Hi Eckard
OK, Forget 'z' (to do with redshift & galactic recession velocity) and just consider a galaxy 'A' moving across the sky left to right with respect to the observer at rest frame 'B', at a velocity 'v' of 0.4c. A gas jet 'C' is moving (right to left from the observers frame) towards the galaxy at velocity 'u', say 0.98c. ('M87' is galaxy Messier87 which has a gas jet recently confirmed at 5 to 6c across Hubbles frame and reputedly doing 0.98c in the local frame).
This is the alternative condition correctly referred to in the web paper as requiring simple velocity addition equation, whether each ones speed is relativised or not. No other logical option exists and it's well known, but the implications are not well understood! You're right about the normal conditions result never being greater than 'c', But relativise 0.98c and 0.4c if you wish, and when you add them together they still come to far greater than 'c'. i.e. the relative speed between A and C is superluminal viewed from B. Remember this is not necessarily when viewed from A or C according to SR, but does give valid apparent superluminal motion from the 3rd frame 'B'. Many relativists would deny this is possible, and that's where it all relied on maths not logic and went wrong.
Now to your sound waves. Consider standing at the centre of a football stadium. With 80,000 people cheering which way do the molecules bump into each other? Wouldn't they get a bit confused? If you're in a soundproof box with a long tube or directional microphone sticking out how could you pick out a sound waves from two people shouting something specific from that direction when there are 79,998 other waves they have to fight their way through messing up the wave pattern? Analog superposition can only go so far! Sound energy propagation, like light, is a lot less simplistic than maths and current science would have us believe. Or am I missing something? (I'm no expert on sound but that one may perhaps allow a relatively easy experiment!?).
Peter