1、 Reference numberISO/IEC 15938-4:2002/Amd.2:2006(E)ISO/IEC 2006Information technology Multimedia content description interface Part 4: Audio AMENDMENT 2: High-level descriptors Technologies de linformation Interface de description du contenu multimdia Partie 4: Audio AMENDEMENT 2: Descripteurs de ha
2、ut niveau Amendment 2:2007 toNational Standard of CanadaCAN/CSA-ISO/IEC 15938-4:04Amendment 2:2006 to International Standard ISO/IEC 15938-4:2002 has been adopted withoutmodification (IDT) as Amendment 2:2007 to CAN/CSA-ISO/IEC 15938-4:04. This Amendment was reviewedby the CSA Technical Committee on
3、 Information Technology (TCIT) under the jurisdiction of the StrategicSteering Committee on Information Technology and deemed accep for use in Canada.August 2007 International Organization for Standardization (ISO), 2006. All rights reserved. International Electrotechnical Commission (IEC), 2006. Al
4、l rights reserved. NOT FOR RESALE. ISO/IEC 15938-4:2002/Amd.2:2006(E) PDF disclaimer This PDF file may contain embedded typefaces. In accordance with Adobes licensing policy, this file may be printed or viewed but shall not be edited unless the typefaces which are embedded are licensed to and instal
5、led on the computer performing the editing. In downloading this file, parties accept therein the responsibility of not infringing Adobes licensing policy. The ISO Central Secretariat accepts no liability in this area. Adobe is a trademark of Adobe Systems Incorporated. Details of the software produc
6、ts used to create this PDF file can be found in the General Info relative to the file; the PDF-creation parameters were optimized for printing. Every care has been taken to ensure that the file is suitable for use by ISO member bodies. In the unlikely event that a problem relating to it is found, pl
7、ease inform the Central Secretariat at the address given below. ISO/IEC 2006 All rights reserved. Unless otherwise specified, 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 w
8、riting from either ISO at the address below or ISOs member body in the country of the requester. ISO copyright office Case postale 56 CH-1211 Geneva 20 Tel. + 41 22 749 01 11 Fax + 41 22 749 09 47 E-mail copyrightiso.org Web www.iso.orgii ISO/IEC 2006 All rights reservedAmendment 2:2007 toCAN/CSA-IS
9、O/IEC 15938-4:04ISO/IEC 15938-4:2002/Amd.2:2006(E) ISO/IEC 2006 All rights reserved iiiForeword ISO (the International Organization for Standardization) and IEC (the International Electrotechnical Commission) form the specialized system for worldwide standardization. National bodies that are members
10、 of ISO or IEC participate in the development of International Standards through technical committees established by the respective organization to deal with particular fields of technical activity. ISO and IEC technical committees collaborate in fields of mutual interest. Other international organi
11、zations, governmental and non-governmental, in liaison with ISO and IEC, also take part in the work. In the field of information technology, ISO and IEC have established a joint technical committee, ISO/IEC JTC 1. International Standards are drafted in accordance with the rules given in the ISO/IEC
12、Directives, Part 2. The main task of the joint technical committee is to prepare International Standards. Draft International Standards adopted by the joint technical committee are circulated to national bodies for voting. Publication as an International Standard requires approval by at least 75 % o
13、f the national bodies casting a vote. Amendment 2 to ISO/IEC 15938-4:2002 was prepared by Joint Technical Committee ISO/IEC JTC 1, Information technology, Subcommittee SC 29, Coding of audio, picture, multimedia and hypermedia information. Amendment 2:2007 toCAN/CSA-ISO/IEC 15938-4:04ISO/IEC 15938-4
14、:2002/Amd.2:2006(E) ISO/IEC 2006 All rights reserved 1Information technology Multimedia content description interface Part 4: Audio AMENDMENT 2: High-level descriptors Remove subclauses 5.2.3. and add following subclauses: 5.2.3 SeriesOfScalarType This descriptor represents a series of scalars, at f
15、ull resolution or scaled. Use this type within descriptor definitions to represent a series of feature values. 5.2.3.1 Syntax 5.2.3.2 Semantics Name Definition SeriesOfScalarType A representation of a series of scalar values of a feature. Raw Series of unscaled samples (full resolution). Use only if
16、 scaling is absent to indicate the entire series. Amendment 2:2007 toCAN/CSA-ISO/IEC 15938-4:04ISO/IEC 15938-4:2002/Amd.2:2006(E) 2 ISO/IEC 2006 All rights reservedMin Series of minima of groups of samples. The value of numOfElements shall equal the length of the vector. This element shall be absent
17、 or empty if the Raw element is present. Max Series of maxima of groups of samples. The value of numOfElements shall equal the length of the vector. This element shall be absent or empty if the Raw element is present. Mean Series of means of groups of samples. The value of numOfElements shall equal
18、the length of the vector. This element shall be absent or empty if the Raw element is present. Random Downsampled series (one sample selected at random from each group of samples). The value of numOfElements shall equal the length of the vector. This element shall be absent or empty if the Raw eleme
19、nt is present. First Downsampled series (first sample selected from each group of samples). The value of numOfElements shall equal the length of the vector. This element shall be absent or empty if the Raw element is present. Last Downsampled series (last sample selected from each group of samples).
20、 The value of numOfElements shall equal the length of the vector. This element shall be absent or empty if the Raw element is present. Variance Series of variances of groups of samples. The value of numOfElements shall equal the length of the vector. This element shall be absent or empty if the Raw
21、element is present. Mean must be present in order for Variance to be present. Weight Optional series of weights. Contrary to other fields, these do not represent values of the descriptor itself, but rather auxiliary weights to control scaling (see below). The value of numOfElements shall equal the l
22、ength of the vector. LogBase In the case, its value is different to the default value, a logarithm has to be performed on the input data, before calculating any series (mean, variance).The value is the base of a logarithm that is performed on the input data. Note that the value of LogBase must be gr
23、eater than 0. Note: Data of a full resolution series (ratio = 1) are stored in the Raw field. Accompanying zero-sized fields (such as Mean) indicate how the series may be scaled, if the need for scaling arises. The data are then stored in the scaled field(s) and the Raw field disappears. In the case
24、, that the value of LogBase is different from its default value, a logarithm must be performed on any input data before series calculation. In case, it is equal to the default value, no logarithm must be performed on the input data. The following formula shows the rule for this calculation. Base con
25、tains the base of the logarithm and is defined in LogBase. In case the logarithmic calculation is invalid (for e.g. log 0) or the calculated output is smaller than -1.0e2the output value is fixed to -1.0e2. )(log inputValueeoutputValubase= Scalable Series allow data to be stored at reduced resolutio
26、n, according to a number of possible scaling operations. The allowable operations are those that are scalable in the following sense. Suppose the original series is scaled by a scale ratio of P, and this scaled series is then rescaled by a factor of Q. The result is the same as if the original serie
27、s had been scaled by a scale ratio of N = PQ. Amendment 2:2007 toCAN/CSA-ISO/IEC 15938-4:04ISO/IEC 15938-4:2002/Amd.2:2006(E) ISO/IEC 2006 All rights reserved 3Figure AMD2.1 illustrates the scalability property. This scaled series can be derived indifferently from the original series by applying the
28、 scaling operation with the ratios shown, or from the scaled Series of Figure AMD2.1 by applying the appropriate rescaling operation. The result is identical. Scaling operations are chosen among those for which this property can be enforced. 6 2 3 3 1 2 3 4 5 6 7 4 2 8 original seriesscaled seriesra
29、tionumOfElementsk (index)totalNumOfSamplesNum 31 Figure AMD2.1 An illustration of the scalability property If the scaling operations are used, they shall be computed as follows. Name Definition Definition if Weight present Min ikNNkikxm)1(1min+= Ignore samples with zero weight. If all have zero weig
30、ht, set to zero by convention. Max Mk= maxi=1+(k 1) NkNxiIgnore samples with zero weight. If all have zero weight, set to zero by convention. Mean +=kNNkiikxNx)1(1)/1( +=+=kNNkiikNNkiiikwxwx)1(1)1(1If all samples have zero weight, set to zero by convention. Random choose at random among N samples Ch
31、oose at random with probabilities proportional to weights. If all samples have zero weight, set to zero by convention. First choose the first of N samples Choose first non-zero-weight sample. If all samples have zero weight, set to zero by convention. Last choose the last of N samples Choose last no
32、n-zero-weight sample. If all samples have zero weight, set to zero by convention. Variance 2)1(122)1(1)/1( )()/1(kkNNkiikkNNkiikxxNxxNz=+=+=zk= wi(xi x k)2i=1+(k 1) NkNwii=1+(k1) NkNIf all samples have zero weight, set to zero by convention. Weight +=kNNkiikwNw)1(1)/1( Amendment 2:2007 toCAN/CSA-ISO
33、/IEC 15938-4:04ISO/IEC 15938-4:2002/Amd.2:2006(E) 4 ISO/IEC 2006 All rights reservedIn these formulae, k is an index in the scaled series, and i an index in the original series. N is the number of samples summarized by each scaled sample. In case logBase is not equal to the default value, X is the l
34、ogarithm of the input data, otherwise the raw input data. The formula for Variance differs from the standard formula for unbiased variance by the presence of N rather than N 1. Unbiased variance is easy to derive from it. If the Weight field is present, the terms of all sums are weighted. Replace su
35、bclause 5.2.5 with the following: 5.2.5 SeriesOfVectorType This descriptor represents a series of vectors. 5.2.5.1 Syntax 5.2.5.2 Semantics Name Definition SeriesOfVectorType A type for scaled series of vectors. Raw Series of unscaled samples (full resolution). Use only if ratio=1 for the entire ser
36、ies. Min Series of minima of groups of samples. Number of rows must equal numOfElements, number of columns must equal vectorSize. This element must be absent or empty if the element Raw is present. Max Series of maxima of groups of samples. Number of rows must equal numOfElements, number of columns
37、must equal vectorSize. This element must be absent or empty if the element Raw is present. Amendment 2:2007 toCAN/CSA-ISO/IEC 15938-4:04ISO/IEC 15938-4:2002/Amd.2:2006(E) ISO/IEC 2006 All rights reserved 5Name Definition Mean Series of means of groups of samples. Number of rows must equal numOfEleme
38、nts, number of columns must equal vectorSize. This element must be absent or empty if the element Raw is present. Random Downsampled series (one sample selected at random from each group of samples). Number of rows must equal numOfElements, number of columns must equal vectorSize. This element must
39、be absent or empty if the element Raw is present. First Downsampled series (first sample selected from each group of samples). Number of rows must equal numOfElements, number of columns must equal vectorSize. This element must be absent or empty if the element Raw is present. Last Downsampled series
40、 (last sample selected from each group of samples). Number of rows must equal numOfElements, number of columns must equal vectorSize. This element must be absent or empty if the element Raw is present. Variance Series of variance vectors of groups of vector samples. Number of rows must equal numOfEl
41、ements, number of columns must equal vectorSize. This element must be absent or empty if the element Raw is present. Mean must be present in order for Variance to be present. LogBase Base of a logarithm that is performed on the input data. If the value is equal the default value, no logarithm is per
42、formed. Note that the value of LogBase must be greater than 0. Covariance Series of covariance matrices of groups of vector samples. This is a three-dimensional matrix. Number of rows must equal numOfElements, number of columns and number of pages must both equal vectorSize. This element must be abs
43、ent or empty if the element Raw is present. Mean must be present in order for Covariance to be present. VarianceSummed Series of summed variance coefficients of groups of samples. Size of the vector must equal numOfElements. This element must be absent or empty if the element Raw is present. Mean mu
44、st be present in order for VarianceSummed to be present. MaxSqDist Maximum Squared Distance (MSD). Series of coefficients representing an upper bound of the distance between groups of samples and their mean. Size of array must equal numOfElements. This element must be absent or empty if the element
45、Raw is present. If MaxSqDist is present, Mean must also be present. Weight Optional series of weights. Weights control downsampling of other fields (see explanation for SeriesOfScalars). Size of array must equal numOfElements. vectorSize The number of elements of each vector within the series. Amend
46、ment 2:2007 toCAN/CSA-ISO/IEC 15938-4:04ISO/IEC 15938-4:2002/Amd.2:2006(E) 6 ISO/IEC 2006 All rights reservedMost of the above operations are straightforward extensions of operations previously defined in section 5.2.3.2 for series of scalars, applied uniformly to each dimension of the vectors. Oper
47、ations that are specific to vectors are defined here: Name Definition Definition if Weight present Covariance )(1)1(1 jjijkNNkijijjkxxxxN=+= +=+=kNNkiijjijkNNkijiijjkwxxxxw)1(1)1(1)( VarianceSummed =+=DjjikNNkijikxxNz12)1(1)()/1( +=+=kNNkiiDjjikNNkijiikwxxwz)1(112)1(1)(If all samples have zero weigh
48、t, set to zero by convention. MaxSqDist 2)1(1maxkikNNkikxxMSD =+=Ignore samples with zero weight. If all samples have zero weight, set to zero by convention In these formulae, k is an index in the scaled series and i an index in the original series. N is the number of vectors summarized by each scal
49、ed vector. D is the size of each vector and j is an index into each vector. In the case, that the value of LogBase is different from the default value, a logarithm must be performed on any input data before series calculation. In case, it is equal to the default value, no logarithm must be performed on the input data. The following formula shows the rule for this calculation. Base contains the base of the logarithm and is defined in LogBase. In case the logarithmic cal