How close to the listener is "near field" here, in your experience?
That seems like an easy thing to play with in a desktop system.
For YOU. And maybe for Earl and others. It is like that in the real world for ME. I hear the direction of bass, so I like stereo bass for playback. If you don't - all the better! You don't have to worry about stereo bass.
I assume the floor and ceiling bounce to come along the same vertical plane as the direct sound. Then the ceiling bounce may mask (horizontal) location cues, but it does not point to a different horizontal direction. If you talk about a slanting ceiling, there would be a problem, yes.
I find that intriguing . . . you should submit yourself for experiment somewhere . . . some grad student in psych could get a thesis out of finding the perceptual mechanism (and then those of us who don't could perhaps learn to). I do know that for myself (once past row G or H in San Jose's California Theater, for example) I cannot . . . and while I doubt it would improve my enjoyment of the music it would nevertheless be a nice skill to have.
Late to this aspect of the thread I know, but I just got a chance to read and try to digest the above paper. 🙂
I have to say its quite interesting, even though I don't fully grasp the explanation given of the interaction at the overlap region of the the velocity node of a lateral mode and the pressure node of front/back mode.
I understand perfectly well how in phase bass will excite the front/back modes and not excite the lateral mode (I've known this for years even though its often overlooked in simplified room mode analysis and discussions pertaining to the "golden ratio" etc) and how out of phase bass will excite the lateral modes but not the front/back modes. This much is easy to understand.
What I don't understand is the mechanism by which the lateral mode is somehow modulating the arrival time difference between the two ears of bass from the front/back mode.
The way it reads is that we "hear" the in phase bass, but that the arrival time of that in phase bass at each ear is alternatively modulated back and forth by the velocity node of the lateral mode creating a "pressure gradient" such that for half of that modes cycle the left ear receives the in phase bass slightly ahead of the other ear and the other half the right ear does. Is that what he's saying ?
If so I assume the mechanism is that sitting in a velocity node for the lateral mode means that the left/right air velocity from that mode is quite high, and thus imparts a sideways skew to the bass propogation from the in phase mode (which would otherwise arrive at both ears simultaneously in a symmetrical setup with two subwoofers) in much the same way that a strong cross wind will carry a sound in the direction it's travelling ?
If so I'm not sure that I buy that there is sufficient lateral velocity even in a velocity node to do what is claimed ? Particularly if the out of phase components captured from reverberation are quite low in level.
On the other hand if we take the paper on face value that it does work as described (whatever the precise mechanism) then I'm quite pleased that my sub in each corner "ideal system" suggestion earlier in the thread would seem to be almost ideally suited to meet Griesingers criteria. 🙂
If we look at the conclusions/summary -
"Successful results depend on: 1. having an input recording that includes at least
two channels where the reverberation is independently recorded, and thus uncorrelated
with the other channels."
Obviously the out of phase content needs to exist in the recording for it to work, which is out of our hands, if it exists in a particular recording it should work, if it doesn't no harm done - the 4 subs work together to provide conventional modal smoothing on mono bass.
"2. The presence of independently driven room modes that
overlap in such a way that the lateral pressure gradient of one mode combines with the
pressure of another. In the case of two channel stereo, the best results usually occur when
an asymmetric lateral mode (driven by the difference signal between the loudspeakers)
creates a pressure gradient at the listening position, and a medial mode (usually a
front/back mode) supplies the pressure."
Four subs, one in each corner does nothing to harm this, so long as the two left subs are driven from the left channel and the two right subs from the right channel.
"In many rooms it can be helpful to place low frequency
drivers at the sides of the listening position rather than at the front of the room."
In a room where the listener is near or at the back of the room (common in small UK living rooms where you're forced to put the speakers on the long wall and sit at the opposite long wall) the two rear subs are actually to either side of the listener - exactly as suggested in the paper to improve the effectiveness of the technique by producing the lateral mode directly to the sides of the listener rather than relying on it to propagate down the room. (with some loss for each reflection to get there)
So it seems depending on the room size and layout a sub in each corner driven in stereo pairs is ideally suited for the above technique.
Add to this the benefits of modal smoothing - a sub in each of 4 corners is one of the known optimal modal smoothing placement solutions. (even if it uses one extra sub than the 3 sub approach)
Finally, symmetrical placement of the front two subs either side of the centre line relatively near the main speakers makes it possible to cross them over higher than normal (say 160Hz) to provide some "fill in" in the 100-160Hz region where a deep notch usually develops which can suck the "warmth" out of the bass.
(A problem not addressed by the asymmetric 3 sub approach where the subs are producing summed mono bass and only one sub is near a main speaker - since localisation of the subs and distortion of image location would occur if you tried to cross them over that high) The rear subs can be limited to 80Hz or so and perform normal modal smoothing below this frequency range.
Seems like an ideal approach on many fronts to me, and one which I will definitely be trying when I get the opportunity...(eg when we move house again later this year allowing me to actually set up a stereo again! 🙂 )
Last edited:
So far the only experimenting has been on myself. 🙂
I don't know that it improves my enjoyment of the music, but it does improve my enjoyment of a recording. Stereo bass sounds more realistic to me than mono bass. I can't judge whether or not it improves my enjoyment of real life, because I can't turn it off to test it!
There is more about the subject over in the Bass Location thread in the Lounge.
I will have to read that article. I would never consider a "crossover" to the subs and certainly not one as high as 160 Hz. My software does a good job of showing that "crossovers" in the modal region are anything but well behaved.
I also had not considered "stereo" bass when I worked out the multiple sub techniques. It adds an interesting twist.
I do know that the four corners are not good locations for four subs in mono. You can see that pretty well in the software (which does work for me on a stand alone machine.)
It would be curious to work out the stereo bass perception vectors as Griesinger is describing them. What I don't get is if there is some integrating of the sound over time such that it would average out the time variations. At HFs there is certainly some averaging and all phase information is lost, but at LFs the signals are more synchronous so there may not be any averaging effect and we may well respond to a rapid change in the perception vector. If this is true then there would be an ideal room shape and listener location that would optimize the modal presentation.
I wasn't suggesting a crossover to the main speakers at 160Hz, sorry if that wasn't clear. I was suggesting the same as in a previous post of mine a couple of pages back:
Main speakers driven unfiltered with a natural bass extension going down to 30-40Hz.
Front corner subwoofers low pass filtered at 160Hz, rear corner subwoofers low pass filtered at 80Hz. (slopes of both filters yet to be determined)
Left subwoofers driven from the left channel, right subwoofers from the right channel to preserve any stereo properties of the bass (as in the quoted paper)
Exact location and distance of the front subs from the main speakers would be optimised to best fill in the notch that typically occurs in the 120-140Hz range, which I'm thinking may turn out to be with the subs a 1/2 wavelength (at the notch frequency) away from the main speakers towards the corners.
At 140Hz that's just over a metre so with typical main speaker positioning that may indeed work out nicely with the sub more or less at the corner.
Interesting, because I'm sure I've read a paper from Harmon (?) who studied this and came to the conclusion that with 4 sub woofers the optimal locations for modal smoothing and minimising seat to seat variation in a shoebox room were either all 4 in the corners or all 4 in the middle of the 4 walls, with both providing a similar degree of modal smoothing but with the corner locations providing greater efficiency in terms of power in to SPL out due to to maximal coupling to the room.
I would need to go back and find the paper again and re-read it to see if I misunderstood it.
The paper quotes the mechanism for detecting direction at bass frequencies as the time difference between each ear of the negative going zero crossings of the waveform. It says that we cannot detect direction of bass for a steady state sine wave in a room however a more impulsive bass note can be localised due to the zero crossing on the near ear being triggered first. (The time delay is the same as it would be for higher frequencies after all, and determined by ear spacing...)
I know you've dismissed this before as being a case of localising the higher frequency transient of a rapidly changing signal, however a changing bass signal doesn't necessarily have to have harmonics that extend out of the bass region - it could be something like a shaped sine burst which has nearby side bands but no significant harmonic components beyond the bass range. In the paper it states (at least in the authors opinion) that a band limited impulsive signal can still be localised despite lacking content above bass frequencies while pure steady state sine waves cannot.
(Pure sine waves are often hard to localise even at high frequencies so that's not surprising to me)
Yep, read the paper - room shapes and listener locations is specifically covered. 🙂
The very quick summary is that a square room is almost entirely bad for the effect he is trying to achieve, (almost non existent) while rectangular rooms of various proportions can work well, with the two common layouts of speakers on the short wall and listener in the room centre, and speakers on the long wall and listener at the centre of the opposite long wall both able to work well depending on the length to width ratio. (More than one ratio works) Some configurations have multiple "good" listening locations which maximise the effect.
Last edited:
That is certainly what I've found in my listening tests. I have difficultly locating sine waves of any frequency, high or low - but sharply filtered noise is much easier. It does not have to have mid frequency content to be heard. He also found that filtered voice was much easier to locate than most signals.
https://nyal-mellor.squarespace.com...cy_optimization_using_multiple_subwoofers.pdf
There are others in the series that cover this as well . . . I'll see if I can find them . . .
I guess this refers to
... to create spatial impression ... We must use the sound source(s) to
cause the time delay across the listener’s head to vary chaotically – at
the time rate of the original hall.
(page 3)
This paragraph in my view is about spaciousness and envelopment, it
is not about localization.
Nevertheless he claims before that the room is "preserving" (some of the)
localization information in the modal pattern excited:
I will not comment on this, because it seems to vague to me, even from
the perceptional point of view.
But concerning the 2 modes (front-back and lateral) example:
It is not assumed, that these example modes
- front-back providing in phase pressure
- lateral providing left-right presure gradient
are at same frequency: They have different frequency.
What he tries to achieve is "chaotic time delay" across the listers
head varying at the same rate as "in the hall" (the recording venue).
I think the effect is more comparable to the "frequency beat" you may
perceive in larger halls and also churches, which make up a "wide, deep spatial
impression". The "rate" he is refering to, is made up by a beat frequency
between front-back and lateral mode:
Those "rates" may even be subsonic (IMO) and the front-back and lateral
bass component need not even originate from the same source in the
"original event":
The components may arise e.g. from the base frequency of one organ pipe
and an overtone of a different organ pipe, which radiate a complex
time varyiing pattern in space of the original venue.
Those (cues to ...) spatially chaotic beats (perceivable in the original venue
like church, hall, etc. ...) cannot usally be reproduced in the home, but (cues)
find themselves on the recording, if the reverb has been picked up at sufficiently
separated places of the original venue and recorded independently.
There's no chance to pickup that with pure intensity stereophony, using coincident
microphones or just pure pan-pot stereophony: But many recording techniques
(microphone configurations) provide such cues.
He is trying to make those cues perceivable in a customized home setup,
that's what the paper is about IMO.
And my personal remark would be: It really can make the difference
between "feeling" in a real large venue during reproduction or not.
It can make e.g. a church recording sounding (more like ...) a church
instead of "(sub-)woofers in my living room".
I am leaving out the "correct reproduction" matter here consciously ... because
i know you guys' different opinions quite well after some time of discussion ... 😉
I do agree (with ScottG e.g.), that the way Griesinger tries to reproduce phase
fluctuations in lower bass at home, might neither be the only way to do it nor
the possibly best.
Kind Regards
Last edited:
"With 5000 subwoofers, modal variation is virtually
eliminated."
Inspiring observation.
Btw, I tried "1/4 from boundaries" approach (4 built-in ceiling subs). At least for flattening out the bass in small cinema it worked quite well.
Last edited:
Thanks for sharing!
It's good to see David's interview sparked this topic back to life again 🙂 . I guess the subject will never cease to be relevant.
By the way, a lot of what I know about audio I've learned here at diyaudio, and this topic in particular. Therefore I guess this may be the most appropriate place to inform you that I recently started an audio company with two other guys: Synergy Acoustics.
We'll be offering a unidirectional monitor loudspeaker, separate bass speakers / subwoofers, a power amp, and a preamp with DSP. In line with what we have been discussing in this topic, Synergy Acoustics emphasizes that good sound starts with the loudspeaker-room interface. Our products are all aimed at providing great sound in 'real' rooms, although of course they'll sound great in dedicated studio's as well. We're currently busy starting up. Only when everything is ready will our website go live.
If it is so, then combined with Harman knowledge it may leave us with the following:
- 4 subs in 4 corners (or even better - in ceiling) connected as stereo;
- delay applied to rear subs to be phase coherent with front subs (1m distance is rougly 3ms delay), as it may be benefitial for flatter "terrain" in higher (near crossover) frequencies (have anyone tried this and investigated effects?);
- for same reason additional delay applied to main fronts in respect to aligned subs may be rewarding to play with, too.
IMO dipoles would be more than enough for a small room with 1-3 person listening area. I liked them more than 1/4 in-ceiling setup. Assembled within 4 hours and fed from 6-channel amp through 3-way 4-th order Linkwitz-Riley crossover and EQed down to 20Hz with FIR-based high-resolution room EQ software (still in beta). The bass was quite evenly aligned throughout the small room. This time bass wasn't leaking into the toilet and making 50Hz modal bump there (what a pity!). Also, neighbours confirmed that this time they didn't get much of the bass, so I assume that room modes were not excited very much (as expected from dipoles). No basstraps, no other treatment. Bass was crossed with symmetrical dipole mids at 190Hz and 2ms delay applied to the latter to flatten x-over area. See picture and L/R measurement and correction curves.
As i was re-reading the Griesinger paper in 2012, i was inspired
to a different (subwoofer) approach using "interchannel dipoles":
Picture1 shows a variant also allowing nearfield listening.
With both channels driven by the same signal, the setup
is like common dipole setups ...
Picture2 shows a variant with dipole saparation equal to
the room's longest axis ...
Kind Regards
Attachments
Last edited:
I'm not sure that delaying the rear subs is the right thing to do. One of the reasons for having a rear set of subs is so that the lowest fundamental length (front/rear) mode of the room is actively cancelled out.
This happens when front and rear are driven in phase and are positioned half a wavelength apart - eg the length of the room apart. This fundamental mode is quite troublesome because it leads to very "bottom heavy" boomy low bass at the rear end of the room, but a lack of low bass half way along the room.
By adding the rear subs you actually reduce the SPL at the fundamental resonance at each end of the room without affecting the response at the middle - after EQ is applied you then get a much more uniform bottom end response along the length of the room that has good low frequency extension without booming at each end of the room.
If you delay the signal to the rear subwoofers you prevent this active mode cancellation from occurring, and if you delayed them enough to bring them into time alignment with the front subs you would end up with the same mode problem you started with but 6dB more output.
By the way I'm not sure where you get 3ms / 1 metre from - I don't know about you but my listening room isn't 1 metre long! 😀 To make the front and rear subs "phase coherent, you would need 18ms for a 6 metre long room...
When subwoofers regularly have over 20ms of group delay at low frequencies I'm not sure that correcting it by only 3ms would be a worthwhile or even audible endeavour ?
Another thing to consider is that if you have front corner and rear corner subs working together in phase, at frequencies where they're less than half a wavelength apart they create a virtual source in the middle of the room, thus the apparent location of the front subs is not behind the main speakers, the apparent location of ALL subs together is actually in front of the main speakers - but since the subs will have quite a bit of group delay of their own relative to the high frequencies from the main speakers this will push the apparent source location further away, at least from a time alignment perspective.
The net result is the 4 sub front/back arrangement may reduce the effective group delay increase at low frequencies for a listener at the back of the room. (Or am I way off the mark here ? Anyone ?)
Last edited:
- Status
- Not open for further replies.