Sound Source Separation via Interchannel Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for separating sound sources from multi-channel audio signals are inefficient due to unclear channel distribution information, leading to noise and deteriorated separation quality.
Innovation Solution
An apparatus and method that determine interchannel correlation parameters, estimate channel distribution values, and calculate membership probabilities using Gaussian mixture models to precisely separate sound sources from multi-channel audio signals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If sound source separation is performed based on empirically selected specific values, then separation can be implemented, but noise occurs and separation quality deteriorates due to sudden signal changes
Solution Approach 1:
The patent changes from using fixed empirically selected specific values to using dynamically determined channel distribution values calculated from interchannel correlation parameters. This allows the separation parameters to adapt to signal conditions, preventing noise and deterioration caused by sudden signal changes while maintaining ease of implementation through automated parameter calculation.
Solution Approach 2:
The patent introduces a feedback mechanism where channel distribution values are continuously calculated from interchannel correlation parameters and used to adjust the separation process in real-time. This feedback loop enables the system to respond to signal changes and maintain high separation quality without manual intervention.
2Device complexity
If channel distribution information is not precisely determined, then separation can be performed simply, but noise occurs and separation quality deteriorates
Solution Approach 1:
The patent replaces manual or empirical determination of channel distribution information with an automated computational system that calculates interchannel correlation parameters and derives channel distribution values algorithmically. This substitution maintains simplicity in operation while achieving high precision through mathematical computation rather than manual adjustment.
3Ease of operation
If fixed specific values are used for separation, then the process is simple, but the system cannot adapt to different signal conditions
Solution Approach 1:
The patent transforms the separation system from using static fixed specific values to using dynamic channel distribution values that are continuously calculated from interchannel correlation parameters. This dynamic approach maintains ease of operation through automated calculation while providing adaptability to different signal conditions and sudden changes.
Data Source
AI summary
Disclosed are an apparatus and a method for separating sound sources capable of learning distributions of corresponding sound sources based on the assumption that specific sound sources have specific distributions based on interchannel correlation parameter in audio signals providing space perception through a plurality of channels to separate an amount corresponding to energy contribution of the corresponding sound sources from mixture signals. Exemplary embodiments of the present invention can more precisely predict the channel distributions of the specific sound sources included in the input mixture signals and more accurately separate sound sources than a method for separating a sound source based on the channel according to the related art, under conditions that general channel distribution information of the specific sound sources are approximately modeled.