Without illustrations or a knowledge of cranked organs I should not
be sticking my oar in, but it sounds interesting.  If the goal is to
maintain a constant air pressure to the organ under varying load,
wouldn't it be just as effective to vary the speed of the pump instead
of its stroke?

Amplifying on this assumption, and further assuming a sealed system
without a 'dump' valve, it would seem that the air pressure should be
proportional to the force needed to depress the pump, and thus the
torque applied to the crank by the operator.  If the mechanism were to
be driven by a belt or chain, this torque could be measured by the
tension in the belt or chain.

Specifically, a "V-belt" drive might be rigged with the sort of
torque-sensing pulley found in some washing machines.  The two sides of
the pulley are separate and, in response to increased tension on the
belt, will separate and thus change the drive ratio.  Similar schemes
are occasionally used on bicycles, but these chain-driven 'automatic
transmissions' are somewhat noisy.

I freely admit that I may have interpreted the problem backwards.
I offer the comment in the hope that it may encourage different
strategies.

Mark Kinsler