MSO- Multi Sub Optimizer Beginners Thread

N

NorCalRP

Full Audioholic
Found out about this reading a app note on the mini DSP website. I searched here to see if there were any dedicated threads, to no avail. It looks really promising- and might make my purchase of Dirac Live a waste of money (I wonder if I could transfer the user license to someone else if I don't download it on my PC).

Edit: We've turned this into a thread for beginners. Once I have measurements I'll start posting progress pics
 
Last edited:
A

andy_c

Audioholic
Found out about this reading a nap note on the mini DSP website. I searched here to see if there were any dedicated threads, to no avail. It looks really promising- and might make my purchase of Dirac Live a waste of money (I wonder if I could transfer the user license to someone else if I don't download it on my PC).
There's hasn't been a dedicated thread for MSO on Audioholics, but the MSO homepage lists some online discussions at other forums. I won't link them directly here, as the forum mods might frown on it. But since you started this one, maybe it should be the Audioholics MSO thread. I'm happy to answer questions about it.
 
N

NorCalRP

Full Audioholic
There's hasn't been a dedicated thread for MSO on Audioholics, but the MSO homepage lists some online discussions at other forums. I won't link them directly here, as the forum mods might frown on it. But since you started this one, maybe it should be the Audioholics MSO thread. I'm happy to answer questions about it.
Thanks for jumping in. I did see that. I've also seen folks from here on some of those other forums. I'm honestly just wondering why there's so little here given all the bass heads. It looks like an awesome program.
 
A

andy_c

Audioholic
Thanks for jumping in. I did see that. I've also seen folks from here on some of those other forums. I'm honestly just wondering why there's so little here given all the bass heads. It looks like an awesome program.
That's an interesting point. Back in 2006, Todd Welti and Allan Devantier of Harman published a classic article in the AES Journal called "Low-Frequency Optimization Using Multiple Subwoofers". This introduced the esoteric subject of multiple-subwoofer optimization, for which they provided a solution called Sound Field Management, or SFM. The trouble is, to take advantage of that technology requires the purchase of some hardware whose cost is north of $20k. Then, you have to hire a Harman-certified technician to do the measurements and run their software to configure the hardware.

Others, including Earl Geddes, became interested in achieving similar goals to SFM at low cost, and after having seen a video by Earl, I was inspired to write MSO.

That's how MSO came about. I suspect the reason MSO hasn't seen wider interest is that people expect a solution to this problem that's "plug and play", wherein one does some measurements, provides the data to the program, and it figures out the best possible answer once and for all without any further input. If you want that, you have the option of paying Harman $20k for the required hardware, then more for their technician to set your system up. If you're willing to put forth some effort, you can achieve similar results at low cost. The bad news is that some experimentation is required to give the best results. The good news is that once you do the required measurements, all experiments are done within the software itself so that no further measurements are required.

So there's a learning curve. The more you're willing to climb it, the better the results you can achieve.
 
lovinthehd

lovinthehd

Audioholic Jedi
Thanks for jumping in. I did see that. I've also seen folks from here on some of those other forums. I'm honestly just wondering why there's so little here given all the bass heads. It looks like an awesome program.
Did all that elsewhere before coming here pretty much. Andyc is one of the guys who helped, too.
 
N

NorCalRP

Full Audioholic
Am I understanding correctly that MSO is the middle link of a REW->MSO->miniDSP chain? If so, would the Dirac Live upgrade I got before, but have not yet received, be rendered useless?
 
N

NorCalRP

Full Audioholic
Btw- I read the first portion of the tutorial day before yesterday, but yesterday was a real ass-kicking at work so I just vegged out when I got home. Working 12 hours a day, 7 days a week can get at you a bit from time to time
 
N

NorCalRP

Full Audioholic
Getting deeper into it and becoming more excited. @lovinthehd , is this the method you use to calibrate?
 
A

andy_c

Audioholic
Am I understanding correctly that MSO is the middle link of a REW->MSO->miniDSP chain? If so, would the Dirac Live upgrade I got before, but have not yet received, be rendered useless?
Yes. Here's the general approach.

1) In REW, measure the frequency response from each sub to each listening position using the REW acoustic timing reference feature. For N subs at M listening positions, that will be N * M measurements. Make sure all the filters in the miniDSP are cleared out, including disabling the crossover, and setting delays to 0 and attenuation to 0 dB before measuring. See the measurements page of the MSO manual for detailed instructions on doing the measurements.

2) Export each of the measurements from REW as a text file using File, Export, Export measurement as text in REW. Use a .FRD extension for ease of use (it's the default extension for MSO file import), and limit the upper frequency value of the exported files to no larger than 400 Hz so the files are not too large and import into MSO quickly.

3) Import the measurement text files into MSO.

4) Set up your project similarly to how it's done in the tutorial, but use only PEQ filters, gain blocks and delay blocks in the sub channels you define. Start with 4 PEQ filters per channel. Don't use any low-pass filters. The tutorial uses low-pass filters, but it was set up to use the Geddes approach, which is unusual in that it avoids using the AVR crossover. You'll be using the low-pass filters that are built into your AVR.

5) Try running an optimization.

6) If you aren't satisfied with the results of the optimization, try different things and/or ask here for help.

7) Once you are satisfied with your optimization results, for each subwoofer channel in MSO, export a biquad text file. For N independently-controlled subs, that will be N biquad files. Keep track of which one belongs with which sub (and miniDSP channel).

8) Import each appropriate biquad text file into the correct output filter channel of the miniDSP. This is described in the 2x4 HD manual in the "Parametric EQ file import (REW integration)" section. This will set up all the filters in each channel in one shot. It doesn't set up attenuation or delay though.

9) Manually set the individual miniDSP channel delay and attenuation as specified in MSO for each miniDSP output channel.

10) Set the AVR sub distance and trim value as specified in the MSO filter report for the configuration.

11) After you're done, it's a good idea to measure your system with all the subs energized at once, for at least one listening position. You can export these measurements as text, then import them into MSO using MSO's File, Import Plot-Only Measurements command. You can then plot these measurements on the same graph as MSO's predictions, allowing for easy comparison. The measured results should agree with MSO's prediction very closely. If the don't, then something is wrong and that problem must be chased down.
 
Last edited:
N

NorCalRP

Full Audioholic
Yes. Here's the general approach.

1) In REW, measure the frequency response from each sub to each listening position using the REW acoustic timing reference feature. For N subs at M listening positions, that will be N * M measurements. Make sure all the filters in the miniDSP are cleared out, including disabling the crossover, and setting delays to 0 and attenuation to 0 dB before measuring. See the measurements page of the MSO manual for detailed instructions on doing the measurements.

2) Export each of the measurements from REW as a text file using File, Export, Export measurement as text in REW. Use a .FRD extension for ease of use (it's the default extension for MSO file import), and limit the upper frequency value of the exported files to no larger than 400 Hz so the files are not too large and import into MSO quickly.

3) Import the measurement text files into MSO.

4) Set up your project similarly to how it's done in the tutorial, but use only PEQ filters, gain blocks and delay blocks in the sub channels you define. Start with 4 PEQ filters per channel. Don't use any low-pass filters. The tutorial uses low-pass filters, but it was set up to use the Geddes approach, which is unusual in that it avoids using the AVR crossover. You'll be using the low-pass filters that are built into your AVR.

5) Try running an optimization.

6) If you aren't satisfied with the results of the optimization, try different things and/or ask here for help.

7) Once you are satisfied with your optimization results, for each subwoofer channel in MSO, export a biquad text file. For N independently-controlled subs, that will be N biquad files. Keep track of which one belongs with which sub (and miniDSP channel).

8) Import each appropriate biquad text file into the correct output filter channel of the miniDSP. This is described in the 2x4 HD manual in the "Parametric EQ file import (REW integration)" section. This will set up all the filters in each channel in one shot. It doesn't set up attenuation or delay though.

9) Manually set the individual miniDSP channel delay and attenuation as specified in MSO for each miniDSP output channel.

10) Set the AVR sub distance and trim value as specified in the MSO filter report for the configuration.

11) After you're done, it's a good idea to measure your system with all the subs energized at once, for at least one listening position. You can export these measurements as text, then import them into MSO using MSO's File, Import Plot-Only Measurements command. You can then plot these measurements on the same graph as MSO's predictions, allowing for easy comparison. The measured results should agree with MSO's prediction very closely. If the don't, then something is wrong and that problem must be chased down.
@andy_c , I really appreciate you taking the time to respond to me.
 
A

andy_c

Audioholic
Btw- I read the first portion of the tutorial day before yesterday, but yesterday was a real ass-kicking at work so I just vegged out when I got home. Working 12 hours a day, 7 days a week can get at you a bit from time to time
Take your time. If you're unsure about anything at all, just ask. That way, anyone else just starting out who might be reading this could benefit from the Q&A.
 
A

andy_c

Audioholic
Generally, and it's been a long while and I need to make some changes so tagging along as a refresher (or learn something!) :)
I got to thinking this might be a good "MSO beginner's thread." The big MSO thread is cluttered with a lot of information, some of which is now obsolete because of changes that have been made to MSO over the years. I can link to this thread in the MSO documentation. I'm thinking I can, on a one-time basis, answer beginner questions with a level of detail that I couldn't in general do for everyone just for reasons of time - a kind of "ask me anything about MSO". Then, having done it once, the thread could be used as a referral for questions from new users.
 
N

NorCalRP

Full Audioholic
I got to thinking this might be a good "MSO beginner's thread." The big MSO thread is cluttered with a lot of information, some of which is now obsolete because of changes that have been made to MSO over the years. I can link to this thread in the MSO documentation. I'm thinking I can, on a one-time basis, answer beginner questions with a level of detail that I couldn't in general do for everyone just for reasons of time - a kind of "ask me anything about MSO". Then, having done it once, the thread could be used as a referral for questions from new users.
I saw that 7000+ reply thread and immediately backed out of it. I'm happy to be the guinea pig. I'm brand new to REW, MSO & miniDSP.
 
A

andy_c

Audioholic
I saw that 7000+ reply thread and immediately backed out of it. I'm happy to be the guinea pig. I'm brand new to REW, MSO & miniDSP.
If you're new to REW, it would be a good idea to download the latest beta, which means signing up at AV Nirvana. The latest one as of Sept 19, 2019 is 5.20 beta 24. This will ensure the REW UI looks the same to you as it does to me.
 
N

NorCalRP

Full Audioholic
If you're new to REW, it would be a good idea to download the latest beta, which means signing up at AV Nirvana. The latest one as of Sept 19, 2019 is 5.20 beta 24. This will ensure the REW UI looks the same to you as it does to me.
I will. I'm signed up there. I tried to download 23 ( I think), and had a bit of an issue, so I've been hitting the cancel button. I'll try tonight
 
N

NorCalRP

Full Audioholic
Should I run a new Audyssey setup before I start taking all my measurements? Also, should I use both sub pre-outs with 2 subs handled by input 1 and the other by input 2?
 
A

andy_c

Audioholic
Let me answer your second question first.

Also, should I use both sub pre-outs with 2 subs handled by input 1 and the other by input 2?
You should configure your AVR to the single-sub mode.

This affects the configuration of your miniDSP. Do you have the 2x4 HD? If so, you'll need to configure its input routing matrix. In the 2x4 HD manual, this is described in section 6.4.3, "Routing". Here's an image that shows the matrix setup for routing a single input to all four outputs. I don't have a miniDSP, but I think they renamed the inputs and outputs to better describe their actual function. See the manual for how to do this. Please let me know if my assumption that you have the 2x4 HD is right, and if not, which one you have.

routing.png


Should I run a new Audyssey setup before I start taking all my measurements?
When using MSO in situations where you'll apply room correction later, you should first disable all room correction, especially dynamic EQ, before doing any REW measurements. This means you'll be using an MSO "sub-only configuration", which means you won't be using MSO to integrate the satellites and subs - at least at this point. You'll only be measuring the subs. For this to work properly, you'll need to set things up so that when you measure the frequency response of the subs, there's a range of frequencies at the high end, say 100-200 Hz, over which the response of the subs is nominally flat, not rolling off. There are two ways you can do this.

1) Apply the input signal to L or R, and temporarily set the crossover frequency of your AVR to as high as it will go (say, 200-250 Hz)
2) Apply the input signal to LFE and temporarily set "LPF of LFE" (or whatever it is called in your AVR's terminology) to as high a frequency as it will go.

Whether you have all input channels available to you in REW depends on whether you're using any additional sound-related software (such as ASIO4ALL) in conjunction with your measurement. Are you using AustinJerry's REW measurement guide? If you set up your measurement system per his instructions, you can apply the input signal to any channel of the 7.1 collection. If not, you'll only be able to apply the input signal to the L or R channel in REW.

Also, make sure your AVR is configured for plain old 7.1 or 5.1 input - no fancy upmixing or other matrix processing.
 
N

NorCalRP

Full Audioholic
I do have the 2x4HD. The last time I ran Audyssey I had just two subs, and used both pre-outs to individually measure them- I'll have to wipe out all that data, because if I don't have my speaker configuration set to 2 subs I'm unable to turn on Audyssey. I currently am just using pre-out 1 (pre-out 2 is disconnected) on the AVR running into input 1 on the dsp, then routing to the 3 outputs.

I'm not using AJ's guide. I'm kind if was just looking to optimize the subs, but having read the tutorial, it seems I have to measure the mains as well?
 

Attachments

A

andy_c

Audioholic
Let me preface my comments by saying I'm a two-channel guy and have never used Audyssey, so I don't know all its ins and outs. But I'm aware of some of its properties.

I do have the 2x4HD. The last time I ran Audyssey I had just two subs, and used both pre-outs to individually measure them- I'll have to wipe out all that data, because if I don't have my speaker configuration set to 2 subs I'm unable to turn on Audyssey. I currently am just using pre-out 1 (pre-out 2 is disconnected) on the AVR running into input 1 on the dsp, then routing to the 3 outputs.
It appears that you're in a kind of limbo state with regard to Audyssey. You first ran it with two subs, run independently from the two AVR sub outs. They you added a miniDSP and a third sub, with all three subs running from the miniDSP outs, and one AVR sub out disconnected. Audyssey applies the same EQ to both sub outs (but it can use different distance settings for the two outputs). So you're running the system now with Audyssey sub correction determined using two subs, but you've added a third without changing any Audyssey sub correction.

To use MSO correctly, it's essential to start from a clean slate. This includes not only clearing out all filters, delays and attenuation from the miniDSP, but also clearing out any Audyssey sub correction (by temporarily turning it off.)

I'm not using AJ's guide. I'm kind if was just looking to optimize the subs, but having read the tutorial, it seems I have to measure the mains as well?
The tutorial was written before the "sub-only" mode was added to MSO. You only need to measure the satellites if you want to use MSO to optimize the integration between subs and satellites.
 
Last edited:
newsletter

  • RBHsound.com
  • BlueJeansCable.com
  • SVS Sound Subwoofers
  • Experience the Martin Logan Montis
Top