Explore random music generator based on Short-Time Fourier Transform

This paper delves into the mechanism of how the Short-Time Fourier Transform (STFT) is used for generating random music. A brief review of the history of stochastic music development is at the outset. The paper contains the principle of audio digitization. This starts with how to convert a continuous audio signal into discrete samples. The Nyquist Theorem plays an important role in that process to preserve signal integrity. The Heisenberg uncertainty principle takes effect as the STFT is applied to covert these samples. It states that when converting the audio signal between the time domain and the frequency domain, the audio signal can only have the properties of one or the other. This paper categorized the audio signals into three categories based on their spectral characteristics. The paper also points out the reason why natural sound effects are always chosen to generate random music due to their inherent complexity and randomness. This paper demonstrates the detail of STFT in generating random music, explaining how to specifically manipulate spectrum analysis to generate random music with practical examples. The paper concludes by discussing the possible future role and direction of STFT in the field of stochastic music.