I know what you mean.
At one moment in time I was involved in noise specifications for car electric motors to operate sunroofs.
We were an international team, and soon got ourselves entangled in psycho-noise and adjacent tertz-bands. I did lots of tests on motor vibration, using a setup with a phonograph needle placed on a small lump of wax on the motor housing. Presto HiFi recordings directly on the noise generator. It took sophisticated gear a factor thousand more expensive to better that.
Using Wave software I could do frequency analisys (Fast Fourrier Transform) and changes in time. I was soon able to isolate bad motors just by looking at the diagrams. This was however not very suitable for production, so in the end everybody agreed to use simple dB mesurements on contact vibration again. Often you have to make a decision fast without taking too many variables into account. Not the best way, but the way everybody can agree on.