Research On Integration Of Audio And Video Quality

Quality integration is a key to multimedia quality assessment. In this paper the important issue of quality integration is concerned, given the unimodal quality of audio and video. Firstly, the quality integration model recommended in G.1070 is evaluated based on experimental results. Theoretical analyses aided with empirical observations suggest that the constant coefficients used in the G.1070 model should actually be piecewise adjusted for different levels of audio and visual quality. Then a piecewise function is proposed to perform quality integration under different levels of the audio and visual quality. Performance parameters are taken into account, including the Pearson correlation coefficient (PCC) and the root mean square error (RMSE) between estimated multimedia qualities and the actual data. Performance gain observed from experimental results substantiates the effectiveness of the proposed model.Moreover, existing multimedia quality assessment methods usually obtain the overall qualities of the audio and video separately at first, and then evaluate the overall multimedia quality using the quality integration model. However, the multimedia quality is in fact different during different time periods since the video and the audio quality vary in time. Therefore, a multimedia quality assessment method is proposed which integrates the unimodal quality of audio and video on the basis of the Group of frame (GOF). Through dividing a multimedia sequence into GOFs with variable lengths, the GOF quality is first evaluated by the quality integration model. Then considering the fact that the perceptual quality of the multimedia is predominantly determined by the worst parts of GOFs, the multimedia quality is estimated by averaging the quality of a few GOFs with the worst quality.