Speech Signal Processing Using Dual Microphone Copying

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current noise reduction methods in electronic verbal communication, such as those used in hands-free vehicle systems, fail to adequately improve speech intelligibility in noisy environments with low signal-to-noise ratios, as they often preserve background noise and struggle to reliably estimate the spectral envelope in noisy parts of the speech signal.

Innovation Solution

The method involves using a second microphone positioned further away from noise sources to enhance the signal from a first microphone by synthesizing noisy parts of the signal based on the spectral envelope and excitation signal from the second microphone, allowing for reliable reconstruction of the speech signal even in highly noisy conditions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Object-affected harmful factors

If noise reduction methods such as Wiener filters or spectral subtraction are applied, then some noise suppression is achieved, but the intelligibility of speech signals is not sufficiently improved when perturbations are relatively strong resulting in low signal-to-noise ratio

Engineering Contradiction:
Improvenoise suppressionVSAvoidspeech intelligibility
Core Design Contradiction:
Object-affected harmful factorsVSReliability

Solution Approach 1:

The patent uses a second microphone to capture a copy of the speech signal that is less affected by noise. This copy is then used to reconstruct and replace the noisy portions of the first microphone signal, effectively copying the clean speech characteristics to fix the corrupted signal.

Inventive Principle:
Principle #26Copying

Solution Approach 2:

The second microphone signal acts as an intermediary that bridges the gap between the noisy first microphone signal and the desired clean speech signal. It provides the necessary information to reconstruct the speech without the noise interference.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Manufacturing precision

If partial speech signal reconstruction is attempted in noisy parts, then speech signal quality may be improved, but the spectral envelope cannot be reliably estimated in noisy parts of the speech signal

Engineering Contradiction:
Improvespectral envelope estimation accuracyVSAvoidnoise interference
Core Design Contradiction:
Manufacturing precisionVSObject-affected harmful factors

Solution Approach 1:

Instead of estimating the spectral envelope directly from the noisy first microphone signal, the patent copies the spectral envelope information from the second microphone signal, which is less affected by noise. This copy provides reliable spectral envelope data even when the original signal is heavily corrupted.

Inventive Principle:
Principle #26Copying

3Device complexity

If a single microphone is used to capture speech signals, then device complexity is reduced, but noise interference from localized sources severely affects speech quality and intelligibility

Engineering Contradiction:
Improvemicrophone configurationVSAvoidnoise interference
Core Design Contradiction:
Device complexityVSObject-affected harmful factors

Solution Approach 1:

The patent merges the signals from two different microphones by combining the first microphone signal (with noise) and the second microphone signal (with less noise). This merging allows the system to leverage the advantages of both microphones to produce a cleaner output signal.

Inventive Principle:
Principle #5Merging (Combining)

Data Source

PatentEP2056295B1Speech signal processing
Publication Date: 2014.01.01 NUANCE COMMUNICATIONS INC
  • EP2056295B1 patent drawingFigure 1
  • EP2056295B1 patent drawingFigure 2
  • EP2056295B1 patent drawing

AI summary

The present invention relates to a method for speech signal processing comprising detecting a speaker's utterance χ1(n) by at least one first microphone positioned at a first distance from a source of interference and in a first direction to the source of interference to obtain a first microphone signal, detecting the speaker's utterance χ̃2(n) by at least one second microphone positioned at a second distance from the source of interference that is larger than the first distance and/or in a second direction to the source of interference in which less sound is transmitted by the source of interference than in the first direction to obtain a second microphone signal, determining a signal-to-noise ratio of the first microphone signal and synthesizing at least one part of the first microphone signal for which the determined signal-to-noise ratio is below a predetermined level based on the second microphone signal.