Maybe I'm getting old, but any exhaust that
causes me to turn my head ( IOW, gets my
attention ) is too loud.
Since perceived volume doubles every three
decibels, 100 db would be 267% louder than 92.
That's not true.

3db is the lowest easily perceived volume change.
TEN DB is TWICE AS LOUD !The 100db exhaust is NOT twice as loud as the 92db exhaust !3dB = twice the power
6dB = twice the amplitude
10dB = twice the perceived volumeAdding up two 12dB noise sources will get you, on average, 15dB (which will not sound twice as loud)
I was working with a sound engineer and asked for a level to be dropped 3db.
Their reply was “so you want it half as loud?” and I said “No, 3db” which was countered with “3db is half the volume”.
So that’s what prompted me to look into decibel changes and how that translates to our real world perception.
It is true that to increase a volume level by three db requires twice the power, which I think is where the confusion is.
A doubling in power does not equal a doubling in audio perception.
It was determined many years ago in controlled audibility testing, that the following rules were generally accurate among the population:
6dB SPL increase is perceived as an approx. 50% increase in volume by a sample group.
10dB SPL increase is perceived as an approx. 100% increase in volume by a sample group.
Another Explanation
A 3dB increase is twice as loud, in that increasing the level by +3dB by definition means twice as much audio energy is now being pumped into the room – well, actually 1.995 times as much, thanks to the wonders of logarithms.
But the human ear’s response is also logarithmic, so twice the energy does not sound like twice the volume.
There is, of course, no clear point where anybody’s going to say “ah, that’s now precisely double as loud as it was before”; there’s no little mental VU meter needle.
But the general rule of thumb is that people tend to call a 10X, or 10dB, increase in audio power “twice as loud”, if you insist that they indicate such a point, and this is backed up by neurological studies.
JJ