Andi,
Before posting this I looked through previous discussions and I was hoping for a different response. AFAICT this is not a viable solution for multiple (and somewhat inter-related) reasons.
Firstly, if you do not explicitly put in a tie in the score then the player will articulate this as a separate stroke. Putting in "legato" or adding in slurs is confusing to the player and may force the player to bow differently than they would want. On a long held note it is typically up to the player to decide when to switch bow direction.
Secondly, if you do not explicitly put in slurs/phrases in a passge, it is understood by the payers that the notes should be separately bowed - i.e., detache. In Sibelius Playback, you can specify the duration of un-slurred notes - the value I typically use is 85%. This is not a perfect mechanism - you get longer gaps in some situations than desired - but it is 'good enough'. But in order for this 're-trigger' technique to work you need to set the duration of unslurred notes to 100%. You then lose the separation between notes every where else in the score.
I tried various work arounds. I set the 'unslurred notes' to 100% in Sibelius, then tried adding detache and non-legato instructions. The notes still played back legato. I could score a passage of 8th notes as dotted 16ths - that gets the desired sound but it is very confusing for the players.
I'm hoping that there is something we can do here I realize that sampling is somewhat of an art form and that compromises have to be made, but this is basic. As a composer - as opposed to someone doing film scoring - the score has to come first so the players know what to play. If we have to make some compromises in the authenticity of the playback, that is an acceptable solution.
Eric