A caveat with recordings:
As a note, when I was recording sound samples of my own keyboards, I found that all the mechanical recordings came out roughly the same level. Cherry MX brown and Alps CM blue sound similarly loud, despite blue complicated Alps being ridiculously noisy in reality. The Tactile Pro 3's Fukkas were also relatively similar in volume in the recordings. (I made them for my TP3 review and decided against posting them, perhaps as they were all so inaccurate when it came to volume.)
The issue may be that my rubbish microphone doesn't pick up enough bass, and ALPS switches are bass-heavy, while Cherry brown switches in FILCO boards have a particularly high treble clack. The microphone's frequency response will have an effect on what you perceive. (The "control" recording of a Dell KB1421 is significantly quieter, despite that being a particularly loud keyboard for a dome – the hefty thock is missing from the recording, but you get all the ping from the Alps switches.)
I also remember that Ripster's recordings of keyboards were made in such a way that all the "underneath" sounds of scratching and scraping were picked up in a way that bore no semblance to how a keyboard sounds to the operator. His recordings made all keyboards sound horrendous.
Sound recording is so hard – you can troll YouTube for recordings of a switch you're familiar with (e.g. Cherry brown) and find so many recordings that sound nothing like your own keyboard. Local acoustics play a part in that as well – desk surface, wall surface etc.
TL;DR Don't trust recordings
(Granted, one that's flexible at extremely low force, but the conductive elements in a Cherry or Alps switch are flexible at fairly low force, too.)
