The overall circuit architecture makes all the difference, not the used devises.
A very key point....... but one that sidesteps and wiggles around the fact that really one cannot use the same circuits with the two types of device..... they are not that identical.... So inevitably there are circuits which are characteristic of tubes, and less usable or totally unusable with "transistors". One can often use circuits of the same "general type', but not ones that are "identical".* Comparisons, DBT or not, can therefore never really separate the device from its usage.
So it comes back around (despite efforts to avoid that) to the difference in devices. Each type device is inevitably used in its characteristic circuits, which then becomes somewhat indistinguishable from the device itself.
It happens that there ARE characteristic differences in harmonic structure between tubes with their circuits, and at least bipolar transistors with their circuits. Both even/odd and harmonic order differences exist, typical of the usages. They don't typically show up in the strict linear range.
That is almost by definition, if you think about the matter logically.
Distortion is the name for any difference between input and output other than overall amplitude..... In the "linear" range there isn't much if any difference between input and output, by definition.
Therefore, if there is a difference in sound, one of the two must produce more distortion, or a different type of distortion, otherwise the outputs would be identical. But in the "linear" range, distortion is quite small, and clearly the opportunity for a difference is also quite small. And, by extension, very hard to find.
But in the real world, input levels are not controlled, outputs are loaded with various unsuitable loads, and clipping is common, at the very least on transients or peaks.
So another really deadly fault in the DBT test is that the units were not used in ways typical of actual guitar usage..... In the linear unclipped and un-distorted range, one SHOULD NOT be surprised that most people cannot distinguish any difference between tubes and transistors..... if the circuits are competently designed, there will BE little difference.
it is totally improper to attempt to spin the test as having relevance past what was actually done, what was actually tested...... Which seems to be that if you do your best to match two circuits in frequency response, and run both at levels which are virtually undistorted, the total audible difference between them is going to be small.
Seamonkey apparently has somewhat of an agenda, evidently an unquestioning faith that any "methodologically valid" test is always golden as proof (or disproof) of a theory which it purports to test, and unfailing energy in promoting his beliefs.
I'm fine with that.... it's really his problem, after all.....
I do not share his apparent beliefs. Neither life not science is that simple. There are ALWAYS other considerations, unforseen issues, and things which the test didn't test.
As in this case.
* The one exception is the devices which Kirkwood Rough had at NAMM many years ago. They were "solid state", of a somewhat obsolete type normally seen only in textbooks now, and were an essentially direct duplicate in performance of a generic 12AX7. They measured same, and sounded same..... because they BEHAVED the same... INHERENTLY.
No other solid state device I have any familiarity with has that sort of performance.
He seems to be a T.B. member, BTW.....
