Phase Vocoder Bandwidth Extension Temporal Alignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Phase vocoder-based bandwidth extension methods face challenges in temporal alignment and cross-frequency coherence, leading to negative impacts on the magnitude response and transient response of audio signals, particularly due to frequency selective time delays and lack of cross-frequency coherence among patches.
Innovation Solution
The solution involves a patch generator that performs time stretching of subband signals and adjusts phases using filterbank-channel dependent phase corrections, ensuring that the center of gravity of transposed Dirac impulses is aligned across patches, and applying additional time delays to minimize transient energy spread and improve vertical alignment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Volume of moving object
If phase vocoder based bandwidth extension methods are used to generate variable number of band limited sub bands, then the overall bandwidth is extended, but frequency selective time delays are introduced that negatively impact transient response
Solution Approach 1:
The patent applies parameter changes by adjusting the phase of individual subband signals based on their center frequencies and the time stretching factor. The phase correction formula φ_n = -nπ(1-α) dynamically modifies phase parameters to compensate for frequency-selective delays, thereby aligning temporal characteristics across different frequency bands while maintaining bandwidth extension.
Solution Approach 2:
The patent uses copying by generating multiple patch signals that are frequency-shifted versions of the original signal. These patches are created by modulating the original signal with complex exponentials at different center frequencies, and then combining them to form the bandwidth extended signal with improved temporal alignment.
2Duration of action of moving object
If synthesis windows are arranged in fixed hop sizes dependent on stretching factor, then time stretching is achieved, but individual patches have different time delays causing frequency selective delay
Solution Approach 1:
The patent applies local quality by applying different phase corrections to different subband signals based on their specific center frequencies. Each subband n receives a tailored phase correction φ_n = -nπ(1-α), where the correction amount depends on the local frequency characteristic (n), thereby addressing the frequency-selective delay problem locally for each subband.
3Volume of moving object
If patches are generated with different center frequencies, then bandwidth extension is achieved, but lack of cross frequency coherence negatively impacts magnitude response
Solution Approach 1:
The patent implements feedback by using the known center frequencies of the subband signals and the time stretching factor to calculate appropriate phase corrections. This feedback mechanism ensures that the phase relationships between different frequency components are maintained coherently, improving the magnitude response through constructive interference rather than destructive cancellation.
Data Source
AI summary
An apparatus for generating a bandwidth extended audio signal from an input signal, includes a patch generator for generating one or more patch signals from the input signal, wherein the patch generator is configured for performing a time stretching of subband signals from an analysis filterbank, and wherein the patch generator further includes a phase adjuster for adjusting phases of the subband signals using a filterbank-channel dependent phase correction.