Not quite. When you have a front end that spits an instruction into micro-ops that's a fixed mapping of 1 to 1-3 instruction and your decoder can handle many of those every clock. When you a micro-coded instruction you can have that instruction turn into a whole subroutine of indefinite length and on most current designs you can only decode one at once and it takes over until it completes.