Go dual Subwoofers!

As long as you can get output from one sub where you have a null for another, you should be ok.
If one sub has a null at 50Hz but the other doesn't, you should be able to use the time alignment to fill the null with the response of the other sub. This is with the caveat that you might create nulls elsewhere. This is why, in this case, more generally does = better. You have more output so less of chance of all subs having a null at the same frequencies, when distributed around the room.

In REW's room sim, don't be disheartened if you add a second sub an it doesn't do much. Make sure you click the "align subs independently" checkbox, and then feel free to play with each sub's delay to see what difference it can make. I've often improved on room sim's dual sub prediction.
 
This is where my single set up sits just now using UMIK and Rew, which I appreciate isn't perfect, but its not bad.

1629471534221 (1).png



Add a second sub and REW simulator goes like this... I guess its the 55Hz null that concerns me, and is putting me off spending another £1k+ :)

1635694158297.png
 

Attachments

  • 1629459158627.png
    1629459158627.png
    484.5 KB · Views: 66
Can you screenshot the full room sim window please so I can replicate and have a play?

It's not a bad response though, I would probably want to correct that bump from 35 to 55Hz and then raise the overall level to match the mains.
 
Can you screenshot the full room sim window please so I can replicate and have a play?

It's not a bad response though, I would probably want to correct that bump from 35 to 55Hz and then raise the overall level to match the mains.
Apologies, I didn't get a chance to get back on the laptop, but this is the room...

Two rows, and my current sub sits behind the rear right as per the diagram :)

I have used the PB2000P in app to tweak the PEQ which gave the response I posted above.

1636036451921.png
 
Personally I think REW can be misleading when looking at SPL with multiple subs, whenever I ve added a sub it has unequivocally added to the system. For example I dont believe the SPL combinations with REW
 
Personally I think REW can be misleading when looking at SPL with multiple subs, whenever I ve added a sub it has unequivocally added to the system. For example I dont believe the SPL combinations with REW

Do you mean in room sim or with actual measurements?
 
Yes, sometime I can see the sim suggest I will get very little additional SPL and if I just look at the graph I would hardly notice losing one of the subs, but in the room the difference is stark and obvious. I could persuade myself on the basis of the graphs to use only 2 subs, but sonically the result is totally different.

I have not SPL checked losing one sub and see how many db I lose, so it may be more or at least party to do with response and impact of multiple drivers working in parallel. For me equalling SPL across the frequencies is part of the target but this doesn't tell me necessarily how it will sound.
 
Ok that’s fair, I don’t bother using it for levels. We know that if things are in phase we get 3dB and if they’re colocated we get 6dB. But understanding if any placement options work or which of a a set of placement combinations is likely to yield the best results is where I find it most valuable.
 
I guess my thoughts are thinking that adding a second sub won't be worthwhile based on REW room simulation might be misleading, as it only gives an indication of how they should combine with peaks and nulls. But we know that assuming you can adjust delay then you can get positive addition at the MLP for a certain frequency range and the combination with front drivers at the crossover can also usually be managed constructively so adding a second sub is nearly always worthwhile (perhaps even always assuming it for some reason doesn't exacerbate a problem ?). also the ability to getter a more even response over a larger ares increases too
 
Agreed, adding a second (or more) subs it very likely to have a positive influence. More point sources means a more even energising of the room which leads to smoother bass all around. We do need to remember that we‘re pretty fortunate and have dedicated rooms with very flexible placement. If that’s not the case then, even with delays, it might be the case for everyone and sometimes the available places don’t give a better response.

For example, both of my side subs have nulls at the same point, no amount of delay will resolve that. I’m lucky in that they’re not my only placement options (or my only subs!) so I can get over that. If they were my only possible locations though, adding a second wouldn’t be worthwhile.
 
Folks not sure if you saw my post in the Multi Sub Optimizer thread, but MSO is a revelation when compared to using REW alone for multiple subs.

MSO calculates the gain and delay required for "Each" sub and the filters required so that you get a flat response at multiple listening positions.

I've had dual subs for a long time and used REW for many years. After using MSO (free download etc) it's incredible what it has done.
 
Folks not sure if you saw my post in the Multi Sub Optimizer thread, but MSO is a revelation when compared to using REW alone for multiple subs.

MSO calculates the gain and delay required for "Each" sub and the filters required so that you get a flat response at multiple listening positions.

I've had dual subs for a long time and used REW for many years. After using MSO (free download etc) it's incredible what it has done.

I'm in the vast, vast minority here but I've never had MSO get me good results! The predicted looks pretty good, but the filters it uses to get there are insane, and I'm never able to apply them in a way that means I get the predicted output. Clearly I'm doing something wrong as others have no problems. I suspect it's in my target settings but I'm not sure.

I have a friend coming over this weekend who's used it successfully so I'll be having him show me how to do it. Once I can get it right I'm sure I'll be joining you in singing its praises.
 
I had a similar issue where the actual results weren't like the predicted. It turned out I wasn't using REW measurements with a Timed Reference signal. I had clicked the "loopback timing reference" but that wasn't correct. I had to do the measurements via the AVR (or PrePro in my case) with all processing turned off.

There is a learning curve with MSO but is quite simple if just doing dual subs and a few listening positions. The main thing to get your head around is that from your initial computation, you can create multiple iterations to help you find the best result. That involves adjusting the target db level up and down to see what it does. Sometime setting it at 80db instead of 75db gave better results, other times 68-70 gave good results.

Second thing was setting the maximum cut level for PEQ at a decent enough level to cope with your issues. I used 20db max cut, with a max limit of 21db.

Thirdly, possibly allowing for small amount of boost. I used 3db maximum, with a 3.1db max limit.

Finally I used the Optimise Gain and Optimise delay feature in my setup as well.
DD4659F7-2CB5-4815-8E26-493FD5C528DC.jpeg
 
Getting the results isn't a problem, it's the application of. I'm pretty sure I'm using a timing reference as all my REW measurements and time alignments there work. I'll know more after the weekend.
 
Oh I see. Do you have a miniDSP?.

I have the 2x4HD and had no issues uploading the filters and typing in the delay and gains into the relevant channels.

Worth noting that the HD version can handle 10 filters on the input and output, The other versions can handle fewer IIRC, and this has to be correctly set in the setup prior to running the calculations. Otherwise the filters won't load properly.
 
Oh I see. Do you have a miniDSP?.

I have the 2x4HD and had no issues uploading the filters and typing in the delay and gains into the relevant channels.

Worth noting that the HD version can handle 10 filters on the input and output, The other versions can handle fewer IIRC, and this has to be correctly set in the setup prior to running the calculations. Otherwise the filters won't load properly.

I have a 2x4HD but I’m probably doing it wrong there.
 
Well make sure 10 filters are selected from the default 4 in the setup wizard. Across a range of 18hz to 199hz I found 10 was better than 4 or 6.

At the end when you click Optimise Gains it will put a value into what's called "Shared Filters". That shared Gain value has to be manually typed in as well, but in the Input section of MiniDSP. The usual gain\delays go in the Output section per subwoofer.

Look forward to hear how you get on.
 
I had very good results with MSO and two subs, doing a subs only optimisation. I've never got a subs and mains run to produce anything that wasn't crazy.

I'm picking up a third sub this weekend so I'll have another go with MSO when that's installed.

My situation is the rear of room 1/4 location is pretty even apart from a huge 20Hz wide null. So subs there on left or right or both is good with minimal EQ but drops off a cliff at 25Hz. If I put one sub anywhere between 1/2 and 1/4 of the front of the room, it's great from DC to ~40Hz but a hopeless mess between 50 and 70Hz. If I time front and back (REW alignment tool, thanks Conrad) I can get a good looking result in REW between 18Hz and crossover but the sweep sounds nasty with phasing effects as the subs interact. This sounds ok in a movie though.

The good MSO run I got a year or so ago, seemed to greatly reduce the phasing effects during a sweep and sounded much better in movies too. I'm not really sure how it did that but maybe the different EQ for the two subs removed the bad interactions and kept the good ones.
 
Ive tended to use the REW alignment tool and see what it recommends at a chosen frequency then have a look with the slider as to what's going on. I have also found on several occasions 'just' going for the best summation at the chosen frequency doesnt necessarily sound the best. I often settle on less delay than the analysis suggests with a worse curve (living with imprefection in the graph) but having a better sonic result. Easy to experiment.

I also try very hard to get as close as possible across the whole seating area and manage to get within 2 db to 35 Hz and 3-6db to 70 Hz across a 2m seating area

Ive never tried MSO, I'll be interested if that might work with a 10x10 HD and 4 subs ?
 
Ive tended to use the REW alignment tool and see what it recommends at a chosen frequency then have a look with the slider as to what's going on. I have also found on several occasions 'just' going for the best summation at the chosen frequency doesnt necessarily sound the best.
Not just me then, interesting.

I often settle on less delay than the analysis suggests with a worse curve (living with imprefection in the graph) but having a better sonic result. Easy to experiment.
I've ended using a much worse looking curve that sounds better. I think the usual technique of phase aligning at crossover (or just tweaking for the nicest trace) often results in significant delays which I find objectionable in some material. Greatest Showman was unwatchable with my last session of trying subwoofers in different locations, far too much advance/delay in the subs channel.

I think what's needed is time and phase alignment around crossover and therefore localisable frequencies and different phase alignment below that, probably PEQ as well. I suspect that is what Dirac's DLBC is all about but I don't have the hardware to find out.

I plan to try an allpass filter optimisation with MSO, it's a newish thing, poor mans DLBC, I think that's what that's all about. I glazed over in the maths for that stuff at college but I think it does something like that.
Ive never tried MSO, I'll be interested if that might work with a 10x10 HD and 4 subs ?
There was someone who had a huge project with >10 subs, the author AndyC asked for a copy of the project as he just wanted to see it for himself, as I understand, it all worked fine.
 
Not just me then, interesting.


I've ended using a much worse looking curve that sounds better. I think the usual technique of phase aligning at crossover (or just tweaking for the nicest trace) often results in significant delays which I find objectionable in some material. Greatest Showman was unwatchable with my last session of trying subwoofers in different locations, far too much advance/delay in the subs channel.

I think what's needed is time and phase alignment around crossover and therefore localisable frequencies and different phase alignment below that, probably PEQ as well. I suspect that is what Dirac's DLBC is all about but I don't have the hardware to find out.

I plan to try an allpass filter optimisation with MSO, it's a newish thing, poor mans DLBC, I think that's what that's all about. I glazed over in the maths for that stuff at college but I think it does something like that.

There was someone who had a huge project with >10 subs, the author AndyC asked for a copy of the project as he just wanted to see it for himself, as I understand, it all worked fine.

Not to derail the thread but another point about multiple subs is not having the subs rolling off and the mains rolling in. I have all speakers full range and add in via the DSP any missing frequencies from the sub group so I'm 8in 4 out at the moment. The advantage is the impact of additional weight across the bass, my fronts, centre and surrounds. They manage anything from 30-40Hz+ very well so seems silly to turn that off !
 
Well make sure 10 filters are selected from the default 4 in the setup wizard. Across a range of 18hz to 199hz I found 10 was better than 4 or 6.

At the end when you click Optimise Gains it will put a value into what's called "Shared Filters". That shared Gain value has to be manually typed in as well, but in the Input section of MiniDSP. The usual gain\delays go in the Output section per subwoofer.

Look forward to hear how you get on.
Just coming back to this, a friend came over and I went through a project with him. We were able to apply the results and get exactly the predicted measurement, so that progress.

The difference is the way I was doing the export for the subs, I think. Previously I'd been copy/pasting the biquads, but this time I did an export for each channel and loaded them in. It's still doing weird things with the levels (it's basically muting two of my subs below about 20Hz with stacked -12dB cuts) but I think that's a levels thing.

At least it's working for me now. Now I can move my subs about, take more measurements and see how it goes.
 
Great to hear you made progress with MSO.

You can prevent stacked filter cuts by setting the Max Cut to your max cut +1db. Same for max gain except you set that to max gain in dB +0.1db.

It’s worth trying different positions of subs. I tried a few but they caused a big null that not even MSO could resolve. So I picked the positions that allowed the software to result in no nulls.
 

The latest video from AVForums

TV Buying Guide - Which TV Is Best For You?
Subscribe to our YouTube channel
Back
Top Bottom