I am using Sigma Studio to create bi-amp loudspeaker EQ files and am finding that Sigma Studio peaking filter response deviates from the expected theoretical filter response. For example, a simple filter with paramters F=1kHz, Ao=+6dB and Q=1 does not match a computer-generated theoretical response (from LEAP software) or the response achieved with a DSP from another manufacturer. The other DSP response matches the theoretical response exactly.
As an example, the measured bandwidth of the filter created by Sigma Studio/Sigma processor is 987.5Hz rather than the expected 1kHz. The Sigma filters are more "peaky" than theoretical. For example, if I enter Q=0.8 into Sigma Studio, I get pretty much the theoretical reponse of Q=1. The error does not appear to be linear, however; simply reducing the Q value by 20% on filters with different parameters does not get the right answer.
Plots of the aforementioned responses are attached.
Has anyone else experienced this? Does anyone at Analog Devices have any information that might be helpful?