Preamps

Clean preamps are perceptually identical#

Every mic preamp that is not broken sounds essentially the same when it is gain-staged cleanly. If you keep it out of its non-linear range, a vintage Neve and a cheap interface's op-amp preamp are not just approximately the same. They are perceptually identical, and you will not tell them apart in an ABX test.

The exception is if there is some residual frequency response deviation coming from somewhere. After you control for that, they will then be identical. If not, I would be interested at a demonstration.

If you drive a preamp into its non-linear range, it can certainly sound very different.

Tracking clean and adding the drive later#

Many engineers track clean. That's the approach I take. Driving preamps while tracking is risky: one loud moment can overload it and ruin the best take, and distortion which gets printed cannot be undone (well, it's complicated - see my distortion reverser video).

For anyone who tracks clean, spending thousands on boutique preamps is a waste of money. A competent budget eight-channel unit usually does the job. Paying more is justified by features, not "character".

For engineers who deliberately push their preamps, a transformer-coupled class A preamp is a legitimate way to get that sound in analogue, so no, preamps are not a scam for them. But a vintage preamp processes the signal now and a plugin processes it later. So the real choice is when the overload gets applied. That is the idea behind the Mic Pre in Tractatus: separate curves for each half of the waveform and a DC bias control, so you capture clean and choose the overload afterwards. The "preamp" in a channel-strip plugin is essentially just distortion imitating a console's graceful overdrive (which is often far less graceful than people think, especially in consoles from the 80s).

Class A is a topology, not a quality grade#

Marketing often uses "class A" as if it meant the best class. It only describes how the amplifier is built.

A single transistor can only swing one way, say from 0 to 10 volts, but audio goes positive and negative. So you pretend 5 volts is zero and bias the transistor around that false centre (the Q point), burning power as heat even when it has no input signal. Something then has to get rid of that DC offset. On the Neve 1073, that is a transformer. A coupling capacitor is another alternative, and API and others used capacitors. Transformers were plentiful at the end of the valve era and Rupert Neve preferred them.

Class B splits the wave between two transistors and suffers crossover distortion. Class AB fixes that with a small bias current. Big consoles eventually moved to class AB because it's significantly more efficient. A console like the Neve VR uses op-amps with no DC bias in the preamp path, so its preamps sound pretty transparent and hardly saturate.

The character is in how they overload#

A class A stage feeding a transformer that saturates magnetically overloads softly. Core saturation is also frequency-dependent, so it sounds different from transistor clipping. A class AB stage with no transformer, or one whose transistors clip before a clean transformer saturates, just clips, like any digital or electronic clipper. When a pushed preamp only clips its transistor stage symmetrically, the result is a colder, square-wave-like clip.

So a transformer only helps if the stage before it delivers clean gain and the transformer saturates first.

Two preamps from the same maker can sound completely different#

Two Neve designs with transformers can sound nothing alike when pushed. The reasons are physical: topology, how hard the transformer is driven, and above all the DC bias, also called the quiescent point. On a 1073 the bias makes one half of the waveform run out of headroom first, and that asymmetry is where the even harmonics come from. A transformer on its own does not clip asymmetrically, so driving it does not magically produce them. The physics is in distortion.

Input impedance#

Polarity inversion and high-pass filters are trivial to do in the mix. Input impedance is different. The preamp does not change the tone, but its impedance loads a dynamic microphone differently and changes the mic's tonality. Condensers buffer this out. In my drum test, low impedance gave a sound I considered desirable on some of the mics.