@alex-z @abdo123 @ernestcarl OK I played around some more and was able to get HDMI etc. Long story short here is my SPL from my LCR and my two subs. This is post EQ correction from Anthem ARC but the uncorrected isn't terribly different. As you can see a sizeable null in the 65hz region that I can't get rid of. My LFE low pass filter is at 120hz and my front crossover is at 80hz. What I don't understand is that my subs seem fine? And when I move the LFE or crossovers to insane values like 40hz or 200hz or even turn some things off the null is still there (or widens). How can I have this null if the subs seem fine and my front speakers shouldn't be playing much around 65hz? I'm obviously missing something here...
View attachment 221003
There may be some phase difference induced cancellation from a not so optimal bass-managed sub + mains xo EQ and time alignment. It would be informative to see the response of the main on their own overlayed with the sub (non-LFE) response. The "time-of-flight" or time alignment between sub and main channels is also not shown in your graphs -- for this you need an IR/step or ETC plots.
You could also use REW's
alignment tool button found in the All SPL tab options menu (gear/sprocket icon) to check for the theoretical sum of your separate measurements...
----------
Alternatively, if you notice the LFE (+10 dB signal) is at odds with the bass-managed subwoofer signal itself -- which is to be subsequently summed with the mains (SPL leveled more in line with the mains so minus the +10 dB boost), it may be that the LFE signal's
non-low-passed response
electronically cancels out with the bass-managed sub signal's low-passed response. The latter can easily be fixed by the addition of an
all-pass filter or pure time delay applied to the LFE channel prior summing with the low-passed, bass-managed signal.
In the examples I have shown
here, for my DESK 7.1c setup, a
200 Hz all-pass filter is needed to be applied to the
LFE signal before electronically summing with the low-passed (-10 dB level and 200 Hz Butterworth 24dB/oct) bass-managed subwoofer signal. Without an all-pass filter applied, the LFE will partially cancel out with the bass-managed sub signal when coherent channel source signals are passed through the processor.
In the case of my COUCH 5.1c (actually more like 4.1 -- without the center channel) setup, I do not apply any additional low-pass filtering to the bass managed sub response at all rendering the use of an all-pass filter moot -- thus it is entirely skipped.
In the three Smaart magnitude and phase response graphs below, you can clearly see that the separate LFE and bass-managed subwoofer
responses (blue phase line) remains acoustically (and likely electronically as well I think) the same regardless of the difference in EQ applied. Depending on how the processor is designed, this kind of correction may or may not be performed.
image examples taken from:
Probably the simplest BM implementation looks like this:
Something a bit more "tailored" may look like this:
Where each speaker set has a different HPF / mixed phase filter xo EQ design...
However, what may be missing is a simple all-pass filter applied to the LFE prior summing:
I do my own rudimentary MCH bass-management processing digitally via JRiver on my HPTC
(edited for clarity)
JRiver's convolution section applies a partially corrected linear phase HPF so that 15 Hz filter for the sub/LFE channels is left disabled above.
Below is yet another set of example steps I performed using REW's basic EQ and Alignment tools:
Frequency dependent window (FDW) 15 cycles applied
Here's a little summary of what was done:
- Make sure the time of flight (or rather arrival as measured at the microphone) for all speakers is as close to "zero" or t=0 in the IR/ETC graphs
- Make sure to adjust the SPL level of all speakers -- if the response you are dealing with is god-awfully non-linear, you might want to apply some preliminary EQ correction beforehand as in the two examples I've shown here.
- Apply HPF if desired in order to match the phase responses better across the relevant xo region. Occasionally an all-pass filter may be called for certain time sensitive applications to match speaker phases from vastly different speaker designs/models. If the processor has the capacity, sometimes a linear phase or mixed phase xo EQ can be applied instead to avoid an increase in group delay using ordinary min-phase HP filters.
- Check the predicted or generated phase response if they align close enough
- Create an "aligned sum" response measurement
- Make changes to your processor and verify the new settings and EQ corrections work as predicted
Hmmn... only thing I skipped was the convolution section part which some will not find relevant.