The article recommends using a 150-Ohm resistor as a "surrogate microphone" when measuring preamp noise, and careful matching of gain levels (set with test tones and a meter, not just by ear) when making comparisons. Those are excellent suggestions.
It is indeed completely senseless to judge a preamp by turning its gain all the way up and listening with nothing connected to its input. You will hear differences among different preamps or recorders that way, but that tells you
literally nothing about how the preamps will behave with a microphone attached, at gain settings that you would actually use for recording. Really, literally nothing. Don't do it.
What the article doesn't mention, though: Equivalent input noise varies considerably at different gain settings. Perhaps unexpectedly, in most preamps it is lowest at high gain settings--often by a lot. That's why most manufacturers specify it at the maximum gain their preamps offer--generally a far higher setting than one would ever use day-to-day. -- Unfortunately, when you only know the EIN of a preamp at one gain setting, you can't infer anything about its noise at other gain settings. The quieter of two preamps as measured at 60 dB gain (typical for a spec sheet) may not be the quieter of the two at 30 or 35 dB gain (typical for actual recording).
Also not mentioned in the article: Most microphones are significantly noisier than either a 150-Ohm resistor
or the input of a good preamp.
- Complication 1: Microphone noise varies considerably at different frequencies, and so does the ear's sensitivity to that noise. Frequency weighting curves are generally applied in microphone noise measurements--but any such weighting is only appropriate for one particular SPL range. When the noise is at a markedly different level--and in good modern microphones, the noise is 25+ dB lower than the "A" curve was designed for. Thus the weighting curve can skew things enough to produce misleading results.
- Complication 2: Smooth, steady noise of a given average level is far more tolerable than noise with the same average level but with prominent, momentary "spikes." Some noise measurements, e.g. "CCIR quasi-peak", respect that difference far more than others ("RMS")--guess which figure most manufacturers publish?
Then finally, in live recording environments the same variables occur--frequency and time distribution of the noise--but especially in live, public situations, environmental noise nearly always "swamps" microphone noise.
Thus the search for the quietest preamp can easily become an exercise in misplaced perfectionism (see
https://en.wikipedia.org/wiki/Streetlight_effect).
--best regards