ΣSigmaStudio Ask AI Русский Log in

Second-Order Medium EQ Filter Output Observations

SigmaDSP & SigmaStudio dns Nov 3, 2025 00:00 3 replies Original on forum ↗
Subscribe
dns Question
Nov 3, 2025 00:00

Hi Team,

We are working on the EVAL-ADAU1452REVBZ board using a second-order filter implemented with a Medium-Size EQ filter in double precision mode (without slew). When we vary the gain, we observed a distorted output. We have attached the code and related waveform below.

code snip:

Module Parameters:

Fc: 80 Hz Q: 0.60

Filter Type: Peaking EQ

Module path:

Test Steps:

Switch the Gain level sequentially: 0 dB → –10 dB → +10 dB → 0 dB.

Observed Behavior:

Please provide your feedback regarding this behavior.

software

4.5

pastedimage1762177356876v3.png
pastedimage1762177356876v3.png
pastedimage1762177337430v2.png
pastedimage1762177337430v2.png
pastedimage1762177559155v5.png
pastedimage1762177559155v5.png
Harishgowtham Reply
Nov 4, 2025 00:00

Hello dns,

Have you tried with the slew version of the filter?

If not, can you please try with the slew version and observe the result?

Regards,

Harish

dns Reply
Nov 4, 2025 00:00

Dear Team,

We have not tried the slew version of this module, as it is not the intended module. We would like to confirm whether the disturbance observed in the output while changing the gain is expected behavior of this module, and if so, what the cause of it.

Regards,
Dns

DaveThib Reply
Nov 4, 2025 00:00

Hello Dns,

I will add to what Harish said.

This is classic zipper noise when you change the gain. What makes this a little different is that it is not just a gain factor change but filter coefficients that need to be changed. So there will be some drastic changes when the filter coefficients are changed. This is exactly why we have the slew versions! So thanks for showing why the slew versions are needed! Depending on your application you will probably need to use the slewed version.

If your application can mute the audio, change the filter and then unmute then use the version you are using. If your application requires real time changes while listening to the audio then you must use the slewed version. What it does is slowly changes the coefficients every sample period until it matches the desired new setting. Making these changes with no audible artifacts.

Dave T