This research focuses on the time-domain identification of modal parameters using impact response excitation from signals with a relatively small dynamic range and high noise contamination (e.g., from high-speed cameras). The information required to identify the modal parameters is limited and is contained mostly at the beginning of the signal. In order to perform an identification from such a response, the following challenges have to be overcome: a good frequency-domain separation (for close modes), a good localisation in the time domain and an over-determination (to reduce uncertainty). To overcome these challenges this research introduces the Morlet-wave modal identification method as an extension of the Morlet-wave damping identification method, which has already proven capable of identifying the damping of short signals. Here, the method is extended to the modal parameters and an over-determination approach is proposed to reduce the uncertainty. The method identifies each mode shape separately from 10 to a maximum of 400 oscillations and at damping levels from 0.02% to 2% with a strong presence of noise in the signal. The method is tested on an experimental example and the results are compared to the classical modal identification methodology.