Spatial Audio Processing: MPEG Surround and Other Applications

9.6 Conclusions

A scheme for joint object coding and mixing was motivated and derived using a stereo mixer with amplitude panning, delay panning, and gains for each source. Note that from a statistical viewpoint, the side information that represents the relative power in sub-bands between the sources (Equation 9.13), is similar to the ICLD used by a BCC (or MPEG Surround) scheme. However, the usage of these parameters is different. In the case of the joint-source coding scheme, the parameters are relative sub-band power values for each source signal without any spatial characteristic. In the case of BCC, the parameters are the inter-channel cues between audio channels that describe the spatial image of a complete auditory scene.

Also, the joint-coding synthesis stage, shown in Figure 9.5, at first sight looks similar to a BCC synthesis scheme. The differences are that the filters do not have the purpose of reproducing an ICC parameter related to a multi-channel signal, but merely the purpose of mimicking the mutual (time invariant) independence of the output source signals. Another important difference to BCC is that the joint-source coding output signals are not intended for listening in isolation but require post-mixing.