The video data (FOV partial datasets) corresponding in time to the selected FOV data sets, including the dynamic transition data between the initial FOV and the target FOV are then streamed to the client in the correct sequence to the client. The already precalculated and stored FOV datasets are accessed here.

FIG. 5 shows how, after the selection of an initial data set 15 and the target FOV dataset 17, a dynamic FOV dataset 16 is inserted in whose segments the direction of view 4 changes successively. The streaming of the datasets 15, 16, 17 to the video client 13 is symbolized by the filmstrips 14.

The video sequences (static FOV datasets and dynamic FOV datasets) are here synchronized in such a way that seamless changes between the video sequences are possible at points with fixed definition in the temporal sequence (transition points). For this purpose, the positions of all the time points n*t_G (36 in FIG. 3) in the videos are registered (e.g. through byte offsets or storage in separate files).

The calculation for concatenation, which requires relatively little computation, of the resulting video (37 in FIG. 3) is carried out for each video client 13 after the call from the CDN/storage server 5 in the backend or by each video client 13 itself.

e) Display of the video

The video client 13 then begins (at time T₀) a session with the replay of an arbitrary, precalculated static FOV dataset 15.