Summary

This blog presents a modification of the Farrow resampler that reduces computational cost by splitting the impulse response into many segments and bank-switching between them. It shows how the bank-switched Farrow trades interpolation accuracy for lower runtime complexity and outlines evaluation and implementation considerations for real-time systems.

Key Takeaways

• Understand the bank-switched Farrow architecture and how it differs from a conventional Farrow structure.
• Estimate the interpolation accuracy vs. computational-complexity trade-offs when increasing the number of impulse-response segments.
• Implement a bank-switched Farrow resampler and adapt it to fixed-point or real-time DSP constraints.
• Compare performance against conventional Farrow and polyphase resamplers to select the appropriate resampling strategy.

Who Should Read This

DSP engineers or researchers experienced in resampling and filter design who need efficient multirate implementations for real-time or embedded systems.

Still RelevantAdvanced

Topics