<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> <HTML><HEAD> <META http-equiv=Content-Type content="text/html; = charset=iso-8859-1"> <META content="MSHTML 6.00.2800.1106" name=GENERATOR> <STYLE></STYLE> </HEAD> <BODY bgColor=#ffffff>It would be difficult to come up with a complete = dynamic model -- one that would predict behavior, given a set of inputs. However it would be easier, and a necessary first step to = the development of the model, to measure this finger pressure during the = entire keystroke, as Phil describes below.  Graphed over time, it = should show lost motion, hammer movement, escape, and aftertouch.  Add the = force of the hammer on the string, and you've got the complete envelope of the = mechanical inputs and outputs of the action (well, save for the dampers). Car magazines publish graphs of car engine performance; = wouldn't it be useful to have graphs like this for a range of test = blows from ppp to fff? They could be used to classify actions into categories = like "light" to "firm", as was suggested. Piano manufacturers have = robotic players to settle the action which could be programmed for a series = of test blows, as well as the various player mechanisms (although the former = might better simulate a real finger pressing a key).  Do manufacturers = take dynamic measurements as well? --Cy Shuster-- Rochester, MN >On Friday, May 2, 2003, at 12:48 PM, Phillip Ford wrote: > As I see it this will be one of the problems that = Stephen Birkett (or > any other contenders) faces in making a dynamic = model of an action. > What input are you going to use?  Does the player = push down with > constant force?  With constant velocity?  With = constant acceleration? > Does he push down with more force initially to = break the static > friction and then back off to a lower force? Does he = attempt to > rapidly accelerate the key in the first part of the movement = and then > maintain a constant velocity for the rest of the > movement? <DIV> </DIV></BODY></HTML>