A life-size mechanical middle ear model and human temporal bones were used to evaluate three different middle ear transducers for implantable hearing aids: the driving rod transducer (DRT), the floating mass transducer (FMT) or vibrant sound bridge, and the contactless transducer (CLT). Results of the experiments with the mechanical model were within the range of the results for human temporal bones. However, results with the mechanical model showed better reproducibility. The handling of the mechanical model was considerably simpler and less time-consuming. Systematic variations of mounting parameters showed that the angle of the rod has virtually no effect on the output of the DRT, the mass loading on the cable of the FMT has a larger impact on the output than does the tightness of crimping, and the output level of the CLT can be increased by 10 dB by optimizing the mounting parameters.