I think it'd be a more useful presentation if they had gone with an x-y chart of temps vs dBA....
It's a nice product comparison, but their testing methodology is very flawed. Those 12V/7V performance runs just don't make much sense. If you look more closely at the noise performance at 7V, you'll notice that the first few ranks are all equipped with PWM fans, which probably dropped to quite low revolutions. Those heatsinks also underperformed in the 7V testings.
Bummer, so much work for not much assertive power.