1、 INTERNATIONAL TELECOMMUNICATION UNION ITU-T P.862.1TELECOMMUNICATION STANDARDIZATION SECTOR OF ITU (11/2003) SERIES P: TELEPHONE TRANSMISSION QUALITY, TELEPHONE INSTALLATIONS, LOCAL LINE NETWORKS Methods for objective and subjective assessment of quality Mapping function for transforming P.862 raw
2、result scores to MOS-LQO ITU-T Recommendation P.862.1 ITU-T P-SERIES RECOMMENDATIONS TELEPHONE TRANSMISSION QUALITY, TELEPHONE INSTALLATIONS, LOCAL LINE NETWORKS Vocabulary and effects of transmission parameters on customer opinion of transmission quality Series P.10 Subscribers lines and sets Serie
3、s P.30 P.300 Transmission standards Series P.40 Objective measuring apparatus Series P.50 P.500 Objective electro-acoustical measurements Series P.60 Measurements related to speech loudness Series P.70 Methods for objective and subjective assessment of quality Series P.80 P.800 Audiovisual quality i
4、n multimedia services Series P.900 For further details, please refer to the list of ITU-T Recommendations. ITU-T Rec. P.862.1 (11/2003) i ITU-T Recommendation P.862.1 Mapping function for transforming P.862 raw result scores to MOS-LQO Summary ITU-T Rec. P.862 provides raw scores in the range 0.5 to
5、 4.5. It is desired to provide a MOS-LQO (P.800.1) score from P.862 to allow a linear comparison with MOS. This Recommendation presents the mapping function and its performance for a single mapping from raw P.862 scores to the MOS-LQO (P.800.1). This will allow MOS-LQO scores from ITU-T Rec. P.862 t
6、o be comparable independent of the implementation of ITU-T Rec. P.862. The given function for transformation presented in this Recommendation has been optimized on a large corpus of subjective data representing different applications and languages. Source ITU-T Recommendation P.862.1 was approved by
7、 ITU-T Study Group 12 (2001-2004) under the ITU-T Recommendation A.8 procedure on 13 November 2003. ii ITU-T Rec. P.862.1 (11/2003) FOREWORD The International Telecommunication Union (ITU) is the United Nations specialized agency in the field of telecommunications. The ITU Telecommunication Standard
8、ization Sector (ITU-T) is a permanent organ of ITU. ITU-T is responsible for studying technical, operating and tariff questions and issuing Recommendations on them with a view to standardizing telecommunications on a worldwide basis. The World Telecommunication Standardization Assembly (WTSA), which
9、 meets every four years, establishes the topics for study by the ITU-T study groups which, in turn, produce Recommendations on these topics. The approval of ITU-T Recommendations is covered by the procedure laid down in WTSA Resolution 1. In some areas of information technology which fall within ITU
10、-Ts purview, the necessary standards are prepared on a collaborative basis with ISO and IEC. NOTE In this Recommendation, the expression “Administration“ is used for conciseness to indicate both a telecommunication administration and a recognized operating agency. Compliance with this Recommendation
11、 is voluntary. However, the Recommendation may contain certain mandatory provisions (to ensure e.g. interoperability or applicability) and compliance with the Recommendation is achieved when all of these mandatory provisions are met. The words “shall“ or some other obligatory language such as “must“
12、 and the negative equivalents are used to express requirements. The use of such words does not suggest that compliance with the Recommendation is required of any party. INTELLECTUAL PROPERTY RIGHTS ITU draws attention to the possibility that the practice or implementation of this Recommendation may
13、involve the use of a claimed Intellectual Property Right. ITU takes no position concerning the evidence, validity or applicability of claimed Intellectual Property Rights, whether asserted by ITU members or others outside of the Recommendation development process. As of the date of approval of this
14、Recommendation, ITU had not received notice of intellectual property, protected by patents, which may be required to implement this Recommendation. However, implementors are cautioned that this may not represent the latest information and are therefore strongly urged to consult the TSB patent databa
15、se. ITU 2004 All rights reserved. No part of this publication may be reproduced, by any means whatsoever, without the prior written permission of ITU. ITU-T Rec. P.862.1 (11/2003) iii CONTENTS Page 1 Introduction 1 2 Description of the training databases and conditions . 1 2.1 Subjective content 1 2
16、.2 Speech material and test procedure 1 3 The mapping function and its performance 2 4 Conclusions and comments 3 ITU-T Rec. P.862.1 (11/2003) 1 ITU-T Recommendation P.862.1 Mapping function for transforming P.862 raw result scores to MOS-LQO 1 Introduction ITU-T Rec. P.862 provides raw scores in th
17、e range 0.5 to 4.5. It is desired to provide a MOS-LQO (P.800.1) score from P.862 to allow a linear comparison with MOS. The aim of this separate Recommendation is to provide a single mapping from raw P.862 scores to MOS-LQO. This will allow MOS-LQO scores from ITU-T Rec. P.862 to be comparable inde
18、pendent of the implementation of ITU-T Rec. P.862. 2 Description of the training databases and conditions The presented mapping function has been trained on a large corpus of test samples that covers VoIP, wireless, fixed and clean conditions/applications. Due to the importance of the applications,
19、different weightings of the conditions/applications have been used (Table 1). Both simulated and field-collected samples and also background conditions have been considered. The field databases exhibit a high percentage of values in the lower end of the MOS scale. The reliability of the calibration
20、for the lower end of the MOS scale is therefore ensured. The simulated databases, especially the clean conditions, guarantee the reliability at the upper end of the MOS scale. Table 1/P.862.1 Database characteristics Database type Weight % (No. of files per database type/ Total No. of files %) Simul
21、ated Field VoIP 19% 64.3% 35.7% Wireless 43% 41.7% 58.3% Fixed 30% 15% 85%Clean 10% 100% N/A The normalization of the experiments has been ensured by the introduction of the certain MNRU conditions in the training process. 2.1 Subjective content The accuracy of the calibration process is strongly de
22、pendent on the MOS panel statistics. The available training databases have been obtained from a large set of experiments. Different training databases are therefore characterized by different individual MOS standard deviation and confidence intervals. A number of at least 24 votes per condition/per
23、file have been used. 2.2 Speech material and test procedure The speech material is represented by sentence-pairs spoken by both female and male genders. In addition, nine languages are considered (British English, American English, Swedish, French, Italian, German, Finnish, Dutch and Japanese). It s
24、hould be noted that in the training databases 85% of the samples represented the western languages and 15% of the samples represented the Japanese language. All the test databases are based on ACR listening quality subjective tests. 2 ITU-T Rec. P.862.1 (11/2003) 3 The mapping function and its perfo
25、rmance The mapping function is given by equation 1. 7660.4*4945.11999.0999.4999.0+=xey (1) The graph of the function is presented in Figure 1. Users may approximate this curve using other means (for example a lookup table) but this should be within the following constraints: The mapped MOS-LQO score
26、 should be within 0.01 absolute of the curve defined in this Recommendation. This constraint should be held for all points, no more than 0.01 apart, over the whole raw P.862 range 0.5 to 4.5. P.862.1_F100.511.522.533.544.55 1 012345P.862MappedP.862Figure 1/P.862.1 P.862 Algorithms mapping function T
27、his constraint will help to ensure that MOS-LQO scores from P.862 will be comparable for all implementation of ITU-T Rec. P.862. The inverse function that allows the transformation from the mapped P.862 scores to the raw P.862 scores is given by equation 2. 4945.1999.0999.4ln6607.4=yyx (2) Three sta
28、tistical measures have been used to analyse the performance of this mapping function: the Pearson correlation coefficient, R, the prediction error, Ep, the mean residual error, Em. In addition, the distribution of the mean residual error has been determined. The performances of this mapping function
29、 are shown in Tables 2 and 3. The performances have been analysed overall corpus of training samples. ITU-T Rec. P.862.1 (11/2003) 3 Table 2/P.862.1 Application Metric P.862 Mapped P.862 R 0.876 0.879 CI 95%-lower limit 0.855 0.86 Ep 0.492 0.441 CI 95%-upper limit 0.501 0.449 Overall Em 0.121 0.031
30、Table 3/P.862.1 MOS bins 0.25 0.5 0.75 1 1.25 1.5 1.75 2 P.862 scores (%) 36.1 66.63 87.44 96.95 99.56 99.96 99.96 100 Mapped P.862 scores (%) 41.92 72.64 91.22 98.4 99.64 99.88 99.96 100 4 Conclusions and comments The P.862 algorithms mapping function presented in this Recommendation has been optim
31、ized on a large corpus of subjective data representing different applications and languages. The overall performance shown in Tables 2 and 3 shows that the presented mapping function performs better than the original P.862 and it is continuous within the whole raw P.862 scale, respectively 0.5 to 4.
32、5. In addition, the mapping ensures a domain rescaling from 0.5 . 4.5 to 1.02 . 4.56. The presented function is, therefore, recommended to be applied for all types of applications in order to get the subjective estimate MOS-LQO from P.862. It should be noted that the presented function has some prac
33、tical limitations: i) The presented mapping function has been optimized on databases that originate from all types of applications. Other mapping functions optimized only for a specific application or language could perform better on that specific application or language than the presented function.
34、 ii) Although the training databases contain a large percentage of samples with scores in the low MOS region, there is a lack of samples within the raw P.862 score range 0.5 to 1. Within this range the mapped P.862 function interpolates and therefore determines a prediction error Ep and a mean resid
35、ual error Em that could be slightly higher than the one presented in Table 2. It is expected however that the error lies within the 95% CI of the Ep, respectively Em. Printed in Switzerland Geneva, 2004 SERIES OF ITU-T RECOMMENDATIONS Series A Organization of the work of ITU-T Series B Means of expr
36、ession: definitions, symbols, classification Series C General telecommunication statistics Series D General tariff principles Series E Overall network operation, telephone service, service operation and human factors Series F Non-telephone telecommunication services Series G Transmission systems and
37、 media, digital systems and networks Series H Audiovisual and multimedia systems Series I Integrated services digital network Series J Cable networks and transmission of television, sound programme and other multimedia signals Series K Protection against interference Series L Construction, installat
38、ion and protection of cables and other elements of outside plant Series M TMN and network maintenance: international transmission systems, telephone circuits, telegraphy, facsimile and leased circuits Series N Maintenance: international sound programme and television transmission circuits Series O S
39、pecifications of measuring equipment Series P Telephone transmission quality, telephone installations, local line networks Series Q Switching and signalling Series R Telegraph transmission Series S Telegraph services terminal equipment Series T Terminals for telematic services Series U Telegraph switching Series V Data communication over the telephone network Series X Data networks and open system communications Series Y Global information infrastructure, Internet protocol aspects and Next Generation Networks Series Z Languages and general software aspects for telecommunication systems
