The situation is somewhat simpler for those in apartment buildings. Economical high rise construction revolves around concrete or steel (encased in concrete for fire protection) frames i.e. columns and floor slabs. This gives you a head start in terms of heavyweight structure, but flanking transmission - the conduction of sound through structural paths - is likely to be the major issue, and is beyond your control unless you really do have more money than sense.

Houses built in the UK before c.1910 generally have single-leaf external walls (no cavity), built of anything from stone to snot-and-horsehair. Since this is a general overview I'll ignore them for now...

How big is the problem?

Before you really consider tearing into things, would buying a decent set of headphones help?

This of course will not help where the problem is noise ingress, but consider when the sound is a nuisance. The normal level of background noise during daylight hours is onehelluva din when you're trying to relax in the evenings. There's something like a 20dB difference between evening and daytime ambient noise levels - subjectively, about four times the loudness, since humans perceive 8-10dB as a doubling of volume. The upshot is that muffling nuisance noise down to tolerable levels during evening-time listening can be really difficult - because it's just so much more obvious.

What to attack: some sweeping generalisations

How effective a solution you can cook up depends on thoughtful analysis, the degree of upheaval you can tolerate and your budget. Look to solve the most direct paths first; in most cases it is transmission through the floors/walls which is the problem - sound energy transfer is an area-weighted thang.

Subjectively the most annoying sound you'll need to stop is bass content (say below 150Hz), for which the only solution is mass. A 102mm brick wall plastered both sides gives you about 40-42dB loss; lightweight partition on the other hand (13mm plasterboard both sides, timber or metal studs) only provides 27-32dB in comparison; 10dB less - which represents 10 times as much energy passing through! So what can be done to mollify a Barrett Box?

Improved wall performance can be had at various levels. Filling a stud partition with mineral wool can improve figures 4-6dB; adding a second layer of plasterboard to both sides (and mastic sealing all the joints) will improve matters to about 42-45dB; and building the partition right in the first place - decoupled leaves of 2 layers of 15mm board (British Gypsum 'Soundbloc' or similar) each side on (separate) staggered studs, mineral wool cavity filling and careful sealing of joints can pull this up to a very good 53-57dB, which represents the limit of reasonable return on your investment. Not an option unless you are building anew though.

All this can be in vain though if the flanking walls are continuous, or the floor above lightweight; the sound will travel through these paths and degrade the isolation achieved.

Floors and ceilings are frequently problematic acoustically - especially impact noise. UK building regulations require a minimum mass of some 300Kg/m² for party floors between occupancies for acoustic reasons. Unless you have a concrete upper floors you won't get any where near this with pugging (sand loading) an existing timber floor, if only because your ceiling is now resting at your feet....

In upgrading an existing floor probably the best approach is to mix of mass and absorption. Briefly it involves installing a second ceiling - losing of about 100 mm of headroom - on framing spanning between the walls, covered on the underside with two layers of plasterboard (second layer laid perpendicular to the first, with sealed joints) with 50 -75mm of fibreglass mat in the space below the original ceiling. Seal the edge of the new ceiling to the walls with mastic. This construction will provide up to 38dB of isolation, which is about as good as it gets. A 'normal' suspended ceiling (eg office-style 'acoustic' tiles in a lay-in grid) will not only not work, it could be even worse at certain frequencies - and might well severely upset/attenuate the bass balance in the room. It could also make the room excessively deadened in the midrange, which apart from killing reproduced music can have a surprisingly negative effect on subjective comfort.

If this sounds like too much bother, and you have suspended timber floor above (eg to the children's bedroom...) with plaster/plasterboard ceiling, you could try a second layer of plasterboard on the ceiling - it's much easier and cleaner than adding sand above (and nearly as dense) - and laying an isolated floor in the bedroom; try 18mm flooring grade chipboard floating on as little as 6mm of foam underlay.

A few other little details

Next down the list in significance (and it is quite a way down) is the in-air path. After you've achieved a lossy (massive) enclosure, you need to get fussy about even tiny holes. The keyhole in a door typically halves the door panels acoustic isolation (in a frequency-dependant way); but there is usually about a hundred times as much free area between the edge of the door and the frame....so brush or magnetic seals to all four edges are a good idea, whether or not your doors are heavy, ie hardwood and/or solid core ( and no, the timber type is irrelevant). Solid doors are advantageous though...mass again

Glazing is a bit anomalous. Glass has intrinsically high density, but is available only in relatively thin panels; it also has low internal damping, so a glass panel has a fairly tightly-defined resonant signature depending on its dimensions, thickness and mounting. Double glazing is better than a single thick pane, but far from perfect. The transmission of sound through double glazing is driven by the coincidence effect (acoustic coupling of the leaves through the intervening gas) and so isolation provided can be quite poor at certain frequencies. In critical applications it's usually the case that you end up with two leaves of very different thickness, spaced 100mm apart or so, in extremis the cavity is filled with sodium hexaflouride - a dense gas - to add damping. Domestically, double glazed units are about as good as it gets (23-34dB); for preference, the two panes should be of differing thicknesses and fairly widely spaced; 18-24mm is typical in commercial units and less than ideal, compromised by thermal considerations; wide cavities suffer from convection currents increasing heat loss. If the glazing is a problem, you could add an extra pane in a separate frame; but be aware that thick glass gets very expensive, very quickly !

Beyond this point, returns plummet. Ventilation noise(s) and decoupling of partitions to minimise transmission through structure are really outside the scope of this brief scribble - another day perhaps....

It's quite a minefield to address the problems of balancing isolation, cost, upheaval and maintaining /improving the acoustic of the listening space, but it can be done - and with only a few comparatively simple measures. Like most things though the law of diminishing returns applies with a high interest rate. There are inherent problems which are truly difficult to deal with and will ultimately limit the isolation achievable, for example masonry walls resonate around 80Hz, so provide minimal isolation around this frequency; flanking transmission through common structure can swamp the loss provided by a-purpose-built partition etc. Keep a sense of proportion - if your house is divided up with lightweight partitions there is only so much you can reasonably achieve.

Further pages on this topic: we'll add more of this stuff including typical building details and materials for acoustic partitions /separating floors as remedial works if anyone is interested....let us know if you are!

Further reading: a bunch of links yet to be posted...

© the twisted pair 2000