Frame-Based IIR Signal Filtering for Smooth Parameter Transitions

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing audio and speech signal filtering technologies face challenges in avoiding discontinuities and maintaining low complexity, particularly when Infinite Impulse Response (IIR) filter parameters change between frames, leading to instability and increased computational complexity.

Innovation Solution

A system and method that employs two filter units to gradually transition IIR filter parameters from a higher to a lower filtering status and vice versa within an initial subinterval of an update interval, allowing for reduced discontinuities and complexity by eliminating the need for cross-fading and separate filter operations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If IIR filter parameters are updated once per frame to adapt to time-varying signals, then filtering adaptability is improved, but discontinuities and instabilities are introduced

Engineering Contradiction:
Improvefiltering adaptabilityVSAvoidsignal stability
Core Design Contradiction:
Adaptability or versatilityVSStability of the object's composition

Solution Approach 1:

The patent applies preliminary action by performing a first filtering operation at the beginning of each frame using parameters from the previous frame before applying the current frame's parameters. This preparatory step ensures a smooth transition between frames, preventing discontinuities while maintaining adaptability to time-varying signals.

Inventive Principle:
Principle #10Preliminary action

2Stability of the object's composition

If cross-fade approach is used to smooth parameter transitions, then discontinuities are reduced, but computational complexity increases

Engineering Contradiction:
Improvesignal continuityVSAvoidcomputational complexity
Core Design Contradiction:
Stability of the object's compositionVSDevice complexity

Solution Approach 1:

The patent merges the transition smoothing function into the filtering operation itself by performing two filtering operations (first with previous frame parameters, then with current frame parameters) rather than separating the smoothing and filtering steps. This consolidation achieves discontinuity reduction without requiring additional cross-fade computational overhead.

Inventive Principle:
Principle #5Merging (Combining)

3Measurement precision

If filter parameters are changed frequently to track signal variations, then filtering precision is improved, but computational complexity increases

Engineering Contradiction:
Improvefiltering precisionVSAvoidcomputational complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies partial action by performing filtering operations only on specific portions of the frame (initial part with previous parameters, remaining part with current parameters) rather than uniformly across the entire frame. This selective approach maintains high filtering precision where needed while reducing unnecessary computational complexity in other regions.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP3707715B1Signal filtering
Publication Date: 2023.01.04 FRAUNHOFER GESELLSCHAFT ZUR FORDERUNG DER ANGEWANDTEN FORSCHUNG EV
  • EP3707715B1 patent drawingFigure 1~1a
  • EP3707715B1 patent drawingFigure 2
  • EP3707715B1 patent drawingFigure 3

AI summary

There are discussed methods and systems for filtering an information input signal (11, 11a, x), divided into different update intervals, according to parameters varying with the update intervals, to obtain a filtered output signal (y, 15). A system (10) may comprise: - a first filter unit (12) to filter a first filter input signal (11, x) at least at an initial subinterval (T i ) in a current update interval (T) to obtain a first filter output signal (y', 13), according to parameters associated to the preceding update interval, the first filter unit (12) being configured to change the parameters along at least the initial subinterval (T i ) from a higher-filtering status to a lower-filtering status; and - a second filter unit (14) to filter a second filter input signal (13), at the initial interval (T i ), according to parameters associated to the current update interval (T) to obtain a second filter output signal (15), the second filter unit (14) being configured to change the parameters along at least the initial subinterval (T i ) from a lower-filtering status to a higher-filtering status. The first filter input signal (11) is based on the information input signal (x), the first filter output signal (13) is an intermediate signal (y'), the second filter input signal is based on the intermediate signal (y'), and the filtered output signal (y) is based on the second filter output signal (15).