Ableton Forum

@x3000:

Indeed, your spectrum shows a declining slope of ~ 24 dB/octave. (-51, -74 between 5 k and 10 k). That's what I would expect. My spectrum gives me ~ 16 dB/octave both for the autofilter and the VCF of the Juno 6. That's at least acceptable.

Now, send the same input to the spectrum of EQ Eight. For me the result is fully different, maybe 8 dB/octave.

Hence, I still think this could be a different scale.

Yes, indeed. The display in the EQ8 does not appear to be technically correct. The reference does not describe the spectrum display exactly.
It could be that the frequency response in EQ8 is displayed in dB(A) and in Spectrum in dB(C). But that would not explain the big difference.

I would guess that Ableton hasn't reinvented the wheel every time, but has integrated standard libraries or algorithms for various cases.

Maybe they forgot to square the values:
log(1/16) = -12dB
log(1²/16²) = - 24.1dB

I wouldn't bother with that. The display of the spectrum in EQ8 is only a rough guide. In the end, you have to adjust by ear.

However, you can see that the stand-alone spectrum analyzer in Live (Spectrum) appears to output consistent values.

I really don't see the point here. These Ableton filter implementations are mathematical textbook variable Q filters and they work exactly as they should.

There is a VST plugin to check frequency response and phase, maybe try that, it'll prove everything.

https://www.bertomaudio.com/eq-curve-analyzer.html

@benmuetsch:

The point is not the filter, but the display of the frequency band in EQ8, which you can easily check with Spectrum.