I noticed I had quoted this earlier, and somehow forgot to put my commentary in after, as I was replying to more than one post at once. Sorry about that.
What happens with the feedback divider is interesting. At low radio frequencies indeed, the divider will attenuate the signal. But a feedback loop in a conventional power amplifier isn't usually a simple resistive potential divider. It is usual to bypass the resistor between the output and the negative feedback input with a small value capacitor. This helps manage loop stability. Also, unless a DC offset servo is used, the leg of the divider to common will be run though a large value DC blocking capacitor. So, as the frequency increases the first resistor reduces in value as the bypass capacitor starts to dominate (this is why it is there) and the DC blocking capacitor - typically an electrolytic - will start to become inductive. The entire loop will look more and more like a direct connection from output to negative input. Eventually, as the frequency rises even higher nothing looks like what it says on the label, and the most important aspects are just the layout of PC traces. By then most audio use active components will have ceased useful operation anyway, so it won't matter.
