Multi-Sub Optimizer (MSO): Lessons Learned, Tips & Tricks

[3ll3d00d]

If you want to try out methods of getting an even response and want to use MSO then I would configure the subs as a mono pair under the stereo 3 ways and do a sub only run in mso. I would use far fewer PEQs and also try including an all pass filter. There are multiple strategies you can use here which may sound different in your setup. You can then compare this to a stereo 4 way setup (unlikely to be using MSO) using prefilters (macro 0) which you design (or use something like rew to design) using measurements taken at multple positions Vs a pure acourate setup (single point measurement)

 

[Keith_W]

Thank you. I will try that.

 

[FrantzM]

Hi

I have used MSO and thought my results were good.. that was until two thing as happened:

1. I re-arranged my furniture's: I added a small table in the back of the room. The subwoofer that was in the back will be moved somewhere else.. in the back.
2. My miniDSP died, thanks to my using the wrong power supply.. Sending 51 VDC to it, courtesy of a similar looking power supply that was for a nearby PoE switch.

I am in favor of presenting an optimized subwoofer, call it the MSO-subwoofer to Audyssey. Let Audyssey do its things of integrating the other speakers (L,R,C, Surrounds, Ceiling Atmos) with this MSO-ed virtual sub. I don't intend to use MSO for mains and sub integration, only for subs and let Audyssey via the Windows app take care of the rest.

I will have to re-measure and ... I need to understand and make sure I do it right this time, because reading through the copious documentation, I might have done it all wrong and was simply lucky, so... I am exposing myself here as an utter ignoramus b/c:
I didn't perform the "Timing Reference" step.
I simply measured each subs at the MLP and luckily the levels were the same.

My questions now.
How do I establish the Timing Reference in my system? Is this even necessary?
What is the best way to match the gains/reference levels of the subs at the MLP positions, via REW, before MSO?
Do I need to match the levels of each subs at the MLP? What is the recommended method: Use the subwoofer gains potentiometer? Or use the miniDSP input gains? I personally prefer to use the subwoofer potentiometer to do so but , you know more, so I am asking.

Let me know and I apologize if this has been covered already. In which case point me toward it

Peace.

 

[kiwifi]

Hi
My questions now.
How do I establish the Timing Reference in my system? Is this even necessary?

You are in luck! I have just completed a re-MSO of my four subs today, so it is fresh in my mind.

Yes the timing reference is necessary. Feed the accoustic reference output from REW to one of your main speakers. If you have separate amps, then go directly into the amp. Otherwise use a direct input on you pre-amp, AUX or multi channel input, disable Audyssey for that input.

What is the best way to match the gains/reference levels of the subs at the MLP positions, via REW, before MSO?
Do I need to match the levels of each subs at the MLP? What is the recommended method: Use the subwoofer gains potentiometer? Or use the miniDSP input gains? I personally prefer to use the subwoofer potentiometer to do so but , you know more, so I am asking.

I use Audyssey to set the individual subwoofer gains to 75dB at the MLP. If you run XT32 with the subs turned off, you get the opportunity to set the sub output level.

I use a combination of attenuation in the DSP plus adjustment of the gain knob on the sub. I like to have -12dB of attenuation at each sub output in the DSP, before entering the MSO filters. The MSO filters will be a combination of gain, delays and PEQ filters.

If you use cut only filters, which you should, to maintain headroom, then you will likely need to add some output gain to restore your 75dB levels for each sub, once the MSO filters are installed. Hence the -12dB starting point.

 

[FrantzM]

You are in luck! I have just completed a re-MSO of my four subs today, so it is fresh in my mind.

Yes the timing reference is necessary. Feed the accoustic reference output from REW to one of your main speakers. If you have separate amps, then go directly into the amp. Otherwise use a direct input on you pre-amp, AUX or multi channel input, disable Audyssey for that input.


I use Audyssey to set the individual subwoofer gains to 75dB at the MLP. If you run XT32 with the subs turned off, you get the opportunity to set the sub output level.

I use a combination of attenuation in the DSP plus adjustment of the gain knob on the sub. I like to have -12dB of attenuation at each sub output in the DSP, before entering the MSO filters. The MSO filters will be a combination of gain, delays and PEQ filters.

If you use cut only filters, which you should, to maintain headroom, then you will likely need to add some output gain to restore your 75dB levels for each sub, once the MSO filters are installed. Hence the -12dB starting point.

Thanks FOr the quick reply. I apologize for being dense , but here are some of my questions:

Yes the timing reference is necessary.

Why? Is the acoustic timing reference important in the context of two subs, not Mains + SUB integration? Want to understand what I am doing, help me refine the process for me and remember it.

Feed the acoustic reference output from REW to one of your main speakers.

How? That is where I am lost: That "Timing Reference" is used how? I go to REW and send this Timing Reference to the AVR? How do I do that? How does that relates to the rest of the process? I send it, how it becomes a reference? If it is not henceforth tied to all my other measurements? What am I missing?

I use Audyssey to set the individual subwoofer gains to 75dB at the MLP.

That I get.. I can do that ... but

If you run XT32 with the subs turned off, you get the opportunity to set the sub output level.

Don't understand... Why would I run XT32 with the subs turned off? My process would have been to massage the subs via MSO. Obtain a super-smooth, super flat subwoofer in the room from 15 to 200 Hz, and present this "ultra sub" to Audyssey to integrate it with the mains? Again What am I missing?

I use a combination of attenuation in the DSP plus adjustment of the gain knob on the sub. I like to have -12dB of attenuation at each sub output in the DSP, before entering the MSO filters. The MSO filters will be a combination of gain, delays and PEQ filters.

That would mean what? -12 dB with respect to the max sub output or ??? I understand that the MSO filters are a combination of, in my case , mostly cuts and some positive gains, not much more than 3 dB in my case, since my subwoofers, so far are cheap and with a 150 watts. I will upgrade to better subs, for now they seem quite adequate for my small room, perhaps to my listening levels...

Sooo many questions. I thank you all in advance, because there will be likely more .

Peace.

 

[JStewart]

How? That is where I am lost: That "Timing Reference" is used how? I go to REW and send this Timing Reference to the AVR? How do I do that? How does that relates to the rest of the process? I send it, how it becomes a reference? If it is not henceforth tied to all my other measurements? What am I missing?

From REW help file:
Using a timing reference allows REW to eliminate the variable propagation delays within the computer and interface so that separate measurements have the same absolute timing.

Or put another way, using a timing reference overcomes the fact that your computer sometimes thinks it has more important things to do than your REW measurement

 

[DaveBoswell]

@FrantzM just select the "Use Accoustic Timing Reference" option on the "Measure" dialog box prior to measuring (see below). You will know it is on if you hear a little chirp at the begining and the end of the sweep. The timing reference should come out of the the L or the R speaker. What many do is temporarily re-wire the the Sub Input to the L Speaker Pre-out (if you have a pre-out), disconnect the L Speaker, assign the timing reference to the R speaker, temporarily set the the L and R speakers to Large, and then Measure using the L speaker instead of the LFE. Doing this provides a scenario where the subs are presented with a full range of frequencies with no cross over / LPF applied to the subs while do raw measurements. Once done measuring all the raw measurements, set everything back to the normal setup (and don't forget to re-set the L+R to Small!!)

 

[kiwifi]

Why? Is the acoustic timing reference important in the context of two subs, not Mains + SUB integration? Want to understand what I am doing, help me refine the process for me and remember it.

The acoustic reference is used to calculated the relative delays between subs

Don't understand... Why would I run XT32 with the subs turned off? My process would have been to massage the subs via MSO. Obtain a super-smooth, super flat subwoofer in the room from 15 to 200 Hz, and present this "ultra sub" to Audyssey to integrate it with the mains? Again What am I missing?

Turning off the subs but telling XT32 that you want to measure one, triggers the XT32 screen that allows you to setup the volume levels. Once you are in that screen you power up the subs one by one and set the levels.

That would mean what? -12 dB with respect to the max sub output or ??? I understand that the MSO filters are a combination of, in my case , mostly cuts and some positive gains, not much more than 3 dB in my case, since my subwoofers, so far are cheap and with a 150 watts. I will upgrade to better subs, for now they seem quite adequate for my small room, perhaps to my listening levels...

It is relative to the max input to the DSP without clipping, which is 0dB. By adding -12dB of digital attenuation to the DSP outputs before you start, you have 12dB of headroom to add your digital gains from MSO. Digital clipping can occur when a 0dB input, results in a greater that 0bB output.

 

[kiwifi]

Hopefully this will provide some inspiration for those still on their MSO journey...

Here are the measured results of my latest subwoofer optimization using MSO. For mesurements the REW test signal (from a Motu M4) is fed directly into the DSP (a pair of QSC DSP-4) and the output is measured at the MLP using a Behringer ECM8000.



I have four subs, as listed in my signature. They are randomly, rather than symetrically distributed around the room and are also at different heights, where possible. Despite this the "after MSO" result looks considerably better than the "before", and sounds better too!

The "after MSO" response is what Audyssey XT32 will see when the input to the DSP is connected to the Sub 1 out from the AVR.

FYI: In MSO I am using five shared (input) PEQ filters limited to +0 -12dB gain, plus 3 x PEQ per sub (output) limited to +0 -6dB gain. No stacking.
I have added a 160Hz Butterworth LPF at the input to the DSP and MSO attempts a fit to that curve across the full measured frequency (10 - 240Hz)

 

[zeismoebner]

@andyc56​

Thank you for your fantastic work and inspiration.
You helped me get a lot closer to audio nirvana. I learned that a tidy basement is very important for the entire spectrum and enjoyment of music.
How can we show our appreciation?
How can we at least buy you a coffee?