the abovementioned graphs were probably created on the grounds of experimental measurements of various records. The science is a little bit different.
There is no doubt that the material itself affects playback and there can be some difference between a lacquer, nickel mothers (first, second etc. copy) and vinyl records due to thoughness of material. Especially, when the same stylus/pickup is used for recording of all three "grooves". There would be different contact surface when the same stylus force is used OR the stylus force should be changed for every type of material.
But diameter loss should be ZERO when cut properly for a comparison - I mean when all the geometric parameters of grooves are inside limits of mechanical technology: angles in grooves to not exceed velocity limits defined by cutting styli shapes, radii in grooves to not exceed limits defined by playback styli shapes (conical/elliptical etc). And acceleration problems generally (often too loud highest frequencies) due to inability of some parts of cutting/playback systems to accelerate so fast. The curves would be almost flat along the diameter axis when test signals are cut at lower levels.
We preffer to use different pick-ups (and styli of course) for lacquers (Shure SC39ED), nickel mothers (Shure M44MC) and vinyl records (various Shure and Ortofon). All should be calibrated for the good tracking of the given medium and ideally also with linearized frequency response (compensated by filters).
I want to say that the measured so called HF loss near the center of records/mothers/lacquers is mostly some kind of distortion. Suprisingly, you can increase feeling of HF presence when some middle range frequencies are boosted to the state of distortion when additional harmonics are created and they are inside audible range.
GZ Vinyl / GZ Media Lodenice