1、ITU-T RECMN*P-83 93 4862593 0593623 7T9 I NTERNATIONAL TELECOMMUNICATION UN ION ITU-T TELECOMMUNICATION STANDARDIZATION SECTOR OF ITU TELEPHONE TRANSMISSION QUALITY SUBJECTIVE OPINION TESTS P.83 (03/93) SUBJECTIVE PERFORMANCE ASSESSMENT DIGITAL CODECS OF TELEPHONE-BAND AND WIDEBAND ITU-T Recommendat
2、ion P.83 (Previously CCIlT Recommendation”) FOREWORD The ITU Telecommunication Standardization Sector (ITU-T) is a permanent organ of the International Telecom- munication Union. The ITU-T is responsible for studying technical, operating and tariff questions and issuing Recommendations on them with
3、a view to standardizing telecommunications on a worldwide basis. The World Telecommunication Standardization Conference (WTSC), which meets every four years, established the topics for study by the ITU-T Study Groups which, in their turn, produce Recommendations on these topics. ITU-T Recommendation
4、 P.83 was prepared by the ITU-T Study Group XII (1988-1993) and was approved by the WTSC (Helsinki, March 1-12, 1993). NOTES 1 As a consequence of a reform process within the Intemational Telecommunication Union (ITU), the CCITT ceased to exist as of 28 February 1993. In its place, the ITU Telecommu
5、nication Standardization Sector (ITU-T) was created as of 1 March 1993. Similarly, in this reform process, the CCIR and the IFRI3 have been replaced by the Radiocommunication Sector. In order not to delay publication of this Recommendation, no change has been made in the text to references containin
6、g the acronyms “CCITT, CCIR or IFRB” or their associated entities such as Plenary Assembly, Secretariat, etc. Future editions of this Recommendation will contain the proper terminology related to the new ITU structure. 2 telecommunication administration and a recognized operating agency. In this Rec
7、ommendation, the expression “Administration” is used for conciseness to indicate both a O ITU 1994 All rights reserved. No part of this publication may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying and microfilm, without permission in writing
8、 from the ITU. . ITU-T RECMNr C O C .- m C u .- 6 C - a, Qz4Ql Tl204040-92ldl i Ll LZ L3 L4 L5 Increasing Q 4 Increasing line bit rate + FIGURE llR.83 Graphical method of deriving Figure 10 from Figures 3 and 8 r, Recommendation P.83 (03/93) 11 ITU-T RECMN+P-83 93 48b2591 059Lb35 293 - IRS response
9、(narrow-band) Annex A Comparison of different SNR definitions (This annex forms an integral part of this Recommendation) Unweighted O O C-message -1.9 -2.0 Psophometric -3.6 -2.3 Table A. 1 shows some calculations using the weightings described in Recommendation 0.41, on known spectra using two diff
10、erent frequency responses. These spectra were chosen to represent those most commonly used in subjective testing. Flat response (w ideband) - White (gaussian) noise consistent with circuit noise-psophometric and C-message designed to measure the effects of this type of noise. Unweighted C-message Ps
11、ophometnc - Hoth (room) noise and vehicle noise are used as environmental noise in subjective tests (see A. 1.1.2.2iP.80) and represent the noise(s), picked up by the microphone, measured at the telephone line terminals. IRS response (wideband) TABLE A.lP.83 Unweighted O O C-message -1.9 -2.1 Psopho
12、metric -3.6 -2.4 White noise Fiat response Unweighted (narrow-band) C-message Psophometric Hoth noise O -3.6 -2.6 O -5.2 -5.9 O -7.3 -6.3 Moving vehicle O -6.4 -4.1 O -3.7 -2.4 O -16.5 -14.0 O 4.4 -3.1 Stationary vehicle O -4.2 -2.8 O -2.3 -2.3 O -12.7 -1 1.3 O -2.5 -2.5 NOTES - - Negative sign mean
13、s quieter. - All values are differences in dB with respect to unweighted. Calculation of narrow-band unweighted used the 1/3rd octave bands, centred at the preferred frequencies as defined in IS0 R.266, from 315 Hz to 3150 Hz inclusive. Calculation of wideband unweighted used the 113rd octave bands,
14、 centred at the preferred frequencies as defined in IS0 R.266, from 100 Hz to 6300 Hz inclusive. Calculation of C-message and psophometnc weightings used 1/3rd octave bands, centred at the preferred frequencies as defined in IS0 R.266, from 100 Hz to 5000 Hz inclusive. Care must be taken when assumi
15、ng that C-message weighting is equivalent to psophometnc weighting - this is only true for gaussian type noise (see Recommendation 0.41). - - - 12 Recommendation P.83 (03/93) Annex B Sets of parameters to determine codec performance (This annex forms an integral part of this Recommendation) B.l Narr
16、ow-band systems (300-3400 Hz) a) Experiment 1 -Effect of errors, input level und listening level Speech input level: 3 (14,26 and 38 dB below overload) Listening level: Errors: 3 (O, 1:10000, 1:lOOO) Transcoding : 1 transcoding Environmental noise (sending): 1 ( 30 dBA) Experiment 2 - Efect of truns
17、codings, input level und listening level Speech input level: 3 (P+ 10, P, P- 10) b) 3 (14,26 and 38 dB below overload) 3 (P + 10, P, P - 10) Listening level: Transcoding : “X” Error: 1 (1:lOOO) Environmental noise (sending): Experiment 3 -Effect of environment noise, room noise, input level und list
18、ening level Speech input level: 1 ( 30 dBA) c) 3 (14,26 and 38 dB below overload) 3 (P + 10, P, P - 10) Listening level: Transcoding : 1 transcoding Error: 1 (1:lOOO) Environmental noise (sending): Room noise: 2 where “P” is the preferred listening level; 2 ( 30 dBA and “y”) “x” “y” “z” is the numbe
19、r of transcoding combinations to be tested; is the sending noise level to be tested; is the number of room noise conditions, typically 2. All experiments must also include the narrow-band MNRU conditions. These three experiments are not intended to be exhaustive and should be supplemented with other
20、 experiments to better characterize the codec. B.2 Wideband systems (100-7000 Hz) a) Experiment I - Effect of bit rute, BER, input level and listening level Speech input level: 2 (20 and 38 dB below overload) 3 (P + 10, P, P - 10) Listening level: Transcoding : 1 transcoding Environmental noise (sen
21、ding): 1 ( 30 dBA) Bit rates: 3 (48,56 and 64 kbit/s) BER: 3 (O, 1:lOOOO and 1:lOOO) Room noise: 1 ( 30 dBA) Recommendation P.83 (03/93) 13 ITU-T RECMN*P-83 93 m 4862591 0593637 Obb m b) Experiment 2 - Effect of transcodings, input level and listening level Speech input level: 2 (20 and 38 dB below
22、overload) 3 (P + 10, P, P- 10) Listening level: Transcoding : Environmental noise (sending): Bit rates: BER: 3 (O, 1:lOOOOand 1:lOOO) Room noise: Experiment 3 - EfSect of mismatch, input level and listening level Speech input level: “x” these include synchronous and asynchronous 1 ( 30 dBA) 3 (48,56
23、 and 64 kbit/s) 1 ( 30 dBA) c) 2 (20 and 38 dB below overload) Listening level: 3 (P + 10, P, P - 10) Transcoding : 1 transcoding Environmental noise (sending): 1 ( 30 dBA) Bit rates: 3 (56 + 64,48 + 56,48 + 64 kbit/s) BER: 2 (O and 1: 1000) Room noise: 1 ( 30 dBA) where “P. is the preferred listeni
24、ng level; “x” is the number of transcoding combinations to be tested. All three experiments must also include the wideband MNRU conditions. References COMBESCURE (P.), et al: Quality evaluation of speech coded at 32 kbit/s by means of degradation category ratings, Proc. ICASP 82 (International Confe
25、rence on Acoustics, Speech and Signal Processing), Vol. 2, Paris, May 1982. CCIT Report of the meeting of Working Party XVIIV2 (Speech processing), COM XVIII-R 28, Annex 1, pp. 13-39, December 1983. GOODMAN (D.J.), NASH (R.D.): Subjective quality of the same speech transmission conditions in seven d
26、ifferent countries, Proc. ICASP 82 (International Conference on Acoustics, Speech and Signal Processing), Vol. 2, Paris, May 1982. MODENA (G.), COLEMAN (A.), USAI (P.) and COVERDALE (P.): Subjective performance evaluation of the 7 kHz audio coder, IEEE Global Telecommunications Conference 1986 (Glob
27、ecom 86) Houston, Texas, 1-4 December 1986. EUROPEAN JEG Subjective testing methodology for the evaluation of low-bit rate codecs for mobile radio, CCITT, COM XII-68, May 1986. SWEDISH TELECOMMUNICATION ADMINISTRATION Report on subjective test on candidate codecs for mobile radio, CCITT, COM XII-147
28、, February 1987. Bibliography CCITT Recommendation G.726 40, 32, 24 and 16 kbith Adaptive Differential Pulse Code Modulation (ADPCM), Vol. III. DAUMER (W.R.) and CAVANAUGH (J.R.): A subjective comparison of selected digital codecs for speech, Bell Systems TechnicaE Journal, Vol. 57, No. 9, November
29、1978. 14 Recommendation P.83 (03/93) RICHARDS (D.L.) and BARNES (G.J.): Pay-off between quantizing distortion and injected circuit noise, Proc. ICASP 82 (International Conference on Acoustics, Speech and Signal Processing), Vol. 2, Paris, May 1982. BOYD (I.) and SOUTHCOTT (C.B.): A speech codec for
30、the Skyphone service, British Telecom TechnoZogy Journal, Vol. 6, No. 2, April 1988. COLEMAN (A.), GLEISS (N.) and USAI (P.): A Subjective Testing Methodology for Evaluating Medium Rate Codecs for Digital Mobile Applications, Speech Communications, Vol. 7, pp. 151-166, June 1988. CROWE (D.P.): Selec
31、tion of Voice Codec for the Aeronautical Satellite Service, European Conference on Speech Communication and Technology, Vol. 2, S37, pp. 320-323, September 1989. COLEMAN (A.), GLEISS (N.), SOTSCHECK (J.), USAI (P.) and SCHEUERMANN (H.): Subjective performance evaluation of the RPE-LTP codec for the Pan-European cellular digital mobile radio system, ZEEE Globecom 89, Dallas, Texas, 27-30 November 1989. WHEDDON (C.) and LINGGARD (R.): Speech and Language Processing, Chapman and Hall, 1990. Recommendation P.83 (03/93) 15
