Anti-Causal Audio Filtering for IIR Phase Linearization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Audio signal processing using Infinite Impulse Response (IIR) filters introduces non-linear phase, leading to poor fidelity and latency issues, which are not adequately addressed by existing solutions that either increase latency or compromise performance.
Innovation Solution
Implementing a phase linearization method using an anti-causal filter in conjunction with an IIR filter, which includes a buffer, filter state estimation, and an anti-causal filter to produce a second filtered audio signal with a more linear phase at the expense of increased latency, allowing for customizable output.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If an IIR filter is used for audio signal processing, then calculations, memory usage, and delay are reduced, but non-linear phase is introduced causing poor fidelity
Solution Approach 1:
The audio signal processing is divided into two separate filtering stages: a forward IIR filter for efficient processing and a backward IIR filter for phase correction. Each filter handles a specific aspect of the processing, allowing the system to achieve both computational efficiency and phase linearity by segmenting the overall filtering function into specialized components.
Solution Approach 2:
The patent applies the filter operation in reverse by processing the signal backward through a second IIR filter after the forward pass. This inversion approach allows the system to correct the non-linear phase introduced by the forward filter while maintaining the computational efficiency of IIR filtering, effectively using the reverse operation to eliminate the harmful effect.
2Manufacturing precision
If a FIR filter is used to achieve linear phase, then phase linearity is improved, but calculations, memory, and delay increase
Solution Approach 1:
The patent replaces the traditional FIR filter mechanical structure with an IIR filter-based system that achieves phase linearity through a different mechanism. Instead of using the inherent linear phase property of FIR filters which requires many coefficients, the invention uses the recursive structure of IIR filters applied in forward and backward passes to achieve phase correction with fewer computational resources.
3Manufacturing precision
If phase linearization is applied using anti-causal filter, then audio fidelity is improved, but latency increases
Solution Approach 1:
The system performs preliminary filtering through the forward IIR filter to prepare the signal for phase correction. By pre-processing the signal with the forward filter and then applying the backward filter, the system establishes a two-stage process where the first stage prepares the signal in advance for the phase-linearizing operation, allowing for optimized latency management.
Solution Approach 2:
The patent implements a dynamic system that can selectively apply the backward filter based on processing needs. The dual-filter architecture allows the system to adjust its operation dynamically, enabling phase linearization when high fidelity is required while potentially bypassing the backward filter stage when lower latency is prioritized, providing flexible trade-off management.
Data Source
AI summary
An audio signal processor includes a digital filter block configured to receive an audio signal and output a first filtered audio signal, and a phase linearization block configured to receive the first filtered audio signal and output a second filtered audio signal with a more linear phase.