Understanding subwoofers

[john61ct]

Yes my take is, whether multi-way or separate enclosures does not fix a given dip caused by the room IF stereo co-located

Smoothing those out requires moving the boxen around, which usually implies mono bass

 

[kyuu]

That region between 80Hz - 400Hz is typically non-minimum-phase in most rooms and thus can not be equalized.

Eh? This would seem to contradict the notion of EQing out room mode peaks below Schroeder, unless I'm misunderstanding you.

I do, and they do indeed. Depending on the room and placement you will still get some irregularities in this area, but typically less than a conventional speaker. It's pretty much a given just by looking at the anechoic measurements / polar plots, as the reduced energy to the sides and rear will give less energy to the reflections that contributes to the cancellations in this area.

Another interesting aspect is that the response is suprisingly similar across different rooms. This is also evident when listening.

Here is an example from different rooms (combined average og left+right measurements), showing both slope and +/-3dB window so it's easier to see how much it deviates. Smoothing is 1/12th on all of these. All are different rooms.

Manta in my room
View attachment 530053

Manta in another room
View attachment 530054


Saranna, measurement from a customer room:
View attachment 530056


And to show that it's not magic, here is Saranna with a customer who insisted on having the speakers almost 1 meter from the wall, then we still got some reduction in energy around 150-300hz, and this is pre-EQ measurement so also clearly a huge peak at 50hz, but that was easily tamed by EQ. And beyond that still quite tidy:
View attachment 530061



No EQ above 100hz in any of these measurements.

It might just be confusion on my end, but this seems to show that cardioid bass does indeed have a benefit in reducing SBIR, which is the more common cause of issues in the ~100-200Hz area. This is reinforced by your showings of positioning the speaker close to the wall. This is unsurprising. I think a lot of people are thinking of room modes when they are talking about "room interaction", though, and given that all these measurements (except for the last one, which shows a big room mode peak) are equalized under 100Hz to address room modes these wouldn't seem to support the notion that the cardioid bass necessarily has a beneficial effect with regards to room modes in particular.

 

[Keith_W]

Eh? This would seem to contradict the notion of EQing out room mode peaks below Schroeder, unless I'm misunderstanding you.

Room mode peaks below Schroder are minimum-phase and can therefore be equalized. Please read what I said again - I said 80-400Hz. You MIGHT get some minphase regions up to 120Hz but this depends on the dimensions of the room. Above that it's mostly non minphase.

 

[ppierre]

which usually implies mono bass

So what? There is one bass drum in a concert, one organ in a church and record are mixed taking in account mono.

When length of low frequency approach the room size, we heard the room not the source.

Our brain is literally integrating the comportement of bass in closed space. Why loose our time with a chimeric stereo bass. Building a double bass array to address decay time is IMHO a greater priority.

 

[levimax]

So what? There is one bass drum in a concert, one organ in a church and record are mixed taking in account mono.

When length of low frequency approach the room size, we heard the room not the source.

Our brain is literally integrating the comportement of bass in closed space. Why loose our time with a chimeric stereo bass. Building a double bass array to address decay time is IMHO a greater priority.

While I agree with you on DBA there are recordings with two drum sets playing.

 

[JRowland]

there are recordings with two drum sets playing.

Just about every Grateful Dead recording ever made to start.

 

[sigbergaudio]

Is that compensated for in recordings or are we all listening to a 60 Hz dip when we listen to stereo?

60hz was a random example. It could be 30hz, 44hz, 97hz. It depends on the dimensions of the room, where the speaker is in the room, and where the listener is in the room.

 

[john61ct]

So what? There is one bass drum in a concert, one organ in a church and record are mixed taking in account mono.

When length of low frequency approach the room size, we heard the room not the source.

Our brain is literally integrating the comportement of bass in closed space. Why loose our time with a chimeric stereo bass. Building a double bass array to address decay time is IMHO a greater priority.

Of course, if you do not care about stereo bass, then my comment does not apply to you.

We all set our own priorities.

But, the point of envelopment is not the direct "sound stage" it is the feeling created by non-correlated REFLECTIONS

even by frequencies where there is no localisation. And this is usually preserved by keeping a stereo mix way down deeper than you may think.

Yes such immersion is a goal apparently difficult to achieve, but for some, that just makes it more compelling.

 

[sigbergaudio]

Eh? This would seem to contradict the notion of EQing out room mode peaks below Schroeder, unless I'm misunderstanding you.

The transition area is a frequency range where it may be possible to EQ, but you need to know what you're doing. The only "safe" place to do anything but broad shelf filters or very low Q notch filters is below this transition area (typically below 100-150hz). As a general rule I don't recommend that customers EQ our systems above 100-150hz, and I don't do it myself either.

It might just be confusion on my end, but this seems to show that cardioid bass does indeed have a benefit in reducing SBIR, which is the more common cause of issues in the ~100-200Hz area. This is reinforced by your showings of positioning the speaker close to the wall. This is unsurprising. I think a lot of people are thinking of room modes when they are talking about "room interaction", though, and given that all these measurements (except for the last one, which shows a big room mode peak) are equalized under 100Hz to address room modes these wouldn't seem to support the notion that the cardioid bass necessarily has a beneficial effect with regards to room modes in particular.

My post wasn't an attempt to prove beneficial effect with room modes below 100hz. That was not the question being asked. The Manta monitor itself actually is cardioid all the way down. But since it is crossed over to omni subwoofers by design, the system as a whole is cardioid to around ~80-100hz. The Saranna is cardioid to around ~130-150hz since the bass section is not cardioid. This means I do not have data to prove anything related to cardioid below 100hz.

I addressed the question I was asked, which was whether it would improve the ~80-400hz area, so the transition range that is often plagued by SBIR effects. The answer to that is yes.

 

[kyuu]

I addressed the question I was asked, which was whether it would improve the ~80-400hz area, so the transition range that is often plagued by SBIR effects. The answer to that is yes.

Okay, so the confusion was on my end as I wasn't sure how the question about "room interaction" was intended or being interpreted (referring specifically to SBIR, room modes, both, or something else I was unaware of).

 

[sigbergaudio]

Okay, so the confusion was on my end as I wasn't sure how the question about "room interaction" was intended or being interpreted (referring specifically to SBIR, room modes, both, or something else I was unaware of).

I may have fallen off the discussion at some point, but as far as I noticed, @Jaxx1138 talked about the 80hz+ range all along, so upper bass / lower midrange, not deep bass (as in below 80-100hz).

 

[Jaxx1138]

Eh? This would seem to contradict the notion of EQing out room mode peaks below Schroeder, unless I'm misunderstanding you.

We need to clarify "room-modes" and "SBIR" are two separate phenomena.

SBIR are known as quarter wave nulls; where a bounce from a wall causes destructive interference by sending a wave back into the direct sound 180 degrees out of phase. These tend to happen in the region of 80-400Hz approximately. They can occur from back or side walls.

Modes are resonant frequencies of an enclosed acoustic space, where sound waves reflect between boundaries and reinforce themselves to create standing waves. This creates high-pressure zones (nodes) and low-pressure zones (antinodes) that remain stationary in space. This will occur when the wavelengths begin to correspond to the dimensions of the room. These will happen below the Shroeder frequency or sometimes called the transition frequency.
Most common modes are ;
Axial Modes(two walls): The strongest and most problematic; these occur between two parallel surfaces ( front and back walls).
Tangential Modes(four walls): Sound reflecting off four surfaces ( four walls).
Oblique Modes(six walls): Sound reflecting off all six surfaces (walls, floor, and ceiling). These are generally weaker due to energy loss at each reflection.

 

[janbth]

My understanding is that a point source creates a single wave front that as it travels through air is inherently better in the time domain that seperated lobes like multiway speakers or separated subs. This is why coaxial drivers are so popular. All this stuff.is very subtle and trying to ABX colocated vs distributed for slam is more or less impossible. I can quick switch from IIR to Linear phase but it has a massive tell when going from IIR to linear phase so can't ABX blind but the switch is quick and they sound noticably different to me.

The problem is that the traveling wave analogy is not a good approximation below the transition frequency (Schroeder frequency). This is the modal region, where standing waves (room modes) may dominate sound propagation. The wavelengths of sound in this region is comparable to the dimensions of the room itself, which is why the location of subs is not relevant for sound localization - up to a point, of course. Things like imaging and sound stage is provided by the range of frequencies above the transition frequency. This is where coaxial drivers do their thing.

 

[janbth]

We need to clarify "room-modes" and "SBIR" are two separate phenomena.

SBIR are known as quarter wave nulls; where a bounce from a wall causes destructive interference by sending a wave back into the direct sound 180 degrees out of phase. These tend to happen in the region of 80-400Hz approximately. They can occur from back or side walls.

Modes are resonant frequencies of an enclosed acoustic space, where sound waves reflect between boundaries and reinforce themselves to create standing waves. This creates high-pressure zones (nodes) and low-pressure zones (antinodes) that remain stationary in space. This will occur when the wavelengths begin to correspond to the dimensions of the room. These will happen below the Shroeder frequency or sometimes called the transition frequency.
Most common modes are ;
Axial Modes(two walls): The strongest and most problematic; these occur between two parallel surfaces ( front and back walls).
Tangential Modes(four walls): Sound reflecting off four surfaces ( four walls).
Oblique Modes(six walls): Sound reflecting off all six surfaces (walls, floor, and ceiling). These are generally weaker due to energy loss at each reflection.

I agree with everything you say, but I believe «node» usually means low-pressure zone (nulls) and «antinode» means high-pressure zone.

 

[Jaxx1138]

I agree with everything you say, but I believe «node» usually means low-pressure zone (nulls) and «antinode» means high-pressure zone.

Sorry, and thank you for the correction.

 

[levimax]

The problem is that the traveling wave analogy is not a good approximation below the transition frequency (Schroeder frequency). This is the modal region, where standing waves (room modes) may dominate sound propagation. The wavelengths of sound in this region is comparable to the dimensions of the room itself, which is why the location of subs is not relevant for sound localization - up to a point, of course. Things like imaging and sound stage is provided by the range of frequencies above the transition frequency. This is where coaxial drivers do their thing.

OK but for bass/ mid bass I am trying to maximise the "impact" defined as dp/dt or another way the rate of pressure change at the LP. I believe it is settled science that a wave front from a point source has higher impact than a wave front from distributed sources. I also know the room messes up point source wave fronts but it also messes up distributed source wave fronts so it still makes sense that a point source wave front will have diminished but still greater impact than a distributed source wave front at the LP. Audibility is the hard part as ABX borders on impossible. I think my 2 "point source" system has better slam than when I had 4 subs but others report better slam with distributed subs, imagine that One thing I don't think is debatable, but I might be surprised, is that 2 colocated subs (or 2 full range speakers) are easier to set up and integrate than a distributed multisub system.

 

[thcdru2k]

Another option is deciding which nulls and SBIR cancellations are actually tolerable within the constraints you are working with, such as placement limitations or lack of access to systems like Dirac ART. If a null is outside the crossover region, relatively mild, or the SBIR cancellation is narrow and well-controlled, it may be preferable to accept it.

 

[levimax]

Another option is deciding which nulls and SBIR cancellations are actually tolerable within the constraints you are working with, such as placement limitations or lack of access to systems like Dirac ART. If a null is outside the crossover region, relatively mild, or the SBIR cancellation is narrow and well-controlled, it may be preferable to accept it.

Agreed, Psychoacoustic smoothing can be very helpful for deciding what to worry about and what to leave alone.

 

[kyuu]

Another option is deciding which nulls and SBIR cancellations are actually tolerable within the constraints you are working with, such as placement limitations or lack of access to systems like Dirac ART. If a null is outside the crossover region, relatively mild, or the SBIR cancellation is narrow and well-controlled, it may be preferable to accept it.

Indeed. I'm a sorry fool with just one measly subwoofer on the front wall (though I do at least have Dirac). There's definitely a null in there, but I have only noticed it in two circumstances: tooling through the bass region with a tone generator, and on one particular electronic song which has a descending bass note that happens to hit the null. Otherwise, it's completely unnoticeable. I could spend a lot of time, effort, and money adding a sub or three and integrating them to address it... or just leave it well enough alone since it's really a non-issue 99.99% of the time.

 

[pablolie]

From all the back and forth in this topic, I'd get the feeling no one really understands subwoofers. :-D

Probably the answer is that subs are versatile tools that can solve different problems - and we all have different priorities.

Do you want to listen very loud in a larger room but you want to keep using bookshelves? Or are you completely satified with your bookshelves just as they are, just want a little extra bass extension? Different priorities, different ways of using subs... and different ways of "understanding subs".