1、INTERNATIONAL TELECOMMUNICATION UNION)45G134 / TELECOMMUNICATIONSTANDARDIZATION SECTOROF ITU30%#) amended at Geneva, 1980, and at Melbourne, 1988)1 PreambleIt is important that the characteristics of quantizing distortion measuring apparatus are specified with sufficientprecision to ensure that all
2、future designs of measuring apparatus conforming to the recommended specification shall becompatible with one another, i.e., they shall be capable of interworking and give results of specified accuracy withoutthe need for any special procedures or corrections to the measurements results. It is consi
3、dered equally important that alldesigns of measuring apparatus conforming to the recommended specification shall be capable of interworking withexisting designs of measuring apparatus already in use by various Administrations, who will thus not be placed at anyeconomic disadvantage. The following sp
4、ecification is based on the proposals studied by Study Group XVIII and isspecifically aimed at the foregoing compatibility objectives.Note The question of interworking between existing designs of quantizing distortion measuring apparatus isnot, in itself, directly relevant to this specification, but
5、 it is worth recording that this topic has been studied by theFederal Republic of Germany and the United Kingdom Post Office. Satisfactory rules have been established to facilitateinterworking between the different existing types of measuring apparatus which use a band-limited pseudo-random noisesou
6、rce.2 Testing method proposedThe method proposed is that described in Method 1 in 9 of Recommendation G.712 1. The proposed noisesource is band-limited pseudo-random noise having a probability density distribution of amplitudes which issubstantially near to a Gaussian distribution1).The signal-to-to
7、tal distortion power, including quantizing distortion, is measured as the ratio of the power ofreceived stimulus in the reference band, to the noise power in the measured band. A correction is included to relate themeasurement to the full PCM speech channel bandwidth._1)The receive measuring apparat
8、us specified in 3.2 may also be used to measure quantizing distortion using a sinusoidal test signalin the frequency range 350-550 Hz (preferably at 420 20 Hz) instead of the pseudo-random noise stimulus. It should be noted,however, that while the measurement is similar to Method 2 described in Reco
9、mmendation G.712 1, the obtained measurementresults are related to a bandwidth of 3.1 kHz and that no noise weighting is provided. It should also be noted that results given bythe pseudo-random noise and sinusoidal methods may not be the same.2 Fascicle IV.4 - Rec. O.131The principle of the measurem
10、ent is illustrated in Figure 1/O.131.d01-scFIGURE 1/O.131.D013 Basic specification clauses3.1 SendThe sending signal is a band-limited pseudo-random noise having the following characteristics:3.1.1 Band limited noise stimulusApproximately Gaussian distribution of the amplitudes within the bandwidth
11、of the send filter. The bandwidthcan have any value from 100 Hz to 200 Hz between the 3-dB points (see 3.1.4 and 3.1.5 below).3.1.2 Number of spectral linesNot less than 25 spectral lines with a spacing not greater than 8 Hz measured at the output of the send filter.3.1.3 Peak-to-r.m.s. ratio10.5 dB
12、. Tolerance 0.5 dB.Note 1 The requirements according to 3.1.1 to 3.1.3 above may be accomplished by a noise stimulusderived from the output of a 17-stage shift register with exclusive OR gating with the outputs of stages 3 and 17 returnedto the input of stage 1. The shift register produces a maximal
13、ly long sequence of (217 1) bits.The shift register is driven at a clock frequency fcHz such that the spectral line spacing fsHz of the outputsignal is less than or equal to 8 Hz.In order to meet the specified limits of the peak-to-r.m.s. ratio of the sent signal as given in 3.1.3 above, theclock fr
14、equency can be adjusted to:fc= fs(217 1) HzTo keep the peak factor within the specified limits, a stability of the clock frequency fcon the order of 1% isrequired.Fascicle IV.4 - Rec. O.131 3Note 2 Instead of using a shift register to generate the noise signal, other principles may be adopted as lon
15、gas the generated signal has the characteristics recommended in 3.1.1 and 3.1.3 above.3.1.4 Frequency position of sent signalBetween 350 and 550 Hz.3.1.5 Sending filter characteristicsThe attenuation of the bandpass filter with reference to minimum attenuations should be as follows:not lower than 35
16、0 Hz 3 dB point at lower frequencynot exceeding 550 Hz 3 dB point at upper frequencybelow 250 Hz greater than 55 dBat 300 Hz greater than 20 dBat 580 Hz greater than 6 dBat 650 Hz greater than 20 dBat 700 Hz greater than 40 dBat 750 Hz greater than 50 dBat and above 800 Hz greater than 60 dBThe resp
17、onse characteristic of a filter designed to these limits should give a bandwidth between 3-dB points ofat least 100 Hz.The performance requirements for the sending filter characteristics conforming to the above limits is given inFigure 2/O.131.3.1.6 Sending reference level range0 dBm0 to at least 55
18、 dBm0 for relative levels according to Recommendation G.232, 11 2 with a settingaccuracy of 0.5 dB.3.1.7 Output impedance (frequency range 300 Hz to 4 kHz) Balanced, earth free (other impedances optional) 600 ohms Return loss . 30 dB Output signal balance. 40 dB3.2 Receive3.2.1 Receive reference fil
19、terNominal bandwidth of reference path 350-550 Hz. (See Note below).The characteristic of the filter is chosen to prevent inaccuracy in the measurement of the received noisestimulus in the presence of quantizing distortion and other system noise conditions. The filter should not diminish thepower of
20、 a noise band between 350 Hz and 550 Hz by more than 0.25 dB.4 Fascicle IV.4 - Rec. O.131d02-scFIGURE 2/O.131.D02Note The receive reference filter ideally restricts the bandwidth of the reference path to respond only to thespectrum of the received noise stimulus. However, the bandwidth of 350-550 Hz
21、 is chosen to allow for the need tointerwork with test apparatus having a noise source bandwidth of up to 200 Hz.3.2.2 Bandwidth of measuring pathAt least 2.4 kHz (with a loss variation of less than 2 dB). The required bandpass characteristic of the filters formeasurement of distortion products is i
22、ndicated below and is such that received noise stimulus does not affectmeasurements. Attenuation with reference to the minimum attenuation:150 Hz and below greater than 60 dB650 Hz greater than 55 dB700 Hz greater than 35 dB750 Hz greater than 20 dB800 Hz 3 dB or greater3.4 kHz 3 dB or greater3.5 kH
23、z greater than 10 dB3.6 kHz greater than 20 dB3.7 kHz greater than 40 dB3.75 kHz greater than 50 dB5.0 kHz and above greater than 60 dBFascicle IV.4 - Rec. O.131 5The performance requirements for the measurement filter characteristic conforming to the above limits is givenin Figure 3/O.131.d03-scFIG
24、URE 3/O.131.D033.2.3 Bandwidth correctionThe calibration of the test apparatus shall include a correction factor of appropriate value to relate the signal tototal distortion power measured to the total distortion power present in the full PCM channel bandwidth of 3100 Hz. Thecorrection factor is giv
25、en by the following expression, which assumes a uniform distribution of distortion power over thechannel bandwidth:10 log103100y(dB)where y is the effective noise bandwidth of the measuring filter in Hz.3.2.4 Input impedance Balanced, earth free (other impedances optional) 600 ohms Return loss . 30
26、dB Input longitudinal interference loss (below 4 kHz) . 46 dB Input longitudinal interference loss (at 40 Hz) 60 dB3.2.5 Input reference level range0 dBm0 to at least 55 dBm0 for relative levels according to Recommendation G.232 2.6 Fascicle IV.4 - Rec. O.1313.2.6 Accuracy of the signal-to-total dis
27、tortion ratio indicationFor reference levels in the range 6 dBm0 to 55 dBm0 and an absolute distortion signal not less than 72 dBm0: Measuring range 10 dB to 40 dB: Accuracy 0.5 dB. Measuring range 00 dB to 10 dB: Accuracy 1.0 dB.For reference levels in the range 0 dBm0 to 6 dBm0: Measuring range 20
28、 dB to 40 dB: Accuracy 1.5 dB. Measuring range 00 dB to 20 dB: Accuracy 2.0 dB.Note 1 These limits include the inaccuracies which are caused by: the effective bandwidth of the measuring filter, the receive reference filter, the attenuator in the measuring path, the characteristics of the indicating
29、circuit.Note 2 For reference level ranges 0 dBm0 to 6 dBm0, the wider tolerances are not only required by themeasuring apparatus but reflect also the characteristics of PCM coders and decoders when operated near the overloadpoint.4 Operating environmentThe electrical performance requirements shall b
30、e met when operating at the climatic conditions as specified inRecommendation O.3, 2.1.References1 CCITT Recommendation Performance characteristics of PCM channels between 4-wire interfaces at voicefrequencies, Vol. III, Rec. G.712.2 CCITT Recommendation 12-channel terminal equipments, Vol. III, Rec. G.232.
