Influence of DC offset on bias

I rescind my post. It's in error for the purpose of this thread.

Hifihvn,
In transistors it's the same story only make plate equivalent to collector(assuming connection with common emitter). There's a purpose to adjust the offset so the output is measured the same across collector and ground on both opposite junction transistors pnp,npn pair (I assume that it's a class B amplifier, but I might be wrong) and another channel accordingly. So the offset might be different in transistors (they're way harder to match than tubes and have larger parameter tolerance) but bias current will be the same and 'sound' the same.
Can you tell what kind of amplifier we're talking about?
Brand? Class of operation? Must be a very vintage one. Most of the current ones would have an auto-bias which is a piece of cake to implement.

The amps in question are a pair of LSR&D Superamp monos with no offset pots, AmpzillaII which is a dcservo amp., Son of Ampzilla with both offset and bias pots, and some Bedinis also with no offset pots, a pair of early Boothroyd Stuart Meridian 105 monos with no offset adjustment. I recently sold a Meridian 559 with auto bias because IMO, it had no high-end characteristics at all. The 105's were fabulous and thought it would be more of the same but it sounded very average with no real redeeming qualities; not particularly fast, no great extension or linearity, imaging, sound stage etc.. I wondered if it was because it is not a fixed bias amp? When offset and/or bias is off, these are exactly the characteristics an amp displays. I guess my question is how much tolerance is engineered into auto bias?

An auto-bias bases is simply using a reverse conductivity of diode. Depending on how large your bias current should be you choose the diode with the same reverse conductivity current. It's all in the parameter list. There are certainly more sophisticated and advanced auto-bias supplies that allow to match upto 10% of parameter tolerance.

Thank you. That explains why I was disappointed with the 559. 10% is huge. I can easily detect anything that's not identical. Maybe I should be "Golden ears". Pinpoint imaging and transient response and therefore detail absolutely depend on a perfect bias adjustment in relation to each channel being the same.