Solar-like oscillations are believed to be excited stochastically by turbulent convection. The stochastic excitation depends on the details of the turbulent spectrum. In Samadi (2001) a generalized formulation for the oscillation power was derived. This formulation first calibrated on a solar model allows one to investigate different assumptions about the stellar turbulent spectrum. We thus explore consequences of the assumed turbulent description for solar-like oscillating stars with different effective temperatures and luminosities. Large differences are found in the amplitude spectrum of a given star when using different turbulent spectra as well as in a star-to-star comparison. We conclude that COROT's performances will be able to discriminate between several turbulent models.