How does one measure the differences between amplifiers when they clearly sound different?
The first and simplest measurements are frequency response and noise levels: frequency response shows us that everything we put in makes it through to the output and noise levels are an obvious thing to measure so we don’t have hiss.
Fancier measurements tools include harmonic and intermodulation distortion analyzers: devices that measure byproducts of the amplification process. These simplistic measurement tools put in a pure tone (or several pure tones) and look to see if anything else is generated by the amplification device that shouldn’t be there. These added tones or harmonics are called distortions and they are measured quite easily: the tool simply removes the known test tone and then measures whatever is left over that shouldn’t be there.
We can also measure impulse and step responses but again, we’re looking to see if anything is added to this impulse or to make sure it looks as close to the original signal as possible.
In each case we’re looking to see if the tones we put into an amplifier come out unscathed at the other end and that we didn’t add anything new along the way.
Unfortunately most amplifiers pass these tests with flying colors – yet can sound significantly different than each other.
After writing a bit about overtones, harmonics and the properties of different instruments that give them their unique sounds, it occurred to me that none of the above tests even hint at how these complex and related overtones are preserved between two pieces of equipment.
If we consider that these complex waveform groups are all so intricately related in terms of phase, attack, decay and amplitude then perhaps it’s no wonder that even minute changes in the amplifier should change these relationships causing us to hear differences in the equipment.
Certainly there exists very sophisticated analysis equipment that can take a type of snapshot of these complex waveforms and then compare the two – but that’s rarely done (to my knowledge) and even if you could examine these in minute details, what would serve as your baseline reference of what’s right?
I remember years ago in Stereophile magazine the Carver Challenge. Bob Carver was convinced that he could narrow down the differences between a tube amp and a solid state amp and to do that he simply devised a difference measuring device. This instrument compared the gain matched output of one power amplifier vs. the other and showed the difference between the two. Carver then hand tuned the solid state amp to have as little difference as possible between the tube amp and itself and then challenged the editors to see if they could tell the difference: many couldn’t.
The point of this was not whether one was right or wrong, just that there was a difference and it could be measured and dialed in to reduce that difference, thus making a solid state amp sound like a tube amp.
I bring this up to point out that today we designers still use the best measurement tools we have available to us: our ear/brain mechanism. However, we would clearly do better if we could establish a reference and then devise a tool to help guide us in our designs -= and that tool should focus on these overtones and their relationships to the original in spatial and phase parameters.
It’d be quite a different measurement device than anything I am familiar with.