I disagree about why I got downvoted, but I don't care enough to argue the point.
You're correct about the temperature/duty cycle curve, but not "the only things that determine maximum fan speed" and what that means in relation to that curve. I'm loath to toot my own horn, but the concepts you're describing and are claiming I don't understand are what I work in professionally. I mentioned earlier I'm an EE - an electrical engineer.
Maximum fan speed is inherent to the fan itself and, put very simplistically, in a brushless DC motor is determined by the torque generated. As the motor spins faster, the total resistance (again simplistically, back EMF and net friction) increases, and once it equals the motor's torque output, the motor has reached its maximum speed. Fans operating at a given voltage will have different maximum speeds based on blade size and geometry and motor design.
PWM fan controllers simply correlate a temperature with a duty cycle percentage, but cannot cause a fan to spin faster than its maximum speed. If one fan is rated for 3500 RPM and another 6000 RPM, they will spin at their respective rating at 100% duty cycle. No more.
Noctua's fans in general tend to have engineering superior to many lower priced alternatives, somewhat reducing the sound generated at a given speed, but the low-noise line of fans for which they are often noted achieves that low noise primarily through reduced speed, and thus reduced output. Similarly, the low noise adapter included with some fans is simply an in-line resistor, also reducing speed for reduced sound.