[fattire]

Multi Sub Optimizer (MSO) is free software written and published by @andyc56. As you may guess from the name, MSO helps us optimize for systems that have more than one subwoofer. What is unique about MSO, is that it does not use a single, global EQ for each sub. MSO individually and automatically optimizes the delay, gain, and EQ of each sub. This approach allows MSO to optimize the bass response for both the main listening position AND one or more secondary listening positions!

If you have a single listening position, or just a single sub, MSO is probably not the right tool for you to use. I would personally recommend the REW-only alignment/EQ method used by many folks. MSO is really targeted at >1 subs with >1 listening positions.

This tutorial was created with a pre-release version of MSO due a little later this year. It provides some huge usability improvements for new users to get up and running much faster. Andy has graciously published a preview version that can be used to follow along. The pre-release can be found here: download link. That link contains a stand-alone executable of MSO, a CHM help file, and a readme.html file. Windows 10 users will need to pay attention to the readme.html in order to fix a display issue with the CHM help file. This is preview software so it is NOT recommended that you replace your current MSO installation with this version just yet. I haven't found any issues, but just in case .

MSO v1.1.0 has been released. It is available for downloads on Andy's site here: MSO v1.1.0. This includes the new measurement import and configuration wizards as well as an all new and improved tutorial (also updated on his site).

Additionally, the tutorial is a sub-only optimization. MSO can help to optimize the subs and integrate with the mains or center. That's out-of-scope for this tutorial.

The intent of the tutorial is to educate, not sway opinions. I'm a very happy MSO user and just want to help put another tool in your home theater tool box to help you get better results.

EDIT: Before jumping in with MSO, you need to take the proper measurements to use for input. I've recorded tutorials on that process and the videos can be found here: Measurements for MSO

PS: This is literally my first YT video ever ... be gentle ;-)

MSO Homepage: Multiple Subwoofers: Optimize Them With Multi-Sub Optimizer Software
Original MSO Thread:Optimizing subwoofers and integration with mains: multi...

EDIT: I uploaded the individual segments I recorded to a playlist format in addition to the full-length video. Might make accessing things a bit easier: View full playlist

 

[fattire]

EDIT: MSO v1.1.0 -> v1.1.6 Update Summary Video
EDIT: MSO Results Comparison

 

[fattire]

RESERVED

 

[Falonious]

Thank you for taking the time to do this
Looks interesting.

 

[fattire]

Thank you for taking the time to do this
Looks interesting.

You're welcome! It's nice to give back a little to the community that has helped me so much. It's long, I know, but if you make it all the way through, please let me know what feedback you've got!

 

[Fxguy1]

The actual method to do this in first place is in Basstraps and location of Subwoofer. Here what you need is not just frequency domain correction but bass decay in the room. If the bass decay is not right you will not feel the body in bass and sounds very muddy. So two things should be in tandem to deal with this one is the locations of the subwoofers and another the basstraps as well. The last 2 or 3db of variation to be fixed in DSP if desired. You can observe InventionAudio Subwoofer location Identification they have done it well with FEA / CFD Analysis for the rooms with not just relatively flatter response in 1/6th octave but also with bass decay.

 

[markus767]

Multi Sub Optimizer (MSO) is free software written and published by @andyc56. As you may guess from the name, MSO helps us optimize for systems that have more than one subwoofer. What is unique about MSO, is that it does not use a single, global EQ for each sub. MSO individually and automatically optimizes the delay, gain, and EQ of each sub. This approach allows MSO to optimize the bass response for both the main listening position AND one or more secondary listening positions!

If you have a single listening position, or just a single sub, MSO is probably not the right tool for you to use. I would personally recommend the REW-only alignment/EQ method used by many folks. MSO is really targeted at >1 subs with >1 listening positions.

This tutorial was created with a pre-release version of MSO due a little later this year. It provides some huge usability improvements for new users to get up and running much faster. Andy has graciously published a preview version that can be used to follow along. The pre-release can be found here: download link. That link contains a stand-alone executable of MSO, a CHM help file, and a readme.html file. Windows 10 users will need to pay attention to the readme.html in order to fix a display issue with the CHM help file. This is preview software so it is NOT recommended that you replace your current MSO installation with this version just yet. I haven't found any issues, but just in case .

Additionally, the tutorial is a sub-only optimization. MSO can help to optimize the subs and integrate with the mains or center. That's out-of-scope for this tutorial.

The intent of the tutorial is to educate, not sway opinions. I'm a very happy MSO user and just want to help put another tool in your home theater tool box to help you get better results.

PS: This is literally my first YT video ever ... be gentle ;-)

MSO Homepage: Multiple Subwoofers: Optimize Them With Multi-Sub Optimizer Software
Original MSO Thread:Optimizing subwoofers and integration with mains: multi...

Great video and thanks for doing this. The video will finally give people a good indication how much of a learning curve MSO is for them.

 

[markus767]

The actual method to do this in first place is in Basstraps and location of Subwoofer. Here what you need is not just frequency domain correction but bass decay in the room. If the bass decay is not right you will not feel the body in bass and sounds very muddy. So two things should be in tandem to deal with this one is the locations of the subwoofers and another the basstraps as well. The last 2 or 3db of variation to be fixed in DSP if desired. You can observe InventionAudio Subwoofer location Identification they have done it well with FEA / CFD Analysis for the rooms with not just relatively flatter response in 1/6th octave but also with bass decay.

Although it is understandable that people would like to get simple answers to complex issues, I agree that the sole focus on a flat response can be misleading. Just because the magnitude response is flat doesn't mean that ringing is effectively reduced or worse, that additional audible ringing has been introduced. How does the response look like at other points within the listening area after MSO? What about excessive group delay?

 

[Fxguy1]

hat the sole focus on a flat response can be mislead

First challenge is to get the THX reference levels which is 115db for the LFE. Now getting that number in a room is relatively harder unless you are using number of subs and a point to note is that that number should be along with the basstraps.

Consider that if you dont have any midband absorbers in the room like having flat walls without any treatment and you can probably achieve 105db without much hardware coz you have reflections contributing to the overall SPL but getting the SPL after the basstraps are placed in the room then you are truely getting the right reverb values. Just like RT60 you have MT60 which is modal decay which also needs to be in between 0.4 to 0.5sec. Now when the basstraps are placed considerable amount of energy is being absorbed in the room just like how mid band absorbers absorb the midrange energy.

In that scenario you need to maximize the energy at the listening position now here is the point to consider that. When you are in dire need of more clean SPL you cannot make subs to fight each other like in out of phase alignments to suck out the energy to fight for normalized output. Using that process one is not maximizing the efficiency of the overall SPLs generated by the susb.

Look if you have 8 subs fighting each other you might get about 4 or 5 subs summation energy but what if the positions are chosen in such a way that the summation and cancellation not to be deliberately done at the subwoofer controls? You are maximizing the energy by minimum higher quality subs. Good deal isn`t it?

 

[markus767]

First challenge is to get the THX reference levels which is 115db for the LFE. Now getting that number in a room is relatively harder unless you are using number of subs and a point to note is that that number should be along with the basstraps.

Consider that if you dont have any midband absorbers in the room like having flat walls without any treatment and you can probably achieve 105db without much hardware coz you have reflections contributing to the overall SPL but getting the SPL after the basstraps are placed in the room then you are truely getting the right reverb values. Just like RT60 you have MT60 which is modal decay which also needs to be in between 0.4 to 0.5sec. Now when the basstraps are placed considerable amount of energy is being absorbed in the room just like how mid band absorbers absorb the midrange energy.

In that scenario you need to maximize the energy at the listening position now here is the point to consider that. When you are in dire need of more clean SPL you cannot make subs to fight each other like in out of phase alignments to suck out the energy to fight for normalized output. Using that process one is not maximizing the efficiency of the overall SPLs generated by the susb.

Look if you have 8 subs fighting each other you might get about 4 or 5 subs summation energy but what if the positions are chosen in such a way that the summation and cancellation not to be deliberately done at the subwoofer controls? You are maximizing the energy by minimum higher quality subs. Good deal isn`t it?

Well, you can't cheat physics and putting subs at the best locations is impractical in living rooms and home theaters.
The best solution is probably a DBA. It is unobtrusive, simple to integrate and gives better performance than any other solution. Half of the energy is "wasted" but that's the price of admission.

 

[Fxguy1]

Living rooms have lots of objects which anyway requires a full DSP method but for dedicated rooms yes its very much possible. We are not defying physics here but rather get best efficiency in the way the summation happens. Hence a new subwoofer is designed which is SUB312 from Invention Audio to tackle this problem where you can not just place the sub in the front but also on the side walls rear as well as the ceiling and also has the depth of the subwoofer is mere 8 inch with 12 inch drivers. HOME THEATER SUB WOOFERS – Invention Audio with these the placement of the subwoofers are breeze so you don`t have to worry about where the optimization system chooses the location in the room be it either any wall or can also be constrained that. If you have choices of only front and back wall so what are the best positions to get as much flatter response as possible.

 

[Fxguy1]

The above are before cases without proper localization.

these are with right localization you can see how well the nulls are pushed outside the listening area as if like seats are protected by arms.
also you can observe that there is symmetrical response on the seats on either left or right and as well as maximum SPL is concentrated in the listening area across all the 6 seats.

 

[markus767]

Living rooms have lots of objects which anyway requires a full DSP method but for dedicated rooms yes its very much possible. We are not defying physics here but rather get best efficiency in the way the summation happens. Hence a new subwoofer is designed which is SUB312 from Invention Audio to tackle this problem where you can not just place the sub in the front but also on the side walls rear as well as the ceiling and also has the depth of the subwoofer is mere 8 inch with 12 inch drivers. HOME THEATER SUB WOOFERS – Invention Audio with these the placement of the subwoofers are breeze so you don`t have to worry about where the optimization system chooses the location in the room be it either any wall or can also be constrained that. If you have choices of only front and back wall so what are the best positions to get as much flatter response as possible.

That's all good but this is a thread about MSO. Why not open a new thread?

 

[fattire]

The actual method to do this in first place is in Basstraps and location of Subwoofer. Here what you need is not just frequency domain correction but bass decay in the room. If the bass decay is not right you will not feel the body in bass and sounds very muddy. So two things should be in tandem to deal with this one is the locations of the subwoofers and another the basstraps as well. The last 2 or 3db of variation to be fixed in DSP if desired. You can observe InventionAudio Subwoofer location Identification they have done it well with FEA / CFD Analysis for the rooms with not just relatively flatter response in 1/6th octave but also with bass decay.

Bass optimization is like real estate. It's all about location, location, location ;-). I agree that folks must take the time to find the locations that work best in their rooms THEN work to optimize those as best they can. The vast majority of folks on this forum don't have access to CFD analysis for their rooms, but it's a fascinating topic that I personally would like to learn more about. If you're the person that ran the simulations you posted, would you mind starting a thread on the tooling and perhaps how to use it over in the "Audio, Theory & Setup" section? I definitely like to learn more, but I want to keep this thread focused on helping people learn how to use MSO.

Great video and thanks for doing this. The video will finally give people a good indication how much of a learning curve MSO is for them.

Thank you Markus, I appreciate it! If you have any feedback, please send it my way. I want to do a follow-up on taking REW measurements the correct way for use with MSO, a video on a subs+mains integration, and perhaps a shorter version of this video (I didn't intend for nearly 2 hours in length, but that's where it ended up). I'm 100% open to annotating this video or reshooting sections if things are incorrect or not clear.

The new, coming version I used is much easier than the current version. It imports the measurements, and sets up the initial config automatically. That was previously a bunch of manual work which, while not difficult, was definitely tedious and error prone in my experience.

 

[Fxguy1]

Very true guys will open up the thread. I have developed the algorithm. It costed a lot though and almost 5 years of research. Can explain what is possible with the solution in a new thread as its a technology worth to be known.

 

[DarinS]

Please post the link when you have it up and running. All of this is very interesting

 

[andyc56]

This is a great video by Jeff. I've linked it on the MSO home page, as well as linking to the shorter video that describes data entry into the miniDSP in that section of the tutorial.

I wanted to talk about the MSO Save BEQ miniDSP Biquad Text for this Channel command, the usage of shared filters in MSO, and how this relates to biquad merging with BEQ. Apparently one of the BEQ programs has a facility for combining BEQ biquads with biquads intended for use in the miniDSP input channel, created by non-BEQ programs. If you use shared sub filters with MSO, these would normally end up in the input filters of the miniDSP, where they would need to be merged with the BEQ filters. By "normally" here, I mean, "in the absence of using the Save BEQ miniDSP Biquad Text for this Channel command".

The entire purpose of the Save BEQ miniDSP Biquad Text for this Channel command is to avoid the need for such merging. It does this by exploiting an equivalence relationship involving shared (input) filters and output filters. Consider this block diagram for which there are 2 input filters (FL1 and FL2) and 2 filters per output channel (FL3-FL10). This situation is illustrated below.

Going from input to each output, we have:
in->out1: FL1->FL2->FL3->FL4
in->out2: FL1->FL2->FL5->FL6
in->out3: FL1->FL2->FL7->FL8
in->out4: FL1->FL2->FL9->FL10

Now suppose we replicate FL1 and FL2 into each output channel, then remove them from the input channel. We end up with the block diagram shown below.

This diagram is exactly equivalent to the diagram above. In terms of the signal flow, we have this:
in->out1: FL1->FL2->FL3->FL4
in->out2: FL1->FL2->FL5->FL6
in->out3: FL1->FL2->FL7->FL8
in->out4: FL1->FL2->FL9->FL10

which is exactly the same as the signal flow of the first diagram. These two systems are equivalent.

So, what happens when you invoke Save BEQ miniDSP Biquad Text for this Channel on, say, the output channel out1? It doesn't just save the two ouptut biquads FL3 and FL4. It saves four biquads for that channel by grabbing the shared biquads FL1 and FL2, along with FL3 and FL4 and including them in the output biquad text file. Doing this for the rest of the channels, you'll end up with four biquad text files, each with four biquads. In this case, saving the shared sub channel to a biquad text file would be unnecessary (and an error in fact), as FL1 and FL2 will have already been placed into each of the output biquad text files.

This entirely eliminates the need for any biquad merging, as MSO is merging the input biquads into the output ones already for you. If you need more than 10 total biquads per channel, you can enable the Use crossover biquads if output biquad limit exceeded option in the Application Options dialog, hardware page. The crossover can hold 8 biquads, so the combined input and per-output biquad count can be as high as 18. IOW, you could have 10 shared PEQs and 8 per-channel PEQs, or any other combination that adds up to a maximum of 18 PEQs (biquads).

There was a bug in the previous version where Save BEQ miniDSP Biquad Text for this Channel was disabled unless Use crossover biquads if output biquad limit exceeded was enabled. This has been fixed in 1.1.0.

 

[fattire]

This entirely eliminates the need for any biquad merging, as MSO is merging the input biquads into the output ones already for you. If you need more than 10 total biquads per channel, you can enable the Use crossover biquads if output biquad limit exceeded option in the Application Options dialog, hardware page. The crossover can hold 8 biquads, so the combined input and per-output biquad count can be as high as 18. IOW, you could have 10 shared PEQs and 8 per-channel PEQs, or any other combination that adds up to a maximum of 18 PEQs (biquads).

There was a bug in the previous version where Save BEQ miniDSP Biquad Text for this Channel was disabled unless Use crossover biquads if output biquad limit exceeded was enabled. This has been fixed in 1.1.0.

Thanks for the shout out and the link! I hope more people find the video useful.

This is an AWESOME feature for BEQ that I wasn't aware of. Or perhaps I was, but had forgotten about it. I actually have an "MSO w/ BEQ" video planned so I'll be sure to cover this. Again - as a user of both MSO and BEQ, this is super cool.

 

[DarinS]

So, this stuff is new to me and I am maybe 6 weeks out (hopefully) from my theater being complete. Quick question after watching about 10 minutes of the MSO video...

The guy presenting as 4 subs. I assume one in each corner, but I guess that doesn't really matter.

I will have dual PSA TV2112's. REW simulation puts the subs at wall midpoints front and rear as my smoothest response after trying many, many simulations and seat positions. In fact, I used that REW sim to actually determine the size of my riser and MLP position. That said, I hope that is the best place for my subs once I get everything in there! Will I need to go about trying a few different sub positions with REW first to determine my best position or will MSO be able to help me out with that also?

 

[fattire]

So, this stuff is new to me and I am maybe 6 weeks out (hopefully) from my theater being complete. Quick question after watching about 10 minutes of the MSO video...

The guy presenting as 4 subs. I assume one in each corner, but I guess that doesn't really matter.

I will have dual PSA TV2112's. REW simulation puts the subs at wall midpoints front and rear as my smoothest response after trying many, many simulations and seat positions. In fact, I used that REW sim to actually determine the size of my riser and MLP position. That said, I hope that is the best place for my subs once I get everything in there! Will I need to go about trying a few different sub positions with REW first to determine my best position or will MSO be able to help me out with that also?

I'm the guy in the video . For reference, my subs are not in the corners. The fronts are at approximately the 1/4 and 3/4 points of the front wall. The "rear" subs are on either side of our couch, directly inline with the listening positions. I don't have a purpose-built room. I used REW to find the locations with the best responses. Then did the integration with MSO.

If your theater is an enclosed rectangle, then the REW simulator is pretty darn accurate; even scary accurate in most cases. However, nothing beats actual in-room measurements. It's a "trust but verify" situation. I would not want to plan on specific locations from a simulation, only to find that the response is different due to room construction. If the wall construction is drastically different from wall-to-wall, room modes can shift positions or frequencies, even in a rectangular room. This is a common construction technique in basements. Two walls might be concrete block backed by earth while the others are timber construction. REW can accommodate for this if you enter the data, but again ... best to verify if you have the flexibility.

Assuming you DO have placement flexibility, I would want to do a sub-crawl, as you suggest, using REW to find the two positions that compliment one another. You are going to have to measure with REW regardless to get the input data for MSO. Might as well be 100% sure you've got the subs in the best spots while you're at it. Furniture sliders make this super easy.

Once you've verified the two best spots, use MSO to get the most even possible response across your listening positions from those sub placements.