Dynamic range compression (DNR) is used for several purposes. One reason (perhaps the major one) is to reduce the impact of background noise on audio signals. Another reason is to reduce the surprise effect of sudden loud sounds, while yet another is to normalise the signal level of a transmitted signal. Such normalisation may be useful for processes such as speech recognition, where a speech recognition process may work better with normalised signal levels. Dynamic range compression may also increase the apparent loudness of an audio source, and some broadcasters like to do this as it feels that it gives them a greater presence and competitive advantage.
Dynamic range compression is often used by broadcasters, particularly commercial broadcasters, as they feel it gives their station a louder and more distinctive presence. Such compression may also improve the perceived subjective sounds if they are listened to in noisy environments. This is because the compression may raise the level of quiet signals above the level of background noise. This can be useful in cars, though there is often no need for the DNR to be done by the broadcaster as this can easily be done at the receiving device under user control.