Digital Signal Processing 2024W1

Digital Signal Processing

• This course, LING 313, introduces digital signal processing, specifically focusing on its application to linguistic phonetics and speech science.
• The course is taught by Jahurul Islam and covers the 2024W1 term.
• Analog Signals: These are continuous signals, representing a value at every point in time, like a clock's hands moving or a temperature reading throughout a day.
• Challenges of Analog Signals:
  • Continuous nature makes storage and analysis difficult as infinite decimal values are possible.
  • Painfully slow to analyze using traditional methods.
• Digitization: The process of transforming an analog signal into a digital one is called analog-to-digital conversion.
  • This is crucial for modern computing to process the signals.
• Sampling/Sampling Rate:
  • Analog-to-digital processes select specific values (samples) from the continuous signal.
  • The sampling rate determines how frequently samples are taken.
  • Crucial for accurate representation: must sample at twice the frequency of highest frequency component as determined by Nyquist's Theorem for no loss of data.
• Nyquist's Theorem:
  • A sinusoidal signal can be exactly reconstructed from its samples if the sampling rate is at least twice the highest frequency of the signal (frequency components).
  • Failure to implement this causes aliasing (distortion or the appearance of a signal with a different frequency than the actual one being sampled).
• Anti-aliasing Filter: A filter often used in digitization, to remove frequencies higher than half the desired sampling rate to avoid aliasing
• Downsampling/Resampling: An anti-aliasing filter should be applied before downsampling, a common process of transforming a signal by reducing the sample rate, for effective representation.
• Quantization:
  • Assigning a discrete value to a sampled signal.
  • The number of quantization levels affects accuracy; more levels produce a more faithful representation.
  • A trade-off between signal accuracy and processing power exists.
• Fourier Transform (FT):
  • A mathematical technique that decomposes a complex wave into its component frequencies, aiding in analysis.
  • Faster versions (FFT) typically utilized in digital implementations for computational efficiency.
• Sine Function/Sine Waves:
  • Sine waves: are fundamental building blocks in many signals.
  • The values are calculated from right-angled triangles for various angles.
  • Sinusoidal values consistently fall within -1 to 1.