I often use 2-3 send busses with a convolution reverb with same settings but different predelay times to model different distances. For example, vocals go to the bus with more predelay, as the vocalist is (virtually) standing in front of you and reflections take more time to return to your ear compared to the original signal. For drums, I often imagine them standing near the back wall (of a stage) and reflections return very shortly after the original signal, therefore I send them to the bus with short predelay time (I often take out bass drum and snare from this rule).
Of course, EQ, pan etc. is used too as suggested before...