Skip to content

References: Audio Signal Processing, Sound Classification, and Speech I/O

  1. Fast Fourier transform - Wikipedia - Explains the FFT algorithm that converts time-domain samples into frequency-domain data in O(n log n) time, the exact transform this chapter uses to turn an audio buffer into a frequency spectrum.

  2. Mel-frequency cepstrum - Wikipedia - Describes how the mel scale reshapes frequency bins to match human pitch perception and how cepstral coefficients are derived from it, grounding this chapter's MFCC feature-extraction step.

  3. Spectrogram - Wikipedia - Covers how stacking successive frequency spectra over time produces a time-frequency-energy image, the visual representation this chapter uses to explain what sound-classification models actually learn from.

  4. Understanding Digital Signal Processing (3rd Edition) - Richard G. Lyons - Prentice Hall (2010) - Lyons is widely credited across the DSP community for explaining the FFT, DFT bins, and complex phasors in plain English with graphical, non-calculus analogies rather than proofs, the same intuition-first approach this chapter takes toward the FFT and frequency spectrum.

  5. Fundamentals of Speech Recognition - Lawrence Rabiner and Biing-Hwang Juang - Prentice Hall (1993) - Rabiner is credited with the "urn-and-ball" analogy for hidden Markov models, first popularized in his celebrated 1989 tutorial and carried into this textbook, making the probabilistic sequence models behind speech-to-text and keyword spotting concrete for students.

  6. Discrete Fourier Transforms (scipy.fft) - SciPy Documentation - Official tutorial on computing the FFT in Python with scipy.fft, showing exactly the frequency-spectrum computation this chapter's np.fft.rfft code example performs on a captured audio buffer.

  7. Mel-frequency Cepstral Coefficients (MFCC) for Speech Recognition - GeeksforGeeks - Step-by-step walkthrough of the nine stages that turn a spectrum into MFCC features, with Python code using librosa, matching this chapter's audio-feature-extraction and preprocessing-pipeline sections.

  8. Simple audio recognition: Recognizing keywords - TensorFlow - Official tutorial that trains a small CNN to recognize spoken keywords from spectrograms, directly illustrating this chapter's keyword-spotting and wake-word-detection pipeline from raw waveform to trained-model decision.

  9. Raspberry Pi Audio - Raspberry Pi Foundation - Official documentation for Raspberry Pi's I2S-based DAC and amplifier HATs, explaining how digital audio travels from the GPIO header to dedicated conversion and amplification hardware, the exact signal chain this chapter's output section describes.

  10. Adafruit MAX98357 I2S Class-D Mono Amp: Pinouts - Adafruit Learning System - Wiring guide for an I2S DAC-and-amplifier breakout board, showing the LRC/BCLK/DIN digital audio pins and speaker terminals that correspond to this chapter's PCM5102 DAC, amplifier module, and speaker output stages.