Spectrogram Analysis: Psychoacoustic Foundations & Signal Processing

This lecture delves into the analysis of spectrograms, exploring tasks such as automatic music transcription, onset detection, and beat tracking. The instructor explains the Discrete Fourier Transform (DFT) and how it is used to interpret complex numbers in the context of musical signals. The lecture also covers the Short-time Fourier Transform (STFT) and its practical applications, including the transformation of energy over time in spectrograms. Additionally, the Constant-Q Transform (CQT) is introduced as a method for achieving consistent frequency resolution. The lecture concludes with a discussion on the importance of spectrograms in audio signal processing and the trade-off between frequency and time resolution.