(to Recommendation E.855)Relationship between
18、speech loss probability and its estimationThe following relationship exists between the speech loss probability (Pi) and its estimator (Pie):lim lim =TPTTDTPie kkKi =1if such limit exists and where T is the observation period of time and TDkis the duration of the k th transmissioninterruption of K t
19、ransmission interruptions over TIt should be noted that there is also the following relationship:PMIDMTBIiiiiL=+=11where, MIDi is the mean time duration of transmission interruption caused by the i th component of a telephoneconnection and MTBIi is the mean time between interruptions caused by the i
20、 th component of the connection, underthe assumption that the transmission interruption duration and the time between transmission interruptions areexponentially distributed (see also Figure B-1/E.855).ANNEX C(to Recommendation E.855)Quality of speech impaired by short interruptionsC.1 MeasureThe su
21、bjective opinion is used as a measure to assess or evaluate speech quality impaired by short interruptionswith durations shorter than 1 s.C.2 Measuring methodRecommendation P.77 was applied for this subjective evaluation with five grade opinion scores (Excellent = 4,Good = 3, Satisfactory = 2, Poor
22、= 1 and Unacceptable = 0). The test procedure was comprised of a 40-second text taperecorded in Japanese spoken by a female, which was listened to by 20 test subjects through indoor test circuits with atransmission interruption generator.6 Fascicle II.3 - Rec. E.855C.3 Test resultsThe relationship b
23、etween frequency and duration of transmission interruptions for a given Mean Opinion Score(MOS) is depicted in Figure C-1/E.855.The dotted line in this figure shows the locus of “frequency duration = 0.5%” which is considered as apermissible limit of the freeze out rate or the percentage of speech loss for designing Digital Speech Interpolation (DSI)and Time Assignment Speech Interpolation (TASI) equipment.Note The product of frequency and duration of the short interruption is identical to P = MID /(MTBI +MID) in Annex B.
