Loudspeaker perception

[ScottG]

The psychoacoustic benefits directivity will bring..


- Elias

I've never been "sold" on this argument for any given freq.. I'm not saying that I haven't benefited from it, or that there haven't been good arguments for it, but rather - was any benefit significantly derived from enhanced directivity OR from some other reason related to the design that produced enhanced directivity?

I *think* there have been, but only midbass freq.s due to side-nulling as it reacts to apparent channel-separation. Perhaps also at higher freq.s with respect to the reflection from the opposite speaker.

Also, a dipole per se isn't exactly "more directive" when compared to more traditional monopoles at higher freq.s. Nor am I sure if this is good or bad, or perhaps more precisely - what is overall better (and why)?

 

[markus76]

Originally posted by Elias Visaton B200 on the wall, how can you bear them like that not pointing to listening spot? They beam like h*ll.

Post #581

 

[Elias]

Hello,

still chewing on the DBA approach – collecting as much data / impressions as I can - not that much to be found on that

I think the DBA is not optimal in it's implementation as I saw it in one of the german forum.

The thing is, if you are going to do such a 'massive implementation', why not to include some psychoacoustics improvements at the same time.

In the original implementation sound comes from front and is electronically killed at the back wall. What you hear is _mono_. Not good for the bass perception. As Griesinger writes about this, usually such a bass in perceived inside a head. It's not externalised. Unnatural.

I think the whole DBA should be turned 90 degree in the room. Put your arrays at the both sides. Feed L signal to the left wall and kill it at the right wall, and vise versa for R signal. This way you will maximise interaural differences. If you have stereo material at bass freqs, you are able to phantom image!

This is something I've also been thinking many years.

- Elias

 

[mige0]

That's just plain wrong. Obviously you missed a couple of studies by Olive that have proven that objective assumptions on the quality of a certain speaker can be made by utilizing high resolution measurements.

Am I the only one that sees an antilogy in refusing to read existing scientific studies and claiming at the same time that there's no scientific aspect in louspeaker building?

You got me !
I'm with you to that point that "assumptions on the quality of a certain speaker can be made"
For sure it can be stated a band pass of 500Hz to 2000Hz to be less than optimal – or something like.

Where we differ is that there is something "objective " like about "perception" in general – and about "loudspeaker perception" in particular.
Maybe its semantic – but as far as I translate it as "WAHRnehmung" (to hold sth. for truth) there simply isn't anything that is normative in a strict scientifically sense.

For example
When you look at a house you seeing clearly - you "hold for truth" it *is* a house.
Looking at a picture of a house (a fake of a house in other words) you usually "hold for truth" that this picture shows a house.

BUT not everyone at every time is able to identify a house at a picture as to be the picture of a house.
Its a learning process to translate all our senses - *and* all the fakes we like to play around with - into something meaningful.

Hence I strongly doubt that even "Olive" can prove the "reality" I'm in when listening to different speakers .

Well the basic idea is to create a planar wave front coming from the front wall and from the back wall. The one at the back wall is phase inverted and time delayed. The delay matches the room length so the two waves meet at the back wall where they neutralize each other. The planar wave is generated by multiple subwoofers evenly distributed on the front and back wall. The more subwoofers the more planar the wavefront stays to higher frequencies. Horizontal and vertical modes are eliminated by the buildup of the planar wave front.

Thanks for summarising the basic operation.
(*This* papers I read )

What I can't see is any interesting discussion ongoing about that "pitch and catch" thing - except the few ones already pointed to.

One would think a pretty practicable way to avoid almost all the big hassles down in the room's mode department would gain some more DIY attraction.

If you look at the FR presented
http://www.avsforum.com/avs-vb/showthread.php?t=837744
http://www.avsforum.com/avs-vb/showpost.php?p=11791341&postcount=76

ain't they stunning?
Even more so that there is *no* room treatment applied AFAIK

As it requires adjustable delay - -maybe simply not enough "digital" DIY'ers around ?

The room-CSD's you presented are very good too – considering it wasn't precisely a text book planar wave generated by the subs placed at the bottom only - and not in wall mounted in a dense mesh structure.

What catches my interest is that "double bass array" virtually creates a "no room" situation – no room modes to deal with – great decay
– no directivity discussions to struggle with either




Michael

 

[markus76]

I think the whole DBA should be turned 90 degree in the room. Put your arrays at the both sides. Feed L signal to the left wall and kill it at the right wall, and vise versa for R signal. This way you will maximise interaural differences. If you have stereo material at bass freqs, you are able to phantom image!

This is something I've also been thinking many years.

Stop thinking - build it and let us know the outcome.

 

[Salas]

There is a catch though. The bass level will be almost constant everywhere in the room when the mains level will not. Either the bass level must be set for listening position and the mismatch tolerated when wandering, or planar mains are due.

 

[mige0]

Hello,



I think the DBA is not optimal in it's implementation as I saw it in one of the german forum.

The thing is, if you are going to do such a 'massive implementation', why not to include some psychoacoustics improvements at the same time.

In the original implementation sound comes from front and is electronically killed at the back wall. What you hear is _mono_. Not good for the bass perception. As Griesinger writes about this, usually such a bass in perceived inside a head. It's not externalised. Unnatural.

I think the whole DBA should be turned 90 degree in the room. Put your arrays at the both sides. Feed L signal to the left wall and kill it at the right wall, and vise versa for R signal. This way you will maximise interaural differences. If you have stereo material at bass freqs, you are able to phantom image!

This is something I've also been thinking many years.

- Elias

Yes, also my take on this – the 90deg turn certainly is an option.

Maybe also a split of the array would do - in practical terms.
For mono (as mostly in this FR) you have full benefit and with stereo signals "pitch and catch" is somewhat compromised but should still be of some benefit.

Markus' measurements indicate that there is some tolerance to sub optimal placement – a big advantage IMO.

There is a catch though. The bass level will be almost constant everywhere in the room when the mains level will not. Either the bass level must be set for listening position and the mismatch tolerated when wandering, or planar mains are due.

Right, but for me - I'm not planning for a complete audience or for walking around the room – and at least it wouldn't keep me from further exploration at the first place.

Integrating with the main speakers may be more difficult regarding to correct overall delay than about level – just a guess.

Stop thinking - build it and let us know the outcome.

Already on the list - but don't hold your breath on it
Train your brain - dont "stop thinking"

 

[el`Ol]

P.S. Visaton B200 on the wall, how can you bear them like that not pointing to listening spot? They beam like h*ll.

As you are active in some other threads you should know much stranger ways of using beamy drivers.

 

[Elias]

Hello,

So, the next step is to use additional synthetic reverberant channels.
...
I have a pair of speakers directly to the side of the listening position,
...
doing a mute/un-mute comparison it's immediately apparent how the presentation expands and fills in with the reverb, but doesn't really result in degraded placement.
...
Future experiments along this route are to add rear channels for a total of 4 reverb channels, and obviously to play with the reverb characteristics.

You are doing some interesting things there. Thumbs up!

Over the years I've come to a conclusion that adding 'surround' spkeakers to a stereo system is one single most significant improment one can possible make to improve the 'realism'.

I think the side position as you heve done gives very good envelopment because it maximises interaural differences.

Adding another set of speakers at the back as you plan will add more realism. However if you already have side speakers the improvement is not as huge as without side speakers.

Getting the reverb setting right for different recordings can be a little bit labour intensive operation. One idea is to write the basic settings on the sticker in the CD cover

But many good recordings contain information of original acoustic space already. Why not to utilise that. Can use L-R signal for the surround channels. Then the task is relaxed, since you only need to decorrelate that for different speakers. That is don't use full artificial reverberant tail but only small set of reflections of the impulse response. Find a system you can tune accordingly.

It will be great

- Elias

 

[Elias]

Hello,

There is a catch though. The bass level will be almost constant everywhere in the room when the mains level will not. Either the bass level must be set for listening position and the mismatch tolerated when wandering, or planar mains are due.

Constant bass level in the room is only a good thing. No more mode problems.

I think line arrays for main speakers would be great with DBA to have least distance effect. Maybe dipole line arrays

Yes, also my take on this – the 90deg turn certainly is an option.

Maybe also a split of the array would do - in practical terms.
For mono (as mostly in this FR) you have full benefit and with stereo signals "pitch and catch" is somewhat compromised but should still be of some benefit.

Integrating with the main speakers may be more difficult regarding to correct overall delay than about level – just a guess.

Simplest test would be putting your current bass speakers at the side and doing the signal processing for them. If you find any improvement then proceed to the full array. Doesn't lose anything doing a test.

- Elias

 

[markus76]

Originally posted by mige0 Markus' measurements indicate that there is some tolerance to sub optimal placement

But only when you compare DBA to SBA Running the suboptimal DBA setup without delay and inverted phase showed an even better performance. That indicates that a multisub setup like the ones proposed by Earl Geddes or Welti work best in a normal living room.

Best, Markus

 

[Elias]

...
a couple of studies by Olive that have proven that objective assumptions on the quality of a certain speaker can be made by utilizing high resolution measurements.

Quality of a speaker? Measurements?

How does Olive know where is the goal set?

What parameters do you measure when considering psychoacoustic issues of a speaker?

Just think about directivity pattern for example, there is no consensus for that. How do you measure if you don't know what you shoud get.

High resolution is no guarantee of science.

- Elias

 

[markus76]

Read? Explore the perspective and knowledge of others? Not good?

It's really frustrating to talk to you. You want to know what psychoacoustic studies have discovered. When naming studies that might be of interest then you refuse to study them. So I'll stop the discussion with you in this thread because it's just a waste of time for me.

 

[mige0]

Hello,

But many good recordings contain information of original acoustic space already. Why not to utilise that. Can use L-R signal for the surround channels. Then the task is relaxed, since you only need to decorrelate that for different speakers. That is don't use full artificial reverberant tail but only small set of reflections of the impulse response. Find a system you can tune accordingly.

It will be great

- Elias

Good point Elias, starting out with the L-R signal for the surround speakers and maybe mix in some slightly artificial reverb to your taste to get a sense of being in a room / your listening room - as assumed by most recording engineers.

I liked my dipol sub - now I have this very expensive Gedlee speaker that I have to love

An externally hosted image should be here but it was not working when we last tested it.

Dipole sub? Where..?

Oh, THAT! It's so small and practical I didn't notice it at first

Markus, listening to a bass speaker in the nearfield – like in the setup you show – where the room itself is usually dominating everything looks like a very clever idea.

How did you enjoy the bass?
Did you get good integration with the main speakers?


John has done some interesting work on the "technical" aspects of such a set up:
http://www.musicanddesign.com/Dipole-axis.html

Constant bass level in the room is only a good thing. No more mode problems.

Simplest test would be putting your current bass speakers at the side and doing the signal processing for them. If you find any improvement then proceed to the full array. Doesn't lose anything doing a test.

- Elias

Sorry not practical.
My main speaker is built to have as much point source character as possible. Not in the sense of directivity but in the sense of all sonic power emerging from one single spot in space (in the lateral direction).
I cant get it as ideal as that though, but moving the bass – crossed at roughly 300Hz – away from the mid / high would be more of a compromise than I want to do.
Would have to build flat enclosures too, as dipole at the wall isn't really working for flat wave form generation and "in wall" mounting or "infinite baffle" isn't an option for an experiment only.

But only when you compare DBA to SBA Running the suboptimal DBA setup without delay and inverted phase showed an even better performance. That indicates that a multisub setup like the ones proposed by Earl Geddes or Welti work best in a normal living room.

Best, Markus

You have tried – to some degree – so I respect your findings.

On the other hand – multi sub isn't able to do anything about decay time of room modes (Though Earl would like people to belief in this now and then – see his point about "active absorbing" some pages back).


Michael

 

[graaf]

I have to admit that I like Markus' older setup a lot
very clever ideas and very elegant implementation

and those Behringers, and AKG headphones - things which I also own

hey Markus!
how would You compare Your older system's performance vs the Nathan's?
what are the most significant improvements? in sound quality of course

because in terms of decor and looks Your older setup is a winner

best!
graaf

 

[markus76]

Originally posted by mige0 Markus, listening to a bass speaker in the nearfield – like in the setup you show – where the room itself is usually dominating everything looks like a very clever idea.

How did you enjoy the bass?

It was the best bass I've heard up to today - very dry. Multisub comes close but has to be complemented by passive absorption. The drawbacks of a nearfield sub is that there is a very defined listening position, it works for one listener only and at a certain volume the modal field of the room takes over.

Best, Markus

 

[markus76]

Originally posted by mige0 multi sub isn't able to do anything about decay time of room modes (Though Earl would like people to belief in this now and then – see his point about "active absorbing" some pages back).

He did? Don't think so. He advocates additional passive absorption.

Active absorption works only as long as it actually does something. In the case of multisub or a DBA the actively generated "anti-modal field" depends on the music signal and does not respond to the actual modal field in the room. So when you switch of the "anti-modal field" the room takes over again. There still will be ringing as active absorption can't dissipate energy. This can only be done with passive absorption.
Toole asked an interesting question: "Do we hear the spectral bump or the temporal ringing?"

Best, Markus

 

[mige0]

It was the best bass I've heard up to today - very dry. Multisub comes close but has to be complemented by passive absorption. The drawbacks of a nearfield sub is that there is a very defined listening position, it works for one listener only and at a certain volume the modal field of the room takes over.

Best, Markus

Wouldn't have thought that it could be SPL dependant ?
No high levels / compression due to limited dipole excursion involved ?

Having an even smaller sweet spot to listen for is a concern I also would have with the 90deg turned double bass array.

He did? Don't think so. He advocates additional passive absorption.

Active absorption works only as long as it actually does something. In the case of multisub or a DBA the actively generated "anti-modal field" depends on the music signal and does not respond to the actual modal field in the room. So when you switch of the "anti-modal field" the room takes over again. There still will be ringing as active absorption can't dissipate energy. This can only be done with passive absorption.
Toole asked an interesting question: "Do we hear the spectral bump or the temporal ringing?"

Best, Markus

Not sure about yet – but I would rather put it slightly different.

DBA (in ideal) avoids room modes by generating a planar sound wave at first hand (something we don't find with usually placed speakers in rooms) - and then – before room modes get excited dies down the original wave front at the back wall.

Basically DBA mimics an infinite tube – and there is no such thing like "modes" connected with infinite tubes

Would you expect kind of leakage into room modes? - sure - with a less than optimal set up - but also when build / measured like
http://www.avsforum.com/avs-vb/showthread.php?t=837744
http://www.avsforum.com/avs-vb/showpost.php?p=11791341&postcount=76

from Nils at AVS?

 

[gedlee]

To clarify about multi-subs versus DBA.

Why is a plane wave desirable? I don't see a reason to assume that.

Multi-sibs CAN act as active absorbers. If I have two subs in a room and I turn on a third and the sound level (averaged arround the room to be precise) goes down (and this is easy to arrainge, just try it), then the third sub IS actively absorbing the sound in the frequency range where the sound level has gone down. This is simply physics.

Active sound sources can decrease the decay along with the levels, but they don't have to. It all depends on how they are setup, phase, amplitude, etc. And we should all remember that it is possible for sound decay in real rooms to not be monotonic. Namely, the sound CAN increase before it falls or fall at different rates over time. And active can affet each of these in different ways. It could actually increase the sound before it decays, etc.

My opinion about Floyds question is its both. Modal decay is a complex thing. For instance, the frequency of modal decay does not occur at the frequency of the excitation. So the decay could be quite dissonant. The longer a sound exists the louder it will seem to be (up to a point) so both the gain from a mode as well as the extended decay will tend to make that frequency pronounced.

 

[markus76]

Maybe the mods should split the thread at this point. We are starting to talk about "Absorption in the modal region".

Originally posted by gedlee Why is a plane wave desirable? I don't see a reason to assume that.

Earl,

If it would be possible to create a plane wave that isn't altered be the room (which is something what happens in real rooms) then there would be no horizontal and vertical modes at all - something very desirable.

Originally posted by gedlee Multi-sibs CAN act as active absorbers. If I have two subs in a room and I turn on a third and the sound level (averaged arround the room to be precise) goes down (and this is easy to arrainge, just try it), then the third sub IS actively absorbing the sound in the frequency range where the sound level has gone down. This is simply physics.

This is something I don't understand. When two waves converge they interfere but no transformation of energy happens. But this is how absorption is defined: transform sound energy into some other type of energy, e.g. into heat as a porous absorber does.
So in our case the energy stays in the room and interacts with it when active absorbtion is switched off.

Best, Markus