Related Experiment Videos
Compensating for window effects in the calculation of spectrographic instantaneous bandwidth
1Department of Electrical Engineering, University of Pittsburgh, PA 15261, USA.
IEEE Transactions on Bio-Medical Engineering
|April 14, 2000
Summary
Windowing effects distort instantaneous bandwidth calculations from spectrograms, especially for non-stationary signals. This study proposes a new technique to reduce these distortions and suggests alternative time-frequency distributions.
Area of Science:
- Signal Processing
- Biomedical Engineering
- Time-Frequency Analysis
Background:
- Spectrogram analysis is crucial for understanding signal dynamics.
- Windowing functions introduce distortions in instantaneous bandwidth computation.
- Previous compensation methods are limited for non-stationary signals.
Purpose of the Study:
- To analyze window-induced distortions in spectrogram-based instantaneous bandwidth calculations.
- To evaluate the accuracy of existing compensation techniques for non-stationary signals.
- To propose an improved method for reducing window distortions.
Main Methods:
- Utilizing exact results from Cohen and Lee for distortion analysis.
- Comparing compensation accuracy for stationary versus non-stationary signals.
- Developing and presenting an alternative technique to mitigate window effects.
Main Results:
- Existing compensation approximations are accurate only for stationary signals.
- Accuracy significantly decreases with increasing signal non-stationarity (e.g., frequency modulations).
- The proposed technique effectively reduces window-induced distortions.
Conclusions:
- Windowing significantly impacts instantaneous bandwidth measurements, particularly for dynamic signals.
- A novel technique offers improved accuracy over existing compensation methods.
- Alternative time-frequency distributions can overcome inherent windowing limitations.