I have not read all this thread but I will throw in my twopenn'orth. I think it is all to do with second harmonic distortion. 2nd harmonic is a musical octave up so it adds brightness, presence and some might say it enhances the soundstage. The push pull topologies used in just about all op amps these days cancel 2nd harmonic distortion so they lack this effect. Instead they add predominatly third harmonic distortion which is a musical fifth. This adds power and loudness to a note but not depth. All the early single ended class A tube circuits and the well known three transistor class A circuits made back in the 70s by the likes of Neve and Helios, all exhibit primarily second harmonic distortion.
Cheers
Ian
I'm in agreement with you Ian.
I'll add that with any comparative test make sure the signal polarity of the devices are the same. Depending on the source material the differences in perceived brightness can be significant if one is inverted relative to the other. Some years ago I posted a test here using a trumpet solo that were identical only one had inverted polarity. I didn't say what the difference was but asked if they sounded the same or different. Several people posted they could clearly hear a difference. One poster - someone very seasoned and experienced IMHO - asked if they were different takes. If the two files were opened in an editor the differences would have been obvious but I asked that no one peek.
Back to Ian's comments...
I built a
distortion-maker which allowed me to control it and find out how various types of distortion sounds.
My thoughts and impressions.
1) Second harmonic adds brightness particularly if its asymmetric with positive pressure at the ear drum. Don't take my word for it. Try it.
2) "Instead they add predominatly third harmonic distortion which is a musical fifth." I think its almost an octave+fifth interval but indeed they do. Op amps, tube amps, saturated transformers etc. can add symmetric odd-order compressive distortion due to non-linearities. I don't understand Abbey's comment "Actually they do not "add" 3rd harmonics, they cancel out 2nd order." Push-pull do both: They cancel second (even) and add third (odd). Third is added due to symmetrical non-linearity.
3) Compressive odd-order distortion in gross amounts can become compressive-sounding. Listen here as the distortion changes from odd-order expansive to odd-order-compressive.
Odd-Order Compressive/Expansive Distortion Demo Radical but illustrative of changes made only by adding distortion.
4) HF IM from even order distortion folds down into the midrange where it may not be masked and can vary widely depending upon HF CM rejection in the Demrow/Cohen active input topology. See:
Component Matching Minimizes IM Distortion in the Demrow-Cohen Balanced Front End - Pro Audio Design Forum Key jangle test? I work with a manufacturer who will not let a preamp design go into production unless it has passed this test. Odd order forms sidebands which, in the midrange, have a thickening effect. (Listen to the other Waveulator demos.)
5) Feedback. My Class-A (up to about 100 mW) headphone amp's distortion signature changes significantly when its at a power level near the exit point of class A. In Class-A at higher power levels with the output stage open loop it has a simple H2 and H3 structure. Switching to overall FB mode lowers overall THD but spreads it out over a higher-order more-complex harmonic structure. If preamps sound different are their distortion signatures radically different?
6) Impedance. I agree that any comparison should present an identical load impedance to the microphone. Otherwise you're not comparing preamps.
7) Why not do a null test with an actual acoustic source? Mult the microphone into both preamps, record a performance and then attempt to null them.
8) (Added.) A phantom-powered mic at high SPLs often produces high-harmonic content common mode (CM) voltage due to signal-induced changes in CM operating current. The phantom source and pick-off resistors (usually 1% tolerance) imbalance the bridge and convert some of that CM voltage to differential at the input which then gets amplified by the preamp's differential gain. It would seem that active preamps with higher differential input impedance would be more likely to see this CM to differential conversion. In addition to conversion to CM at the input, some of the CM passed through the front-end can also get converted to differential due to finite CM rejection downstream. I've seen this effect at the input with a scope but never actually quantified it. One could argue that bench measurements of the effect might be low enough to provide masking but in the real world EQ - sometimes large amounts of it - are applied.
9) (Added.) If one of the preamps being compared has an input transformer, is its core magnetized by a prior phantom fault? That will add H2. I've had clients who de-Gauss'd their input transformers with high level LF tone every once-in-awhile.