Peak Picking Noise Removal in Time-Frequency Signal Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for noise and interference reduction in speech signals, such as adaptive filters and spectral subtraction, face challenges in accurately estimating instantaneous frequencies and separating signal and interference components, especially for non-stationary narrowband interference and broadband noise.
Innovation Solution
A method involving a short-time Fourier transform (STFT) is used to convert the signal into a joint time-frequency domain, where instantaneous frequencies are estimated and modified to redistribute elements, allowing for the identification and elimination of noise and interference by focusing on peak values, resulting in a noise-free and interference-free signal.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If adaptive filter is used to remove non-stationary narrowband interference, then the interfering signal can be removed, but accurate estimation of instantaneous frequency at each time is difficult and the filter cannot remove noise
Solution Approach 1:
The patent transforms the signal from time domain to joint time-frequency domain using STFT, adding a frequency dimension to the analysis. This allows simultaneous observation of both time and frequency characteristics, making it possible to identify and remove interference without requiring precise instantaneous frequency estimation at each time point. The time-frequency representation provides a broader perspective that resolves the estimation difficulty.
Solution Approach 2:
The patent introduces an intermediary peak picking process between the STFT and signal reconstruction. Instead of directly filtering in time domain or using complex adaptive algorithms, the method uses peak detection in the time-frequency domain as an intermediary step to identify signal components. This intermediary approach simplifies the interference removal process while maintaining accuracy.
2Object-affected harmful factors
If spectral subtraction is used to remove broadband noise, then noise components can be reduced, but the clean signal is only an approximate solution requiring ad hoc inversion criteria
Solution Approach 1:
The patent extracts only the peak components from the time-frequency representation rather than attempting to reconstruct the entire signal. By taking out only the significant peak values that represent the actual signal components and discarding the rest (noise and interference), the method achieves cleaner signal estimation without requiring complex inversion processes or ad hoc criteria.
Solution Approach 2:
Instead of processing the entire time-frequency spectrum, the patent applies partial action by selectively processing only the peak components. This partial approach focuses computational effort on the most important signal elements while ignoring less significant components, thereby improving both efficiency and reliability of the signal estimation.
3Object-affected harmful factors
If notch filter is used for stationary narrowband interference, then the interference can be removed, but the filter cannot remove noise or interference whose frequencies change with time
Solution Approach 1:
The patent employs a dynamic approach by using short-time Fourier transform with overlapping windows, allowing the filter characteristics to adapt continuously to changing signal conditions. The STFT breaks the signal into short stationary segments, applies filtering to each segment, and reconstructs the overall signal. This dynamic segmentation allows the system to handle both stationary and non-stationary interference effectively.
Solution Approach 2:
The patent segments the continuous signal into overlapping short-time frames for independent analysis. Each segment is processed separately in the time-frequency domain, allowing localized interference removal. This segmentation enables the system to adapt to frequency changes over time while maintaining effective noise and interference removal in each local segment.
4Measurement precision
If STFT is used to convert signal to joint time-frequency domain, then signal components can be concentrated along instantaneous frequency curves, but computational complexity increases
Solution Approach 1:
The patent extracts only the peak components from the full time-frequency representation rather than processing all frequency bins at all time points. This extraction approach maintains the benefits of accurate time-frequency representation while dramatically reducing computational complexity by focusing only on the most significant signal elements.
Solution Approach 2:
The patent applies partial processing by performing peak picking on only a subset of the time-frequency data rather than exhaustive processing of all components. This partial action reduces the computational burden while preserving the essential signal characteristics needed for effective noise and interference removal.
Data Source
AI summary
Removing noise and interference from a signal by calculating a joint time-frequency domain of the signal, estimating instantaneous frequencies of the joint time-frequency domain, modifying each estimated instantaneous frequency, if necessary, to correspond to a frequency of the joint time-frequency domain to which it most closely compares, redistributing elements within the joint time-frequency domain according to the modified instantaneous frequencies, computing a magnitude for each element in the redistributed joint time-frequency domain, plotting the results, identifying peak values, eliminating from the redistributed joint time-frequency domain elements that do not correspond to the peak values, identifying noise and interference in the peak values, eliminating the noise and the interference from the redistributed joint time-frequency domain elements, and recovering a signal devoid of noise and interference from the modified redistributed joint time-frequency domain.