The GMEM (National Center for Musical Creation) has created an integrated microsound synthesis system called GMU (“GMEM Microsound Universe”) inside the Max/MSP environment. GMU works in real time and its center is made of a few microsound synthesis objects (written in C at GMEM). Around these objects, we have elaborated sophisticated sound control strategies, with high-level interactive access inputs. They can be deterministics or stochastics. The whole system offers many connections with tools for Composition Aided by Computers (written in Common Lisp and Java) and tools for sound analysis. Before describing our system and its applications, we will show how the synthesis and transformations of microsound have already been significant in music nowadays.