The most obvious reason for decanting a wine is to separate it from any sediment that has formed in the bottle which not only looks unappetizing in the glass, but usually tastes bitter and/or astringent. Before wine-makers mastered the art of clarification this was necessary for all wines. Today such a justification of the decanting process effectively limits it to those wines outlined in ageing as capable of development in bottle, in most of which some solids are precipitated as part of the maturation process. Vintage and crusted ports in particular always throw a heavy deposit (since they are bottled so early in their evolution), as do red wines made with no or minimal filtration. It is rare for inexpensive, everyday table wines to throw a deposit, and most large retailers insist on such heavy filtration that a deposit is unlikely (although not unknown in older, higher-quality reds). To check whether a wine bottle contains any sediment it should be stood upright for at least 24 hours and then carefully held up to the light for inspection at the base (although some bottles are too dark for this exercise to be effective).
Another, traditional but disputed, reason for decanting is to promote aeration and therefore encourage the development of the wine's bouquet. Authorities as scientifically respectable as Professor Émile Peynaud argue that this is oenologically (see oenology) indefensible: that the action of oxygen dissolved in a sound wine is usually detrimental and that the longer it is prolonged-i.e. the longer before serving a wine is decanted-the more diffuse its aroma and the less marked its sensory attributes. His advice is to decant only wines with a sediment, and then only just before serving. If they need aeration because of some wine fault (see faults in wines) such as reduction or mercaptans, then the taster can simply aerate the wine by agitating it in the glass. His argument is that from the moment the wine is fully exposed to air (which happens when it is poured, but not to any significant extent during so-called 'breathing') some of its sensory impressions may be lost, and that decanting immediately before serving gives the taster maximum control.
It is certainly wise advice to decant fully mature wines only just before serving, since some are so fragile that they can withstand oxygen for only a few minutes before succumbing to oxidation. And it is also true that the aeration process of an individual glass of wine can be controlled by the person drinking out of it. However, there are certain types of wines, Barolo most obviously, which may not have been included in Professor Peynaud's experiments with decanting regimes, which can be so concentrated and tannic in youth that to lose some of their initial sensory impressions is a positive benefit.
There is also the very practical fact that many hosts find it more convenient to decant before a meal is served rather than in the middle of it. There are also people who enjoy the sight of (perhaps both red and white) wine in a decanter so much that they are prepared to sacrifice the potential reduction in gustatory impact.