FIR Filter Coefficient Selection for Low-Delay Passband Power

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Traditional FIR filters face a trade-off between maintaining power within the passband and minimizing delay time, where increasing power within the passband often results in longer delay times, which can be perceived by users, especially in audio signal processing where more than 5 ms delay is typically required.

Innovation Solution

The proposed solution involves selecting a subset of filter coefficients from a larger set of 2L+1 coefficients, specifically from the (L−N+1)th to the (2L+1)th coefficient, to maintain a delay length of N samples while ensuring the power within the passband remains close to the input signal power, using a signal processing apparatus that samples and processes audio signals.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If the power within the passband is increased to maintain signal quality, then the frequency selectivity is improved, but the delay time increases beyond acceptable limits

Engineering Contradiction:
Improvefrequency selectivityVSAvoiddelay time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent extracts only the necessary portion of filter coefficients from the complete set of 2L+1 coefficients. Specifically, it uses a subset of N+L+1 coefficients starting from the (L−N+1)th coefficient, removing redundant coefficients that contribute to delay while maintaining the essential filtering function and passband power characteristics.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent changes the parameter of filter coefficient selection by using a specific subset of coefficients rather than the complete set. This parameter change in the coefficient selection strategy allows the filter to achieve the desired frequency selectivity with reduced delay time, resolving the technical contradiction between these two parameters.

Inventive Principle:
Principle #35Parameter changes

2Reliability

If traditional FIR filter coefficients are used to maintain passband power, then the signal quality is preserved, but the delay time exceeds 10 milliseconds causing perceptible audio delay

Engineering Contradiction:
Improvesignal qualityVSAvoiddelay time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent extracts a optimized subset of N+L+1 filter coefficients from the traditional 2L+1 coefficients, specifically selecting coefficients from position (L−N+1) to (2L+1). This extraction removes redundant coefficients that cause excessive delay while preserving the coefficients necessary for maintaining signal quality and passband power characteristics.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

Instead of using all coefficients from the beginning (0 to 2L), the patent inverts the approach by selecting coefficients from a later position (L−N+1 to 2L+1). This inverted selection strategy achieves the same filtering function with reduced delay, resolving the contradiction between signal quality and delay time.

Inventive Principle:
Principle #13The other way round (Inversion)

Data Source

PatentUS11259118B2Signal processing apparatus, signal processing method and non-transitory computer-readable recording medium
Publication Date: 2022.02.22 ACER INC
  • US11259118B2 patent drawing
  • US11259118B2 patent drawing
  • US11259118B2 patent drawing

AI summary

A Finite Impulse Response (FIR) filter is configured to minimize delay and maximize passband power by adjusting the filter coefficients applied to the sampled values. The FIR filter obtains an input signal and samples the input signal to generate a set of sampled input values. The FIR filter generates a set of filter coefficients, with each filter coefficient based on a corresponding sampled input value in the set of sample input values. The FIR filter selects a subset of sampled input values that have been most recently sampled from the input signal, and selects a subset of filter coefficients corresponding to sampled input values that are not the most recently sampled. The subset of sampled input values is combined with the subset of filter coefficients to generate an output value for the FIR filter.