Symmetrical OP-Amp

Good one. In reality you don't need so much current mirrors when working with discrete components, IC's use current mirrors, because it's physically hard and space inefficient to incorporate resistors in the IC die.

I prefer it with current mirrors. The mirrors used as constant current sources for the differential parts of the amp increase the input resistance to unbelievably high values. In fact, because the components are ideal in the sim, the input resistance is infinite, and the only way for current to get in the inputs is via the capacitance of the transistors themselves. Also, if I get rid of the current mirrors that are part of the differential parts of the amp, it switches from efficient class-AB operation to to inefficient class-A operation. And yes, it's class-AB, not class-B. Class-B is where there is very high crossover distortion because the bases of the push pull transistors are pretty much hooked up directly to each other, making it such that neither transistor is active at all when the input is at zero. Class-AB uses two diodes to create enough voltage to turn the transistors partially on, so that both transistors are NEVER off at the same time. This mostly removes the crossover distortion, but there still is some very small amount, but it shouldn't be noticeable. You can look it up if you want. Please tell me if I'm wrong, and if I am wrong, give me the link to the site that disproves me (I can't follow or copy/paste the link, but I can write it down on paper and copy from there). Also, please note that I said that the transistor bases are "pretty much" directly hooked up to each other. There might be resistors in between them, but the effect should still be there.

You can look at my "Class-B vs. Class-AB" circuit if you want. Also, the long delays with comment-uploads/circuit-uploads-and-updates you sometimes mention, are probably because you don't use the "refresh" button (the one with two arrows forming a circle).

Look class B and class AB shouldn't be viewed as separate classes. Even the worst engineer would design their class B amp with a slight idle current, basically making it a class AB amp. The major difference is the amount of idle current you allow through the amp. For instance a class B amp would allow maybe 10-20mA to run through the output transistors, when idle. A class AB amp might allow for a 100mA for instance, but both of them have an active region in the class A mode. For instance your amp now consumes 1.57W at any given moment, with an output power of 0.98W. That gives you a good efficiency value of ~63%. If this were a class A amp, you'd have a constant consumption of 2.47W and the same output of ~1W, with an efficiency of ~41% which for a low power amp, such as this, is an ideal value. If you plan on using this as a high fidelity preamp, keep it in class A. If you plan on using it as a power amplifier stick with class B. Also, one more thing I'd like to add. Shy away from high impedance inputs, unless you really really need one. They tend to pick up noise all over, and worsen the overall stability and fidelity of the amp. Other than that I really like the circuit. Keep it up ;)

Did you look at my "Class-B vs Class-AB" circuit, and followed the steps I laid out in the description to get to a website that showed class-B as having no idle current? Also, about that noise thing, I don't think someone would leave the inputs of their OP-Amp, or any amp, floating.

I just found out that with this circuit, if you multiply the impedance of the load by 1000, you get the input impedance, so, with the load having an impedance of 50ohms, the input impedance is 50kOhms.

This means it has a current amplification of exactly 1000.

No man, even class B ampa have some idle current. Indeed much less than class AB, but at least 10-20mA or so is desirable

You must know by now that usually to overcome the crossover distortion, a negative feedback is employed. But there is no op amp (or any other error amp) to compensate for the crossover in an infinitely small time, and there is also the overshoot problem as well, but that's another thing altogether. So you drive the amp very slightly into bias (10-15mA for instance) to avoid a full compensation.

Well, just yesterday, I read somewhere that pure class-B audio amps are very rare due to the fact that they don't have any bias current, making it so that they have a very high crossover distortion. So, whenever you mention class-B amps, you must not mean pure class-B amps, but halfway in between the two classes. So, a class-ABB amp? Two parts B, one part A?

Ideally a class B amp does not have any idle current, but in reality, any sensible and self respecting engineering biases it at a few mA or so. Leaving it completely unbiased, even with a very deep negative feedback, causes crossover distortion.