Hummmm... This is really much more complex than what is seen at first blush. The practice of noise reduction must address the needs of the operator first - IMHO.
There are essentially three areas of cencern for noise level reduction: One is masking or hiding the report of the shot at the target (or areas in front of the weapon). For most readers of this board, this would never be a concern and deserves no further discussion.
The two other locations are to the side of the shooter (noise polution) and at the shooter's ear -- When we speak of these "areas", these are physical locations, the difference between the perceived noise level can be very different at each location.
Attenuation of side noise is nice and a concern for military and law enforcement operators -- The exposure to these noise levels by those "on line" is a major concern, as well is the possible reflection of noise back to the operator -- But, as I assume that most are interested in a suppressor for the safety and comfort of the shooter (and the cool factor)... I would offer the following.
First... all bullets have flight noise, does not matter if the round is so slow you can see it or if it is going 4000fps -- all bullets have flight noise. True, subsonic speeds produce noise levels that are so minimal that they are of no concern for this discussion, but when a round becomes transsonic, the flight noise increases greatly... to a point around 1300fps, where the report levels off for supersonic (above transsonic) and even hypersonic flight.
So, what does that mean? To the shooter, very little -- Bullet flight noise is a radial emission and with the exception of ground reflection and reflection from objects in front of and to the side, the shooter will perceive very little of the bullet flight noise. To anyone standing to the side of the bullet path, it will be much louder -- But even this gets a little tricky, especially from a hearing preservation stand point.
Just say that Joe's rifle has a muzzle report of 160dB -- the perceived report is actually the combination of 3 different sounds over a very short time... The bullet noise (or "crack"), the muzzle signature and ground reflection. All this is scrunched into about 30mS.
If we attenuate the muzzle report to below 140dB, the bullet noise will indeed be the loudest part of the new report, keeping the actual dB level at about 140dB. But, to the observer, it will "sound" even quieter, if the new levels are taken using a weighted scale dB(A), the difference can be seen... also, the duration of the new signature (above a painful level)will be much shorter (less than a few mS) making it seem "quieter" again. The best way to actually track the signature is to record and graph it, peak dB levels can not show nearly as much as time graphed frequency and kilopascal levels.
The safety concern is that the observer is still being subjected to damaging noise levels, but is beeing "fooled" that they are not as bad as they really are -- This is compounded in real life by the observer "beleiveing" the signature is not as loud, because they know it has been suppressed.
Crap... I am rambling, cutting to the chase: If you want to screw a can on your centerfire, high-velocity rifle, it will still be loud. To the shooter, it will most likely be a "have to raise your voice to get over it loud", but some say that the suppressor can replace hearing protection for the shooter... maybe not. It is a viable solution for reducing the level to below a sudden and permanent damaging level.
If you expect that really quiet, mouse pissing on cotton "phitt" sound, you need a system that matches the weapon, ammuniton and suppressor for that effect -- and there are trade offs, naturally.