Yes, Pianoteq is the French piano modeling program, and I am running it as a standalone application in AVL Linux alongside ecasound and Jack. It is for playing piano live, with low latency.  (Also I recently read about  vsi plugin version of Pianoteq , and I'm wondering if it might be better to run it that way; some posts say so.) I will read these responses more carefully, but it seemed to me I want ecasound to be a Slave to the earlier parts of the pipeline so everything runs in lockstep, processing the 64 sample buffers with maximum efficiency.  Seems advantageous to me that I'm avoiding d/a conversion until final step b4 the amps.