ADAU1452 Q factor value will affect the Crossover result?
Hello guys,
I meet one problem which is the "Higher Q factor" affect the crossover output result.
Here is mine part of the DSP structure:
from the above figure, the 10 bands EQ are before the crossover, and the final result will output Low path(red line in AP) and the High path(blue line in AP)
if I use the lower Q factor value, the output looks fine, like below:
1KHz boost 5 dB and the Q is 0.7
20KHz boost 5 dB and the Q is 0.7,
from the AP Frequency Response result:
the result looks fine on both path, all smooth.
but if I change the 1KHz Q factor to 12 and the 20KHz Q factor is same as the 1KHz,
the result will like below:
the crossover frequency is at 2.1KHz, if Q factor value is 12,
the low path frequency line(red line) result is correct, but the high path frequency(blue line) will be affected by the low path EQ boost, the blue line appear a little peak on the 1KHz.
from the 20KHz result is same as the 1KHz, the red line will appear a little peak.
Does the Q factor has a maximum value to avoid affecting the other path(to get the smooth result)?
or is there some limitation when the EQ tuning before the Crossover?
Thanks!
Alvis.
Audio Signal Processors
adau1452
adau1452 sigmastudio
SigmaDSP Audio Processors
crossover
Q factor
peaking filter
Show More
Hello,
The peaks are appearing in both the Low and Hi crossover outputs because you have placed the equalizer BEFORE the crossover. This is not a problem in the DSP, or a limitation of the algorithms.
The crossover applies a smooth 2nd or 4th order roll-off between the woofer and tweeter depending on what settings you choose. In the frequency domain, you will find that each speaker output signal is the sum of the transfer functions of the equalizer and crossover.
When you increased the Q to 20, the equalizer's contribution to the combined transfer function became visible on the plot. I don't believe that this is a problem. The spurious peaks are more than 35dB down, which is typical for this kind of crossover.
With a lower Q the equalizer will still contribute to the combined transfer function in the same way, but the crossover filter dominates so it is difficult to see the difference on a plot.
Joshua
Hi Joshua,
Thank you for your reply,
after checking some article and make some experiment,
I found the problem what you said!
I will re-designed my structure, thanks!
BR,
Alvis.
Please log in to reply.




