What makes speakers "disappear " and can it be measured?

[Sancus]

The band is always playing in front of you, even when you're in the front row at a stadium concert. I don't know why I'd want to hear the bass guitar coming from the side speaker. Maybe for karaoke...

I mean, this is a matter of taste. But there's no reason to limit yourself to physical or historical limitations that have nothing to do with the actual medium. The technique isn't new or anything either, Pink Floyd was doing it 50 years ago. Artists can and should put whatever they want in whatever channels they want.

 

[JWAmerica]

If the music were engineered that way, it could sound good. But too often the soundstage is simply stretched to wrap around the listener, not much better than using an upmixing DSP.

 

[Tangband]

One thing that for sure helps to make the two loudspeakers dissapear in the soundstage is when you :

1.- optimize so the acoustical level between tweeter and midrange is within 0,5 dB at the crossover frequency, for both loudspeakers. ( very important )

2. - do time-alignement of tweeter and midrange so the phase is exactly right at the crossover frequency. ( less important )

The importance of those things is easy to realize with an active system with a dsp crossover unit like dbx pa2 or minidsp, and a good microphone. Its great fun to try different settings in DIY, the changes in sound quality are clear for everyone to hear. A difference of 0,5 dB at the crossover range is enough to make a change in the loudspeaker sound, from good to very good.

Peter Snell did some of those things with his passive loudspeakers in the early -80:s , and indeed, his speakers did stereo-image very well.

This is something that can be optimised and bettered with almost every commercial passive loudspeaker one can buy, including Revel or Kef.

 

[Sancus]

If the music were engineered that way, it could sound good. But too often the soundstage is simply stretched to wrap around the listener, not much better than using an upmixing DSP.

Yes, it can be done well, and poorly. Depends on the mixer and the recording quality. 2L's catalog are the best examples of it being done fully acoustically in a space, as opposed to multitracked(which can also be good, but requires good mixing). Interestingly it seems they are finally selling some of their surround music directly in fully digital files... been waiting for that.

 

[Thomas_A]

One thing that for sure helps to make the two loudspeakers dissapear in the soundstage is when you :

1.- optimize so the acoustical level between tweeter and midrange is within 0,5 dB at the crossover frequency, for both loudspeakers. ( very important )

2. - do time-alignement of tweeter and midrange so the phase is exactly right at the crossover frequency. ( less important )

The importance of those things is easy to realize with an active system with a dsp crossover unit like dbx pa2 or minidsp, and a good microphone. Its great fun to try different settings in DIY, the changes in sound quality are clear for everyone to hear. A difference of 0,5 dB at the crossover range is enough to make a change in the loudspeaker sound, from good to very good.

Peter Snell did some of those things with his passive loudspeakers in the -80:s , and indeed, his speakers did stereo-image very well.

This is something that can be optimised and bettered with almost every commercial passive loudspeaker one can buy, including Revel or Kef.

If you sit in front of one loudspeaker at LP and hear a central phantom image still coming from the centre, I think thiis is a clear sign of speaker ”dissappering”.

Speakers devoid of resonances are one key, another is the ratio of direct vs reflected sound, flloor reflection(?) and balance between left and right speaker, toe in etc.

 

[Kal Rubinson]

Interestingly it seems they are finally selling some of their surround music directly in fully digital files... been waiting for that.

????? I've been downloading from them for years (although there was a recent interruption of direct availability).

 

[Sancus]

????? I've been downloading from them for years (although there was a recent interruption of direct availability).

Sorry what I meant was the Auro3D/Atmos files. Those are relatively new afaik. At least, I hadn't seen them available anywhere before.

 

[youngho]

If you sit in front of one loudspeaker at LP and hear a central phantom image still coming from the centre, I think thiis is a clear sign of speaker ”dissappering”.

what does “sit in front of one loudspeaker at LP” mean?

 

[richard12511]

I'm a big fan of music upmixed to 3ch, but not 5ch (or more). 5ch sounds okay for orchestras, but for regular band recordings I can't get used to it, even when it's mastered in multichannel. The band is always playing in front of you, even when you're in the front row at a stadium concert. I don't know why I'd want to hear the bass guitar coming from the side speaker. Maybe for karaoke...

This is why Auro3D is the only upmixer I like. It's the only one that doesn't put instruments and other elements of the mix to beside/behind/above you. Pretty much all it does is expand the stereo image horizontally and vertically(at least with low enough effect power). Requires 9 channels, though, and Auro2D unfortunately does the same as the other upmixers. I have this same bias, but as for immersing yourself in the other style, imagine you're a king with a custom band playing just for you with you in the center .

 

[Elkios]

We have wall mounted Anthony Gallo Acoustics Strada2 main speakers and matching subs in our mountain house, and the speakers totally disappear sonically. Wall proximity matters not, with a proper loudspeaker design.

I have owned ref 3.1 and support what you hear. Great passive speaker's.

 

[puppet]

The only enjoyable multichannel music recordings I've liked have used the crowd noise and reflections/reverb from the venue for the side channels. All music came from the front three.

This sound intriguing. Do you have any examples a guy could look into?

 

[Tangband]

If you sit in front of one loudspeaker at LP and hear a central phantom image still coming from the centre, I think thiis is a clear sign of speaker ”dissappering”.

Speakers devoid of resonances are one key, another is the ratio of direct vs reflected sound, flloor reflection(?) and balance between left and right speaker, toe in etc.

If the tweeter are to high in level on one of the loudspeaker then the soundimage will shift.
You can have variations much more than one dB in loudspeakers that you buy, or build.
It can be corrected in dsp with active loudspeakers, or by changing values in the crossover in passives.

 

[pogo]

If the tweeter are to high in level on one of the loudspeaker then the soundimage will shift.
You can have variations much more than one dB in loudspeakers that you buy, or build.
It can be corrected in dsp with active loudspeakers, or by changing values in the crossover in passives.

Before using a DSP, it is better to use matched h/w (passive components, speaker chassis, ...).

 

[Tangband]

Before using a DSP, it is better to use matched h/w (passive components, speaker chassis, ...).

In real ”active” mode driving the loudspeakers from a digital crossover filter, the active digital filter is a part of the dsp.
In that case you dont need any sounddegrading passive components at all between amplifier and loudspeaker units.
If its done right its the best approach for good sound.

You can ofcourse combine a dsp with analogue crossovers, the question is - why ?
You can correct for room resonances with the dsp, but you still have a ( potential ) sounddegrading passive crossover between your amplifier and loudspeaker drivers. Its much better use to let the dsp do the crossover also. That can be done in minidsp.

ofcourse, you can go all analogue and fine tune the passive crossovers in left and right loudspeaker, just like Peter Snell did at Snell acoustics. Using different values of resistors and a lot of measurements.

 

[BenB]

Many of the concepts that have been considered here are actually testable. If you start with a speaker that disappears, you can sometimes take away that speaker's special qualities and see if the disappearing act continues.

For example, if a speaker that disappears well is designed to be linear phase (at least across the passband), you can simply add phase distortion to mimic the typical phase response of speakers that aren't linear phase, and see if it changes people's perceptions.

Similarly, you could use DSP to add compression to see if that kills the disappearing act.

It's much harder to address the directivity related theories. Some people have tried to test this by treating reflection points, but those treatments are usually poor substitutes for actually having directivity that reduces the reflection across a wide bandwidth. I have considered modifying one of my multi-way line arrays so that I can turn the top half of the array on and off at will.

At any rate, things that can be tested should be tested. Personally, I'm convinced that typical phase distortion makes no difference in this regard, and is completely imperceptible at frequencies above 1500 Hz or so.

 

[Frgirard]

The phase distortion? The moon phase is better.

 

[nerdoldnerdith]

I own and have heard many of the best speakers mentioned on thus forum. To this day I have heard no pair of speakers that disappears more than my Morrison Audio Model 19.1's. When you listen to them it does not sound like there is anything coming from them at all, even when you get up an walk around. You have to get within a couple feet of a speaker to be able to tell that it is the source of the sound.

The next closest thing would be the Dutch&Dutch 8C.

What both of these speakers have in common is that they have very even 360 degree dispersion. The Morrisons have perfect horizontal dispersion and have the same frequency response in every direction. The 8C's off-axis frequency response to +/-180° is very close to the on-axis frequency response from 100Hz on up. Very few speakers have this capability, and at best have good dispersion to +/-90°.

This leads me to believe that dispersion in every direction including behind the speaker is crucial to how the speaker will image and whether or not it will disappear.

 

[Thomas_A]

If the tweeter are to high in level on one of the loudspeaker then the soundimage will shift.
You can have variations much more than one dB in loudspeakers that you buy, or build.
It can be corrected in dsp with active loudspeakers, or by changing values in the crossover in passives.

Yes, matching of drivers an tweeters are important, at least within 0.5 dB.

 

[Thomas_A]

In real ”active” mode driving the loudspeakers from a digital crossover filter, the active digital filter is a part of the dsp.
In that case you dont need any sounddegrading passive components at all between amplifier and loudspeaker units.
If its done right its the best approach for good sound.

You can ofcourse combine a dsp with analogue crossovers, the question is - why ?
You can correct for room resonances with the dsp, but you still have a ( potential ) sounddegrading passive crossover between your amplifier and loudspeaker drivers. Its much better use to let the dsp do the crossover also. That can be done in minidsp.

ofcourse, you can go all analogue and fine tune the passive crossovers in left and right loudspeaker, just like Peter Snell did at Snell acoustics. Using different values of resistors and a lot of measurements.

With respect to active speakers, you can have a range of solutions. I used to have active filters for both 3 kHz and 80 Hz, but have left the 3 kHz active solution many years due to space saving and simplicity. A good passive filter is good enough, IMO. I think a real benefit of active comes when you adapt the amplifier for the specific drivers such as ACE-Bass, current drive or servo feedback. I don't know how many manufacturers use this now.

https://www.current-drive.info/disto/68

 

[Thomas_A]

I own and have heard many of the best speakers mentioned on thus forum. To this day I have heard no pair of speakers that disappears more than my Morrison Audio Model 19.1's. When you listen to them it does not sound like there is anything coming from them at all, even when you get up an walk around. You have to get within a couple feet of a speaker to be able to tell that it is the source of the sound.

The next closest thing would be the Dutch&Dutch 8C.

What both of these speakers have in common is that they have very even 360 degree dispersion. The Morrisons have perfect horizontal dispersion and have the same frequency response in every direction. The 8C's off-axis frequency response to +/-180° is very close to the on-axis frequency response from 100Hz on up. Very few speakers have this capability, and at best have good dispersion to +/-90°.

This leads me to believe that dispersion in every direction including behind the speaker is crucial to how the speaker will image and whether or not it will disappear.

Wide dispersion could be one thing, it also increases the reflected:direct sound ratio in the room. Wide dispersion however reduces the time-intensity trading which requires rather narrow dispersion.