These two Op-Amps accomplishes the signal phase inversion and the balanced drive signal. The second Op-Amp works as both as a phase splitter and as a driver for the second output tube. The first Op-Amp is configured as a non-inverting amplifier with a gain of 31. In contrast, the second Op-Amp is configured as a inverting amplifier with also a gain of 31.
A third circuit possibility is to use only current feedback from the output stage. The circuit below includes the output tubes in its feedback loop, but only in terms of current. In other words, the output impedance of this amplifier will be extremely high, as the amplifier defines a current source rather than a voltage source. This will un-tune many bass reflex loudspeakers, as the speaker will only find its own resistance shunting its extraneous movements. But if the speaker has a truly flat impedance curve as some of the planer and ribbon speakers do, the high output impedance will not matter.
An EE friend who claimed that he knew why tube amplifiers sound better than solid-state amplifiers made an experiment to prove it. He argued that the higher output impedance of tube amplifiers actually helped the speaker. He bolted four 1 ohm power resistors to an aluminum plate and the resistors attached to two terminal blocks. This arrangement was then inserted in between his solid-state amplifier and his speakers. Effectively, the solid-state amplifier's damping factor fell to 4, as the added 2 ohms of resistance swamped out the .05 ohms output impedance of the amplifier. How did it sound? On his speakers it improved the sound. While it was still too mechanical sounding, it did have a warmer and more relaxed sound with an improved stereo image. I am not sure what was going on. Maybe the resistors isolated the amplifier from the airborne electromagnetic radiation we live in. or maybe it shielded the amplifier from the reactive kickbacks from the crossover. Or maybe, most speakers do need to see a higher impedance from the amplifier.