Thanks!
Interesting software! I tried it quickly. How do you know what results depend on the soundcard and what on the mic-pre? Is it possible to calibrate for the soundcard?
Sunyata
You can test a converter for noise and distortion the same way you test a preamp. But you should be able to rely on your converter's specs. You can also use your PC & soundcard as a signal generator to test a preamp. A straightforward way to calibrate that is to loop the output straight back to the input and measure.
When you are testing a preamp, you are mainly concerned with distortion and noise specs. Preamp noise is measured as equivalent input noise (EIN), which is output noise less gain. You can measure the amount of gain with your converter (comparing to the converter loop). A good preamp would have lower EIN than many converters can manage, but if you add a minimal amount of gain (say 20 to 30dB), then the output noise of the preamp should exceed the input noise of your converter, and can thus be easily measured. The standard measurement for noise uses a 150 ohm resistor connected to the preamp input, and then the thermal noise of the resistor is subtracted. The resistor's noise is very low, so you can probably safely ignore that step in your noise analysis.
Distortion is much simpler, just use a multimeter to calibrate the output level of the converter's test signal (1VRMS would be a standard) and send that to your preamp. If the preamp has too much minimum gain (which you measured above), it could be too hot of an input signal, so attenuate the converter output until the preamp doesn't clip. Then you know what maximum input level of the preamp is. Typically this would be stated at 1% THD, which is below clipping--but how much below depends on the preamp. So just use test signals at various amplitudes until you have a full understanding of its distortion behavior.
Also you will see distortion figures as THD at a given input level, typically a nominal level, say -50dBV or so. A reasonable preamp should maintain low THD until the input signal gets near the maximum input.
You can measure frequency response with white noise. It shouldn't be any trouble at all for a preamp to test flat 20Hz-20kHz.
Then there are other test signals you can try; looking for intermodulation distortion using a combined 18.5kHz and 19kHz signal is one that I like. You can also try to model various transient signals.
I think you will find that the differences between different quality opamps are hard to measure in a real-world swap situation, and even more elusive to hear, especially if the ultimate performance is dominated by the surrounding components.