Frequency-Selective Audio Measurement Using Segmented Signal Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for measuring frequency-selective properties in audio devices, such as mobile phones and hearing aids, face challenges in accurately determining transmission properties due to signal distortion from voice codecs and frequency-dependent compressors, and noise gates often mute background noise, making it difficult to measure quiet frequency ranges and noise levels.
Innovation Solution
A measuring device and method that transform signals into the frequency domain, divide them into time-frequency partial signals, and select suitable signals for measurement, ensuring accurate comparison and synchronization to determine frequency responses and noise levels, even in challenging frequency ranges.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional uninterrupted spectral transformations are used for measurement, then the measurement process is simple, but measurement precision deteriorates in frequency ranges affected by noise gates and signal distortion
Solution Approach 1:
The patent divides the continuous measurement signal into multiple overlapping time windows, transforming a single continuous spectrum into multiple segmented spectral transformations. Each segment is processed independently and then combined, allowing selective use of high-quality segments while excluding those corrupted by noise gates or distortion, thereby improving measurement precision without excessive complexity
Solution Approach 2:
The patent uses more spectral transformations than traditionally required (excessive action), processing multiple overlapping time windows and selecting only the suitable portions (partial action) for final averaging. This approach ensures that even if some segments are corrupted, sufficient good-quality data remains to achieve accurate measurements in difficult frequency ranges
2Reliability
If noise gates are used to reduce background noise, then signal-to-noise ratio improves, but measurement precision deteriorates in quiet frequency ranges where noise gates mute the output
Solution Approach 1:
By segmenting the measurement into multiple time windows and analyzing each separately, the system can identify and exclude segments where noise gates are active, while combining results from segments where the noise gate is inactive. This segmentation allows reliable noise level measurement in frequency ranges that would otherwise be muted by continuous noise gate operation
Solution Approach 2:
The patent performs preliminary analysis of each spectral transformation to determine its suitability before combining results. By pre-identifying and excluding corrupted segments before final averaging, the system maintains high signal-to-noise ratio while preserving measurement accuracy in quiet frequency ranges
3Adaptability or versatility
If voice codecs and frequency-dependent compressors are used in audio devices, then adaptability to speech signals improves, but measurement precision deteriorates due to signal distortion
Solution Approach 1:
The patent segments the measurement signal into multiple overlapping time windows, allowing the system to capture transient characteristics of codec and compressor behavior. By analyzing multiple segments and selecting suitable ones, the measurement can account for time-varying distortion characteristics while maintaining precision in determining overall transmission properties
Solution Approach 2:
The patent uses periodic sweeping signals (sine sweeps or multi-sine tone mixtures) as measurement inputs, which systematically excite different frequency ranges over time. Combined with overlapping time-window analysis, this periodic excitation allows the system to measure transmission properties across the full frequency spectrum while accounting for non-linear distortion introduced by codecs and compressors
Data Source
AI summary
A measuring device has a first transformation apparatus (41), a first time-quantization apparatus (42) and a selection apparatus (43). The first transformation apparatus (41) carries out the transformation of at least one signal (46) in the frequency range. The first time-quantization apparatus (42) divides at least one signal (47) into a plurality of chronologically sequential signals (48). The selection apparatus (43) carries out a selection of partial signals suitable for a measurement.