School of Engineering Department of Electronic and Computer Engineering 158 Department of Electronic and Computer Engineering Projects in Audio Signal Processing Supervisor: Kevin CHAU / ECE Student: JANG Jae Won / ELEC Course: UROP 1100, Fall Audio signal processing is a diverse field dedicated to the manipulation and analysis of audio signals. This report delves into the contemporary practices of audio signal watermarking, a method that discreetly embeds information within signals for purposes such as copyright protection, content authentication, and tamper detection. While numerous watermarking techniques have been researched and implemented, current methodologies are not without limitations. Each watermarking method studied and implemented comes with its own limitations, and therefore, there is not yet a set standard or universal method of watermarking and other methods are still being researched. This report will investigate the Spread Spectrum watermarking technique and evaluate its robustness against several potential attacks. Projects in Audio Signal Processing Supervisor: Kevin CHAU / ECE Student: KWON Jiyoon / ELEC Course: UROP 1000, Summer This project explores the use of low-level spectral and perceptual audio features for characterizing and differentiating between musical genres. Utilizing the GTZAN genre collection — comprising 50 thirty-second audio samples from ten distinct genres — five key features were extracted using Python’s librosa library: spectral centroid, spectral bandwidth, spectral contrast, spectral roll-off, and Mel-frequency cepstral coefficients (MFCCs). These features were averaged per song and analyzed both numerically and visually using bar charts, line plots, and heatmaps. The analysis revealed consistent spectral patterns within genres and clear separability across them, particularly in the frequency content and timbral characteristics captured by MFCCs. The findings demonstrate that these features not only provide intuitive insights into genrespecific acoustic properties but also offer a solid foundation for future classification tasks using machine learning techniques. Projects in Audio Signal Processing Supervisor: Kevin CHAU / ECE Student: LI Zongcheng / CPEG Course: UROP 1000, Summer Head-related transfer functions (HRTFs) have emerged as a popular technology in modern audio signal processing, enabling immersive three-dimensional sound experiences through binaural output devices such as headphones. This report is split into two main sections. The first section presents basic audio signal processing methods learned throughout the project, including Discrete Fourier Transform (DFT), Short Time Fourier Transform (STFT), and advanced modelling approaches such as sinusoidal, harmonic, and stochastic models. These techniques form the foundation for understanding complex audio processing systems. The second section covers the definition, methodology, and applications of head-related transfer function, with some evaluation of possible improvements in the future.
RkJQdWJsaXNoZXIy NDk5Njg=