Bug with calculation of peak eq with 5.20.4 ?

[Leif Boehme]

Hello,

with a beta version of 5.20 (is there a possibility to see the name of the beta version in REW?) I simulate the following alignment / EQ for a tweeter.



This matches exactly with the later measured response using various Analog Devices DSPs programmed with Sigma Studio.
When I load this simulation into REW 5.20.4 the calculated reponse changes as follows.



The filter at 20 kHz is the responsible filter were it's predicted behaviour changed with the version, the rest of the filters act as before.
Could you please fix this?

Thank you!

 

[sm52]

In the two pictures, the scale of the Y-axis is different. Make the scale the same and check.

 

[Leif Boehme]

In the two pictures, the scale of the Y-axis is different. Make the scale the same and check.

I updated the second picture to have the same y-axis scaling as the first one.

 

[sm52]

fixed y-axis scaling

 

[Leif Boehme]

I cleaned up the posts to make them more clear. Independent on the scaling you can see a clear 5 db bump around 12 kHz in the second picture at the response with EQ which was not there with the older version = first picture.

 

[sm52]

Yes, now well it is visible. Why so, I do not know. John's Mulcahy participation Is required.

 

[John Mulcahy]

The original simulation you posted was using an equaliser running at 48 kHz (equaliser sample rate is in brackets after the name). From RC13 onwards there was this change:

* Software equalisers (currently the Generic, TMREQ and AU N-Band EQ options) now run at the sample rate of the measurement they are applied to rather than a fixed rate

If you are running a fixed rate equaliser, rather than one that operates at the sample rate of the data, measure at the sample rate the system will run at, or you could select one of the miniDSP options using that rate.