Table of Contents

 1     Scope        
 2     References             
 3     Definitions
        3.1     Terms defined elsewhere      
 4     Abbreviations and acronyms           
 5     Conventions           
 6     Overview of the ITU-T P.863 algorithm        
 7     Comparison between objective and subjective scores           
 8     Speech material    
        8.1     Recommendations on source speech material
        8.2     Insertion of source speech material into the system under test
        8.3     Recommendations on processed and degraded speech material          
        8.4     Special requirements for acoustical captured speech material
        8.5     Acoustical insertion/capture for loudspeaker phones  
        8.6     Technical requirements on signals to be processed by ITU-T P.863 reference source code as given in Annex C    
        8.7     Predicted scores by the model            
 9     Description of the ITU-T P.863 algorithm     
        9.1     Overview     
        9.2     Temporal alignment 
        9.3     Joining sections with constant delay   
        9.4     Sample rate ratio detection   
        9.5     Resampling  
        9.6     Level, frequency response and time alignment pre-processing
        9.7     Perceptual model     
10     Conclusions           
Annex A – Conformance data and conformance tests    
        A.1     List of files provided for conformance validation         
        A.2     Conformance tests   
        A.3     Conversion of sampling rates
        A.4     Digital attachments  
Annex B – Specification source code of the ITU-T P.863 algorithm in C++ notation    
Appendix I – Reporting of the performance results for the ITU-T P.863 algorithm based on the rmse* metric    
        I.1               Purpose of this appendix            
        I.2               Overview          
        I.3               Performance results for the ITU-T P.863 algorithm          
        I.4               Calculation of rmse*     
        I.5               Scatter plots     
Appendix II – Description of the "full-scale" subjective tests in a superwideband context conducted for the ITU-T P.863 algorithm training and validation    
       II.1     Database structure and subjects requirement 
       II.2     Anchor conditions      
       II.3     Design rules of test conditions for full-scale mandatory tests    
       II.4     Reference and degraded speech material       
       II.5     Transmission and capturing capture of speech material superimposed interlaced with background noises    
       II.6     Transmission and capturing capture of speech material under time warping conditions 
       II.7     Subjective test set up for assessing superwideband speech quality         
       II.8     Limitations in subjective test results    
Bibliography