Subwoofer / Low Frequency Optimization

[DonH56]

So, I get what you're saying here and I agree with a lot of it, but most speakers that we use as mains do go just about as low as subwoofers. They just don't do it as efficiently.

Or as cleanly (distortion-wise). And speaker placement for the best imaging and such is rarely the best place for smoothest low-frequency response since room modes tend to occur around the listening area and not where the main speakers are placed.

You can high pass your mains, but that doesn't get us away from phase problems, especially at the crossover region.

I never said that; I have repeatedly said you must optimize phase at the crossover frequency to provide a cohesive wavefront.

Also, I'm not saying that correcting for frequency response causes phase problems. I'm. saying that having multiple subwoofers can flatten frequency response but increase phase problems. You can correct for the nulls, but what happens to the spaces that were working properly. My concern is that you can add more problems than you fix, but again, I've never tried 3 subwoofers in a room. I'm merely trying to understand why this is a better approach.

You always have to optimize for the listening area. Multiple subs provide the flexibility to optimize a broader area. If you have one sub and correct the response in one spot, then other spots could be worse. If you EQ a peak for the listening spot, then other areas in the room will have less bass. As you add subs, or any speakers, overlapping sound fields become more complex, but provide greater ability to smooth the response across a wider area (you have more parameters to play with). In a room, church, or stadium you still have to time-align the speakers to provide a cohesive sound to the listeners.

You keep saying multiple subs can increase phase problems, and of course they can, but the idea is to use the tools we have to correct the phase (delay) to provide the best response. It is a better approach because it adds additional sources and thus parameters to use in correcting the response. If you have a single sub it is almost impossible to correct all the modes in a room. Different dimensions lead to different modal frequencies; speaker, listener placement cause differing peaks and valleys, the furnishings in the room can affect the response, etc.

Toole, Geddes, Welti et. al. explain this much better than I.

The reality is that I would have to hear it and see if I like it. In my experience with multiple subwoofers (mostly in arenas and other systems where we use sub arcs to decrease modes and create a more even low end response, we often sacrifice punch and clarity.) I'm wondering if that's what happens in small rooms also.

If you need to prove it to yourself there's not much point in talking (writing) more about it! I'm from Missouri and have a healthy streak of "show me!" but if you mistrust everything it is a lot of work to figure it out on your own. I have to defer to experts when it is outside my field.

Punch and clarity is usually a function of frequencies above the subwoofers' range, at least in HT applications. For setting up an arena or other large venue, which is decades in the past for me, alignment was often difficult and constrained by not only the tools I had (no DSP then) but cost and placement constraints of the venue. If aligning deep bass caused loss in "punch and clarity" my first reaction would be that the sound system was not properly integrated and/or the subs had too much overlap with the other speakers. Of course, in rooms large or small it is virtually impossible to get the best sound in every seat, though enough drivers and processing power can come close.

I question applying different EQ to mains and subwoofers when they're operating in the same frequency range in the same room. Wouldn't they have the same problems in the room?

Because they are generally at different places in the room so they exhibit different in-room frequency response. For example, if you place the subs to help cancel (mitigate) room modes, they will be in a different place and different physical environment than the other speakers and thus require different EQ.

My basis here is that flat frequency is a great goal, but not if it sacrifices the listening experience in other ways (which I generally describe as phase problems). I see the same problem in some crossover designs where drivers sum flat but not in phase. When you add a subwoofer that flattens the response by canceling peaks but summing the nodes, does it not also effect other areas of the frequency spectrum and listening experience?

The goal is to utilize crossovers and processing to optimize integration no matter how many subs are in play and minimize interaction with the mains. For example, my rather modest room has a primary mode (null) around 30 Hz, so anything I do to fix that has little impact on the main speakers using a 4th-order crossover at 80 Hz. My additional subs are not really "summing nodes", they are placed to drive the nodes (minima) and thus physically not in position to increase the antinodes (peaks).

I do not believe anything I say will be meaningful to you so will drop this. For me, in my room, multiple subs were the only practical solution to smoothing the response, and I have not heard any negative effects as I went from one sub, to two, to four -- just better bass. No impact I could discern on higher frequencies, and the additional subs provided better bass across a wider listener area.

 

[kevinh]

This article cites research papers in phase distortion and it's audibility, with a comment by Floyd Toole. Not sure if i low frequencies phase is much of an issue in room settings a 80hz wavelength is ~ 14ft, so a 1ft misalignment between subs would not IMO be audible. OTOH the multiple subs seeking to attenuate room modes would be VERY beneficial.

Human Hearing - Phase Distortion Audibility Part 2

This article discusses how sensitive the human ear is to phase distortion and how it determines the listening quality we identify with.

www.audioholics.com

 

[ThatSoundsGood]

Or as cleanly (distortion-wise). And speaker placement for the best imaging and such is rarely the best place for smoothest low-frequency response since room modes tend to occur around the listening area and not where the main speakers are placed.

I never said that; I have repeatedly said you must optimize phase at the crossover frequency to provide a cohesive wavefront.


You always have to optimize for the listening area. Multiple subs provide the flexibility to optimize a broader area. If you have one sub and correct the response in one spot, then other spots could be worse. If you EQ a peak for the listening spot, then other areas in the room will have less bass. As you add subs, or any speakers, overlapping sound fields become more complex, but provide greater ability to smooth the response across a wider area (you have more parameters to play with). In a room, church, or stadium you still have to time-align the speakers to provide a cohesive sound to the listeners.

You keep saying multiple subs can increase phase problems, and of course they can, but the idea is to use the tools we have to correct the phase (delay) to provide the best response. It is a better approach because it adds additional sources and thus parameters to use in correcting the response. If you have a single sub it is almost impossible to correct all the modes in a room. Different dimensions lead to different modal frequencies; speaker, listener placement cause differing peaks and valleys, the furnishings in the room can affect the response, etc.

Toole, Geddes, Welti et. al. explain this much better than I.

If you need to prove it to yourself there's not much point in talking (writing) more about it! I'm from Missouri and have a healthy streak of "show me!" but if you mistrust everything it is a lot of work to figure it out on your own. I have to defer to experts when it is outside my field.

Punch and clarity is usually a function of frequencies above the subwoofers' range, at least in HT applications. For setting up an arena or other large venue, which is decades in the past for me, alignment was often difficult and constrained by not only the tools I had (no DSP then) but cost and placement constraints of the venue. If aligning deep bass caused loss in "punch and clarity" my first reaction would be that the sound system was not properly integrated and/or the subs had too much overlap with the other speakers. Of course, in rooms large or small it is virtually impossible to get the best sound in every seat, though enough drivers and processing power can come close.


Because they are generally at different places in the room so they exhibit different in-room frequency response. For example, if you place the subs to help cancel (mitigate) room modes, they will be in a different place and different physical environment than the other speakers and thus require different EQ.


The goal is to utilize crossovers and processing to optimize integration no matter how many subs are in play and minimize interaction with the mains. For example, my rather modest room has a primary mode (null) around 30 Hz, so anything I do to fix that has little impact on the main speakers using a 4th-order crossover at 80 Hz. My additional subs are not really "summing nodes", they are placed to drive the nodes (minima) and thus physically not in position to increase the antinodes (peaks).

I do not believe anything I say will be meaningful to you so will drop this. For me, in my room, multiple subs were the only practical solution to smoothing the response, and I have not heard any negative effects as I went from one sub, to two, to four -- just better bass. No impact I could discern on higher frequencies, and the additional subs provided better bass across a wider listener area.

I really appreciate your explanation. Everything you're saying here makes total sense and I'm intrigued to give it a try to now.

 

[DonH56]

I really appreciate your explanation. Everything you're saying here makes total sense and I'm intrigued to give it a try to now.

Thanks!

As you know, "worth trying" does not mean "easy", but the end results can be very satisfying.

 

[ThatSoundsGood]

Thanks!

As you know, "worth trying" does not mean "easy", but the end results can be very satisfying.

If it was easy, it would probably be boring. The challenge is the fun. Cheers!

 

[Lttlwing16]

you must optimize phase at the crossover frequency to provide a cohesive wavefront.

My question is it best to apply a 6 cycle frequency dependent window to "remove the room" prior to aligning the phase at the crossover?

 

[DonH56]

My question is it best to apply a 6 cycle frequency dependent window to "remove the room" prior to aligning the phase at the crossover?

You can try... The usual quick scheme is to pick the speaker furthest from the listening position, then adjust the delay of the closer speaker(s) until frequency response (amplitude) is maximized, on the assumption that the minimal delay needed to do that is the point at which everything is "in phase". Software now allows more sophisticated measurements using time-domain (step, "impulse") response that can be helpful in isolating the initial wave front from room (etc.) reflections.

The number of cycles, gating delay, or time windowing would depend upon frequency, room size, and distances involved (e.g. speaker to reflecting surfaces and the listener).

This is not really my area of expertise so other members may (will) have more insight.

 

[elmura]

In my experience, having multiple subs complicates things. The papers by Toole and Geddes are simplistic and don't translate well to real world measurements as implied by their recommendations.
You can try REW Room Sim to experiment with their ideas in your own room. It's surprisingly close to a measured result as long as you ensure all room and speaker specs are accurate.
In my dedicated home theatre room, I've spent many hours experimenting and measuring, with REW assistance, plus various audio techniques.

I have found that to achieve a smooth response works better when you are placing full range speakers and subwoofer in places that are a division of the room length and or width. I learnt this approach many years ago from an audio expert.

eg. In my room of 6m L x 4m W, I have one subwoofer on the front wall at 1/3 room width. And a second subwoofer on the side wall at 1/4 room length.
Then I invert one sub, apply PEQ to tame the peaks on each, and apply delay (distance) to try fill out the nulls. I used REW simulator to start, then measurements to finalise.

It's not flat response, but it's reasonable.

However, I'm not convinced that it's optimised as the bass seems a bit dull. I think that timing issues are causing cancellations in a way that take the impact away. When I turn one sub off, there's more bass presence.

My AV processor does not have room correction software, but it does have individual 5 band PEQ for every speaker/sub (Classe SSP-800)

Certainly, it's not as simple as what the Toole & Geddes papers imply.

 

[Head_Unit]

My AV processor does not have room correction software, but it does have individual 5 band PEQ for every speaker/sub (Classe SSP-800)

Does it also have individual time delays? Otherwise we can easily imagine the timing issues you hypothesize. Supposedly the very low frequencies are so large it doesn't matter but that's just also a hypothesis based on one particular view of the physics, not sure anyone has really researched this.

 

[kevinh]

When you compare the wavelengths at low frequencies in small rooms IMO phase is not a big deal at all. The idea is to get rid of resonances and suck out in a room freq response. That is where multi subs are a benefit Phase is much more important as we go up in the freq range due to wavelength. At 100hz you are looking at ~3.5M wavelength. at 1khz, wheras at 1,000hz it is ~.34M to pohase would be more noticeable, IO believe that Toole and Geddes take this into account.

 

[Andysu]

When you compare the wavelengths at low frequencies in small rooms IMO phase is not a big deal at all. The idea is to get rid of resonances and suck out in a room freq response. That is where multi subs are a benefit Phase is much more important as we go up in the freq range due to wavelength. At 100hz you are looking at ~3.5M wavelength. at 1khz, wheras at 1,000hz it is ~.34M to pohase would be more noticeable, IO believe that Toole and Geddes take this into account.

picture of your room i may see way you can add extra dB or more

 

[kevinh]

picture of your room i may see way you can add extra dB or more

Just commenting on people clutching pearls about phase problems with subs.

 

[elmura]

Does it also have individual time delays?

Yes. Distance for each sub on the AV processor creates the time delays.

When you compare the wavelengths at low frequencies in small rooms IMO phase is not a big deal at all.

Just commenting on people clutching pearls about phase problems with subs.

Well, I can confirm that it is not imaginary. And if you do some measurements, you will change your opinion. Just as a phase switch / dial on subs help the sub bass integrate with the main speakers bass at the crossover frequency, so too does phase adjustment affect the way the subs operate across the same frequency band.

Remember, each sub is at a different location so their individual outputs will have a different response at the listening position. Adjusting the phase on either sub (in a multi sub arrangement) has a big effect on the combined sub-bass frequency response. The AV processor's distance/delay has a similar effect.

 

[elmura]

I've attached a set of measurements I made in January. Front 3 speakers are identical floor standing speakers, placed equidistant from the MLP. Each speaker's crossover point was adjusted and distance set, so were subs.

Note how the subs alone produce a smoother response than adding additional bass sources!

Looking at these now, I'm thinking to revise the crossover point for the orange and green speaker traces to a 55Hz to avoid their cancellation effect at 48Hz.
Another idea is to set all the front speaker crossovers to 110Hz and avoid their interaction! This goes completely against the forementioned papers of having as many sub-bass sources as possible to smoothen out the frequency response!

What do other pros think?

 

[jhaider]

I've attached a set of measurements I made in January. Front 3 speakers are identical floor standing speakers, placed equidistant from the MLP. Each speaker's crossover point was adjusted and distance set, so were subs.

I don’t see where you attempted any optimization. You need to play with relative levels and phase/delay (“distance” is a variable not a value); use whatever works for each sub for smoothest steady state response. This guide is a good starting point for a blended mains-subs approach:

Multiple Subwoofer after Geddes

 

[Head_Unit]

When you compare the wavelengths at low frequencies in small rooms IMO phase is not a big deal at all...At 100hz you are looking at ~3.5M wavelength.

That's the physics view, which is correct. However if you had say two subs somewhat "out of phase" at some low frequency...I'm not sure how to express this well, but the pressure rise from one would arrive sooner than the other. I dimly recall research that we are not sensitive to phase or group delay at low frequencies but I wonder if what was tested would really be the same as testing our sensitivity to initial transient waveforms. Or something like that.

 

[kevinh]

This is the idea of multiple subs. Geddes version if you say have 2 full range speakers the 3 subs are placed to eliminate nulls even to the extent that the recommends 1 sub 1/3rd of the way up the wall. So the goal it to minimize the peaks and sickouts. in the freq response. IIRC he did listening test and found out hearing is relatively insensitive to directional clues at those frequencies we identify in terms of harmonics, where our ears are more sensitive to directional ques

 

[Bob from Florida]

I added a second sub because sometimes I could localize the sound from the single sub. Once added that localization went away. Main speakers fire across a 12 foot deep room - 35 feet wide. Living room one side and dining other side. Subs are behind curved 4 seat couch on each end about 10 feet apart. Both are SVS 3000 Micro's and are easily adjusted via phone app. Audiotools on IOS provides sweeps to stereo and measurements from listening position via phone microphone. First image is subs off, 2nd is closer to ideal but still fairly influenced by room nodes, and 3rd is preferred bass subjectively.

I have more adjustments to make, but I am wondering how common it is to prefer a bit more bass versus more flat?

 

[AdamG]

I added a second sub because sometimes I could localize the sound from the single sub. Once added that localization went away. Main speakers fire across a 12 foot deep room - 35 feet wide. Living room one side and dining other side. Subs are behind curved 4 seat couch on each end about 10 feet apart. Both are SVS 3000 Micro's and are easily adjusted via phone app. Audiotools on IOS provides sweeps to stereo and measurements from listening position via phone microphone. First image is subs off, 2nd is closer to ideal but still fairly influenced by room nodes, and 3rd is preferred bass subjectively.

View attachment 347551
View attachment 347552
View attachment 347553

I have more adjustments to make, but I am wondering how common it is to prefer a bit more bass versus more flat?

It’s pretty common to desire a bit more bass than flat. Most of the preference curves developed in Scientific Studies actually supports this bass boost curve. I run my subs about 4-5 dB hotter than the rest of the speakers. But in the end there is really no right or wrong here. They are all named preference curves because they are the result of people’s preferences. If you like more bass have at it. If you like less do that. Dial your audio system to fit your personal preferences and sit back and enjoy. Don’t worry too much about what others like. It’s your audio and it should be tailored for your tastes.

 

[Julf]

It’s pretty common to desire a bit more bass than flat.

Unfortunately all too true. I am having a hard time getting my car stereo not to boom horribly, as it seems to be designed for average consumer taste rather than transparency.