Stereo Bass using subwoofers

[abdo123]

No worries .

I only play around with sealed woofers which makes things a bit simpler, there are a few different methods and in my experience they all mostly line up.

1) Measure in enclosure T/S parameters with Dayton DATS. Specifically with QTC and Fc you can predict the 2nd order high pass behavior.
2) Measure acoustic response with mic very near to woofer dustcap.
3) Rely on measurements of others. For example Linkwitz has done outdoor / ground plane measurements of the LXmini indicating LR2 @ 60 Hz roll off. Dave Reite reproduced these when design the LXmini+ add on subs.

Once you understand the roll off you can also modify the roll off behavior using a Linkwitz transform and/or additional electrical high pass to obtain your desired roll off. For example I could use a Linkwitz transform to change the LR2 (Q = 0.5) @ 60 Hz roll off to BW2 (Q = 0.707) @ 60 Hz and then apply a BW2 @ 60 Hz electrical HPF and this would result in LR4 @ 60 Hz roll off.

The other variable to consider is that delay as a result of path length differences between the woofer and subwoofer will change phase response and impact how the crossover sums. Linkwitz discusses this here -> https://www.linkwitzlab.com/frontiers_5.htm. In general this analysis treats the subwoofer as just another driver that requires proper integration with the woofer. I also consider the high pass behavior of the subwoofer in the crossover design by applying a HPF equivalent to subwoofer rolloff to the woofer for better phase matching.

To me it is not so much that the crossover negatively impacts phase response it is more that if you don't start from a properly summed crossover in addition to the room you now have the added variable of frequency response issues created by the crossover. For modestly sloped crossovers phase distortion does not seem like a major issue, it is possible to use linear phase crossovers to eliminate the phase distortion but that requires a lot of processing power at low frequencies and has other drawbacks.

Of course YMMV, everyone has different rooms, speakers and listening preferences. I find integrating subwoofers as described above works well for me, it may not be the best solution for you.

Michael

Okay so apparently my mains follow a BW2 at 108 Hz, I know a BW crossover does not sum as flat but rather as +3dB at the crossover frequency. So what's the next step from here?

Also is it normal for a 6.5 inch woofer's HF roll-off to start so early? or am i missing something here?

 

[pjug]

Okay so apparently my mains follow a BW2 at 108 Hz, I know a BW crossover does not sum as flat but rather as +3dB at the crossover frequency. So what's the next step from here?

Also is it normal for a 6.5 inch woofer's HF roll-off to start so early? or am i missing something here?

View attachment 155498

[Edit: Oops, you said HF rolloff. New measurement below. The SB is in the 3-way BMR though, so the crossover rolls it off]
A couple speakers with 6 inch drivers I have for comparison. The Usher is a little larger, maybe 6.5 inches. I am surprised how similar the LF roll off is between these two.

I wouldn't mind if you start a new thread on integrating subs with mains running full range. It is an interesting subject.

 

[abdo123]

I wouldn't mind if you start a new thread on integrating subs with mains running full range. It is an interesting subject.

I've been planning on doing a mega subwoofer integration Guide for this forum since it's such a repeatedly discussed topic. However every few weeks something comes up that pushes me to take my own integration to 'the next level'.

If i would turn back time I would probably buy LS50 Meta and seal their ports, their sublime directivity and natural high LF cuttoff (5.5 inch woofer) would probably make it the best speaker out there for use with a subwoofer.

 

[Tangband]

I've been planning on doing a mega subwoofer integration Guide for this forum since it's such a repeatedly discussed topic. However every few weeks something comes up that pushes me to take my own integration to 'the next level'.

If i would turn back time I would probably buy LS50 Meta and seal their ports, their sublime directivity and natural high LF cuttoff (5.5 inch woofer) would probably make it the best speaker out there for use with a subwoofer.

Why is a 5,5 inch woofer the best to use when crossing to a subwoofer ? Its gonna distort in the bass already at 200 Hz if you play a little loud. Its to small to play down to 80 Hz . And crossing higher, for example at 150 Hz would make the two stereo -subwoofer placement critical, where you have to place the two stereo subwoofers very near the main speakers ?
A 6,5 inch woofer would do the job much better I think.

Every loudspeaker has a ”natural” roll off, - a closed box is 12 dB/oct and a bass reflex construction is 24 dB/oct.
But in a real room you have room gain and the closed box loudspeakers may be almost flat in frequency response deep down i the bass. Instead off dropping off 12/dB oct theoretical, it can be 5 dB/oct in a real room.

The crossover frequency and the distance between subwoofers and main loudspeakers is closely connected. At 80 Hz or lower you are way below the rooms schroeder frequency and you get a bit more freedom to place the subwoofers . Crossing higher in frequency and you must follow certain rules.

A good rule is : The distance between subwoofer and main speaker should be less than 1/4 wavelengt of the crossover frequency. At 100 Hz its only 85 cm.….

The picture shows correct placement of two stereo subwoofer placed near the main loudspeakers. The distance is short enough when crossing at 100 Hz ( red arrows ) .
Using mono-coupled subs in the same way gonna have

several phase distortion because of the distance between the right mono-subwoofer and the left main loudspeaker.

 

[Soundmixer]

heres is much confusion here.
What is music? Its not sine-sweeps or a tone at a single frequence. Music is ordered sounds in the time/frequency domain.

One simple explanation:

An electric bass plays frequencies from about 41 Hz to 8 KHz . Each tone the bass player plays has information from about 41 Hz and up to 8 KHz . If you change the spectra at 5 kHz just a couple of Hz, the sound from the electrical bass will start to sound slightly out of tune. The perceived pitch of the tunes the bass player plays, has become worse.

I am not sure what confusion you are referring to, but I am not confused about this at all. This comment doesn't even come close to the topic at hand. We are strictly talking about BASS frequencies.

After counting , is the placement of the two subwoofers important ? Should the subwoofers be placed near the main speakers, or far away ? Maybe you´re now saying it dont make any sense because you´re crossing at 80 Hz - a much lower frequency than 5 kHz ? But hear this….

I am baffled as heck. I don't know how this is related to anything I have said. Are you confusing me with someone else?

There is no such thing as a perfectly steep crossover, because the subwoofer itself is making noise above the crossover region. Always. Its mecanical distortion from the drivers and the loudspeakerbox AND the room . The last thing is very important to understand.
A subwoofer placed against a wall thats not very rigid is gonna distort the wall- and its very audible. The floor also distort the sound.

If that wall is poorly built, you are correct. Two of my four subs are in the rear corners and I cannot hear the wall vibrate. They have subwoofer platforms and spikes to lift your sub off the ground. Your statement is a worse possible case scenario that not very many folks will experience.

When one understands this , its clear that two subwoofers in stereo placed near the main speaker is the best option, if we are talking two channel sound. This is also very easy to verify with listening tests with real music.

What if the wall behind the two stereo speakers vibrates? Wouldn't that be audible as well? Also, there are distance tools in DSP software, in AVR's and pre-pro's, and some two-channel receivers/streamers that allow you to electronically align your speakers to get the proper time alignment of the mains and subs so you can put your mains in the best possible location for imaging, and place your sub(s) where it can get the best bass response. It is not necessary to place the subs right next to the mains in light of this. Just another perspective.

 

[GimeDsp]

I have yet to see any test/paper that was designed properly to examine this issue. And Have read Welti and Toole.

As discussed here the test to prove LF can't be localized is near impossible.

1.To localize subs you need sufficient samples
2. To find sufficient sound samples you prove LF can be localized.

In such a test like this one must first go to the extremes and see

1. Is it POSSIBLE to localize LF
then
2. Is it probable and under what conditions.

 

[Kvalsvoll]

The picture shows correct placement of two stereo subwoofer placed near the main loudspeakers.

Then I hope they are configured mono, because from what can be seen in the picture, it is unlikely both subwoofers have a similar frequency response.

 

[Kvalsvoll]

I have yet to see any test/paper that was designed properly to examine this issue. And Have read Welti and Toole.

As discussed here the test to prove LF can't be localized is near impossible.

1.To localize subs you need sufficient samples
2. To find sufficient sound samples you prove LF can be localized.

In such a test like this one must first go to the extremes and see

1. Is it POSSIBLE to localize LF
then
2. Is it probable and under what conditions.

Here is a mixture of panned mono drums, then stereo drums.

Panning mono'd drums hard left and hard right shows the extreme of non stereo localization.

Most decorrelated LF content you find will not be "hard-panned" it will be things like timpani where the attack is more centered and the room sound/sustain dominated the decorrelated content.

Below is the sample I found of drum kit.

It is possible to test for this, and it has been done, and the results are that yes, it can be possible to detect direction to some degree at low frequencies down to around 60Hz - below which it becomes impossible due to how hearing actually works. BUT - in all practical situations in a practical sound system, all directional and spatial information that hearing uses to determine location and perceived size of a sound object comes from higher frequencies. It is the harmonics and the leading edge of transients - which all are higher frequency - that is important for location.

To do a proper test, a controlled listening experiment must be designed, using test signals made for this exact purpose.

Bass in small rooms can be challenging, but with current knowledge it is possible to solve most of the issues, and get a quite decent result. Focusing on properties that matters most for bass sound quality and leaving what is not significant makes it much easier.

 

[GimeDsp]

I found some other sample videos I'll post later.

Truth is we never localize a sound only by its LF content.

 

[DonH56]

Re. placement: Every AVR I have used for the past twenty or so years and my current subs as well as others have delay and/or phase adjustment to allow one to place them where they work best, mono or stereo. For that matter, I included an (analog) all-pass filter in my DIY servo sub back around 1980-something to allow me to integrate the sub with the mains irrespective of where they were placed.

 

[abdo123]

Why is a 5,5 inch woofer the best to use when crossing to a subwoofer ? Its gonna distort in the bass already at 200 Hz if you play a little loud. Its to small to play down to 80 Hz . And crossing higher, for example at 150 Hz would make the two stereo -subwoofer placement critical, where you have to place the two stereo subwoofers very near the main speakers ?
A 6,5 inch woofer would do the job much better I think.

Every loudspeaker has a ”natural” roll off, - a closed box is 12 dB/oct and a bass reflex construction is 24 dB/oct.
But in a real room you have room gain and the closed box loudspeakers may be almost flat in frequency response deep down i the bass. Instead off dropping off 12/dB oct theoretical, it can be 5 dB/oct in a real room.

The crossover frequency and the distance between subwoofers and main loudspeakers is closely connected. At 80 Hz or lower you are way below the rooms schroeder frequency and you get a bit more freedom to place the subwoofers . Crossing higher in frequency and you must follow certain rules.

A good rule is : The distance between subwoofer and main speaker should be less than 1/4 wavelengt of the crossover frequency. At 100 Hz its only 85 cm.….

The picture shows correct placement of two stereo subwoofer placed near the main loudspeakers. The distance is short enough when crossing at 100 Hz ( red arrows ) .
Using mono-coupled subs in the same way gonna have View attachment 155567several phase distortion because of the distance between the right mono-subwoofer and the left main loudspeaker.

the thing is a subwoofer with a BW2 low pass at 108 Hz will already have meaningful output till ~500Hz or so. So the important question becomes whether you want the perfect speaker or the speaker with more headroom. In my particular situation I rather have small speakers with cleaner design than extra headroom. People might have different conclusions.

 

[GimeDsp]

To do a proper test, a controlled listening experiment must be designed, using test signals made for this exact purpose.

Here is where this becomes a self-refuting idea.

For a test signal to be sufficient to locate LF below 60hz you have to prove that it can be used to locate LF below 60hz

So in order to prove the signal is sufficient and valid for the purpose you refute any idea opposed to the facts.

What people are missing is this. Sounds span many octaves and our brain never processes them in isolation. So a break down like this could still be "averaged" by the person to provide accurate location.

Floor tom
1. attack-50hz,150, 200hz the "thwack" is easily localized
2 . Major resonance(80-150hz) this is where the response blooms, the "body" of the drum and is still localized
3 . Lesser sustain-this can be the lowest LF content of the sound as the drum head/drum body resolves and drops in pitch.

So even if percentage of localization looks like this

1. 80%
2. 18%
3. 2%

Just because the lowest part of a complex sound is only localized to 2% doesn't mean it doesn't add just enough information to solidify the sound stage.

But regardless of the mechanics and theory, I have pointed out the self-refuting claim countless times.
Proving LF can't be localized at any frequency is impossible, it's akin to proving a negative. You would literally have to do a controlled test with EVERY known sound in existence to prove that non were sufficient, and even at that you could only make the claim that "with all known signals LF has not been localized."

Here is a demonstration of the exact same logic,

"We have not located any pink unicorns"

"Do you have the tools needed to locate pink unicorns"

"Well sir, we don't know since we've never located any with our tools."

 

[sigbergaudio]

Why is a 5,5 inch woofer the best to use when crossing to a subwoofer ? Its gonna distort in the bass already at 200 Hz if you play a little loud. Its to small to play down to 80 Hz . And crossing higher, for example at 150 Hz would make the two stereo -subwoofer placement critical, where you have to place the two stereo subwoofers very near the main speakers ?
A 6,5 inch woofer would do the job much better I think.

Are these your assumptions, or did you actually test? A 5.5" driver would need less than 1.5mm excursion to play 100dB at 200hz.

 

[Kvalsvoll]

Here is where this becomes a self-refuting idea.

For a test signal to be sufficient to locate LF below 60hz you have to prove that it can be used to locate LF below 60hz

So in order to prove the signal is sufficient and valid for the purpose you refute any idea opposed to the facts.

What people are missing is this. Sounds span many octaves and our brain never processes them in isolation. So a break down like this could still be "averaged" by the person to provide accurate location.

Floor tom
1. attack-50hz,150, 200hz the "thwack" is easily localized
2 . Major resonance(80-150hz) this is where the response blooms, the "body" of the drum and is still localized
3 . Lesser sustain-this can be the lowest LF content of the sound as the drum head/drum body resolves and drops in pitch.

So even if percentage of localization looks like this

1. 80%
2. 18%
3. 2%

Just because the lowest part of a complex sound is only localized to 2% doesn't mean it doesn't add just enough information to solidify the sound stage.

But regardless of the mechanics and theory, I have pointed out the self-refuting claim countless times.
Proving LF can't be localized at any frequency is impossible, it's akin to proving a negative. You would literally have to do a controlled test with EVERY known sound in existence to prove that non were sufficient, and even at that you could only make the claim that "with all known signals LF has not been localized."

Here is a demonstration of the exact same logic,

"We have not located any pink unicorns"

"Do you have the tools needed to locate pink unicorns"

"Well sir, we don't know since we've never located any with our tools."

It is possible to make suitable test signals for this, and the test is not very difficult to do. But it requires some work in the preparation, and then the test must be executed - requiring equipment and personnel.

But there is already sufficient knowledge on bass and localisation, so why use resources on such a test. Nothing new will be discovered, no new knowledge will be learned.

 

[GimeDsp]

It is possible to make suitable test signals for this, and the test is not very difficult to do. But it requires some work in the preparation, and then the test must be executed - requiring equipment and personnel.

But there is already sufficient knowledge on bass and localisation, so why use resources on such a test. Nothing new will be discovered, no new knowledge will be learned.

I have read Welti and he was a bust. Please post a study if you make a claim based on a study.

I am sorry you dont seem to understand logic. For a conclusion to be true its logic must be valid regardless of so called data.

The validity of the test is predicated on the sufficiency of the samples for the intended purpose. As such, in order to prove the samples sufficancy you also prove that LF can be located at such frequencies with said signals.

Self refuting idea.

 

[mdsimon2]

Okay so apparently my mains follow a BW2 at 108 Hz, I know a BW crossover does not sum as flat but rather as +3dB at the crossover frequency. So what's the next step from here?

Also is it normal for a 6.5 inch woofer's HF roll-off to start so early? or am i missing something here?

View attachment 155498

Roll off will depend on the driver Qts, Fs and Vas as well as the size of your enclosure. This is one of the beauties of DIY, you can understand the implications of different enclosure sizes for your particular driver and have a good idea of what EQ you need to achieve your desired roll off.

You can use a Linkwitz Transform to lower Fc and change Qtc but the trade off is you need to send more power to the speaker so you may reach amplifier and/or speaker limits.

The roll off you describe is relatively normal. For example a Dayton RS180-8 (6” woofer) will have a Fc of about 100 Hz at Qtc of 0.707.

I’ve attached some papers from Richard Small that talk about sealed enclosure analysis and design for reference. There are a variety of calculators on the internet that will calculate Fc and Qtc for given T/S parameters and enclosure size.

Michael

 

[abdo123]

Roll off will depend on the driver Qts, Fs and Vas as well as the size of your enclosure. This is one of the beauties of DIY, you can understand the implications of different enclosure sizes for your particular driver and have a good idea of what EQ you need to achieve your desired roll off.

You can use a Linkwitz Transform to lower Fc and change Qtc but the trade off is you need to send more power to the speaker so you may reach amplifier and/or speaker limits.

The roll off you describe is relatively normal. For example a Dayton RS180-8 (6” woofer) will have a Fc of about 100 Hz at Qtc of 0.707.

I’ve attached some papers from Richard Small that talk about sealed enclosure analysis and design for reference. There are a variety of calculators on the internet that will calculate Fc and Qtc for given T/S parameters and enclosure size.

Michael

Thank you so much for sharing the documents, i will take a look at it.

Silly me thought that the measurement is the 'anechoic' bass response of the whole speaker.

 

[Tangband]

Are these your assumptions, or did you actually test? A 5.5" driver would need less than 1.5mm excursion to play 100dB at 200hz.

Thats true, but if the crossover slope is just 6 or 12 dB/oct at 200 Hz , theres gonna be much energi even at 100 Hz.
And much worse, letting the 5,5 inch woofer roll of naturaly without HP filtering its gonna have severe distortion at 100 Hz. Even if the 5 1/2 driver manages to play music with 1,5 mm excursion, at 200 Hz, it gonna do it with more distortion than an equally good 6 1/2 inch woofer.

If you look at Amirms loudspeaker measurements regarding distortion, you can se that all speakers with 5 1/2 inch woofer has high distortion numbers lower than 150- 200 Hz if playing at 96 dB.

Using subwoofers at floor level is only usable up to about 100 Hz anyway - my experience. Crossing higher , you have to build the subwoofer into the same cabinet as the mid-woofer.

( Edit : a three way active loudspeaker containing subwoofers is a bigger compromise than a 4-way, but if you cross it at 80 Hz and 2 kHz the problems are manageable, using for example a 12 inch subwoofer, 6 1/2 inch woofer and a tweeter with a waveguide.
An active loudspeaker with a 5 inch mid- woofer would have to have the subwoofer built in to the same cabinet for high spl and short distance between subwoofer and midwoofer, crossing at 125 -250 Hz instead of 80 Hz - my opinion.)

 

[abdo123]

Thats true, but if the crossover slope is just 6 or 12 dB/oct at 200 Hz , theres gonna be much energi even at 100 Hz.
And much worse, letting the 5,5 inch woofer roll of naturaly without HP filtering its gonna have severe distortion at 100 Hz. Even if the 5 1/2 driver manages to play music with 1,5 mm excursion, at 200 Hz, it gonna do it with more distortion than an equally good 6 1/2 inch woofer.

If you look at Amirms loudspeaker measurements regarding distortion, you can se that all speakers with 5 1/2 inch woofer has high distortion numbers lower than 150- 200 Hz if playing at 96 dB.

Using subwoofers at floor level is only usable up to about 100 Hz anyway - my experience.

I would be all in favor of a no compromise design, unfortunately coaxials don't do well with 6.5 inch woofers.

There was no positive correlation between low distortion and listener preference when Dr. Olive did the research for the preference model, so when you're given the choice of smooth directivity and listening window vs distortion you should always pick the earlier unless the distortion is unrealistically high due to distance from speakers .etc

 

[GimeDsp]

It is possible to make suitable test signals for this, and the test is not very difficult to do. But it requires some work in the preparation, and then the test must be executed - requiring equipment and personnel.

But there is already sufficient knowledge on bass and localisation, so why use resources on such a test. Nothing new will be discovered, no new knowledge will be learned.

Im not trying so sound like a know it all.
I've spent tbe last 8 weeks almost full time setting up studio systems with stereo/mono subs and gathering data for symmetrical and asymmetrical rooms. We have used numerous music and test signals and have noted stereo subs have a large impact on sound location.

When you say "studies" you have to cute the study, otherwise its worthless.

The Welti study fails on all accounts.
Is you have a legit study, please post it.