I've just been browsing a few of these articles.
I found one that discusses an "extended audiogram," which consists of 14 data points across a range up to 16,000 Hz. That means out of a possible 16,000 data points, only 14 are actually measured.
The core issue here is that far too few data...