Downmixer Phase Magnitude Segmentation Audio Interference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional downmixing methods often result in unwanted interferences, such as constructive and destructive interferences, leading to unfavorable changes in audio quality, including Flanger-like effects and signal cancellations, which current technologies struggle to address effectively while maintaining computational efficiency.
Innovation Solution
A downmixer that determines the magnitude value of the spectral domain signal based on loudness information and computes the phase value separately, allowing for a scalar representation of the downmix signal, which is then combined to achieve a complex-valued number representation, thereby avoiding numerical instabilities and artifacts associated with complex-valued additions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If conventional downmixing methods are used to combine multiple audio signals, then the computational complexity is reduced and the downmix signal is obtained, but unwanted interferences (constructive and destructive) occur leading to unfavorable changes in audio quality
Solution Approach 1:
The patent segments the downmixing process into separate magnitude calculation and phase calculation steps. The magnitude is calculated from the absolute values of input signals, while the phase is calculated from the complex-valued sum. This segmentation avoids the interference problems of conventional simultaneous combination while maintaining computational efficiency.
Solution Approach 2:
The patent introduces an intermediary approach by calculating magnitude and phase separately before combining them. The magnitude component uses absolute values to avoid destructive interference, while the phase component uses complex summation to capture interference patterns. This intermediary separation resolves the contradiction between simplicity and quality.
2Manufacturing precision
If energy-preserving corrections are applied in the spectral domain, then the spectrum is equalized and desired spectral characteristics are achieved, but the correction values become sluggish in reaction and amplify constructive or attenuate destructive interferences
Solution Approach 1:
The patent performs preliminary separation of magnitude and phase calculations before final combination. By pre-calculating magnitude from absolute values and phase from complex sums independently, the system achieves immediate response without the sluggishness of iterative energy-preserving corrections, while still achieving spectral equalization through the separate magnitude scaling step.
3Ease of manufacture
If simple addition of N channels is used to form M channel signal, then the downmix process is straightforward and computationally efficient, but strong destructive interference occurs when vectors have similar magnitudes and 180° angular difference
Solution Approach 1:
The patent segments the combination process into magnitude combination (using absolute values) and phase combination (using complex sums). This segmentation allows simple addition-like operations while avoiding the harmful destructive interference that occurs in conventional simultaneous complex addition, as the magnitude component ignores phase information that causes cancellation.
Solution Approach 2:
The patent converts the harmful destructive interference effect into a beneficial feature by separately calculating phase from complex sums. The phase calculation naturally captures interference patterns and phase relationships, allowing the system to preserve useful phase information while the magnitude calculation avoids destructive cancellation by using absolute values.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A downmixer for providing a downmix signal on the basis of a plurality of input signals is configured to determine a magnitude value of a spectral domain value of the downmix signal on the basis of a loudness information of the input signals. The downmixer is configured to determine a phase value of the spectral domain value of the downmix signal and the downmixer is configured to apply the phase value in order to obtain a complex valued number representation of the spectral domain value of the downmix signal on the basis of the magnitude value of the spectral domain value of the downmix signal. An audio encoder uses such a downmixer. A method for downmixing and a computer program are also described.