The challenge with this question is that it requires one to suspend the laws of physics and invent new ones. There's no concept of "reversing a sound" in real physics. Sound always travels in the direction sound travels. Because of this, we would need to define what it means for a sound to be reversed. The answers would vary based on the definition.
To provide one such definition, consider the case of simply reversing the vibrations. To do this, we're going to have to locally reverse entropy, which is legal (you can't reverse the effects of entropy globally, but you can do it locally).
In such a definition, the first thing you would notice is your eardrums getting warm. This is the thermal energy that was dissipated from your eardrums as they stopped vibrating from the sound. In the normal sound direction, this occurred after the sound, but in the reversed version we need to have it precede the sound, because we'll need its energy in the next step.
Next, a spontaneous oscillation forms on your eardrum. This oscillation was forced by the vibrations in the air in the forward direction, but must be spontaneous in the reverse direction.
These vibrations produce vibrations in the air which join vibrations from all over the room, focusing in on the wood. These vibrations would eventually knock on the knuckles.
You would hear exactly what you hear in "synthetic" reversed sound, because all you would hear is what the oscillations in your eardrum tell you that you hear. In this case, they are spontaneous oscillations that are the exact reverse of the sound you heard in the forward direction.
Oh, and the knuckles are really confused. As far as the hand would be concerned, it would feel like the door reached out and smacked it across the knuckles. That's what happens when your run physics backwards.