An ideal hi-fi speaker should have the following requirements:
A single transducer that contemporarily meets all these requirements has not been invented yet. A speaker able to correctly reproduce 20Hz cannot correctly reproduce 20Khz but this is true even if we tighten the range from 100Hz to 10Khz, at least until we are speaking of hi-fi. In fact to produce high frequencies it is necessary to move a small amount of air but with high linear acceleration values, and this requires a small and extremely light transducer. On the contrary, for low frequencies it is necessary to move greater and greater volumes of air the lower the frequency. This involves the use of bigger membranes — therefore heavier — and capable of high excursions. It is plain to see that we are speaking of antithetical demands.
The only way to have your cake and eat it is to split the audio spectrum in at least two frequency bands, one including the lowest frequencies only, the other the highest ones. This way the low frequencies can be played by a speaker designed for this purpose and the high frequencies by another specific transducer. These frequency bands are called ways and such a system is called two-ways.
If the spectrum is splitted in three bands then we'll get a three-ways and we'll need three dedicated speakers: a woofer for the lows, a tweeter for the highs and a midrange for the mids, which can be dome or cone according to the project choices. And so on.
The number of speakers always equals the number of ways. However, this is not inversely true. In fact it is possible that one or more ways end in two or more speakers, each reproducing the same frequency interval.
To operate this delicate splitting in frequency bands just described are called really them, the filters