We are aware the ANSTI machines do not match a human. That is why we spent a lot more than an ANSTI machine on one that does match a human, including the bronchial resistance.

This may, or quite likely may not, be what the EU Committee intended to achieve, and the woolly wording of where the reference is other than its hydrostatic position, does not help one bit. They may have also come up with the calibration orifice by mistake: transparency in the process would be most welcome. It a tester were to take the reference point directly outside the mouthpiece, one can chop almost 0.8J/L from the graph I posted: that is how large these differences in measurement method can be. DL reads EN14143 as a safety document for human safety, and given the indicator on the calibration Lissajou they use, then it appears to be correct to do the proper safety test and measure relative to the moving volume with the right breathing machine. NORSOK is more explicit with this, in that the breathing simulator should simulate a human.