The goal of signal recreation is linearity. Testing for non-linearity is the purpose of THD measurement. It is an inherent feature of performing Fourier analysis. Our hearing having non-linear processing elements has nothing to do with analog to digital and digital to analog conversion and Shannon-Nyquist theorems. The two are totally unrelated.
You miss my point....I dont contest the value of this Shannon theorem... 😁😊
The goal in circuit design is linearity for sure... It is electronical design goal...Noise has a meaning here which is not the same than for the hearing process itself...
You dont seems conscious that all electrical measures has an interpretative meaning ONLY in a theoretical framework refering to our actual understanding about hearing...
Some use the concept "accurate" and "noise" in a confused way...They conflated the two possible meanings of the word "accurate" for a non linear detection system like the ears or for a linear tool detection system and they confused the two ways the ratio signals/noise can work for a linear detection system and for a non linear detection system...
The method by which scientist can study the way introduction of noise can help non linear detection system is called: stochastic resonance method...
The way the cochlea is non linearly structured make it able to use an Hopf bifurcation tool analysis inherent in the small fibers cells of the internal ears...
The ears are not a PASSIVE detector system but an ACTIVE non linear one able to amplify ...Then his ability to resolve information exceed many hundredth of times any passive system...
It is the reason why electrical partial set of measures ALONE cannot determine out of listening experiments what will be the sound quality of gear...Like some few deluded ASR disciples falsely claim...
The two groups are deluded not only audiophiles... Sorry....But one at least know that learning how to listen and hearing are FUNDAMENTAL....