Multi-Sub Optimizer - EQ For Multiple Subs

I might have missed it but what's wrong with the result you get?

I mean MSO will vary to some degree over time, this is inherent to the way it works
 
I did a quick (5min?) run of your project and got

upload_2019-5-8_23-5-0.png


seems ok to me
 
I have been getting results like this:

3532c7c772.png

Is each of those curves the combined result of all subs plus main speakers, with each curve being at a different seat location?
 
Is each of those curves the combined result of all subs plus main speakers, with each curve being at a different seat location?

Not subs & mains just subs, at four different seating locations.
 
Last edited:
Not subs & mains just subs, at four different seating locations.

What crossover frequency are you using and how steep is the roll off? Just thinking your mains may be significant at the region you're worrying about the dip.
 
What crossover frequency are you using and how steep is the roll off? Just thinking your mains may be significant at the region you're worrying about the dip.

Probably will be using either a 80Hz or 100Hz crossover. Not sure what crossover my AVR uses to be honest.
 
Probably will be using either a 80Hz or 100Hz crossover. Not sure what crossover my AVR uses to be honest.

I've not done what you're trying but my gut feeling is you want to be making measurements of both the sub and mains with the crossover active, and running the optimisation with all speakers included. I'm very open to being corrected by others though.

Are you able to change your sub positions? If you are then you might be able to help flatten out your dip around 70 Hz if you could identify what room mode is causing it and placing a sub at a null of this mode.

Is any averaging applied to your curves?
 
I just had a quick nose at the MSO website and the example they show is for a combined optimisation of four subs plus main speakers so I don't think I was leading you astray above.
 
It is only possible if you have the ability to apply the required eq to all channels which is often not the case for someone using an avr.

A sub only config is still useful as you then present that combined result to whatever comes next. I remember commenting earlier though that this means you only really care deeply about the response to maybe 20Hz below the crossover as the overall response will change once you add the mains.
 
Running an AVR based optimisation including the mains would seem the logical next step.
 
It is only possible if you have the ability to apply the required eq to all channels which is often not the case for someone using an avr.

A sub only config is still useful as you then present that combined result to whatever comes next. I remember commenting earlier though that this means you only really care deeply about the response to maybe 20Hz below the crossover as the overall response will change once you add the mains.

Actually, I thought MSO let you run an optimisation that applied adjustments just to the subs but factoring in the contribution from the mains?
 
So after running M.S.O again I got more consistent results, no idea what caused the issues before.

555c13b630.png
 
You can do that too if you like

What would be the downside? I may well be missing something but it sounds like the better option to me.
 
What would be the downside? I may well be missing something but it sounds like the better option to me.
I'd have thought increased complexity is the downside, both in terms of measurements and the need to use the right filters for the crossover (information that isn't generally available so ideally you'd do a loopback measurement of processor). It's not generally necessary because you're aiming for a decent response from the subs alone and if you do get a dip like the above then you're quite probably going to fill it in anyway.

Bear in mind this is usually a multipass approach so you don't need something perfect from the first pass, you're just reducing the work to do in the remaining passes.

Ultimately MSO is just a toolbox really so you can use it in any way you see fit :)
 
I'd have thought increased complexity is the downside, both in terms of measurements and the need to use the right filters for the crossover (information that isn't generally available so ideally you'd do a loopback measurement of processor).

I'm not with you. Surely you just apply the crossover on the AVR as you would normally and then measure the mains and the subs. I don't see why you need to know what the filter response actually is.

Edit: or is it not possible to apply the crossover independently of built-in EQ on modern AVRs?
 
One of the variables you can control is that crossover frequency so you may get better results by letting it choose. Alternatively you can measure with some existing DSP in place and then add more on top. As I said, it's a toolbox so you can do things in many ways depending on your particular situation.
 
One of the variables you can control is that crossover frequency so you may get better results by letting it choose.

This was the factor I was missing - that an AVR might adjust the crossover as part of its own optimisation.
 
Was watching The Old Man & The Gun last night & the bass was very booming. It was only music but the bass seemed to take over everything else.

Not measured anything yet but I was using a +4dB boost with my target curve in Dirac. Got me thinking do you guys think I should not let Dirac EQ from 10Hz to 100Hz?

My thinking is I have already EQed these frequencies with M.S.O & no need for Dirac to do it again?
 
I've decided that I'm going to give MSO another go. I've tried it before but without a MiniDSP and I didn't get great results. My set-up has changed so I'm going to try again, but I need some help.

If this should be my own thread, can a mod please move it?

I'm running a 7.4 set up (7.3 at the moment, new sub arrives on Friday).
My source is a PC and I use Jriver as my player.
My processor has 5 band PEQ per channel and obviously level controls and distance per channel (2" increments).
My processor has three sub outputs, but I'm only using one, that goes in to the minidsp 2x4HD and then on to the subs.
I don't have any auto-EQ.
I only have a single listening position. It's a two seater dedicated room and there's very little difference between the two seats. Not enough for my wife to care!
Subs are all set to max crossover, zero delay, matched gain and the room size control on two is set as large.

Currently my crossover is set to 80Hz with a 24dB/Octave slope. On the processor I have the option to go up to 140Hz and I can choose a 12dB/Octave or 24dB/Octave slope.

I'm working in a new config in both the processor and the minidsp so everything's at 10' distance and 0 volume, no crossovers or EQ on the mDSP on either the inputs or the outputs. Stupidly I left the EQ enabled on the processor, but I can re-measure. Hopefully that won't matter to get answers to my questions.

Here are my MDAT, MSOP and MSOW files:

The (L) in the REW measurement name just means that I was using the left channel with the amp turned off instead of the LFE channel.

So my questions are:
- Should I be doing the subs alone first, or should I try and do both the EQ of the subs and the integration of the subs and mains at the same time?
- Should I be doing the subs to flat first? If flat, how do I do that? Is 140Hz with a 12dB/Oct good enough? I could hook the subs up to another channel's output to remove the crossover, is that wise?

If you look at my MSOP file what I did was to measure each of my LCR and each sub and create them as three measurement groups. My thinking was this would give me the best balance and integration with the individual main speakers. Is this a valid approach or is there a better way to do this?

I've read the MSOP manual multiple times but it just references "mains" with no detail about what that measurement should be.

Thanks in advance
 
Following on from the above, I started catching up with the MSO thread on AVS and I found this page: Optimization Strategies

This details what I think I want to do. It still just refers to “mains” though.

Can I use individual speaker measurements and group them as I’ve done (one main and all subs per measurement group) or should I be doing something different ?
 
something seems odd given the gigantic boost that 2 of the subs are getting >70Hz

you are also attempting to optimise for 3 different main channels at once which seems an unoptimisable target, see general considerations in Taking Measurements With Your Measurement Software for your options. For cinema use, you probably want to use C+ subs as your optimisation target & take measurements at the different positions.
 
I have both a length and width mode at that frequency. Both subs are the same distance from the listening position and the same distance from the side walls, so they're all summing at that point.

Having read more of the reference docs and the AVS thread, I don't think I have enough positions to warrant using MSO. Recommended best practice is more positions than channels to be filtered which, in my case, will be at least 5 positions. I just don't have that kind of room in my room, and having them too close is detrimental. Is that your understanding?
 

Attachments

  • RoomSim.jpg
    RoomSim.jpg
    129.6 KB · Views: 53
I haven't noticed that recommendation, I don't follow that myself and mso is useful to me :)
 

The latest video from AVForums

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