Please advise, one virtual sub with mini dsp or use AVR dual independent sub

[Usher99]

I am brand new to REW and mini dsp HD 2X4 . I got my passive sub out of retirement and have mini dsp HD partially configured. The passive sub is dramatically less efficient than my powered sub and cranked up the gain a bit but not matched yet.
I am using an NAD T757 AVR but added Atmos speakers and in lieu of the upgrade boards am getting a new Anthem. I had no clue it would come with dual independent sub outputs. The subs have an 8 ft differential form the MLP so will clearly require time
alignment. The AVR--not shipped yet is supposed to take care of that with some timing pulse or something to that effect. Should I go ahead and create one virtual sub with mini dsp or wait for new AVR and see how it does? I suppose it is impossible to know how well the Anthem will do, but others have probably tried dual independent sub setups with other AVR's. My suspicion given how flexible the mini dsp is, it might be able to do a better job, but not clear if a newbie can easily achieve that. There seems to be good physics behind the adjustments but a touch of black magic only acquired by experience mixed in too. Appreciate any comments.

 

[Sonnie Parker]

I have independent sub outputs on my HTP-1... however, I choose to use REW and time-align my subs and then combine them all into one sub pre-out on the HTP-1, simply because I feel like I can do a better job of it than Dirac, as it does not give me as smooth of a response nor get the alignment right when I keep them separate.

You can time align your two subs with the miniDSP. However... I believe your issue is the subs won't have a smooth response when you crank them on up. The output between the two will change as you adjust the volume, unlike two subs that are identical. You are using the summed response for equalization. The volume at which you calibrate the subs will match, but as you vary the volume, the summed response will not be as smooth.

 

[Usher99]

Thanks so much for your comments. Anthem arc seems to enjoy a good reputation but I am leaning towards the one virtual sub. With the new furniture arrangement the second sub should help even out the response in the room even if imperfect. I dug out an old nad amp I had and bridged it for mono, so the whole second sub didn’t cost much. I hope I can figure out what house curve to use for the mini dsp.

 

[JStewart]

I hope I can figure out what house curve to use for the mini dsp.

You might try just phase aligning the subs with the MiniDSP and letting ARC handle the EQ and room curve.

 

[Usher99]

Hmm,
That is an interesting idea. Thank you for that suggestion. Perhaps I can post the curves so those more experienced can comment if I should try manually. I think the mini dsp has 10 parametric filters and REW can fit them to a curve automagically but I have never done it. Anthem would not reveal anything about arc when I asked.

 

[JStewart]

The idea behind the suggestion is to eq them as one sub after aligning. Same concept as for Dirac room correction (without add-on module) that Sonnie said he prefers.

Here’s a video of the sub alignment process using a MiniDSP:

After completing the alignment ARC can EQ the subs group as one sub.

 

[Usher99]

This is my very first ride on this rodeo and appreciate any input. I limited the plots from REW to three.

I first gain matched the passive sub with the powered sub--ultra cube 10 using a
55 Hz test tone. I just used my phone as SPL as didn't want to get out the
calibrated Mic yet. The passive sub is powered by
by an ancient NAD amplifier bridged for mono. I had to up the gain to +12 dB
which kept the passive sub a tad below the UC-10. I had no clue I happened to
pick a peak on this sub.

The powered sub has two 9" passive radiators to keep the box smaller than
adding a port.

The passive sub response curve is light brown curve, #1, the UC-10 #2 green and
the best i could do in time alignment --#10 blue with both subs going.

Curiously the passive sub is 8ft closer to the MLP (main listening position)
than the UC-10 one would predict about 8 msec & change delay for the passive sub would
be about right---it was actually 4.7 msec to get the best output!! The passive
sub seems to smooth the response at least when I move the mic to different
positions. I did make an attempt to construct a "House Curve" and REW enabled
them to be directly loaded into the mini-dsp with a click.. I had to enable REW (room eq wizard)
to use narrow filters and allow it to boost up to 6db to in theory ameliorate
the null at about 60hz. The predicted response looks decent to my
inexperienced eyes but my subs seem to be mediocre compared to those that make the video. If the 60ish Hz dip is a room null, might be like throwing a few
handfuls of dirt down a sinkhole, but REW predicted it is filled in. The CAO
(chief aesthetics officer) largely determined sub placement so am up the creek
on that solution.

I ran out of time to load the filters and play further as wanted the current
data to sink in and post here. I expected the passive sub to perform way
better.

Any suggestions? Should I have cut the UC-10 output in the gain match a
bit--already at 12dB boost on the minidsp for the passive guy? As suggested in the video I kept the UC-10 powered sub at mid point on the gain control on the back of it. The woofer in the passive sub is ancient---wonder if a more newly designed driver would help? The passive sub is ported and a guestimate of the interior volume is 11 X 14.5 X 21 inches. Must say the REW software is amazing and the mini dsp very nice.

Thanks for any comments.

 

[JStewart]

The passive sub is powered by
by an ancient NAD amplifier bridged for mono. I had to up the gain to +12 dB
which kept the passive sub a tad below the UC-10.

Looks like you made some good headway, but the 12dB gain seems odd. If I understand correctly that’s a 15 fold increase in the voltage output and therefore input to the amp. Typically AVR output voltage is 1 to 2V and a typical consumer amp with rca inputs typically requires 1 to 2V input to drive it to full power before clipping.
Are you hearing any obvious distortion with the passive sub? Could you play some music with low bass and listen to each sub separately to compare and see?
EDIT: can also compare distortion using the REW sweeps.
Not saying anything is definitely wrong, just seems like something may be. I’d at least check specs on the NAD amp for bridged mode and the specs for the passive sub if possible. Assuming you don’t already hear something off.

I had to enable REW (room eq wizard)
to use narrow filters and allow it to boost up to 6db to in theory ameliorate
the null at about 60hz.

Did that work?
Also 6dB is a further doubling of the power.

 

[Usher99]

Looks like you made some good headway, but the 12dB gain seems odd. If I understand correctly that’s a 15 fold increase in the voltage output and therefore input to the amp. Typically AVR output voltage is 1 to 2V and a typical consumer amp with rca inputs typically requires 1 to 2V input to drive it to full power before clipping.
Are you hearing any obvious distortion with the passive sub? Could you play some music with low bass and listen to each sub separately to compare and see?
EDIT: can also compare distortion using the REW sweeps.
Not saying anything is definitely wrong, just seems like something may be. I’d at least check specs on the NAD amp for bridged mode and the specs for the passive sub if possible. Assuming you don’t already hear something off.



Did that work?
Also 6dB is a further doubling of the power.

Very much appreciate the comments. Just coming up for air at work. I have not had a chance to load the filters to see what happened in a new sweep. I hopefully will have time this week.

There are no specs for the passive sub as it is a one-of. I did not notice an significant distortion with both subs going but that was before I tweaked the gain etc. My mains are offline now as being repaired so can't easily listen now but the CC is available. I have not ventured forth with the REW distortion tab yet, but I hope that is saved data with previous sweeps. If not I can re-sweep them and see. Appreciate any advice on how to do that as this is new to me.

The closest speaker repair person wanted no part of trying a new woofer for the passive sub. I only have a superficial knowledge of using T-S parameters and avoiding chuffing. The passive sub seem to be a slot ported design with the opening in front on top. It clearly isn't the best thing since sliced bread but adds a bit to some frequencies and evens out the seat to seat response some. I saw a thread in this forum that paring very dissimilar subs can lead to suboptimal results and sometimes is worse than using the best sub by itself. I can filter the passive sub as required with the minidsp if it starts to misbehave at some frequencies. I just want a bit of performance gain and better seat to seat consistency with limited downside effects by using both subs.

 

[JStewart]

but I hope that is saved data with previous sweeps

Yes, it'll be there.

 

[Usher99]

Clearly your suspicions were spot on. Thank you for thinking of looking at the distortion. The noise floor goes up a bunch at lower freq but the passive sub distortion skyrockets. I likely didn't hear it as the gain was 12 dB lower. I highlighted the THD expressed as percent. Not surprisingly the UC-10 power sub is better behaved. Perhaps I should put a brick wall filter at ?34Hz or so on the output of the passive sub?



Passive sub distortion:

Paradigm UC-10 below, THD highlighted with noise floor showing:

 

[JStewart]

I likely didn't hear it as the gain was 12 dB lower.

That and its supposedly very difficult to hear distortion at those frequencies.

Perhaps I should put a brick wall filter at ?34Hz or so on the output of the passive sub?

I really don't know what will happen. Can try it and see.

 

[Usher99]

It seemed to work well. THD in the combined output much better. Appreciate the suggestions. The THD did not get above 20% for the passive sub with 48 dB/octave high pass filter starting at 34 Hz on the passive sub only. I redid the filters and allowed it to use narrow filters and more boost if required though almost all the filters were cuts as I read is best.

The combined time aligned EQ response at MLP is highlighted with the raw sub outputs included--passive in red and Paradigm UC-10 in green. Decent? Looks OK from my novice perspective. I have to wait for my new AVR to arrive and my main speaker repair before try integrating the virtual sub. I almost regret finishing.

 

[Usher99]

I did not look at other room positions but have now. Both subs were optimized for MLP. I sometimes sit on the couch for movies but not music.

Here is the optimized SPL plot from the same settings as above but from the couch.

Certainly not as good. Is it worth trying MSO for something like this but restrict it to the subs? I can have multiple configurations saved in the minidsp. But then I need to have different presets for ARC if possible as the new single "virtual sub" will differ. Appreciate any comments and am thankful for the comments to date.

 

[JStewart]

Certainly not as good. Is it worth trying MSO for something like this but restrict it to the subs?

Some thoughts:

- MSO was originally designed for more than two subs. AndyC has since said it can work with 2 subs but YMMV.

- Do you have a need to get the best response for anyone other than yourself? If not the two configurations in the MiniDSP with ARC run on top may be your best bet vs
learning MSO. That said, if you're the kind of person that likes learning then go for it.

- The second response is not a bad response for music. Gives up some in the lowest frequencies where there's not much in most music to be had. An ARC calibration
for the "music" position using the same MiniDSP settings may also be worth a try.

 

[Usher99]

Thank you so much for the comments. I sometimes sit at the "mother-in-law" seat on the couch too. MSO just looks so elegant. It appears to to have trouble with a lesser "helper sub" unless the gain is turned up. I have a lot of reading to do before I try it if I go that way.

 

[Usher99]

I hope to try MSO this week for subs only. I can't quite figure out the timing reference. Can MSO optimize the delay for the closest sub with measurements I took with the previous technique as outlined in EP 7 of Hometheaterguru? From reading a bit more it seems a timing reference may be necessary and that is a step I have never done. I have my mains offline for repair but could use the ctr channel if required if I can figure out what to do. The help files for MSO state to clear out the Xover section of the mini DSP--I put in a 48 dB/oct high pass for my passive sub at 34 Hz and that cut the distortion way down. Can I leave that in? There is mention about gain block for the helper sub if using dissimilar subs. I have +12db as it is for the passive sub and -2db for the UC-10 to gain match at 50 Hz or so.

Working With Dissimilar Subwoofers (diy-audio-engineering.org)
There must be a setting in MSO for this.

I then have to figure out how to put in the house curve in a way MSO prefers. The one I used last time is below in green. Not as aggressive as Hometheaterguru but seems OK. Any help appreciated. I hope I can pull this off for config #2 in minidsp.

 

[JStewart]

Can MSO optimize the delay for the closest sub with measurements I took with the previous technique as outlined in EP 7 of Hometheaterguru? From reading a bit more it seems a timing reference may be necessary and that is a step I have never done. I have my mains offline for repair but could use the ctr channel if required if I can figure out what to do.

The Acoustic Timing reference is a high frequency chirp so Center will work. Sub will not.
Might be able to help. Would need to know what REW is connected to and how its connected though.

The help files for MSO state to clear out the Xover section of the mini DSP--I put in a 48 dB/oct high pass for my passive sub at 34 Hz and that cut the distortion way down. Can I leave that in?

Yes. And you can limit the range of correction that MSO will use for that sub as well.

There is mention about gain block for the helper sub if using dissimilar subs. I have +12db as it is for the passive sub and -2db for the UC-10 to gain match at 50 Hz or so.

MSO will allow or not allow a gain adjustment to find the best result. Each sub can also be restricted to any range of gain adjustment, like +/- 3dB, or -8 dB to 0dB, for example.

I then have to figure out how to put in the house curve in a way MSO prefers.

Piece of cake. It's just a text file with frequency and gain. Here's an example:

 

[Usher99]

The Acoustic Timing reference is a high frequency chirp so Center will work. Sub will not.
Might be able to help. Would need to know what REW is connected to and how its connected though.



Yes. And you can limit the range of correction that MSO will use for that sub as well.



MSO will allow or not allow a gain adjustment to find the best result. Each sub can also be restricted to any range of gain adjustment, like +/- 3dB, or -8 dB to 0dB, for example.




Piece of cake. It's just a text file with frequency and gain. Here's an example:

Thank you. Looks like freq then dB inc or dec. My house curve is in green. It looks like I should just say go to from 20 Hz to 120Hz or so every 10 Hz and put in the db plus or minus from say 75 dB?


<<<<<MSO will allow or not allow a gain adjustment to find the best result.

If I have the passive sub to gain match at +12 dB on the minidsp there isn't any room to increase if MSO wants that. I could put in more negative gain on the minidsp for the UC-10 to allow some room. I don't want to change the gain on the UC-10 on the sub (midpoint) if not absolutely necessary as would have to redo configuration 1.

<<<The Acoustic Timing reference is a high frequency chirp so Center will work. Sub will not.
<<<Might be able to help. Would need to know what REW is connected to and how its connected though.

Daunting but enjoyable for the non engineer. I was spoiled by the step by step youtube video for my first pass--still a learning curve to get good results but the procedure was crystal clear.

Nothing special on my connections AFAIK. I have UMIK-1 plugged into USB on my laptop with REW (win 10), usb to minidsp, HDMI from laptop to my NAD T757 (Anthem delivery delayed due to COVID). I used java drivers as in video) set mains to small and cross over at 200 HZ. I think I turned CTR channel off but I can turn on in a second. Timing measurement is new to me. I appreciate any assistance! I hope can figure out this amazing software.

 

[JStewart]

Using REW, can you select the center (or LFE channel for that matter) to send a signal to it? If not the options are:
Connect a speaker to either the right or left output on the AVR so it can be used as a timing reference.
—- OR —-
Install ASIO drivers on the laptop and use the ASIO drivers in REW to directly select any of the HDMI channels.

http://www.asio4all.org/

Once one or the other is done, or unnecessary, you will be able to select on the REW measurement screen which channel to send the sweep to and which to send the timing reference to. Whichever speaker gets the sweep should be physically disconnect so only the sub gets a signal. XOs are not brick wall. Depending on xo slope in the AVR even with an xo setting of 200 you will have output from the main at 100 and perhaps as little as 10dB down.

Regarding gains, in MSO you will disallow a gain adjustment to the larger sub, I.e. it will have no gain block. For the smaller sub keep your MiniDSP adjustment and in MSO set it’s gain block to allow a decrease but no increase.

 

[Usher99]

Thanks for the additional help. I can't select the channel with the java drivers. I'll have to jump into the ASIO pool and hope I don't drown. So it seems MSO requires a sweep (unclear if the same sweep technique used before is OK) for each sub at each position and the timing reference for ?each sub at each position. I'll have the UC-10 and the passive sub for 3 positions so the matrix is small. The timing reference should only change when changing positions though. I didn't see an example for subs only to emulate. My New Year's resolution is to figure this out.

 

[JStewart]

So it seems MSO requires a sweep (unclear if the same sweep technique used before is OK) for each sub at each position and the timing reference for ?each sub at each position.

Same measurement sweep, just with a timing reference activated on REW’s measurement screen. The channel chosen for the timing reference remains the same for all sweeps. One sweep for each sub at each measurement location i.e. seat. So 2 subs and 3 locations is 6 total sweeps.

 

[Usher99]

Thank you for clarifying. I misunderstood and was thinking the timing reference was a separate item.

 

[Usher99]

Asio4all drivers fail to see my NAD T757. So can't try MSO yet. I tried checking asio4all config to 16 bit and resample which is reported to help but no change. Output device list "not connected" only. I tried rebooting everything and letting display load fist, then AVR, then laptop but no joy. Java drivers see the NAD AVR fine but not sure I can use the time referencing from my center channel with them. I am trading in my AVR soon so perhaps I can retry with new one. I was hoping I would luck out and all would work the first time. Any advice appreciated.

 

[Usher99]

As a last ditch effort as have to send in my T757 in as a trade in I uninstalled the asio4all and tried flexasio. It did see the T57 so perhaps plays better with my laptop. I could get a test signal using check levels through my subs but it doesn't see the subs or ctr channel. I wanted to get the measurements at least for MSO but looks like I'm thwarted. I'll try again with new AVR but I was concerned REW might not like it as just released. Appreciate any advice.