If a copy is created, the original file is not touched anymore.
Otherwise the original file is used and the post-processing process is not repeatable anymore, or at least the results might get unpredictable
This is necessary, because there are models, who stream on different
sites with the same name as mentioned in #141. In that case it can
happen that a recording for each site would be started within the same
minute and one recording would overwrite the other.
Split recordings didn't work, because splitRecStartTime had been removed
by accident. Also the splitting now does not start a new recording, but
switches the output file in Streamer. This is a much cleaner and
smoother approach, because it is much faster and no segments are missed