Sound Converter Adaptation Using Sweep Deconvolution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for adapting a sound converter to a reference sound converter are inefficient and unpredictable, requiring multiple iterations to minimize differences in frequency responses, making it difficult to determine the required time for adaptation.
Innovation Solution
A deterministic method using a sound converter with first and second linear transfer functions and a trivial nonlinearity, where a sweep with an exponential frequency curve is used to determine frequency spectra at low and high input levels, allowing deconvolution and division to establish accurate frequency responses, and optionally phase responses, for precise adaptation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple iterations are used to minimize the difference between frequency responses, then the adaptation precision is improved, but the adaptation time becomes unpredictable and extends
Solution Approach 1:
The patent applies preliminary action by pre-calculating and storing the inverse frequency response characteristics of the reference sound converter. Instead of iteratively adjusting the sound converter to match the reference, the system pre-computes the compensation filter that will achieve the desired matching, eliminating the need for multiple iterative adjustments and making the adaptation time predictable.
Solution Approach 2:
The patent uses copying by creating a digital model (frequency response profile) of the reference sound converter and using this model to generate the compensation filter. The inverse frequency response is copied and stored as a lookup table, allowing the system to quickly apply the correct compensation without re-measuring or re-iterating during actual adaptation.
2Loss of time
If a deterministic method is used with sweep signals and deconvolution, then the adaptation time becomes predictable, but the complexity of the measurement procedure increases
Solution Approach 1:
The patent replaces complex iterative mechanical adjustment procedures with a deterministic signal processing approach. Instead of manually or algorithmically adjusting parameters through multiple iterations, the system uses mathematical deconvolution of sweep signals to directly calculate the frequency response and generate the compensation filter in a single deterministic step.
Solution Approach 2:
The patent changes the measurement approach by using sweep signals with exponential frequency curves instead of traditional sine waves or noise signals. This parameter change in the test signal type enables deterministic deconvolution and direct calculation of the frequency response, making the adaptation time predictable while simplifying the overall process despite the sophisticated signal used.
3Measurement precision
If the intrinsic profile is influenced by nonlinearity at high input levels, then the frequency response measurement becomes inaccurate, but increasing the input level is necessary to characterize the full range
Solution Approach 1:
The patent applies segmentation by dividing the frequency response measurement into two separate segments: one at low input levels where the system operates linearly, and another at high input levels where nonlinearity is present. Each segment is measured and compensated independently, allowing accurate characterization of the full input level range while avoiding the problem of nonlinearity corrupting the overall measurement.
Solution Approach 2:
The patent uses an intermediary approach by introducing a trivial nonlinearity element into the measurement chain that acts as a mediator. This controlled nonlinearity allows the system to separate linear and nonlinear effects, enabling accurate measurement of both low-level (linear) and high-level (nonlinear) frequency responses through the use of sweep signals and deconvolution techniques.
Data Source
AI summary
The method for adapting a sound converter to a reference sound converter includes the sound converter having a first linear transfer function with a first frequency response, a second linear transfer function with a second frequency response, and a trivial nonlinearity. The sound converter has a non-linear transfer function corresponding to the frequency response from combination of the first linear transfer function, the trivial nonlinearity, and the second linear transfer function. A first frequency spectrum of the reference sound converter is determined at a low input level. A second frequency spectrum of the reference sound converter is determined at a high input level. The second determined frequency spectrum is used as the second frequency response in the second linear transfer function, and the division of the first frequency spectrum by the second frequency spectrum is used as the first frequency response in the first linear transfer function.