It is very obvious from the PCB layout the
speaker current is being measured as there is a 0.1ohm
sensing resistor in series with the output.
For it to 'read the room' as a microphone is bollocks. Bob does seem to use the back EMF but most likely this is to compensate for the high-ish output resistance so the amp output is probably
less impedance dependent.
The fact that one can use a speaker as a microphone In this case the speaker must be connected to a high impedance microphone input and not simultaneously to a near 0 ohm amplifier output. Back EMF and 'microphone' signal as well as applied signal can not be 'pulled apart' without complex math and fully characterizing the speaker.
Unfortunately this isn't verified measured as only a resistive load is used. A complex load could shed some more light but this doesn't have back EMF. It should be measured with a speaker and on top of that it should also be measured with no signal and a second speaker with a sweep on it directly in front of the speaker and then measure at the primary of the transformer what the amp does with it.
I voted poor because of
safety features. It should have had a different mains socket (without a pin), a double insulated sign on the back and everything mains related should have had some shrink sleeving on it or the ground pin should have been connected.
One thing to keep in mind, given the power remains the same at various impedances, is that while it does not have spades of power available it does output 75W but will sound louder than a 75W SS amp (with any speaker but Maggies).
Most 'power' is needed in the lows for music and the impedance is highest there in speakers.
So a bass note into say 16 ohm will get 75W of power = 34V while at 1kHz (assuming no XO there) will only have 17V (assuming 4 ohm nominal)
A SS amp rated at 4 ohm would need to be rated at 290W in 4 ohm.
This is why a 40W rated tube amp can sound more powerful than a 40W rated SS amp.
Of course, this is cornering it a bit short as there are limits to the max impedance of a speaker where this will work and the speaker would have to peak at the frequencies you want to play louder.
In short... the claim Bob makes has not been verified. The claim of reading the room is bogus.
The reflected sounds into speaker will be immensely small compared to the back EMF and the applied signal and
cannot be told apart from the back EMF due to the huge
level difference of back EMF and 'room reflections'. It simply is physically not possible.
Back EMF itself can be used (and most likely is)
but this was not measured.
@amirm could measure it by 'injecting' a signal (using a SS amp output via a dummy-load) into the output of the Crimson and measuring on the speaker out terminals what signal remains. Then do the same thing with the current measurement resistor shorted.
This is a lot of extra work though and he has been quite busy with other (far more important) stuff.