0
$\begingroup$

I'm not really an audiophile, so I can't determine the difference. I was playing around with Matlab's Graphic EQ and I noticed that there is a difference in the Magnitude response for the same gain between parallel and cascade. I understand that there is a difference graph-wise because in cascade, the magnitude response adjusts around its center frequency according to the command gain but is ~unity otherwise. Parallel however has a resonance at the center frequency and has low gain at other frequencies. But when I listen to the audio, I can't seem to hear the difference?

So if the difference in sound is negligible for non-audiophiles like me, what would be other advantages if I were to choose what would I use? Are they also equal in processing delay, complexity in implementation or number of calculations, and/or memory? Or any other factors?

Thanks in advance!

$\endgroup$
1
2
$\begingroup$

I very much prefer cascade where the individual bands are implemented as peaking filters. This way the EQ is truly flat if all gains are 0 dB and it's very cheap to implement (one biquad per band). On the downside, they are a little harder to update especially if it's supposed to be "click free" .

Parallel EQs often end up with residual pass band ripple and significant group delay distortion even if they are nominally flat. It's also more expensive to implement.

$\endgroup$
1
  • $\begingroup$ thanks for the reply. Can I ask for a few clarifications? What do you mean by "update" and "click free"? Also, when you say cheaper or more expensive, is that because the one biquad per band will ultimately use less memory than parallel? Or does it mean it can simply more efficiently process the data or it doesn't have much delay and thus save on resources? Lastly, do you think if I changed my mind about using cascade, I can reuse the filters I will design to parallel? I apologize for the questions, I'm still new to this subject. Thanks in advance! $\endgroup$ – Minchu Jun 2 at 16:32

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.