In fairness, if they're tested on different road surfaces the numbers aren't really comparable. Road surface type, and age, can more significantly affect the level of noise being generated. 6-7dB difference would appear to be easy to achieve with different grades of asphalt.
I also note the deviation occurring in his measurements...
... in just the 26 seconds of the recording in the e-Tron, he's seeing 2dB variance in all three speed ranges. This is not to discredit him, just to point out that a couple of dB difference in the test results being discussed probably doesn't translate to much in the real world.