For my experiments under steady-state, GAPDH, HPRT house keeping genes are more consistent, while actin and B2M are not. I use both genes and take an average. Some controls are stable under steady-state, but may change while some treatments. Hence, I advice you to standardize and pick a right house keeping gene/ s as control.
There is no general reference gene for all situations. The best is alway to select several and experimentally determine in your system the best reference genes. You could pick up some of the most widely used as starting point, but at the end the trustable ones are the ones which expression does not change under your experimental conditions. It is always a good advise to normalize the expression against multiple internal control genes being the geNorm a very good tool for it.