2
$\begingroup$

I know overlap save and overlap add are used for long data sequence filtering. Are there any other similar or better techniques like these?

$\endgroup$
  • 2
    $\begingroup$ What exactly do you want to do and what does 'better' mean? $\endgroup$ – Matt L. Apr 21 '13 at 19:03
  • $\begingroup$ Hi Matt, I want to know if there are any other techniques other than overlap save/add. Other easy ways to filter long sequences. $\endgroup$ – Harvey Apr 21 '13 at 19:05
2
$\begingroup$

The main alternative that I can think of is the hybrid method proposed by Bill Gardner and patented by Lake DSP (now part of Dolby). There appears to be a copy of Gardner's paper here.

$\endgroup$
3
$\begingroup$

The traditional overlap-add technique cannot be used as is in musical applications (reverberation) where latency is critical. It is a common practice to break down the impulse response into non-uniform chunks, and to run in parallel different convolvers for each segment of the impulse response - some of them being naive FIR implementations (for the short blocks at the "head" of the response), and some of them being high-latency overlap-add FFT convolvers.

See this paper.

$\endgroup$
1
$\begingroup$

Overlap-add and overlap-save are block processing algorithms for implementing FIR (finite impulse response) filters. Both of them use an FFT algorithm to compute the output samples, which makes these methods computationally efficient. In principle, any other filter implementation can also be used, even with long input sequences. IIR (infinite impulse response) filters can be more efficient than any overlap-add or overlap-save method because they need much fewer coefficients than FIR filters to achieve comparable frequency responses.

$\endgroup$
  • $\begingroup$ Say i have a fixed FIR filter(M taps). I have an infinitely incoming sequence. The sequence is taken in blocks, overlap save or overlap can be used to filter and output. Are these the only techniques for FIR filtering? $\endgroup$ – Harvey Apr 21 '13 at 19:46
  • 1
    $\begingroup$ No, you can simply use the direct form implementation of the FIR filter if computational complexity is no big issue. $\endgroup$ – Matt L. Apr 22 '13 at 7:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.