The use of common DSP is well defined by decades of linear music and sound production. EQ, Delay, Distortion, Compression, among others have found common use in most Digital Audio Workstations for decades, but have only recently (within the last 10 years) become efficient enough to use at runtime in order to affect the mix. It's not just the use of fixed settings that makes this area of growth so exciting, but the dynamic and interactive nature of manipulating DSP using parameters from the game to produce special effects that would be impossible to achieve using other techniques.
Effects used at runtime will always use up CPU power, but being aware of the different costs can help you use them more efficiently. If an effect (or effects) are applied to only a few mono instances of a sound object, it is more efficient to apply these at the Actor-Mixer level. If many simultaneous sounds need to be processed by a given effect, it would be more efficient to apply the effect(s) on an audio bus that will be mixed before applying the effect (in stereo or 5.1 depending on the speaker configuration).