Science on low frequency quality?

[sigbergaudio]

Sure, there are members with the right equipment and time to run the tests. The question for me is what is the result. Subs have to complement the speakers already in use. So there is a huge amount of different speakers and rooms. What is good for a specific room and speaker model may not work at another setup. And as I wrote, an equilizer does not change the room, only the frequency response which may make a better sound. Since the bass hearing impression is dependent on the volume it is not a constant. If linearized the room acoustic then still will the sound of music change with the listening volume. Which turns into the situation to listen with a certain personal standard volume.

A good subwoofer is not room dependent or speaker dependent. But it is of course not just to plug it in and then it's perfectly integrated. It takes work and competence.

With regards to getting the right volume in the bass dependent on your volume (if I understand you correctly), I'm not sure I see how this is directly relevant. A linear bass response with good extension will still be way better than the alternative, and one typically adjusts the volume in the bass to the most typical listening volume. Some AVR/processors also offer dynamic EQ to counteract this effect.

 

[SSS]

Yes, I agree. If the adjustments can be made properly. For this I admit don't have regarding subs models an overview of the possible adjustment capabilities. If delay, phase and high-cut slope of the sub may be adjustable to match the opposite declining slope of the low frequency end of the main speaker then at least it may be as if the main speaker had a lower frequency cut off tested e.g. in an anechoic chamber. Adjusting additionally to the room behaviour seems to me an extra challenge. Of course the selling of subs is a business and the ads pretend that it is an easy task.

 

[sigbergaudio]

Yes, I agree. If the adjustments can be made properly. For this I admit don't have regarding subs models an overview of the possible adjustment capabilities. If delay, phase and high-cut slope of the sub may be adjustable to match the opposite declining slope of the low frequency end of the main speaker then at least it may be as if the main speaker had a lower frequency cut off tested e.g. in an anechoic chamber. Adjusting additionally to the room behaviour seems to me an extra challenge. Of course the selling of subs is a business and the ads pretend that it is an easy task.

There are devices that do this automatically for you, and do it pretty well.

 

[jsilvela]

I’ve been testing out different configurations with my two subs.

I’m understanding some of the advice on null-filling better … and also I’m understanding some of it less…

There seem to be two schools of thought regarding multi-sub. I’ll risk summarizing (any mis-representation not intentional)

The Welti approach:
2 or 4 subs, in highly symmetrical configurations, to help eliminate or reduce room modes.

The Geddes approach:
Subs in UN-correlated locations to even out. Suggests trying 1 sub in front corner, 1 in back, 1 in side wall.

The Geddes approach of un-correlated locations seemed to me a more intuitive explanation of the “fill in nulls” advice I see often at ASR to justify multi-sub.
The Welti approach proposes *very* correlated locations. With correlated locations, I reasoned, two subs would likely produce the same nulls.

I tried the Geddes approach first, and always had a wide dip in 20-30Hz, which I assumed was the sign that I needed MSO, or perhaps a third sub + MSO.
Then I tried the Welti mid-front, mid-back configuration, which was suggested in this thread by DonH56 as a good beginning. And… much better.

I’ve started playing with REW Room Sim. My room is not a proper rectangle, but I’m using approximate room dimensions, and approximate location of the main listening position plus two extra positions at its sides.
With Room Sim, I’ve started to understand the dips and peaks in a very different light. What an amazing tool. I made a donation to the project; getting so much value from it.

My main take-aways:

• reasoning about the joint effect of 2 subs by looking at the individual FR’s is not a good idea. The FR does not tell you about the phase, which is critical
• the “symmetric subs to cancel modes” approach seems to work better for me. The nulls are “filled” by damping modes

I’ve been fretting about the complexity of sub and multi-sub, and turns out that 2 subs at symmetric locations are making things much less fussy in my case. Huh.

The Griesinger video brought by youngho was very interesting. Griesinger makes two tangential trips that caught my attention. One is that in his view it is a good idea to raise subs to 1/2 room height, to avoid exciting vertical modes. The other is just how non-chalant he was when showing his home installation. He used 2 subs in symmetric locations and dropped “these things are cheap and work great, about $300 each”.

 

[Willem]

In the meantime I have been working on my own subs. Initially I had two subs, the big one in one front corner, and the smaller one in the other front corner. No room treatment and no dsp eq: my small Kef Kube8 gave off loud switching thuds through my ANtimode 8033, so I could not use that. Instead, I have bought a miniDSP 2x4HD to be used with MSO, but have not yet done the measurements. I decided that this was the moment to add a third sub, so I bought an SVS SB2000. Since I only had some short cables, I first connected all three subs with what I had, i.e. at quite close distance from each other. The resulting sound was booming and pretty awful. When my new cables arrived I located the three subs as per Geddes, i.e. the biggest sub (the SB2000) in the front corner, the medium sized B&W PV1d in the back of the room, and the small Kef Kube8 along the side wall. The sound is now pretty convincing, and very much cleaner. I must add that this is in a large open plan living space, so any kind of symmetrical placement is impossible. Yesterday evening I played Gergiev's live recordings of the Firebird and the Rite of Spring and it was already pretty stunning. The next step will be to perform the measurements for MSO.

 

[FeddyLost]

reasoning about the joint effect of 2 subs by looking at the individual FR’s is not a good idea. The FR does not tell you about the phase, which is critical

Unless you make a 2.2 system.
If N subs are tuned as single LFE emitter, for sure, their single-sub FR is not a valid prediction of end result.

the “symmetric subs to cancel modes” approach seems to work better for me. The nulls are “filled” by damping modes

Or "symmetric subs to not excite modes".
I've used 2.2 with subs in middles of side walls and MLP in middle longitudinal axis. With existing passive absorption FR looked really nice.

Soon I'll have to move them to frontal side and according to Room Sim EQing will be inevitably required.

 

[jsilvela]

Unless you make a 2.2 system.
If N subs are tuned as single LFE emitter, for sure, their single-sub FR is not a valid prediction of end result.

Or "symmetric subs to not excite modes".
I've used 2.2 with subs in middles of side walls and MLP in middle longitudinal axis. With existing passive absorption FR looked really nice.

Soon I'll have to move them to frontal side and according to Room Sim EQing will be inevitably required.

sorry to ask, what is a 2.2 system? Was imagining stereo mains + 2 subs, but I don't think that's what you mean?

BTW absorption is one area I haven't gotten into at all. Seems interesting.

 

[FeddyLost]

what is a 2.2 system?

Left and right front and left and right subs.
Imagine small bookshelf speakers standing on pair of subs - something like this. With time aligned subs it sounds seamless.

My room arrangement allowed only one "acoustically proficient" placement of pair of subs, but still current circumstances require their placement in frontal part.

 

[youngho]

The Welti approach:
2 or 4 subs, in highly symmetrical configurations, to help eliminate or reduce room modes.
...
The Geddes approach of un-correlated locations seemed to me a more intuitive explanation of the “fill in nulls” advice I see often at ASR to justify multi-sub.
The Welti approach proposes *very* correlated locations. With correlated locations, I reasoned, two subs would likely produce the same nulls.
...
reasoning about the joint effect of 2 subs by looking at the individual FR’s is not a good idea. The FR does not tell you about the phase, which is critical

This is not quite accurate. Careful subwoofer placement in the node of a mode can avoid exciting that mode. Symmetric placement of multiple subwoofers with respect to a mode can cancel out the odd-order modes, so "two subs would likely produce the same nulls" is not correct. You may find it helpful to review http://www.wghwoodworking.com/audio/LoudspeakersandRoomsPt3.pdf. My very limited understanding about SFM and MSO (https://www.audioholics.com/room-ac...i-sub-sfm/the-birth-of-sound-field-management and https://www.andyc.diy-audio-engineering.org/mso/html/tutorial/index.html#how_err_calc) is that they do measure each subwoofer individually.

The Griesinger video brought by youngho was very interesting. Griesinger makes two tangential trips that caught my attention. One is that in his view it is a good idea to raise subs to 1/2 room height, to avoid exciting vertical modes. The other is just how non-chalant he was when showing his home installation. He used 2 subs in symmetric locations and dropped “these things are cheap and work great, about $300 each”.

Griesinger placed the subs to the left and right of the listening position. His goal (so-called spatial bass or "bassiousness") are different from those of SFM or MSO (which are different from each other, as well).

 

[jsilvela]

This is not quite accurate. Careful subwoofer placement in the node of a mode can avoid exciting that mode. Symmetric placement of multiple subwoofers with respect to a mode can cancel out the odd-order modes, so "two subs would likely produce the same nulls" is not correct. You may find it helpful to review http://www.wghwoodworking.com/audio/LoudspeakersandRoomsPt3.pdf. My very limited understanding about SFM and MSO (https://www.audioholics.com/room-ac...i-sub-sfm/the-birth-of-sound-field-management and https://www.andyc.diy-audio-engineering.org/mso/html/tutorial/index.html#how_err_calc) is that they do measure each subwoofer individually.

Griesinger placed the subs to the left and right of the listening position. His goal (so-called spatial bass or "bassiousness") are different from those of SFM or MSO (which are different from each other, as well).

I could see my reasoning about null filling was not right. That reference on LoudspeakersAndRoomPt3 ... had seen it before.
I need to re-read again, now I'm sure I'll get new things out of it.

Yes, good point on Griesinger having a different goal in mind.

 

[jsilvela]

Left and right front and left and right subs.
Imagine small bookshelf speakers standing on pair of subs - something like this. With time aligned subs it sounds seamless.

My room arrangement allowed only one "acoustically proficient" placement of pair of subs, but still current circumstances require their placement in frontal part.

In that case, I'm not sure I followed your previous message

Unless you make a 2.2 system.
If N subs are tuned as single LFE emitter, for sure, their single-sub FR is not a valid prediction of end result.

So the difference of the 2.2 system with N subs tuned as a single LF emitter, then,
do you mean that in your case the two subs are tuned with individual delays, possibly individual EQ profiles?
And doing that, then, the response of a single sub does give you an idea of the composite response?

 

[gnarly]

Hi all, may I offer some random thoughts/observations on MSO.....with the upfront realization i am far from an expert, just a guy who has tried a lot of sub deployments.

I like to think there are two basic strategies on indoor sub placement: one is to minimize modes, the other is to maximize them.
Subs outdoors tell me what minimization sounds like, and MSO indoors tells me what maximization sounds like.

Must say I prefer minimization by a wide margin....but is even more difficult to achieve indoors than properly implementing MSO.
A simple single sub in a corner is often the best minimization I can find. So MSO is often a well chosen compromise .

The measurement reciprocity technique, where you put a sub at listening position, and go around the room with a mic to find the smoothest reponse, is the best way I've found to determine optimal sub placements. That's held true for either a single sub, or multiples placing them in sequence one at a time.
Usually the first sub goes in the corner. After the first sub is placed, I'll apply a polarity switch and/or delays EQ on the second sub in the center of the room, again going around the room looking for now-combined response. It gets real confusing fast to try to vary multiples simultaneously ime. (BTW, subs don't really have the ability to adjust phase; polarity is almost always all the have wrt phase.)

There was a great question about can we feel air pressure waves from subs.....and i say without a doubt YES.
There was also a great follow-up reply distinguishing steady tones from impulses like come from a port.
Personally, I love the feel of a transient bass attack. And is a big part of why i prefer a single sub, with mode minimization when possible.

This may sound a little extreme, but an experiment that proves a physics difference to me, is playing an outrageous bass drop on a single strong sub, and then on multiples of the same sub. Picture an LCR setup where all three stacks have the same sub.

The measurement device is an empty beer can on a table about 15ft away from the center of the LCR setup.
The single center speaker can blow the beer can off the table. All three LCR subs make it jump and dance more or less in place, not leaving the table.

Ok i know that is extreme, but I personally think we feel this 'air impulse effect' all the time and is a part of our total hearing, which i think integrates bodily sensations with the ear.
But i digress...sorry..

Oh, If you're trying to figure out how much sub in terms of size and numbers you need, this handy piston calculator is a good starting place

Piston Excursion calculator

www.baudline.com

 

[FeddyLost]

do you mean that in your case the two subs are tuned with individual delays, possibly individual EQ profiles?

It depends. If setup is not completely symmetrical, it may be required.
In my exact case difference between "made up" channels at MLP was relatively small, so I did not bother EQing them.
It's not a mastering studio anyway.

And doing that, then, the response of a single sub does give you an idea of the composite response?

Of course not, but it's not the case of DBA, where in fact negative interference used for fight with axial modes.
I placed 2 subs symmetrically as good as I could, sit between them and they are not interfering if bass is panned in center.
Complete system measured flat in LF.

 

[jsilvela]

Complete system measured flat in LF.

wow, getting flat from placement seems like the best one could hope for

 

[FeddyLost]

getting flat from placement seems like the best one could hope for

Exactly
That was really profitable
No smoothing and no EQing subs besides forming slope (I've made crossover slopes by series of EQs by Jriver to avoid any signal processing in subwoofers)

But i need to insist that this flatness is only when 2 subs active.
And this works nice without EQing each sub just because most of music in fact have mono bass and that's not a mastering studio.
L and R are different and not flat at all with deviation of 8-9 db peak-to-peak.
Nothing catastrophic, but not even close to neat upper picture.

So, this is an example of placement "not to excite modes" - symmetrical, in phase, subs are mid-sides, listener is on middle axis. In fact, dedicated room and MLP.
Works nice only for single position. If you will need to have flat LF at 3-4 seats in a row, most probably you'll have to use MSO.

 

[valve_integrator]

I placed 2 subs symmetrically as good as I could, sit between them and they are not interfering if bass is panned in center.
Complete system measured flat in LF.

I figured cone area is everything, so the bass freqs with roll-off are handled by 2 huge speakers, L & R.
235L capacity each and 2 x 12 and 2 x 15 drivers.
It was all designed and made in house.
They are so heavy you can hardly move them, never mind lift them.
I use the 5 way system for monitoring and policing live recordings.
They don't measure flat in room, and I use a DSP.

For Organ music I have never heard anything like it.
You even can hear a bass drum, -
- so can the neighbours, as the concrete and brick building vibrates a bit.

 

[Willem]

The next step will be to perform the measurements for MSO.

My power amp needed to be repaired in the meantime, so I could only start measuring yesterday. Julian Krause's Youtube video helped me a lot to get the first measurements done:

All I did was measure the in-room response of the entire system, and it showed that you cannot do without measurements. There was a sizeable dip around 200 Hz, i.e. unrelated to the subwoofers. The transition between the main speakers and the subs was smooth and unproblematic, which was a nice surprise, but below that reponse was not only ragged, but bass response in the subwoofer range was also far too strong. I had had some suspicions, but not that it was this bad. Measurements really are indispensible.
The next step before making the measurements for MSO will be to remove all filters in the RME ADI-2, and disable tone and loudness controls.

 

[jsilvela]

My power amp needed to be repaired in the meantime, so I could only start measuring yesterday. Julian Krause's Youtube video helped me a lot to get the first measurements done:

All I did was measure the in-room response of the entire system, and it showed that you cannot do without measurements. There was a sizeable dip around 200 Hz, i.e. unrelated to the subwoofers. The transition between the main speakers and the subs was smooth and unproblematic, which was a nice surprise, but below that reponse was not only ragged, but bass response in the subwoofer range was also far too strong. I had had some suspicions, but not that it was this bad. Measurements really are indispensible.
The next step before making the measurements for MSO will be to remove all filters in the RME ADI-2, and disable tone and loudness controls.

luck with all that, Willem.

Been meaning to ask you: you're a strong proponent of the Geddes un-correlated subs approach.
I imagine because your room is asymmetric?
Have you tried other approaches before that did not work? Which?

 

[Willem]

The main listening area is rectangular, but it opens in two directions to other spaces, so in effect the whole space is highly irregular. For years, I had just one sub (the PV1d) equalized by an Antimode 8033. The result was actually pretty good. I then decided that I did not really need the little KEF Kube8 that I had in my study, so I added that to the main system. Unfortunately the KEF produced a pretty loud switching thud through the other sub. I discovered that this did not happen when I removed the Antimode and connected the subs directly to my ADI-2. So I wondered what would happen if I used a minidsp 2x4HD: would it suffer from the same grounding issue, or not? It turned out not. Since I was now en route to using MSO, I decided I might as well go one step further and add a third and somewhat more powerful sub. And that is where I am now, with a lot of work to do.

 

[jsilvela]

And that is where I am now, with a lot of work to do.

hahaha, what we do for sound. Godspeed!

Unfortunately the KEF produced a pretty loud switching thud through the other sub.

I may have misunderstood, but I think I've had a similar experience. Now I have a Kef 10b and a REL T/5x.
There are often loud, low frequency pops on the REL. This happens when the Kef goes in/out of standby.
The Kef has a much shorter time-out to go to standby than the REL.

I figured that this was really down to my AVR not isolating the sub-outs properly.
It improved quite a bit after I turned down the gain knobs on both subs, as the AVR was giving me a -9dB recommendation anyway.
I turned the REL gain about 5dB down.