Examples of Measured Distortion
All the flaws of voltage drive summarized are quantifiable and of patently detrimental level. Many of these effects also manifest themselves as clearly observable harmonic or modulation distortion mainly as a consequence of the corrupted voltage/current conversion occurring in the driver.
Though the results speak for themselves, it should be minded that measurable nonlinear distortions are only part of the whole picture of ill effects pertaining to voltage drive. There are also various feedback mechanisms that are more related to the temporal accuracy and comparable e.g. to cabinet edge diffractions whose harmfulness is widely recognized but which do not necessarily produce extraneous frequencies in the amplitude spectrum.
More samples and pertinent conclusions can be found in this article at Archives of Acoustics.
Example of contact resistance distortion
The acoustic measurements have been made using a dynamic cardioid microphone (Sennheiser e815 S) at 20 cm distance from the diaphragm and with appropriate frequency compensation. This kind of close-distance distortion measurements should never be made with a budget electret microphone (such as Dayton EMM-6 or Behringer ECM8000) that has a distortion-introducing FET stage in the capsule. With such mics, the real differences cannot be found out. Unfortunately, microphone manufacturers do not specify distortions for their products. To test a microphone's suitability, one can vary the measurement distance with the full modulation test signal. If the distortion decreases with increasing distance, the mic is useless for the purpose.
Also, there is no point in making the HD comparisons by frequency sweeps because the signal levels will differ between the operation modes, making the result next to worthless.
According to W. Klippel, the most published and best known expert in loudspeaker distortions, and CEO of Klippel GmbH, the modulation distortions presented above should not even exist! This becomes evident from his landmark tutorial Loudpeaker Nonlinearities—Causes, Parametres, Symptoms (JAES, vol 54, No 10, Oct 2006) (available online here) like also in a later Klippel nonlinearity poster.
The chart below shows the Klippel view of the behavior of the two-tone IM distortion caused by the L(i ) nonlinearity (p. 41) when sweeping the lower-frequency tone f1 while keeping f2 constant. According to the graph, the distortion should vanish soon after f1 exceeds 2fs, where fs is the system resonance frequency. No explanation is given for this steep decay.
The added orange line indicates how the drivers measured as above roughly behaved. There was not any major decrease in the IMD for f1 values up to the kilohertz region.
Thus, the most authoritative source on the subject ignores the predominant and most detrimental distortion effect in the mid-to-high frequency range, where most of sonic information and the ear's most sensitive region reside. Does anybody feel comfortable with that?