Speech Signal Generation Using Non-Reverberant Reference Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing speech communication systems face challenges in providing optimal speech quality, especially in hands-free scenarios with multiple microphones, due to reverberation and noise, and current methods for combining microphone signals are either suboptimal or complex.
Innovation Solution
An apparatus and method that determine a speech similarity indication for each microphone signal by comparing its properties to those of non-reverberant speech, allowing for efficient combination of signals to generate a high-quality speech signal, even with microphones distributed randomly, without requiring centralized access or dedicated devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If multiple microphones are used to capture speech in hands-free scenarios, then speech coverage and flexibility are improved, but reverberation and noise increase reducing speech quality
Solution Approach 1:
The patent compares spectral parameters of microphone signals against reference properties of non-reverberant speech to determine similarity indications. By analyzing spectral characteristics and comparing them to known non-reverberant speech patterns, the system identifies and selects signals with minimal reverberation, thus improving speech quality while maintaining hands-free coverage
Solution Approach 2:
The patent uses reference properties derived from non-reverberant speech samples as a template or copy of ideal speech characteristics. By comparing actual microphone signals against these reference copies, the system can identify which microphones capture speech most similarly to the ideal non-reverberant condition, enabling selection of the best signals
2Adaptability or versatility
If conventional signal combination methods are used to merge multiple microphone signals, then speech coverage is maintained, but speech quality deteriorates due to reverberation and complexity increases
Solution Approach 1:
The system creates a copy or representation of non-reverberant speech characteristics through reference properties. By comparing actual microphone signals against these reference copies using spectral analysis, the system determines similarity indications that guide signal selection, achieving quality combination without the complexity of conventional methods
Solution Approach 2:
The patent transforms the signal combination problem into a parameter comparison task. By extracting spectral parameters from microphone signals and comparing them to reference parameters of non-reverberant speech, the system selects signals based on parameter similarity rather than using conventional summation or weighting methods, thereby reducing reverberation in the combined output
3Object-affected harmful factors
If sophisticated signal processing methods are applied to reduce reverberation, then speech quality improves, but system complexity and computational resources increase
Solution Approach 1:
Instead of applying complex reverberation reduction algorithms, the patent uses a simpler approach by copying or comparing against pre-established reference properties of non-reverberant speech. This reference-based comparison method achieves reverberation identification and signal selection with significantly reduced computational complexity while maintaining effective reverberation reduction
Solution Approach 2:
The system enables each device to independently determine speech similarity indications by comparing its own microphone signals against reference properties. This self-service approach eliminates the need for centralized processing of all microphone signals, reducing overall system complexity and computational resource requirements while achieving effective reverberation reduction
Data Source
AI summary
An apparatus includes microphone receivers configured to receive microphone signals from a plurality of microphones. A comparator configured to determine a speech similarity indication indicative of a similarity between the microphone signal and non-reverberant speech for each microphone signal. The determination is in response to a comparison of a property derived from the microphone signal to a reference property for non-reverberant speech. In some embodiments, the comparator is configured to determine the similarity indication by comparing to reference properties for speech samples of a set of non-reverberant speech samples. A generator is configured to generate a speech signal by combining the microphone signals in response to the similarity indications. The apparatus may be distributed over a plurality of devices each containing a microphone, and the approach may determine the most suited microphone for generating the speech signal.