Headphone Speaker Recognition Using Inside Microphone
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing digital signal processing techniques for speaker recognition in headsets and headphones are ineffective in noisy environments and may not accurately differentiate between the registered owner and an imposter, especially when using outside microphones that are prone to ambient noise and lack unique ear canal characteristics.
Innovation Solution
The method utilizes the inside acoustic microphone of a headphone to capture speech signals, which are processed to create a unique model reflecting vocal, nasal tract, and ear canal characteristics, enhancing speaker recognition accuracy by employing a speaker recognition algorithm trained with these signals, and optionally using a ratio of inside to outside microphone signals for feature extraction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If outside microphones are used for speaker recognition, then speech can be captured, but ambient noise reduces recognition accuracy
Solution Approach 1:
The patent segments the audio signal into two distinct components: speech signal captured by the outside microphone and ambient noise captured by the inside microphone. This segmentation allows the system to process and separate the desired speech from the harmful ambient noise, improving speaker recognition accuracy in noisy environments.
Solution Approach 2:
The inside microphone acts as an intermediary that captures ambient noise which would otherwise contaminate the speech signal from the outside microphone. By using the inside microphone signal as a reference for noise cancellation, the system can subtract the ambient noise component from the outside microphone signal, isolating the speech signal for accurate speaker recognition.
2Measurement precision
If outside microphones are used for speaker recognition, then speech capture is possible, but unique ear canal characteristics are not obtained
Solution Approach 1:
The patent merges the signals from two microphones positioned at different locations: the outside microphone captures speech while the inside microphone captures both speech and ambient noise. By combining these signals through spectral subtraction, the system extracts speech with enhanced characteristics that include both vocal tract information and unique ear canal acoustic properties, improving speaker recognition precision.
3Measurement precision
If inside microphone is used alone, then ear canal characteristics are captured, but speech signal quality is reduced due to ambient noise
Solution Approach 1:
The outside microphone signal serves as an intermediary that provides a clean reference of the speech signal without ambient noise contamination. This reference signal is used in spectral subtraction to remove the ambient noise component from the inside microphone signal, preserving the unique ear canal characteristics while eliminating the harmful ambient noise effect.
Data Source
AI summary
A speaker recognition algorithm is trained (one or more of its models are tuned) with samples of a microphone signal produced by an inside microphone of a headphone, while the headphone is worn by a speaker. The trained speaker recognition algorithm then tests other samples of the inside microphone signal and produces multiple speaker identification scores for its given models, or a single speaker verification likelihood score for a single given model. Other embodiments are also described and claimed.