Coursera Spring Sale
40% Off Coursera Plus Annual!
Grab it
Explore the critical limitations of Shannon-Nyquist sampling theory in this 45-minute conference talk from ADC 2025, where Professor Emerita Jamie Angus-Whiteoak reveals why understanding the actual waveform behind digital samples is essential for accurate audio processing. Discover how common assumptions about linear signal processing break down in real-world applications, using the seemingly simple task of measuring audio level as a key example where sample values may not represent the true peak levels that occur between samples. Learn about the fundamental issues involved in deriving true signal values from sampled audio and understand why reconstructing the analogue waveform is sometimes necessary for accurate signal processing. Examine various methods for measuring peak values, including their advantages and disadvantages, while exploring alternative approaches that provide accuracy with minimal computational overhead and delay. Delve into efficient techniques for obtaining real waveform values and discover how these principles apply to frequency domain processing as well. Gain insights into how knowledge of actual sample values becomes crucial for non-linear operations, with practical methods for achieving accurate results. This accessible presentation is designed for audio developers with diverse backgrounds, from music applications and game audio to embedded systems and digital signal processing, providing essential knowledge for anyone working with sampled audio systems.
