If soongsc has a large head, then he can probably hear group delay at low frequencies better than we can. 
</quip>
Here's what I still don't understand: These room modes exist, and now we're trying to excite the room in such a way that these modes are naturally cancelled. Is there any reason why multiband parametric equalization cannot be used to achieve the same effect?
Whether we start with one subwoofer or many, we're still exciting resonant modes of the room, though with multi-sub schema, we're simply using some resonant energy to cancel out other resonant energy; we're not in effect making the room non-resonant -- it's still a big box of air with surfaces that continue to be very reflective at low frequencies, despite extensive treatments to prevent reverberation at upper frequencies.
Would it be better simply to approximate the rise time of a particular room mode through calculations and then apply a DSP algorithm that decreases the subwoofer's output vs. time as the room mode is excited? Perhaps this could be testing with a rack-mount pro audio compressor unit set for a very long 'attack' phase, with the gain reduction ratio set to equal the level of energy contributed by the room mode? The ideal device would be a multi-band parametric equalizer with a compressor that also activates within the defined band -- which may be similar in its effect to the method developed by EAW's Dave Gunness for compensating for time smear in typical prosound loudspeakers. Ideally, a single DSP chip would be programmed to do this. However, something similar could be kludged together from several cascaded 4-in/8-out pro audio DSP units, with their outputs routed through multi-channel compressor units before they enter the amplifier.
</quip>
Here's what I still don't understand: These room modes exist, and now we're trying to excite the room in such a way that these modes are naturally cancelled. Is there any reason why multiband parametric equalization cannot be used to achieve the same effect?
Whether we start with one subwoofer or many, we're still exciting resonant modes of the room, though with multi-sub schema, we're simply using some resonant energy to cancel out other resonant energy; we're not in effect making the room non-resonant -- it's still a big box of air with surfaces that continue to be very reflective at low frequencies, despite extensive treatments to prevent reverberation at upper frequencies.
Would it be better simply to approximate the rise time of a particular room mode through calculations and then apply a DSP algorithm that decreases the subwoofer's output vs. time as the room mode is excited? Perhaps this could be testing with a rack-mount pro audio compressor unit set for a very long 'attack' phase, with the gain reduction ratio set to equal the level of energy contributed by the room mode? The ideal device would be a multi-band parametric equalizer with a compressor that also activates within the defined band -- which may be similar in its effect to the method developed by EAW's Dave Gunness for compensating for time smear in typical prosound loudspeakers. Ideally, a single DSP chip would be programmed to do this. However, something similar could be kludged together from several cascaded 4-in/8-out pro audio DSP units, with their outputs routed through multi-channel compressor units before they enter the amplifier.
A short impulse like a gun shot is a wide band signal with content all across the frequency spectrum.Pan said:
We typically separate the room's response into two regions: the region where modes are so dense that their combined reverberation is a better description, and the bass region where discrete modes are problematic. Unless your acousticians gun is a cannon for the 1812 overture it is mostly exciting the room above the schroeder frequency, that is the region where we only take a statistical view and characterize reverberation.
Another thing I don't get is the idea of using dissimilar subwoofers to achieve the desired room loading effect. Wouldn't this simply add even more variability to the real application that we'd need to account for with theory?
If you use a combination of vented enclosures and sealed enclosures, you get a system that suffers to some extent from the drawbacks of both enclosure types in terms of their LF extension and time domain performance -- you can't just use a sealed sub and a vented sub in the same space and expect to get only the best of both worlds. And by using a 'spread spectrum' approach to reflex enclosure tuning, you only incur the time-domain penalties of vented enclosures across a range of frequencies instead of at one frequency.
It seems to me that it would be far better to simply standardize on one type of subwoofer that provides the desired time-domain characteristics, and then equalize its output (by now, we're talking about an "EAS" - electronically assisted sealed - design or a number of infinite baffle subwoofers) until the desired low frequency response criteria are met, and then use a distributed setup comprised of only that type of subwoofer.
To me, mixing subwoofer types in the same setup by choice instead of by necessity falls under the heading of "D'oh!" - by using the same type of subwoofer at all locations, we keep things simple enough that we can better predict the overall performance of the system. Don't go dragging your 10" Dayton cheapie out of the closet just yet.
If you use a combination of vented enclosures and sealed enclosures, you get a system that suffers to some extent from the drawbacks of both enclosure types in terms of their LF extension and time domain performance -- you can't just use a sealed sub and a vented sub in the same space and expect to get only the best of both worlds. And by using a 'spread spectrum' approach to reflex enclosure tuning, you only incur the time-domain penalties of vented enclosures across a range of frequencies instead of at one frequency.
It seems to me that it would be far better to simply standardize on one type of subwoofer that provides the desired time-domain characteristics, and then equalize its output (by now, we're talking about an "EAS" - electronically assisted sealed - design or a number of infinite baffle subwoofers) until the desired low frequency response criteria are met, and then use a distributed setup comprised of only that type of subwoofer.
To me, mixing subwoofer types in the same setup by choice instead of by necessity falls under the heading of "D'oh!" - by using the same type of subwoofer at all locations, we keep things simple enough that we can better predict the overall performance of the system. Don't go dragging your 10" Dayton cheapie out of the closet just yet.
Taterworks said:
Assuming the LF system is minimum-phase (Pan pointed this out earlier and I've read this many times before) then yes, you can use a parametric EQ to fix the response. But this approach is limited to one listening position only because response is very different around the room! This is part of the motivation for multiple subwoofers in Welti's paper. Multiple subwoofers can reduce seat-to-seat variation considerably and enable successfull usage of EQ to flatten the response for selected listening area.
Taterworks said:
Parametric equalization and signal processing will work well for a single listener or a small listening position. Multiple subwoofer are needed to deliver similar bass experiences to multiple listeners. Period.
Taterworks said:
I think the key is that spatially averaged measurements are an integral part of Geddes' approach in setting up the subwoofers, whichever ones are chosen.
The interactions of modes vary throughout the room. You could equalize for a single position but it becomes less effective for regions. For example, if you add a FIR/IIR filter to compensate for a particular mode, then at that mode's peak regions you'll have improved things. However at the nulls of the mode the correction signal will be at a higher level than the mode itself, potentially making the correction signal audible as an aberration.Taterworks said:If soongsc has a large head, then he can probably hear group delay at low frequencies better than we can.
</quip>
Here's what I still don't understand: These room modes exist, and now we're trying to excite the room in such a way that these modes are naturally cancelled. Is there any reason why multiband parametric equalization cannot be used to achieve the same effect?
This is why Welti et all focus on placements that reduce variation over the listening area, and then use dsp to correct the average response across the listening area.
The exception to this would be wavefront reconstruction techniques that use arrays of loud speakers. Here we avoid the spatial problem because we can recreate the actual wavefronts of the modal wave. This sort of processing is perhaps the future of audio, but is somewhat impractical owing to the requirement to lots of channels each with individualized signal processing dependent on the layout of the room. The double bass array concept mentioned here recently is a clever simplification of this approach for bass frequencies in rectangular rooms.
I tire of semantic debates. Transients have bandwidth, you can't make blanket statements without referring more specifically to what you're talking about. A gunshot will clearly have little to no content in the subwoofer region.
Transient response is a property of a system, one that modes in the subwoofer frequencies have minimal impact on.
Transient response is a property of a system, one that modes in the subwoofer frequencies have minimal impact on.
As for EQ'ing room response. Yes it works to some degree since the EQ will be a inverse function more or less of the room resonance which will correct both time and frequency. No compressor is needed for this.
Problem with EQ is that you can not get the nulls up by applying gain. The amplifiers will be severly taxed and loudspeaker distortion will be excessive. And also you will get horrible peaks at other locations than the one you are correcting for.
The only way to avoid nulls is the use of multiple sources or absorbtion in the room. This will reduce/eliminate the build up of the resonances in the first place.
Practicall experience has shown that a combination of multiple sources, absorbers* and some EQ gives excellent results.. or so I read in a old suspect paper in some foreign language that mysteriously went up in smoke jsut as I had finsihed reading it..
*Absorber can be tube traps, helmholtz traps, panel/membrane absorbers or of course suitable design of the elements that makes up the listening room itself.
Someone did critique my approach since "four subs can't deal with all modes" or something like that, well we don't necessarily need "perfect" results, just good enough to get the systems performance up to some level. Some modes can be avoided by adjusting teh listening positions for example and again, the last bit may be EQ'd.
It's also worth to mention that the peaks (as always) is much more annoying than nulls. So by only using an EQ with a standard set up, taking down the extreme peaks will give much improved results.
/Peter
Problem with EQ is that you can not get the nulls up by applying gain. The amplifiers will be severly taxed and loudspeaker distortion will be excessive. And also you will get horrible peaks at other locations than the one you are correcting for.
The only way to avoid nulls is the use of multiple sources or absorbtion in the room. This will reduce/eliminate the build up of the resonances in the first place.
Practicall experience has shown that a combination of multiple sources, absorbers* and some EQ gives excellent results.. or so I read in a old suspect paper in some foreign language that mysteriously went up in smoke jsut as I had finsihed reading it..
*Absorber can be tube traps, helmholtz traps, panel/membrane absorbers or of course suitable design of the elements that makes up the listening room itself.
Someone did critique my approach since "four subs can't deal with all modes" or something like that, well we don't necessarily need "perfect" results, just good enough to get the systems performance up to some level. Some modes can be avoided by adjusting teh listening positions for example and again, the last bit may be EQ'd.
It's also worth to mention that the peaks (as always) is much more annoying than nulls. So by only using an EQ with a standard set up, taking down the extreme peaks will give much improved results.
/Peter
jason_watkins said:
Well that's exactly my point in my response to Janneman! ;-)
If not transients is applicable to woofers one might 1) wonder why the whole business is talking about transient response of various LF alignments and 2) where we draw the limit for transient vs. non-transient.
Shoud we say 179.2Hz? ;-)
/Peter, not very found of semantics either!
Pan said:
Peter
IMO you are a major source of the problem here. As I have pointed out before, I will not respond to your posts because of your attitude. You could start by not tying to push the problem onto other people and learn to show some respect. Of all the people who post here it is you that I find to be the most inaccurate and misleading and yet you speak with such authority. I have trouble resolving this in my mind. Either I misunderstand the situation or you do. I'll leave that for you and the other readers to sort out.
As to the use of multiple subs, Youngho and a few others have it exactly right
There is not much that I can add here except that if you don't agree with the above statement then you misunderstand something.
youngho, I guess you misunderstood my comments.
I should have written "rate" instead of "time". If you switch of the excitation signal all modes will still be there. They will decay at the same rate independend of the excitation signal's level.
The modal field is an attribute of a room's physical properties. You can couple energy to modes in different ways so the frequency response becomes smooth but you can't get rid of modes (without absorption).
I'm always looking at ALL low frequency modes up to let's say 100 Hz. If you "cancel" one mode you will excite another one.
Talking of "cancellation" and Toole, take a look at the picture on page 221. When you place a sub at a null, that particular mode is "cancelled". But that means that we will hear nothing at all. I don't think that this is what we want. What we want is to smooth the whole low frequency band for a given listening position or area.
Best, Markus
P.S. Still searching for the journal describing practical methods to achieve cold fusion and world peace.
P.P.S. I have the feeling that some people do not know that modes can be found in the whole audible frequency band and not only at low frequencies. In fact there are thousands of modes above the Schroeder frequency.
The number of modes in an idealized room measuring 4m x 5.2m x 2.4m:
youngho said:
I should have written "rate" instead of "time". If you switch of the excitation signal all modes will still be there. They will decay at the same rate independend of the excitation signal's level.
The modal field is an attribute of a room's physical properties. You can couple energy to modes in different ways so the frequency response becomes smooth but you can't get rid of modes (without absorption).
I'm always looking at ALL low frequency modes up to let's say 100 Hz. If you "cancel" one mode you will excite another one.
Talking of "cancellation" and Toole, take a look at the picture on page 221. When you place a sub at a null, that particular mode is "cancelled". But that means that we will hear nothing at all. I don't think that this is what we want. What we want is to smooth the whole low frequency band for a given listening position or area.
Best, Markus
P.S. Still searching for the journal describing practical methods to achieve cold fusion and world peace.
P.P.S. I have the feeling that some people do not know that modes can be found in the whole audible frequency band and not only at low frequencies. In fact there are thousands of modes above the Schroeder frequency.
The number of modes in an idealized room measuring 4m x 5.2m x 2.4m:
An externally hosted image should be here but it was not working when we last tested it.
markus76 said:
Markus, if you drive woofers positioned for maxium cancellation of a standing wave between two surfaces the mode will NOT be there, it is NOT excited.
Again that is not correct. When you use multiple suibs you can view the room as being divided into several new small rooms.. sort of. Makes sense?
You can not get smooth frequency response wihtout improvement in the time domain. Time domain and frequency domain are the same thing looking from two different angles when we deal with minimum phase devices.
Nope! With a corner placement of a single sub you excite all axial modes, by moving the sub to a null you side step the excitation of that particular mode but do not excite a new mode that was not there before. Using multiple subs is taking this further.
And that was the "essential" part I mentioned earlier. That's not the way it works. How do you think we can hear deep bass in headphones? In a car? Outdoors?
Placing a sub in a null does not mean there will be no output, it only means we will free ourselfs from that particular mode.
P.S. Still searching for the journal describing practical methods to achieve cold fusion and world peace.
I burned it, would cause to much of a buzz and confusion!
Not only in the audible band but as far up as you can count. The thing is that what we are most sensitive about is modal density. When there's a wide gap between the modes the resonances stand out much more than if they are tightly packed. That's the reason the first axial modes is the most important ones to fight.
/Peter
markus76 said:
Yes, rate and time are different. The rate of decay for a bass mode is essentially the same as for other frequencies in the bass region, so the way it was written was confusing.
I guess I interpreted "modal field", which was otherwise undefined at the time, as the complex frequency-dependent pressure distribution field determined by the room dimensions and shape and stimulated by a sound source. Yes, you're right. I was taking the Zen position: if a mode is not stimulated or if it is cancelled, does it exist for the listener?
I think what you're trying to write is that if you cancel some modes without addressing others, the remaining modes become more obvious and problematic?
That is incorrect. If you place a subwoofer at the node for a mode and have it play the frequency for that mode, you will hear that frequency, but the mode will not be activated. The relative level will be low because it's not reinforced by standing waves, but you will still hear it. But I agree, one would like "to smooth the whole low frequency bandwidth for a given listening position or area as much as possible."
P.S. Still searching for the journal describing practical methods to achieve cold fusion and world peace.
I'm glad the joke was appreciated.
I'm sorry if I have been focussed on modes in the bass region. The thread title is about multiple small subs, so I was only thinking about modes that the subs would be energizing. Unfortunately, users may perceive many problems with bass reproduction due to modes above 80 Hz and below the transition frequency/region causing issues with harmonics and timbre. However, this isn't a problem with the setup of subwoofer(s) itself.
Anyway, I've learned a few things reading this thread and participating a little. Thanks to all, no semantics for me, happy listening, and have a great weekend!
I still only see the single mode/node discussion as being misleading. Even at a node a source will excite other nearby modes since all modes are excited by all frequencies to some extent. Then there is the direct sound, but careful study will show that this direct sound is in fact the contribution of all the modes (Welti got this wrong in his paper). This goes back to something that I said a long time ago that nobody accepted. Free space has to be thought of as a continuum where the modal density goes to infinity - not zero. Hence, even outdoors in free space the direct field is carried by the modes (which are now infinite in density) just as it is in a small room. But in a small room the modes get sparse and hence the ability for them to carry the energy goes down.
The sound from a LF source does not - let me repeat - does not travel in all directions away from the source. (There is what is called an evanescent wave sent out, but this disapates in time and space exponentialy so it is a very small factor. The sound wave can travel only in a discrete number of directions defined by the modes that it excites. This means that the energy emitted by this source in a real room is not the same as the energy emmitted by this source in free space. This can be seen in the radiation impedance for a small room which is NOT the same as that for a source in free space.
People want to think that you take a free space source and bring it into a room and that it emites sound the same way, but that this gets amplified by the modes at certain frequencies. This is not correct. The presence of the room changes everything and not until the source sees a high model density does it begin to behave as it does in free space.
The sound from a LF source does not - let me repeat - does not travel in all directions away from the source. (There is what is called an evanescent wave sent out, but this disapates in time and space exponentialy so it is a very small factor. The sound wave can travel only in a discrete number of directions defined by the modes that it excites. This means that the energy emitted by this source in a real room is not the same as the energy emmitted by this source in free space. This can be seen in the radiation impedance for a small room which is NOT the same as that for a source in free space.
People want to think that you take a free space source and bring it into a room and that it emites sound the same way, but that this gets amplified by the modes at certain frequencies. This is not correct. The presence of the room changes everything and not until the source sees a high model density does it begin to behave as it does in free space.