Using Multi Subs & MSO To Get Even Bass In Room & Across Sitting Positions

[Snoopy8]

27 minutes ago, Satanica said:

Righto, is that both subs or just the L (Left)?

Is MSO choosing or was that you?

 

So at 130Hz at 24dB\Octave the subwoofers will already be 6b down.

I don't have something to simulate with in front of me but I don't think you'll have much EQ potential above 125Hz and of course won't be able to EQ the Left and Right channels independently.

 

At the moment I think you'll have little to no EQ applied above about 125Hz and will have two mono subwoofers that are being crossed over at frequencies far higher than is typical.
You might not have noticeable localisation issues if your two subwoofers are somewhat symmetrically laid out as that would place those frequencies centre between them anyway rather than to one side, I think.

Have LPF for both subs, and letting MSO choose for me.  Often, but not always, MSO maxes out.  I could of course play with different LPF values and see what MSO does.  Have no issues bove 125 Hz.  My left sub is not on the speaker plane, skewed left.  Because of that, I pay close attention to it.

 

31 minutes ago, Satanica said:

I have to mention again I think some of the KII THREE cardioid will be overridden by the subwoofers working higher than 50Hz.

From memory @gibbo9000 has his subwoofer low passed and hi KII THREE's high passed at 50Hz.

My next project is looking at doing this.  I cannot hear any impact MSO with subs, have on sound stage, details. I can instantly mute, unmute the subs through my music player and have not heard any differences.  Does not mean that cardiod pattern has not been affected, just not audible.

[Satanica]

1 minute ago, Snoopy8 said:

My next project is looking at doing this.  I cannot hear any impact MSO with subs, have on sound stage, details. I can instantly mute, unmute the subs through my music player and have not heard any differences.  Does not mean that cardiod pattern has not been affected, just not audible.

 

For sure, its very easy to overstate how audible if at all artifacts can be.

 

Anyway I think this is a good discussion.

I guess I'm just of the mind that the improvement in my system might be subtle at best; I don't know.

If I were to use it it would be round 1 of correction on my two subwoofers working range (<80Hz) before running Dirac Live over the entire frequency range and from what I gather that would get a tick of approval.

[davewantsmoore]

5 hours ago, Satanica said:

Lets say it is 80Hz and if so then how is EQ being applied all the way up 200Hz without EQ applied to your mains?

 

Send full range signal to subwoofer (measure subwoofer response)

Correct subwoofer(s) response to be "flat" up to 200Hz (or whatever Hz you like)

Add lowpass filter at 80Hz (or whatever Hz you like).

Add complimentary high-pass filter to mains (maybe, optional, but let's keep it simple)

[davewantsmoore]

3 hours ago, Satanica said:

Anyway I think this is a good discussion.

I guess I'm just of the mind that the improvement in my system might be subtle at best; I don't know.

If I were to use it it would be round 1 of correction on my two subwoofers working range (<80Hz) before running Dirac Live over the entire frequency range and from what I gather that would get a tick of approval.

 

1. Position subwoofers so their un-EQed combined response is flattest
2. EQ subwoofers (individually) so their combined response is flattest (MSO can help optimise this) .... You should do this at least to an octave above their intended cutoff.
3. Level matching with mains / Add LPF to subwoofer
4. Should now have a "flat" response for subs + mains, including through XO region
5. EQ combined mains/subs, to both correct errors, and to "taste" .... something like Dirac is great.

 

In step 1 and 2... if there is to be significant overlap with subs and mains...... then the "combined response" should be of all bass sources, ie. subs and mains.

[gibbo9000]

5 hours ago, Satanica said:

From memory @gibbo9000 has his subwoofer low passed and hi KII THREE's high passed at 50Hz.

Thats pretty close.  After reading through as much as I could find on the Kii's I took the approach (right or wrong) of trying to protect the Kii cardiod distribution while shifting some of the low frequency load from Kii to Sub to free up the Kii's bottom end SPL limit.  Free from the @Snoopy8 constraint of protecting DSD, it was easy to use the miniDSP for preset x-overs at 30, 50 and 80 Hz using LPF for sub and matching HPF for Kii's.  For normal listening I use 30Hz with a Dirac curve that just irons out the bottom end bumps (below 200Hz).  For a louder, heavier bass fix I use 50Hz with 2db boost below 200Hz, or 80Hz with 4db!  And (as is now obvious to me!), if I use the 30, 50 or 80Hz x-overs with my standard Dirac target curve they all sound the same - which is Dirac doing its job!

 

I did play a little using putting the HPF on Kii's themselves and LPF on Sub as @Snoopy8 is contemplating, but found it harder to set a 'zero gain' x-over and ultimately took the convenient path of using miniDSP. 

[Snoopy8]

1 hour ago, gibbo9000 said:

Thats pretty close.  After reading through as much as I could find on the Kii's I took the approach (right or wrong) of trying to protect the Kii cardiod distribution while shifting some of the low frequency load from Kii to Sub to free up the Kii's bottom end SPL limit.  Free from the @Snoopy8 constraint of protecting DSD, it was easy to use the miniDSP for preset x-overs at 30, 50 and 80 Hz using LPF for sub and matching HPF for Kii's.  For normal listening I use 30Hz with a Dirac curve that just irons out the bottom end bumps (below 200Hz).  For a louder, heavier bass fix I use 50Hz with 2db boost below 200Hz, or 80Hz with 4db!  And (as is now obvious to me!), if I use the 30, 50 or 80Hz x-overs with my standard Dirac target curve they all sound the same - which is Dirac doing its job!

Thank you.  Your SHD adventure was one the reasons why I started my own bass integration journey.  That particular DSD requirement caused much angst and I was tempted a number of time to go for a SHD Studio or maybe a Anti-Mode X4, both of which would have been much easier to implement.  My final solution looked relatively simple now, but it required finding new ways to do things.

 

Dirac delivers excellent results, without requiring detailed DSP knowledge.  That is its strength. 

 

1 hour ago, gibbo9000 said:

I did play a little using putting the HPF on Kii's themselves and LPF on Sub as @Snoopy8 is contemplating, but found it harder to set a 'zero gain' x-over and ultimately took the convenient path of using miniDSP. 

Was thinking of a combination of lowering the boundary setting and apply the tone control to the Kii or apply the LP filter.  I will let MSO determine the XO for the subs. 

[WasM]

@Snoopy8 Is your target setup for stereo?

 

Have you tried the multiple satellite measurement optimisation?

[Snoopy8]

5 minutes ago, WasM said:

@Snoopy8 Is your target setup for stereo?

 

Have you tried the multiple satellite measurement optimisation?

I use it for stereo with dual subs. Also using dual subs in MSO subs only config for HT.

 

Sorry, have not done multi satellites, was developed in the past 12 months when I was away from MSO.

[Satanica]

14 hours ago, Snoopy8 said:

Dirac delivers excellent results, without requiring detailed DSP knowledge.

 

I don't really agree. Unless I'm mistaken you've only used it in a surround setup.

 

In a 2ch setup with the miniDSP SHD there is no subwoofer object(s) but rather just two speakers so you really have to do the following:

 

1. Decide crossover points and slopes of subwoofers and mains and set in SHD.

2. Decide gain values and set in SHD (probably with REW).

3. Potentially time align subwoofers to mains (probably with REW) and in @gibbo9000's case a seperate miniDSP unit is needed to provide the appropriate "time wastage" for the KII THREE's (90 milliseconds?).

https://www.minidsp.com/applications/acoustic-measurements/292-rew-timing-ht

Set in SHD.

 

All of the above are done independetly before even using Dirac.

 

4. Decide polarity of subwoofers and set in SHD based on intitial measurements (probably with Dirac).

5. Use Dirac to measure one or more positions.

6. Experiment with Dirac room curves for best subjective results.

[Snoopy8]

3 hours ago, Satanica said:

 

I don't really agree. Unless I'm mistaken you've only used it in a surround setup.

 

In a 2ch setup with the miniDSP SHD there is no subwoofer object(s) but rather just two speakers so you really have to do the following:

 

1. Decide crossover points and slopes of subwoofers and mains and set in SHD.

2. Decide gain values and set in SHD (probably with REW).

3. Potentially time align subwoofers to mains (probably with REW) and in @gibbo9000's case a seperate miniDSP unit is needed to provide the appropriate "time wastage" for the KII THREE's (90 milliseconds?).

https://www.minidsp.com/applications/acoustic-measurements/292-rew-timing-ht

Set in SHD.

 

All of the above are done independetly before even using Dirac.

 

4. Decide polarity of subwoofers and set in SHD based on intitial measurements (probably with Dirac).

5. Use Dirac to measure one or more positions.

6. Experiment with Dirac room curves for best subjective results.

Your process is complicated because you chose to do it with the SHD in all its complexity.

 

For most people, it is a matter of setting the box, connect the speakers & subs and then follow the Dirac process and it is done. Setting up Dirac with an AVR for HT or in the NAD C658 for stereo is no different. Yes, you can tweak the target curves, but that is optional and does not need understanding of DSP.

[Satanica]

2 minutes ago, Snoopy8 said:

Your process is complicated because you chose to do it with the SHD in all its complexity.

 

For most people, it is a matter of setting the box, connect the speakers & subs and then follow the Dirac process and it is done. Setting up Dirac with an AVR for HT or in the NAD C658 for stereo is no different. Yes, you can tweak the target curves, but that is optional and does not need understanding of DSP.

 

I've simply done the miniDSP way virtually to the letter as per the documentation. I don't consider it trivial. 

[Snoopy8]

1 minute ago, Satanica said:

I've simply done the miniDSP way virtually to the letter as per the documentation. I don't consider it trivial. 

The miniDP SHD was designed for people like you who need and desire the functionality and know some DSP. Thus their processes reflect that and I agree it is not trivial.

 

But the SHD can also be used simply by connecting the speakers & subs and running Dirac.  It is in one way, a waste of the SHD capabilities, but the results may be enough to satisfy many people....

[Snoopy8]

@Satanica, @gibbo9000 , have some new information.

 

Added a high pass filter to the Kii at 80 Hz, and remeasured everything and re-ran MSO.  However, the Kii with high pass filter + Subs affected the SQ.  The high pass filter reduced the dynamics, sound stage.  I may experiment with other filters at different cutoff frequency but right now, this particular approach is not looking good.

[Satanica]

16 minutes ago, Snoopy8 said:

@Satanica, @gibbo9000 , have some new information.

 

Added a high pass filter to the Kii at 80 Hz, and remeasured everything and re-ran MSO.  However, the Kii with high pass filter + Subs affected the SQ.  The high pass filter reduced the dynamics, sound stage.  I may experiment with other filters at different cutoff frequency but right now, this particular approach is not looking good.

 

Interesting but I'm not that surprised by this. If memory serves me right you have a 13 inch and a 12 inch SVS sealed subwoofer so your overall sound could have been benefitting from the Kii's lending a hand down low. 

 

My first thoughts of experimentation was for you to set an 80Hz low pass filter for your subs most importantly and perhaps a high pass to the Kii's at 50Hz. But there are so many combinations it is mind boggling. This got me thinking does anyone know what the BXT module addition does presumably with a settings change or upload to lighten the load to the main modules? I'm quite interested to know and whatever the change perhaps this could be somewhat replicated with subwoofers. 

 

My system comprises a couple of large SVS ported subwoofers and I don't feel like they need the extra effort from my mains below 80Hz but I do wonder if they could help as extra low bass sources ala Geddes style <shrug>. 

[gibbo9000]

3 hours ago, Snoopy8 said:

The high pass filter reduced the dynamics, sound stage.

Interesting - not sure I have seen, but where are your subs positioned?  Along with the Kii's facing forward, or somewhere else?  Kii noted in one of the discussions that they prefer x-over at 30Hz and with subs facing forward alongside Kii's so they notionally contribute to the Cardiod distribution and don;t drive refelections off rear/side walls etc.  (not sure that stands up in practice - but is what they suggested)

 

3 hours ago, Snoopy8 said:

The miniDP SHD was designed for people like you who need and desire the functionality and know some DSP.

At risk of getting off topic, the SHD Studio is, as picked up here, multiple devices in one.  Powerful DSP, Digital Preamp and Dirac Live processor with overlay of 4 pairs of DSP Presets matched to Dirac optimisations.  Can be used in 'simple' mode with Dirac Standard curve alone and will be fine for most (but there are exceptions - and I was one - turned the Kii's overly bright).  The further you dig the more flexibility / options, and the greater measurement and setting management discipline required.

 

But this thread is about MSO and what it can add.  Interested to hear how the experimenting goes between the Kii's full bandwidth and Subs just filling gaps, versus HPF for Kii's and sub doing work below that.  

 

 

 

[Satanica]

I found this interesting regarding the KII's from: https://www.hifinews.com/content/kii-audio-three-bxt-loudspeaker

 

 

"Working In Unison
There is no crossover as such between the THREE and BXT. Rather, the BXT supplements the THREE below 250Hz, mimicking its behaviour while dividing the LF output among a larger number of drive units. As with the THREE, the THREE BXT radiates omnidirectionally up to 50Hz before transitioning to cardioid radiation above 80Hz. This is maintained up to 1kHz, at which point the directivity becomes dominated by that of the tweeter. It's this surrendering of DSP directivity control at HF which makes the THREE so much less complex than the BeoLab 90.

The benefit of cardioid directivity is simple – it radiates 4.8dB less energy into the room for the same on-axis SPL than an omnidirectional radiation pattern, which means you hear more of the loudspeaker and less of the room. Arguably as important as the constrained radiation angle is that it's essentially constant from 80Hz to 1kHz, which means that – over this frequency range at least – early reflections from room boundaries will have the same spectral content as the direct sound, which makes it easier for us to recognise them as reflections and fuse them with the speaker's forward output."

 

 

So from what I can gather between about 50Hz to 80Hz there is a transition zone and not becoming a true cardioid until about 80Hz.

To me this makes sense as frequencies between 50Hz to about 80Hz are still omni-directional and are seemingly only somewhat controllable with the side woofers.

So this to me confirms that a subwoofer crossover at about 80Hz (or lower) should have minimal to no noticeable damage to the radiation properties of the KII's.

 

So this is what I'd be tempted to try:

50Hz High Pass to KII's and allow MSO to EQ them as far as comfortable (200-300Hz?)

80Hz Low Pass to Subwoofers and obviously allow MSO to EQ them.

Mulitple target curves using MSO (i.e. Flat, Bass Boost 1dB, Bass Boost 2dB, Bass Boost 3dB or whatever).

Edited April 9, 2021 by Satanica

[Snoopy8]

18 hours ago, Satanica said:

Interesting but I'm not that surprised by this. If memory serves me right you have a 13 inch and a 12 inch SVS sealed subwoofer so your overall sound could have been benefitting from the Kii's lending a hand down low. 

 

My first thoughts of experimentation was for you to set an 80Hz low pass filter for your subs most importantly and perhaps a high pass to the Kii's at 50Hz. But there are so many combinations it is mind boggling.

I think the Kii filter that is causing the SQ problem.  I did try a number of filters and measured them and selected what I thought was a relatively tame filter.  It was very time consuming to play with filters and then combine with MSO filters, so abandoning this way..

 

15 hours ago, gibbo9000 said:

Interesting - not sure I have seen, but where are your subs positioned?  Along with the Kii's facing forward, or somewhere else?  Kii noted in one of the discussions that they prefer x-over at 30Hz and with subs facing forward alongside Kii's so they notionally contribute to the Cardiod distribution and don;t drive refelections off rear/side walls etc.  (not sure that stands up in practice - but is what they suggested)

According to this review, the side and rear woofers are working in tandem to provide the cardiod pattern from 50 to 250 Hz.  Hence, not surprised that Kii suggested placing the subs on the same plane and facing forward. 

 

One of my subs off the speaker plane and to the left of the left speaker.     However, I managed get them to play nicely after tweaking MSO, but this was time consuming to get things just right. 

-----

Back to basics....

 

My objective is to integrate dual subs with the Kii and at the same time, minimise the impact on the cardiod pattern.  I know that using a 50 Hz LPF in MSO  had no impact on the bass.  After some systematic change and test, found that 70 to 90 Hz LPF wil give the bass I wanted.  So next step is fine tuning.

 

21 minutes ago, Satanica said:

So from what I can gather between about 50Hz to 80Hz there is a transition zone and not becoming a true cardioid until about 80Hz.

To me this makes sense as frequencies between 50Hz to about 80Hz are still omni-directional and are seemingly only somewhat controllable with the side woofers.

So this to me confirms that a subwoofer crossover at about 80Hz (or lower) should have minimal to no noticeable damage to the radiation properties of the KII's.

 

So this is what I'd be tempted to try:

50Hz High Pass to KII's and allow MSO to EQ them as far as comfortable (200-300Hz?)

80Hz Low Pass to Subwoofers and obviously allow MSO to EQ them.

Mulitple target curves using MSO (i.e. Flat, Bass Boost 1dB, Bass Boost 2dB, Bass Boost 3dB or whatever).

Thank you.  I was writing the above when you posted this.  It confirms that I am now on the right track.  And I do not need to play with any high pass filters. 

 

[Snoopy8]

To date, have only mentioned in passing, MSO's ability to integrate multiple subs in a home theater environment. It is especially helpful in a multi seat, multi row HT setup because some people will be seated where the bass is not that good. 

 

 

In my shared room, there were 2 subs, one in the front and another to the right side of the sofa. When combined, they appeared as the red line in the graph below.

 

 

Responses were offset to improve clarity. The combined response was OK, within a 20 dB range and no big troughs.  The left of the sofa had the more even response because it was furthest away from the side sub and was lucky that the subs mostly complimented one another.

 

 

 

After MSO, the bass response was more even across all seating positions.  You could then use this integrated bass response with any  room correction software.

 

My setup was relatively simple, and I could have manually integrated the dual subs.  With multiple subs  and more seats across rows, it would been a big challenge.  MSO makes it much easier to get even bass across multiple seating positions.

 

If you know how to use REW to measure using a timing reference, and have a decent DSP for the subs, I can help get your subs integrated.

Edited April 15, 2021 by Snoopy8
Typos, improved clarity.

[Satanica]

OK, so I gave it a go on the weekend, yes I did.

It certainly provided a much flatter starting point for Dirac Live 3.14 to work with afterwards.

 

I will need to do it all again this weekend as I made some mistakes.

I'll need to balance my woofers (by turnng them down) to my subs better as it maxed out the gain of one of the subs to +12dB. I'm surprised that there was a 4.5dB difference between my front sub (+12dB) and the rear one (+7.4).

 

Question, to effectively fix a crossover at 80Hz one needs to change Optimization allowed = False?

I also changed the Minimum Value to 80 and Maximum Value to 80.1 but I presume this was redundant?

I've supplied a screen-shot.

 

 

Question, by default for EQ under Boost (dB) the Maximum value = 0 and I didn't change it and Minimum value is -15.00. So I presume this means it cannot actually directly boost?

I've supplied a screen-shot.

 

Edited April 27, 2021 by Satanica

[Snoopy8]

5 minutes ago, Satanica said:

Question, to effectively fix a crossover at 80Hz one needs to change Optimization allowed = False?

I also changed the Minimum Value to 80 and Maximum Value to 80.1 but I presume this was redundant?

I've supplied a screen-shot.

Correct.  When you say optmization allowed = false, you are telling MSO to use your value of 80.  Min and Max are no longer relevant.

 

7 minutes ago, Satanica said:

Question, by default for EQ under Boost (dB) the Maximum value = 0 and I didn't change it and Minimum value is -15.00. So I presume this means it cannot actually directly boost?

Correct.  Boosting will not fix a null.  Some people do use a small boost for minor troughs, but I have avoided it.

[Satanica]

7 minutes ago, Snoopy8 said:

Correct.  Boosting will not fix a null.  Some people do use a small boost for minor troughs, but I have avoided it.

 

Cool, I will leave it to Dirac to clean up any small troughs then.

[Satanica]

I gave it another go on the weekend from scratch.

 

After doing all of the measurements and setting everything up I experimented a fair bit with:

Polarity Inversion Blocks and ...

As flat as possible without additional global EQ versus Minimize seat-to-seat variations

 

I found that no Polarity Inversion Blocks and As flat as possible without additional global EQ gave the best results.
I allowed 7 PEQ's although I think a miniDSP (which I have) can handle 8.

 

And then I ran Dirac Live 3.14 from scratch over my system again.

 

This all took quite a bit of work and head scratching at times but I'm very happy with the outcome.

So thanks @Snoopy8 for all of the info and your perseverance.

[davewantsmoore]

On 09/04/2021 at 11:33 AM, Satanica said:

So from what I can gather between about 50Hz to 80Hz there is a transition zone and not becoming a true cardioid until about 80Hz.

 

It's a monopole below 80hz .... as is everything that is inside a "small" room.

 

On 09/04/2021 at 11:33 AM, Satanica said:

So this to me confirms that a subwoofer crossover at about 80Hz (or lower) should have minimal to no noticeable damage to the radiation properties of the KII's.

 

Yep.... or even a bit higher if the xo is steepish.

 

On 09/04/2021 at 11:33 AM, Satanica said:

So this is what I'd be tempted to try:

50Hz High Pass to KII's and allow MSO to EQ them as far as comfortable (200-300Hz?)

80Hz Low Pass to Subwoofers and obviously allow MSO to EQ them.

 

It would be better to use filters which had a matching phase response  ..... but the overall levels is what ultimately matters (and what MSO will help you fix) 

[Mark_A_W]

After a battle, I've achieved reasonable results with MSO.

 

Setup:   

- 2 x Infinite Baffle subs with 2 x 15" Adire Tempests each, in manifolds in the front corners of the floor.  4 drivers, but two subs.

- Additional smaller sub in one back corner.   Decent Focal driver in a junk car audio box, more for testing the process.     Will be 2 additional subs, once I get proper boxes built.   So, 3 subs now, heading to 4 soon.   The additional pair will probably be deliberated rolled off at about 30hz, as they can't do 15hz at a bazzillion dB like the Infinite Baffle subs.    3 ways with Minidsp 8x8 (Surrounds are active too, on another Minidsp 8x8)....20 channels of amplification heading to 21.

 

MSO run against the front 3 channels on the 7.1 (7.3?) setup.    So three Speakers and 3 subs.   3 measurement positions....lots of measurements!   

Look Ma! No Subs!!

 

 

BEFORE:  Massive peaks and dips.



 

The Battle:

MSO kept EQing one of my main IB subs down to -20 or -30 dB, over pretty much its whole range.    Either with Gain, or by stacking PEQs on top of each other.  Basically turning it off.  It seemed to get the level wrong, and reduce the subs by an incredible amount (like 30+ dB) which was more than it needed - closer to 15dB.

 

What I did: 

- Changed the target to be a specific dB level, rather than auto detect.  It was getting it "wrong" and trying to take about 20dB total off my response.  Yes, 200-300hz is too low in my system, but it seemed to be massively over-compensating.  It decided the target was about 90dB.   My response needed the subs dropped a bit, but nothing like that.   A Target Curve would have worked too, but just punching in 105dB (!!) as the target did the job.  I want a rising bass response, not flat.  I used 105dB in the end.

Note: These numbers seem really high, I didn't measure *that* loud.   I must have calibrated REW incorrectly against the SPL meter.  Doesn't matter, absolute is irrelevant, it's all relative, all measurements were done in one sitting with the same settings.

- Change the optimiser to minimise seat to seat variations (I did say "minimise differences between the subs" in error).   Anthem Room Correction is coming next, so a unified "single sub" is really my goal here.

 

- Constrained the Gain and Boost somewhat.

 

Eventually I got this:

 

AFTER:

- 

 

With these filters:

 

 

Then I ran Anthem Room Correction over the top.   Green is After.   Red is Before.

Came out ok.    I still have 2 Minidsp Biquads available per front channel if I want to tweak more. 

 

It sounds good.   Smooth, warm.  But powerful too.  MUCH BETTER THAN BEFORE.   The 50hz mode was overpowering.

 

 

 

 

 

 

   

 

Edited May 6, 2021 by Mark_A_W

[davewantsmoore]

11 hours ago, Mark_A_W said:

 The 50hz mode was overpowering.

 

 

Eeek.   You have modes at what seems like 30, 50, 70.... making the 50 look huge(er than it really is).

Did a v. good job under the circumstances.

 

This is the sort of thing PaulS was alluding to in the other thread, about placement, etc.