Solar-like oscillations are stochastically excited by turbulent convection. In this work we investigate changes in the acoustic oscillation power spectrum of solar-type stars by varying the treatment of convection in the equilibrium structure and the properties of the stochastic excitation model. We consider different stellar models computed with the standard mixing-length description by Bohm-Vitense (1958) and with a generalized formulation of the mixing-length approach by Gough (1976, 1977). We calculate the acoustic power generated by the turbulent convection which is injected stochastically into the acoustic pulsation modes. Large differences in the oscillation powers are obtained depending on the choice of the assumed convection formulation. We show that the high-quality data Eddington will provide, will allow us to distinguish between theoretical predictions of acoustic power spectra obtained with different convection models.