Many of the picture prediction structure figures in 6.7.5. Metadata scalability semantics are confusing. For example, in the figure for structure S2T2 the labels I0 and S0 are unclear.
An example of how the figures may be made more clear can be found in Scalable Video Coding (SVC) Extension for WebRTC in the figures in the scalability mode dependency diagrams (https://www.w3.org/TR/webrtc-svc/#dependencydiagrams*).