Sparse-Signal Frequency Detection for Short-Cycle Harmonic Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audio signal analyzers require long measurement times to accurately quantify harmonic distortions, limiting their effectiveness in identifying small changes in harmonics, and are sensitive to signal noise.
Innovation Solution
A machine learning-based technique using sparse input signals derived from audio signals around zero-crossings, extremums, or steepest portions, combined with neural networks, allows for rapid and accurate detection of added harmonics in audio signals, achieving high accuracy within a few cycles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional audio signal analyzers are used to accurately quantify harmonic distortions, then measurement precision is improved, but measurement time increases significantly
Solution Approach 1:
The system performs preliminary action by pre-processing the audio signal to create a sparse signal representation that captures only the critical information (zero-crossings, extremums, or steepest portions) before feeding it to the ML model. This preliminary extraction of essential features enables the ML model to achieve high measurement precision in harmonic distortion quantification within a fraction of a second, resolving the contradiction between accuracy and measurement time.
Solution Approach 2:
The patent replaces conventional mechanical/analog signal analysis methods with a machine learning-based digital processing system. The ML model, trained on sparse signal representations, substitutes for traditional time-consuming spectral analysis methods, enabling rapid and accurate quantification of harmonic distortions without requiring lengthy measurement durations.
2Measurement precision
If conventional analyzers measure audio signals for extended durations, then measurement precision improves, but sensitivity to signal noise increases
Solution Approach 1:
The system extracts only the essential information from the audio signal by identifying and retaining specific features (zero-crossings, extremums, or steepest portions) while discarding redundant information. This extraction creates a sparse signal representation that maintains the critical harmonic distortion characteristics while eliminating noise-sensitive portions of the signal, thereby achieving high measurement precision without increased noise sensitivity.
Solution Approach 2:
The patent changes the parameter representation from continuous time-domain signals to discrete sparse signal representations characterized by specific parameters (zero-crossing points, extremum locations, or steepest slope positions). This parameter transformation enables the ML model to analyze harmonic distortions based on these transformed parameters, which are less susceptible to noise while maintaining measurement precision.
3Productivity
If machine learning models are trained on sparse input signals, then processing speed increases, but device complexity increases
Solution Approach 1:
The system segments the audio signal processing into distinct functional stages: (1) sparse signal generation module that extracts zero-crossings, extremums, or steepest portions; (2) ML model inference module that processes the sparse representation; and (3) harmonic distortion quantification module. This segmentation allows each component to be optimized independently, achieving high processing speed while managing system complexity through modular architecture.
Data Source
AI summary
A system includes a memory and processor. The memory is configured to store a machine learning (ML) model that is trained to estimate values of frequencies added (FA) in sparse input signals that have been derived from respective input audio signals, the sparse input signals being indicative of one or more FA in the corresponding input audio signals. The processor is configured to (i) receive an input audio signal, (ii) derive from the input audio signal a sparse input signal indicative of the FA in the input audio signal, and (iii) estimate the values of the FA in the input audio signal by applying the trained ML model to the sparse input signal.