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����