The increasing penetration of power-electronic-interfaced devices (PEIDs) is introducing new oscillatory phenomena that extend beyond the frequency range traditionally monitored using phasor-based techniques. These oscillations may exhibit low amplitudes, time-varying characteristics, and intermittent behaviour, making reliable detection challenging. This paper presents a wide-band oscillation detection methodology based directly on waveform measurements. The proposed approach combines short-time Fourier transform (STFT) analysis for spectral estimation, quasi-peak filtering inspired by CISPR 16 measurement techniques, and frequency-dependent trigger logic. By evaluating both the magnitude and persistence of spectral components, the method distinguishes sustained oscillatory behaviour from transient disturbances, harmonics, and measurement noise.

The methodology is first demonstrated using synthetic waveform data containing sustained oscillations, transients, sags, impulse disturbances, and additive noise. Results show that weak oscillatory components can be reliably detected while maintaining robustness against short-duration events. The method is subsequently applied to measurements from a historical HVDC-related oscillation event, where it successfully identifies oscillatory behaviour in the presence of significant harmonic content. Finally, implementation aspects related to waveform measurement units, selective waveform storage, and wide-area monitoring applications are discussed. The results demonstrate that waveform-based monitoring can provide a practical approach for detecting oscillations in a converter-dominated power system.