First there was the vacuum tube, then there was the transistor. Yes, it was progress – the tube is bulky, expensive, it wears out, it doesn't come in different types optimized to positive and negative working.
It does have the saving grace though that it is linear, meaning that it is intrinsically pretty much distortion free.
Now, the transistor – it is small, cheap, doesn't wear out, and it does come in complementary pairs.
But it is not linear – in fact it is about as far from linear as you can get.
With a tube, if you put a little more signal in, you get a little more out. Put a lot more in and get a lot more out.
But if you feed a transistor with a small voltage, it gives nothing. A little more voltage – nothing. A little more…
Go even a smidge over the threshold that lies around 0.6 volts and so much current flows through the transistor that it burns out almost instantly!
Even when well-controlled, the transistor can be considered to be a very good distortion generator. It is very good as a switch too – fully on or fully off. Yes transistors are great at that.
The reason why transistors can be used at all is that they can have a high gain, and gain can be exchanged, via a process known as negative feedback, for low distortion.
But you have to wonder whether using an inherently non-linear device, and then removing the distortion, could possibly be a good idea.
The problem is that things have changed since the days of the vacuum tube. Back then, they knew the tube had its limitations, so they invented the transistor.
But now, the problem is that the transistor has significant limitations too, but no-one seems to be bothered about finding a replacement.
Where are the promising technologies? The search for possible avenues of exploration?
What we need is a device that will give gain, be adaptable to a range of input impedances, quiet, and above all linear and distortion-free.