[_scotty_]

ekovalsky, I have to disagree about the potential for improvement possible in line arrays as a result of using a digital crossover.  All line arrays using a high order [ie.24db/oct slope] crossover exhibit a degree of horizontal combfilter effects and polar response errors and lobing. They also cannot be made transient perfect, time aligned or phase coherent with a passive network
of the type necessary to assure good polar response and flat amplitude.
You can have it all with a digital crossover,a near perfect horizontal polar response pattern, the ability to reproduce an acoustical squarewave only limited by the drivers mechanical phaseshift which may also be amiable to DSP correction. This technology can remove the last design limitations dictated by  passive crossover technology.  The remaining limiting factors would be the choice of drivers used in the arrays.  It is only a matter of time before someone applies this approach to line arrays at the home consumer level.  This technology is already being used in prosound applications involving digitally controlled phased line arrays whose polar response and directivity can be controlled from a DSP console.  Imagine
fine tuning the sound stage from the listening position with a remote.
The future looks bright indeed, Scotty

[John Ashman]

I was largely rehashing what Derek Wilson of Overkill had told me.  Some of what he said did make sense, but no doubt it is mostly marketing talk.

Most likely.  Yes, you can improve a driver's distortion levels if you can completely damp or basically smooth out any standing waves inside the cabinet, but it won't take a 5% distortion driver and make it .5%, let alone .2%.  I don't know how the Manger measurements were taken, so the box may have added a tiny amount of distortion due to standing waves, but not *that* much!   It's very important to start with a driver that has a linear motor and very rigid driver.  I think the only other place besides audio that you can the kind of BS that exists in audio is at your local used car dealership

The Salk HT3 definitely makes a nice target for DSP (be it with the TacT or DEQX).  I am still waiting to hear from Rick and others how things work out with line arrays.  The more reading I've done, it seems line arrays may be best with passive crossovers for the mid/bass cones and tweeter ribbons and limit DSP for the signal division between the arrays and separate subs.

There is no reason at all not to use DSP crossovers with line arrays.  If you replace a 1st order crossover, you can get lower distortion and you can substantially improve horizontal dispersion.  If you replace a higher order crossover, you bring back time/phase alignment and bypass a lot of crossover parts giving you more clarity and low level detail.  Go even steeper and you improve the sweetspot and get other small improvements. Of course, just being able to manage the bass in the room makes DSP worth it to begin with.

Line arrays have great strengths in terms of dynamic range, efficiency and the ability to fill a large room evenly with sound.  DEQX can help a much smaller system have these abilities where they couldn't before.  In a way, DEQX makes a line array less necessary to achieve certain capabilities, but it will play to a line arrays strengths and make it even better at what it does well and make some notable improvements in other areas.  It won't be as *dramatic* of a difference as what it will do with smaller point source designs, but it will still be significant and sort of do the last fine tuning so that you can get everything out of it.  

The only comment I've had is that line arrays are harder to measure properly.  They are big, they almost need to be measured in a huge room with the mic further away.  The closer you are, the easier it is to get a precise measurement with a point source, but with a line array, you need to be further away to get the drivers aligned at your ear for less lobing and a better measurement.  So a professionally measured and corrected line array will largely be easier to deal with.  If Selah corrects their system, they could always e-mail the file, saving you a boatload of hassle.  

I think digital systems will largely obsolete most every design besides point source monopoles and line arrays.  A monopole line array on DEQX will do better than a horn or dipole line array. A point source monopole on DEQX would do better than a single driver speakers, electrostats and other designs.  Point sources and line arrays will become the dominate speaker designs in the future, unless some other new technology comes out.

[Marbles]

I think digital systems will largely obsolete most every design besides point source monopoles and line arrays. ...

I disagree for the same reason that MP3's are far more popular than SACD/DVD-A.

People in general are less concerned with fidelity these days and the additional costs of digital systems with regards to the cost of the DSP unit and additional amps will make passive crossover unit speakers popular for decades to come IMO.

If you want to say that for AUDIOPHILES who want nothing but the highest quality speaker, then "I think digital systems will largely obsolete most every design besides point source monopoles and line arrays. ." than I might agree with you, but for the population as a whole...I disagree with your statement.

[John Ashman]

No, I'm just saying that there won't really be any need for horns, dipoles, ribbons, electrostats, 1-ways, etc.  Sure, they might still cling to a niche here and there, but since digital is more appropriate for line arrays and point sources, it will pretty well give them top dog status in performance for most people.  

As for the masses, it's just going to be a matter of time.  I can guarantee that Bose will be one of the first to put pretty sophisticated DSP in their speakers, mainly to make 35¢ drivers sound better than they should.  As digital amps come down in price and DSP gets more powerful and cheaper, it will be in just about every speaker.  Expect HDMI to include some sort of standardized volume protocol in the future too.  It soon won't be much more expensive to build a DSP/amp card than it will be to build a decent crossover card.  And it will allow mass market companies to cheat like crazy on cabinents and speaker design.  Drop in some drivers, do a quick fix in DSP, bring to market in a month or two, rather than a year or two.  One DSP/amp card could work for every speaker in the line up - no more designing passive crossover for every single speaker with months of tweaking and testing and crap.  They'll just churn out product like crazy.  Retail cost will be higher, but then again, you'll have your amps included.  Not everything will have the sophistication of the DEQX for sure.  Most will just be very basic DSP units, but it will sell gear too.  People will feel the need to upgrade their speakers no matter what.  And there will soon be more analog speakers out there than there will be uses for them.  I'd give that about 5 years and used analog speakers will be like a plague and within 10, there will be remarkably few, if any, in production.

[Davey]

hehehe.

You are TOO funny John.  

Davey.

[John Ashman]

Hmmmm, I don't know how to take that exactly, but I am right

Of course, most niche designs are now so small because of HT taking over, but, for instance, NHTs have never been amazing at classical music, certainly compared to Genesis or other dipole speakers, sounding closed in and "dry" by comparison, but the Xd system just blew away the Genesis 501s I had sitting next to them in detail, soundstage, imaging, bass quality (that one surprised me!) and coherence.  It really sounded stunning with classical music - it had that sense of envelopment and detail I'd only heard before with dipole ribbons, but did a *lot* of other things far better.  Horns claim "dynamics", but a good line array will crush them in this.  A digital 2-way can have the coherence of a 1-way, but with dispersion and, gasp, treble.   Anyway, hmmmm, "too funny".  I didn't think you could be "too funny".  That's like "too detailed", isn't it?  

[ekovalsky]

Well I am anxious to see a production of a DEQX line array.  Build it and it may find its way into my listening room !

[John Casler]

Correct me if I'm wrong, (and Lord Knows, I am not a speaker designing expert), but isn't the problem in a "conventional line array" (a line of cones and a line of ribbons or planars) that the cones will have a phase/alignment problem depending on the listening distance from the speaker?

That is, the top and bottom cone drivers will be farther away than the center drivers.  This then will cause a probelm for the single DEQX channel to both phase and align the total line???

Now I think with the planar, or ribbon line (if multiple drivers are used) that the limited vertical dispersion will ameliorate this problem, and be more like a point source.

I have heard that the "outer cone drivers" can be wired to be early to be in phase coherance, with the center drivers, but wouldn't that still only be accurate "if" you know the "exact" listening distance?  Since more nearfeild or farfeild would throw it of quite a bit?

Obviously this would only be of concern in the "longer arrays", and with more nearfeild listening.

Am I "way off base" in this idea??

[John Ashman]

Correct, DEQX wouldn't fix the delayed issue of the top/bottom drivers (unless you used separate amps and digital delays).  One company is curving their tower which would make it better at closer listening point, but not sure if this would damage the way a line array is supposed to function.  DEQX would help in the horizontal domain though.  When you measure a line array with DEQX, the measurement will look more like it was made with a magic marker than a fine-tip pen.  That's why you need to measure from a longer distance, which requires a big anechoic chamber to do well.  Or, you can measure a single driver/tweeter as a point source, then apply that measurement curve to the entire array.  This is why point sources will still likely be better in smaller rooms at close distances.

[_scotty_]

John the drivers can be arranged in a vertical arc the radius of which would define the distance from the center of the array that corresponds to the point
at which perfect phase coherence exists.  In any phase coherent multi-driver speaker system there is only one axis that will exhibit correct phase response.
In the case of a phased array there is only one axis and one distance where
a condition of phase coherence exists. The Dunlavy SCVs are one example
of a focused phased array .
Scotty

[John Ashman]

All true. The problem there is getting your head at the exact right height. Otherwise, the sound will degrade above and below with an inch or two difference.  That was the problem with the Dunlavys.  WAY too beamy and a sweetspot about 1"x1".  Okay, maybe not *that* small, but close.  Great designs, just purposely limited for the single audiophile.  It's really wonderful to have true phase/time alignment and great dispersion!

With a line array, your brain is picking up the first signal in and largely, but not totally ignores the delayed signal.  This is why the image tracks you as you sit or stand.  If you're 6" off the ground, so will be the image.  With a point source, it stays at the same height as the speaker.  Different strokes, different apps, different customers.

[ekovalsky]

John,

This isn't a problem with the ribbon tweeters.  Because of their limited vertical dispersion you will hear only one or two adjacent drivers at a particular listening height.

With cone mid/bass drivers there will be some time delay from the drivers at the ends of the array compared to those near the center, which will be a bit closer to the listening positioning.  The ear largely ignores this delayed sound arrival.  Also, most of the taller designs incorporate power tapering so output of the drivers at the top and bottom of the array is diminished by 3dB or so compared with the drivers nearer to the center.

There are ways to design the speaker to overcome this issue.  One would be to build a concave baffle to maintain equidistant distance between listening position and each driver.  The Gryphon Poseidon can be yours for only $130,000  Problem with a focused array like this is the sweet spot will be limited to one chair, preferably one with a head vise!

Another solution is to step the baffle so that the top and bottom drivers are more forward than those near the center.  Though not line arrays, see the PBN Montana speakers are built with this type of baffle.  Unfortunately the steps introduce diffraction artefacts and also will require increase in driver center-center spacing which results comb filtering at lower frequencies.  I've yet to see a line array designed this way, probably for good reason.

A conventional line array with a flat baffle could be designed so that a small time delay is introduced to the upper and lower mid/bass drivers compared to those near the center.  But this would require additional DEQX / TacT devices and amps and the added expensive is probably not justified.  

Dr. Jim Griffin's white paper on line array design is required reading on this subject.  Basically if center-center spacing of the mid/bass drivers is kept to a minimum and crossover points are properly chosen, a flat baffle line array launches an effective cylindrical wavefront within the nearfield and the phase/alignment problem does not really develop.

What I see as a bigger issue with conventional line array designs is that they are by nature two way designs requiring a crossover around 2khz.  Smallish mid/bass drivers must be used to keep driver center-center distances to a minimum, and this limits low frequency cutoff.  Big cabinets and vents help but still none of the big line arrays are true full range systems.  

A really great line array design is the Martin Logan Statement E2.  From 200hz up there is a curved, continuous electrostatic dipole array 68" tall.  Since the panels don't move a ton of air, mid bass is supplied by eight 7" cones operating in dipole from 50-200hz.  Below 50hz bipole corner woofer towers (8-12" drivers per side) take over.  The 200hz crossover is a passive design built into the dipole towers (along with the transformer for the electrostatic panels) but the 50hz crossover is an external active unit, for which a DEQX or TacT could be substituted probably to very good effect.

Too bad those ML bipole woofer towers aren't available separately, they would be a ideal mate for the SoundLab Ultimate-1 with a DEQX or TacT providing a 40-50hz crossover.

Correct me if I'm wrong, (and Lord Knows, I am not a speaker designing expert), but isn't the problem in a "conventional line array" (a line of cones and a line of ribbons or planars) that the cones will have a phase/alignment problem depending on the listening distance from the speaker?

That is, the top and bottom cone drivers will be farther away than the center drivers.  This then will cause a probelm for the single DEQX channel to both phase and align the total line???

Now I think with the planar, ...

[John Ashman]

Ed,  you're kind of right, in theory, but tweeters are also far more problematic in an array, despite being *slightly* more directional (not *nearly* directional enough!), because the wavelenths are so small, the damage would be far greater for the same distance differential.  

In *theory*, you're right, the sound from the top and bottom of the array would never reach you, it would be *pushed* above you and below you and you would only hear the drivers at ear level.  Of course, it wouldn't work exactly that way because it's not a true line source, but it's close enough for government work.  That's why line arrays aren't as problematic as a phased array when it might be easy to assume otherwise.  Still, I think their advantages are really best utilized in a largish room.

[Marbles]

Who is "Ed"?

[John Ashman]

ekovalsky?!?  I thought I picked up that his name was Ed somewhere in the thread, apologies if thas is incorrect.

[ekovalsky]

Eric, actually.  But who cares  

[John Ashman]

Doh!!!  Sorry!

[John Casler]

John,

This isn't a problem with the ribbon tweeters.  Because of their limited vertical dispersion you will hear only one or two adjacent drivers at a particular listening height.
 ...

Hi Eric,

Obviously I'm very interested in this.  In my interest, I'm not looking at ribbon tweeters, but midrange drivers, and this dispersion character is the same as the ribbon tweeters.  

With cone mid/bass drivers there will be some time delay from the drivers at the ends of the array compared to those near the center, which will be a bit closer to the listening positioning. The ear largely ignores this delayed sound arrival. ...

I wonder how and who measured or concluded, that the ear ignores these signals.  I know a simple coffee table in front of your listening area totally destroys soundstage and imaging with less amplitude and delay???

 There are ways to design the speaker to overcome this issue. One would be to build a concave baffle to maintain equidistant distance between listening position and each driver. The Gryphon Poseidon can be yours for only $130,000  Problem with a focused array like this is the sweet spot will be limited to one chair, preferably one with a head vise!
...

Heh, heh, While I don't have a problem with "limited" sweet spots (in fact I embrace them in the lateral side to side plane), I think that the concave solution "adds" a positioning parameter that many would have a hard time living with, and that is a "front to back" specific position.  While the Poseidons are probably pretty incredible for $130K, it appears that with "TWO" curved arrays that listening position would be so precise as to make them almost un-usable

If you were too close it would be wrong, and too far away it would be wrong

Even electrically staggering a straight array would seem to offer the exact same limitation.  As John mentions and you recognize, the only solution for a cone dispersion line of drivers, is a DEQX channel for each driver  

 What I see as a bigger issue with conventional line array designs is that they are by nature two way designs requiring a crossover around 2khz. Smallish mid/bass drivers must be used to keep driver center-center distances to a minimum, and this limits low frequency cutoff. Big cabinets and vents help but still none of the big line arrays are true full range systems.

A really great line array design is the Martin Logan Statement E2. From 200hz up there is a curved, continuous electrostatic dipole array 68" tall. Since the panels don't move a ton of air, mid bass is supplied by eight 7" cones operating in dipole from 50-200hz. Below 50hz bipole corner woofer towers (8-12" drivers per side) take over. The 200hz crossover is a passive design built into the dipole towers (along with the transformer for the electrostatic panels) but the 50hz crossover is an external active unit, for which a DEQX or TacT could be substituted probably to very good effect.
...

Hmmmmm... do you really want a dynamically limited, bipolar speaker from 200Hz up?

I wonder if there are any other driver choices, that have limited dispersion, can be arranged in an array, have good dynamics, and operate from the 200Hz up region???      

[ekovalsky]

Hmmmmm... do you really want a dynamically limited, bipolar speaker from 200Hz up?

I wonder if there are any other driver choices, that have limited dispersion, can be arranged in an array, have good dynamics, and operate from the 200Hz up region???      

Not that I can think of  

I do know of one planar unit which has some uncontrolled output in the 1-2khz range and lacks dynamics at its bottom end because it cannot move much air.  It also has requires mechanical modifications to function properly.  But it is a bargain in that it only costs about $25/unit in quantity.  
         

B&G has a new planar driver coming to market with a corrugated drive element and a suspension (allowing for true pistonic motion) that may adequately cover 100hz to ribbon tweeter range.  Watch out for this one...    a flagship system using it may be available by the end of the year.

I would not consider a 68" tall, 16" wide electrostatic panel dynamically limited either.  It won't move enough air in the low-mid bass on its own to have ideal impact which is why a woofer array takes over below 200hz.  And it is a dipole, not a bipole.  Big difference -- dipoles have basically no lateral output since the front and back waves cancel to the sides of the driver element.

What I'd really like is to mate a SoundLab Ultimate-1 (which I can assure you is not dynamically limited above 35hz) to the ML bipole bass towers.  Mmmmmmm.....

[John Casler]

Not that I can think of  

I do know of one planar unit which has some uncontrolled output in the 1-2khz range and lacks dynamics at its bottom end because it cannot move much air.  It also has requires mechanical modifications to function properly.  But it is a bargain in that it only costs about $25/unit in quantity.  
         

..

I wonder what that could be

Not sure about the 1-2khz control, or even the lack of dynamics some claim since a frequency at a specific SPL either is or isn't produced at that SPL.

But I think in the near future, we might see a new design of "driver" that can possibly do what some feel is needed