Microphonic Noise Compensation Model Retraining for Clear Radio Audio
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Communication devices for first responders suffer from microphonic noise, which is introduced by vibrations at the voltage-controlled oscillator (VCO) and interferes with audio output, making it difficult for users to hear mission-critical information, and existing microphonic noise compensation models may become inadequate over time or be challenged by device settings.
Innovation Solution
A computing device retraining a microphonic noise compensation model by selecting an appropriate model based on the type of noise indicated by the communication device, using audio samples to retrain the model, and deploying it when successful compensation is achieved, or adjusting device settings to mitigate noise.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If microphonic noise compensation models are deployed at communication devices, then microphonic noise is compensated, but the compensation models become inadequate over time as noise characteristics change
Solution Approach 1:
The patent implements dynamic retraining of compensation models by collecting audio samples over time and periodically retraining models with new data. This transforms the static compensation model into a dynamic system that adapts to changing microphonic noise characteristics, resolving the contradiction between initial compensation effectiveness and long-term model validity.
Solution Approach 2:
The system continuously monitors audio samples for microphonic noise characteristics and uses this feedback to determine when retraining is needed. By establishing feedback loops that track noise changes and trigger model updates, the system maintains compensation effectiveness over extended periods despite changing noise patterns.
2Illumination intensity
If volume settings are increased to improve audio output, then audio clarity is improved, but microphonic noise increases due to greater vibrations at the VCO
Solution Approach 1:
The patent converts the harmful effect of increased vibrations at higher volumes into useful information by using audio samples collected at different volume levels to train compensation models. The system learns to compensate for volume-dependent microphonic noise characteristics, allowing clear audio output even at high volumes where noise would normally be problematic.
3Reliability
If device settings are adjusted to reduce microphonic noise, then noise compensation is improved, but audio output quality may be degraded
Solution Approach 1:
The system dynamically adjusts compensation parameters based on analyzed audio samples rather than using fixed settings. By changing compensation parameters adaptively according to actual noise characteristics observed in usage, the system maintains audio output quality while effectively reducing microphonic noise, avoiding the trade-off between noise reduction and audio quality.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Effectively reduces microphonic noise to below a given level, ensuring clear audio output by compensating for noise changes over time and device settings, thereby enhancing communication clarity.
Implementation Method 1
processed audio output by a speaker causes vibrations at the VCO, which may translate into noise introduced into the audio output from the RF mixer
Data Source
AI summary
A computing device: receives, from a communication device at which microphonic noise is occurring: a microphonic indicator indicating a type of the microphonic noise; and audio sample(s) that include the microphonic noise; and selects, using the indicator, a microphonic noise compensation model pretrained for compensating for the type of the microphonic noise. The computing device retrains, using the audio sample(s) that include the microphonic noise, the compensation model to compensate for the type of the microphonic noise, and applies the compensation model, as retrained, to at least one of: further audio sample(s) received from the communication device that include the microphonic noise; and an audio test set that includes the type of the microphonic noise. When the compensation model compensates for the microphonic noise in at least one of the further audio sample(s) and the audio test set, the compensation model, as retrained, is deployed to the communication device.