Originally Posted by steakhouse_
I have one little question. I have 3 subs (two on the front and one left rear behind my sofa) and did MSO quite sucessfully. Their summed response (per MSO and measured by REW) is nearly perfect. However, I have one problem. I cannot get them to blend with my mains no matter what delay setting I use. The difference in phase is just to large and fuzzy within the x-over region.
The results below were actually the best I could achieve after trying several delays. I am using a DDRC88-BM with a 5.1 setup where 3 channels are used for the subs. I have to mention that the sub left behind my sofa is a different one with a different roll of at the higher frequencies. The subs in frnt of me are a bit older with a non defeatable LPF ant 125Hz, the sub behind me is an Arendal Sub 1v. MSO did actually get a quite good result anyway but could this be a potential problem with integrating them with the mains? If so I would exchang my older subs with another 2 Arendal Sub 1v.
I'm not sure what you've tried so far with MSO, but is it fair to say you began with a sub-only configuration, measured with the DDRC88-BM crossover disabled, and were able to optimize it successfully, then went with a subs+mains configuration and are now having trouble? If this is so, then you've probably added a shared delay block for the sub channels. One thing you may need to do is adjust the limits of this shared delay block to allow negative delays. For typical AVRs, a negative shared sub delay in MSO corresponds to increasing the sub distance relative to what it was during the measurement. I'm not familiar with the details of the DDRC88-BM, but with some hardware that doesn't fit the pattern I just described, it means that if a negative sub delay is required, say -10 msec, then you would actually have zero shared sub delay, and add 10 msec delay to all your satellites. AVRs do something like this internally when altering the sub distance.
Another thing to consider is if your hardware provides both 2nd-order and 4th-order Linkwitz-Riley crossover options, which I think the DDRC88-BM does. The 2nd-order LR crossovers require the sub channels to be inverted relative to the satellites, while the 4th-order ones do not (assuming in-phase behavior of satellites and subs in the absence of the LR filters). I'd recommend the 4th-order ones. But even still, a shared inverter on the subs might be an option. "Clone Configuration" in MSO can be helpful in such situations. Cloning duplicates an existing configuration along with its graphs., and allows you to preserve the various options you've tried out (e.g. subs inverted and not).
In any case, after doing the sub-only configuration, you'll want to lock all sub filter parameters so MSO doesn't have to start from scratch in the new subs+mains configuration. You can use your originally measured data taken with the crossover disabled, then add the crossover LPF and HPF in MSO with cutoff frequencies all locked to the crossover frequency you want to use. You don't have to measure again with the crossover filters in place (Edit: except to verify the results after optimization).