Table of Contents

 1     Scope            
 2     References   
 3     Definitions   
 4     Abbreviations and acronyms  
 5     Conventions
 6     Subjective test and objective algorithms               
        6.1     Aspects related to subjective testing            
        6.2     Aspects related to objective algorithms       
 7     Evaluation framework              
        7.1     Data preparation             
        7.2     Analysis types  
        7.3     Prediction on a numerical quality scale       
        7.4     Uncertainty of subjective results  
        7.5     Statistical evaluation metrics        
        7.6     Statistical significance evaluation
        7.7     Statistical evaluation in the context of subjective uncertainty: epsilon insensitive rmse and its statistical significance    
        7.8     Statistical evaluation of the overall performance     
 8     Guidance on algorithm selection             
        8.1     Per experiment performance         
        8.2     Overall figure of merit     
        8.3     Worst performance cases               
        8.4      Averaging statistical metrics across experiments      
 9     Special cases               
        9.1     Evaluation of algorithms with more than one output              
        9.2     Evaluation of algorithms against pre-defined minimum performance requirements        
10     Demonstration cases
Appendix I – Algorithm mapping to the subjective scale    
Appendix II – The impact of the third order versus first order mapping    
       II.1     Application of third order and first order mappings  
       II.2     Gain of third order mapping          
Appendix III – Confidence intervals calculation    
      III.1     The standard deviation for file-based analysis         
      III.2     The standard deviation for condition-based analysis              
      III.3     Exceptional cases           
Appendix IV – Normality test    
Appendix V – Statistical significance of the rmse_tot* across all experiments    
Bibliography