Audio Downmix Scaling for Low-Frequency Volume Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing surround-to-stereo downmix arrangements often reproduce low-frequency signal components at the wrong volume due to comb-filter compensation, leading to an unbalanced sound field, where the bass range is quieter than the rest of the music, and this issue is difficult to quantify as it depends on the reproduction area's characteristics.
Innovation Solution
A frequency-dependent scaling method is introduced in the downmix process, where the scaling factor is constant for mid-range to high frequencies and gradually reduced for low frequencies, creating a continuous transition from power summation to customary addition, allowing for controlled comb-filter compensation and differentiation between correlated and anticorrelated spectral components.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If comb-filter compensation using power summation is applied to eliminate sound coloration, then mid-range to high frequencies achieve balanced reproduction, but low frequency components are reproduced with wrong volume and become quieter than the rest of the music
Solution Approach 1:
The patent applies different mixing rules to different frequency ranges: power summation for mid-range to high frequencies to eliminate comb-filter effects, and customary addition for low frequencies to preserve volume. This local differentiation resolves the contradiction by optimizing each frequency range according to its specific acoustic characteristics.
Solution Approach 2:
The patent introduces a frequency-dependent scaling factor that transitions from 1.0 at low frequencies to 0.5 at high frequencies. This continuous parameter change enables smooth transition between different mixing strategies, maintaining low frequency volume while achieving comb-filter compensation in the mid-range and high frequencies.
2Device complexity
If a constant scaling factor is used in downmixing, then the mixing process is simple, but the perceived volume balance across different frequencies is incorrect
Solution Approach 1:
The patent transforms the static constant scaling factor into a dynamic frequency-dependent scaling factor. This dynamic approach allows the scaling factor to vary continuously across the frequency spectrum, achieving accurate perceived volume balance while maintaining reasonable computational complexity through efficient frequency-dependent calculation.
3Quantity of substance
If customary addition is used for low frequency mixing, then the bass range volume is correct, but comb-filter notches remain uncorrected in the low frequency range
Solution Approach 1:
The patent identifies that low frequency components have different acoustic characteristics (strong correlation, phase coherence) compared to mid-range and high frequencies. By applying customary addition specifically to low frequencies and power summation to higher frequencies, the patent locally optimizes each frequency range for its specific characteristics, resolving the contradiction between volume preservation and comb-filter compensation.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
In order to realize a correction for changes in the reproduction loudness at low frequencies in a downmix-arrangement, a mixing arrangement is proposed for mixing at least two audio signals, which mixing arrangement is provided with - a first unit (104) for deriving a first power signal, which is a measure for the power of the first audio signal, - a second unit (105) for deriving a second power signal, which is a measure for the power of the second audio signal, - a cross-correlation unit (103) for deriving a cross-correlation signal, which is a measure for a cross-correlation between the first and the second audio signal, - a unit (106) for deriving multiplication parameters from the first and second power signals and the cross-correlation signal, and - a multiplication and combination unit (107) for carrying out a signal processing on the first and second audio signals and combining them. The unit (106) for deriving the multiplication parameters is provided with a combination unit (110) for deriving a combination signal which is a measure for a combination of the first and second power signals and the cross-correlation signal, and is provided with a scaling unit (109) for scaling one of the signals in the unit for deriving the multiplication parameters with a scaling signal. The scaling signal (D[k]) has a frequency characteristic which is a substantially constant below a first frequency value (kL), is increasing between the first frequency value (kL) and a second higher frequency value (kU) and is again substantially constant above the second frequency value. The unit (106) for deriving the multiplication parameters is further adapted to derive a single multiplication parameter m[k]) from the combination signal, and the multiplications and combination unit (107) is adapted to carry out a signal processing on the first and second audio signals which is equivalent to - multiplying the first and second audio signals with this single multiplication parameter, and - combining the so multiplied first and second audio signals.