Additive manufacturing (AM) parts retain a certain degree of individuality and could suffer from a combination of different defect types, and is therefore the non-destructive evaluation on AM parts remains a challenging task. Engineering non-contact and non-destructive real-time inspection and in-situ quality assurance of AM parts would be a net improvement compared to current quality control methods that are conducted post-production. Here, the authors propose to combine the use of a laser vibrometer with a compression-driven shock tube to assess the quality of AM parts through the evaluation of the vibration spectra of the part. An AM of a cylindrical part was selected for the study, along with different defect types and sizes. These defects include internal voids of different sizes at different locations, local changes in thickness (infill), and local changes in melting temperatures. A numerical model was created and validated using experimental data to conduct model assisted probability of detection (MAPOD). Results were analyzed by evaluating correlation matrices between different models. Results showed that vibration spectra induced by a shock wave were sensitive to different types and sizes of defects under the studied geometry. The defect index yielded an approximately linear relationship with respect to defect void severity. MAPOD curve studies revealed a minimum detectable void defect 0.039% of the AM parts volume.