The transformers I was talking about were used in power converter applications - granted they were mostly nonlinear applications such as resonant topologies and voltage source inverters but it is still clear to see what impact is on the the waveform. Transformer design is very complex - much more so that I made it seem in my over-simplified general comment above. I agree in some instances, you can get the input very close to the output but in general, the B-H curve of a transformer is totally nonlinear. There are some very linear cores but they are more often than not only inductors - similar to McIntosh's autoformer where the action is self-inductive. My comment didn't apply to that but rather complete transformers with a ferrite core, primary, and secondary. In this case, you are at the mercy of the B-H characteristics which are far from linear unless you operate in a very specific range but that never happens when you are talking AC voltage of course. Many of the commercial isolation transformers are undersized and so you run too close to the saturation region where the flux density starts to level off (i.e., nonlinear). Oversizing the transformer costs lots of $$$ and we all know that engineers are forced by the bean counters to keep cost in check - this means you don't add any more core than is absolutely necessary. If the end result is for electrical equipment in a medical lab, the core WILL NOT BE OPERATING LINEARLY because it will be made as small, light, and cheap as possible. It will be just fine for electrical equipment but not necessarily for human ears. Some audiophile companies may do it right but again, I would check it in the lab before betting my life on it.
Having said this, some people may actually like a nonlinear sound eventhough it is totally inaccurate from a strict electrical sense. One more example of "if you like it, keep it." All I am saying is you need to be careful when using a transformer. It isn't a simple thing.
I looked for the waveforms I was talking about but couldn't locate them off hand (we may not have even saved them). There are many examples online about it if you are interested. I will look for a good example and paste the link. The problem we were having was that the slope of the B-H curve changes depending on the direction of the applied voltage becuse that little bit of loss results in "fattening" of the curve which makes the slope different no matter how linear you make it. This is a problem.
Finally, be sure you realize that an autotransformer is NOT the same as a transformer. The two words are not interchangeable. Arthur
