This report presents the results of an experimental investigation
of the effects of using different source measurement techniques in
implementing the coherence function for source identification purposes.
Frequency domain measurements leading to the coherence
function are developed with regard to providing a causal relation
for a single input, single output, linear system. The use of a
near microphone or accelerometer for source measurements is
presented with particular interest paid to the effect of introducing
noise into the input measurement. Experimental results show
that while a near microphone provides a more accurate coherence
statement than an accelerometer, the input of correlated noise may
be a problem. The possibility of an empirical criteria for the
source to noise signal difference at the near microphone is raised
as a means to help avoid this problem.