Summary

This paper presents a straightforward method to substantially reduce the number of multiplies per input sample for conventional IIR lowpass and highpass digital filters. Readers will learn the structural changes and practical trade-offs that enable lower computational cost while preserving filter response and suitability for low-latency, real-time applications.

Key Takeaways

• Apply the reduced-delay IIR structure to conventional lowpass and highpass designs to cut per-sample multiplies.
• Quantify computational savings and compare multiply counts versus standard IIR realizations (Direct Form, cascade).
• Analyze and manage stability, numerical precision, and group-delay implications introduced by the reduced-delay arrangement.
• Implement the approach in fixed-point and embedded real-time DSP environments to achieve low-latency filtering.
• Evaluate trade-offs between complexity reduction and filter performance (magnitude/phase) for audio and communications use cases.

Who Should Read This

DSP engineers and developers with some experience in digital filter design who need lower-complexity, low-latency IIR filters for audio, communications, or embedded real-time systems.

Still RelevantIntermediate

Topics