I have been told (and repeated here) that it's an easier task to get signal on to tape well than to get it off again, which makes some sense when you think about it. Reading that tiny AC fluctuating magnetic signal off the head accurately and amplifying it without adding noise or distortion is not that trivial of a task.
That having been said, Ben, there were consumer full stereo tape recorders with microphones and amps and speakers built in the 60s that used 5 TUBES ! total ! And one of those was the rectifier! ( Don't ask me how I know these things - but you could guess) That's right. Two tubes per channel for everything! One was typically a 12A*7 of some sort which took care of mic/head preamp chores with clever switching and the other was typically a 6BM8 triode/power pentode which provided one more stage before the power output section. The bias oscillator was done by converting one of the output amplifiers to that duty during record with more clever switching. So, 3 triode stages for everything excluding driving the speaker. Mic./line to head and on playback head to line out, then through the power pentode to the speaker.
The reason for all this "minimalist high end triode" and "single ended class A output" stuff was pure COST :) They were too cheap to put in more stages.
That became much easier with transistors and then multiple transistors then op amps. So we went very far the other direction in the 70s.
The balance and trade off is always trying to feed the signal through a minimal amount of stages, but with careful design so as to achieve the best quality.
