Sampling Rate Difference Determination Using Signal Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining sampling rate differences between undisturbed and noisy information signals are inadequate, particularly for short signals and in the presence of spectral interference, as they require large window lengths and are not robust against outliers.
Innovation Solution
A device and method that measure sampling rate differences using offset values between information signals, allowing for reliable determination even with short signals and small sampling rate deviations, by subdividing signals into segments, calculating offset values, and statistically evaluating these values using histogram or regression analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If spectral component shift detection method is used to determine sampling rate difference, then measurement precision can be achieved, but very large window lengths (seconds range) are required and the method becomes unusable when temporal and spectral interference is present
Solution Approach 1:
The patent divides the signal into multiple short frames instead of using one long window. Each frame is processed independently to extract offset values, and then statistical evaluation is performed on the collection of offset values from all frames. This segmentation allows precise measurement without requiring long window lengths, directly resolving the contradiction between measurement precision and window length.
Solution Approach 2:
The patent employs statistical evaluation (histogram analysis, regression analysis) of offset values from multiple frames to determine the sampling rate difference. This feedback mechanism aggregates information from many short frames to achieve the precision that would otherwise require a single long window, while also providing robustness against interference in individual frames.
2Measurement precision
If spectral component shift detection method is used, then sampling rate difference can be determined, but the method is not robust against spectral interference and temporal interference from transmission systems
Solution Approach 1:
By dividing the signal into multiple short frames and processing each independently, the patent isolates the measurement process from interference. If interference affects one frame, it does not compromise the entire measurement, as other frames provide redundant information that can be used through statistical evaluation.
Solution Approach 2:
The patent changes the approach from spectral domain analysis to time-domain offset measurement. Instead of analyzing spectral component shifts which are sensitive to interference, the method measures temporal offsets between corresponding signal features in the time domain, which are more robust against spectral interference and transmission artifacts.
3Ease of operation
If conventional methods use Fourier transformation to calculate sampling rate difference, then measurement can be performed, but there is not enough data available for reliable estimation with short signals (e.g., 10s duration)
Solution Approach 1:
The patent segments short signals into multiple overlapping or non-overlapping frames, extracting offset values from each frame. By aggregating these offset values through statistical evaluation, the method achieves reliable estimation even with limited total signal duration, overcoming the data insufficiency problem of conventional Fourier-based methods.
Solution Approach 2:
The patent extracts and evaluates offset values from multiple frames, using more measurement samples than a single Fourier transform would provide. This excessive sampling approach ensures reliable estimation by accumulating sufficient statistical information from multiple partial measurements across the short signal duration.
Data Source
Figure 1A~1B
Figure 2
Figure 3A~3C
AI summary
In embodiments, a device for determining a sampling rate difference between a first information signal and a second information signal is shown, comprising an offset determining unit for determining related offset values for each of a plurality of segments of the first information signal, said values temporally aligning the plurality of segments with respect to the second information signal. A unit for calculating the sampling rate difference on the basis of the offset values is also shown.