Oh, now I think I get what you mean. Absorption does NOT affect direct sound. For higher frequencies, direct sound is comparatively stronger, hence the tilt.Was it this statement? I still stand by it. Above 10k air attenuation also affects reflections that travel a longer way.
There is a drawing in Toole's book. I don't have it here but can give you the page indication later.
Maybe it was not formulated well?
I meant from around 6-10k the fall off no longer follows a straight line.
Nevertheless, if you take away even more of the higher frequencies, you get a plateau because the direct sound is just much stronger.
Let's take an extreme case. In an anechoic chamber, the response is only direct and thus flat. In a purely reflective environment, far from the speakers, the indirect sound prevails, thus tilted. Was this the misunderstanding?