That design has some strange combinations of parts & values.
R5 is 47 Ohms, with ideally 5.4V across it idle or average, so just over 100mA standing current in the drive circuit. The average power in the driver transistor would be less than 1W at that, so why a 3055 which is 15A and low gain?
No adjustable bias means it's probably class B rather than AB, but for an oscillator that probably does not matter.
There is no bootstrap on the driver/bias chain so it's not going to be very linear, the positive peaks will distort or clip easily.
And worst thing, R9 will totally mess up the bias and dc feedback on the driver, Q1. It should have a capacitor in there.
Some time back I did a series of sequential mods to improve the generic internet design people keep using; it's supposed to be part of a youtube tutorial but I've not got all the actual amps built to match yet, so I've not completed the video..
Stage 0 is the original, then with bootstrap, adjustable bias & midpoint, higher gain power stage then differential input. All the transistor types are generic for the drawings.