@blueskull: Thanks for the translation -- this clarifies some details.
It also explains why Rigol connected all four channels to each of the two digitizing engines. As it seems, each of the digitizers is capable of 10GSa/s. If the new DSO is capable of 20GSa/s and 4GHz, it quickly becomes obvious that these are the figures for single channel operation and also why each input channels needs an analog connection to both digitizing engines.
Moreover, this means that the 4GHz figure is barely feasible for two channel operation and won't be possible in four channel mode (unless they use equivalent time sampling which seems a little obsolete in instruments of this class nowadays). If this is a problem or not for a particular application and if the instrument is a real competitor in the global market at the pricing yet to be disclosed, the customer will finally decide. At least, it's not going to be a mainstream instrument with hunderd thousands of sold units. In this instrument class, probing already becomes a challenege and a lot of speciality knowledge / experience is required to take and interpret measurements properly.
I definitely wish Rigol luck with their new design since a further competitor can only be good for the global T&M instrument market.
Cheers,
Thomas