Recommended DSP?

[Bill Fitzmaurice]

I thought one argument for having very low extension even if below musical content was that the phase can be made more linear at the low end. This would result in less group delay distortion.
- Eric

As soon as anyone starts talking about phase and group delay below 40 Hz your BS meter should be pegging hard into the red zone. Both are non-issues where human beings are concerned, as both are undetectable down that low. Quite literally your ears are too close together. Even at 80 Hz it takes a massive amount of phase shift and/or group delay to be heard, so much that it would be difficult to engineer in that much if you tried. Above 300 Hz the scenario is vastly different, but what is audible with four foot wavelengths simply is not with even twenty foot wavelengths.

[eweitzman]

I take it you are talking about directionality when you bring up ear spacing and frequency? There's no question that phase is not an issue at really low frequencies, but that phase/time is primarily how we localize sound from about 100Hz to 800Hz.

My understanding is that group delay distortion affects the harmonic structure of instruments, given that the fundamental and overtones will have different arrival times if phase change is not a linear function of frequency. This gets into the question of whether a fundamental and overtone with incorrect phase is audible to us, with cochlea that act as analyzers in the frequency domain. I don't know if this is a settled question, but have seen demonstrations where combining the fundamentals and odd overtones yield either square waves or triangle waves depending on the relative phase.

I'm pretty much convinced that poor time alignment of the frequency components of, say, the attack of bass notes alters the character of the bass instruments. Group delay distortion is the only explanation I have for what we call "slow" bass.

- Eric

[Bill Fitzmaurice]

I'm pretty much convinced that poor time alignment of the frequency components of, say, the attack of bass notes alters the character of the bass instruments.

- Eric

True, but all of it taking place above 80 Hz. The easiest way to confirm this is to play music though a sub only with a brickwall filter set at 80 Hz. Not only is there no attack or definition, you can't even distinguish pitch. As for 'slow' bass, that's a meaningless term as far as engineering is concerned, because it doesn't define anything. What the vast majority of loudspeaker and acoustic engineers do agree on though is that group delay is a non-issue below 80 Hz, as whatever influence it may potentially have is totally overwhelmed by room interactions.

[eweitzman]

I have done exactly what you have suggested, yes, there's no "attack" in the bass, but that proves nothing about the subjective differences caused by <insert phenomenon here> that vary from one sub enclosure to the next. The driver I asked about in the other thread was first in an EBS enclosure with three tuned passives and sounded subjectively much worse than it does in a sealed box. 'Slow' is merely a descriptive term -- what is the cause of the difference between the two enclosures with the same driver? It's neither extension nor distortion nor frequency response, so that leaves phase and group delay.

As far as being swamped by room modes, I am fortunate to have a pair of Linkwitz Orions which are fulllrange dipoles and a listening space which is very large, open, and irregular. There are no room modes discernible to the ear, and two measured narrow dips in the upper bass that are fairly innocuous. My sub (when turned on) takes over from the Orions below 50Hz. I don't have it in a corner or sidewall loaded, so as to not introduce modal peaks. So you can see why I'm interested in horn loading for more SPL at less power.

Still, the quality of the bass, or perhaps better said the character of low bass instruments (think massed double basses and large drums) takes a bit of a hit when switching in the sub versus full range dipoles.

AFAIK, minimal group delay distortion seems to be the key behind dipole performance and realistic bass.

- Eric

[Dave Non-Zero]

I know nothing about this sort of stuff, but i did think that putting the same driver in different enclosures would change the frequency response. Is that not the whole point of horn loaded enclosures? Sealed, ported, Xth order bandpass, they all alter the frequency response of the driver surely?

[LelandCrooks]

Linkwitz Orions
- Eric

SET amplification? Pursuit of the perfect sound will make you insane. Fun hobby though.

[Bill Fitzmaurice]

Still, the quality of the bass, or perhaps better said the character of low bass instruments (think massed double basses and large drums) takes a bit of a hit when switching in the sub versus full range dipoles.

AFAIK, minimal group delay distortion seems to be the key behind dipole performance and realistic bass.

- Eric

To tell the truth, I can't speak to this from personal experience. I've never owned a non-horn loaded sub, and never had a complaint about the bass in my systems. Nor have I ever heard a non horn-loaded sub that even vaguely approaches their performance, so I have no desire for one.

[Art Coates]

I've studied Linkwizt's site extensively. And I've built my own dipoles. I think dipoles do have merit for home audio - but it's just "another way to skin a cat" in a universe full of options of varying costs and benefits. To more specifically critique Linkwitz in achieving his own stated goals: I think he over-designed certain aspects of the Orion to the detriment of real world performance, and at an increased build cost. For example, there are drivers available which meet or beat his specifications in performance, including linear and nonlinear distortion, of the Seas and Peerless models he selected, and which are available at 1/3 the cost. With the availability of inexpensive digital processing, there is also little need for the expensive analog crossover. I realize he has arguments against digital processing - but this is a bit far-fetched as all modern recordings on any retail format (CD, DVD) have been processed digitally - with much improvement in a myriad of factors - only one of which is fidelity. Redemption from grossly overpriced components is somewhat achieved by the Pluto design, and by his own admission performance is essentially equal to that of the Orion in its intended bandwidth and output level. Unfortunately, it is also impeded by a needlessly expensive and proprietary analog crossover.

As for dipole bass - one only has to consider that increased cone excursion also increases distortion, and that a dipole sub would require much greater cone excursion than a horn enclosure would.

I use the AutoTuba with my dipoles, and with proper EQ correction they work great with absolutely no performance hit in timbre or muddiness.

To sum up: High end audio is a textbook case of the law of diminishing returns - mostly the case of expenditure and with returns in ego inflation. I admire his work, I really do, but the Orion isn't the last word in audio.

[eweitzman]

For example, there are drivers available which meet or beat his specifications in performance, including linear and nonlinear distortion, of the Seas and Peerless models he selected, and which are available at 1/3 the cost.

I would be interested to know which drivers you refer to. Are the specs measured at the high excursions that dipoles require? And also (...) if the tweeter meets the distortion specs way down at the 1.5kHz crossover. I respectfully suggest that Linkwitz probably looked at whatever it is you are thinking of. He's on record saying that the only thing that would cause him to revisit the Orion design is some improvement in driver technology. The introduction of the Peerless XLS drivers -- with their low distortion at high Xmax -- was what enabled him to design the Orions. [Note: smallish form factor is one of his design constraints, so 12's and 15's need not apply.]

I realize he has arguments against digital processing - but this is a bit far-fetched as all modern recordings on any retail format (CD, DVD) have been processed digitally

No recordings need the 6db/octave gain boost that a crossover for a dipole has, whether digital or analog. You'll either run out of gain in the bass or lose computational accuracy and introduce noise in the highs with digital processing.

I have seen one recent four channel OEM digital crossover board with separate analog gain stages following each digital channel. This may be usable for an active dipole crossover.

Redemption from grossly overpriced components is somewhat achieved by the Pluto design, and by his own admission performance is essentially equal to that of the Orion in its intended bandwidth and output level.

The Pluto has many of the same sonic characteristics as the Orion, but in spite of the neutral portrayal of the two on his website and the generally positive reception of the Pluto, it comes nowhere near the Orion in terms of realistic output levels and clarity and everything that follows from these two.

Linkwitz is no analog luddite BTW. There's no conceivable reason for him to drive *up* the price of the system with an expensive analog crossover or drivers.

As for dipole bass - one only has to consider that increased cone excursion also increases distortion, and that a dipole sub would require much greater cone excursion than a horn enclosure would.

I may not have been clear in an earlier post: I'm not looking for dipole subs!

As you can guess, I'm very partial to the Orions. This thread shouldn't become an ad for the speakers so I'll stop here.

- Eric

[Sydney]

BILL

I've never owned a non-horn loaded sub,

Frankly I am surprised.
I can appreciate your fondness for horns
Do you make a recommendation for minimum separation distance for 2 David's?
I heard, in the old days Klipsch did not recommended Klipshorns in small rooms.

I have also experimented with mounting woofers in alternate cabs: Large Infinite Baffles ( monopole? ), Large Bass Reflex, 4th order bandpass boxes, Passive and Semi Passive radiators, and servo controlled bass.
eweitzman wrote:

Still, the quality of the bass, or perhaps better said the character of low bass instruments (think massed double basses and large drums) takes a bit of a hit when switching in the sub versus full range dipoles.

How close are your subs to your dipoles? To my ears the closer I had my subs to my other drivers the better it sounded with respect to definition and "tightness" of bass.
As I am unfamiliar with Orions, at what frequency do front and back wave cancel? The baffle seems small compared to other dipoles.

[Bill Fitzmaurice]

Do you make a recommendation for minimum separation distance for 2 David's?

No, because in home-sized rooms where you'll use them distances between walls are so short that room modes and cabin gain are going to dominate anyway, so just fiddle with them until you find what works best. My first horn loaded woofer was a KHorn clone I built in 1971, and once you've heard what a horn loaded woofer sounds like you're not going to settle for less.

[Art Coates]

As you can guess, I'm very partial to the Orions.

Admittedly, my tone was harsher than I intended in my earlier post.

They can be improved upon. I stand behind all my claims, including admiring Linkwitz for his all his work, including the Orions and Plutos. Agree to disagree?

-Art

[Ron K]

Do you make a recommendation for minimum separation distance for 2 David's?

No, because in home-sized rooms where you'll use them distances between walls are so short that room modes and cabin gain are going to dominate anyway, so just fiddle with them until you find what works best. My first horn loaded woofer was a KHorn clone I built in 1971, and once you've heard what a horn loaded woofer sounds like you're not going to settle for less.

Very much agreed. I started listen to Aletc A7s and RCA double 15inch theater horns.Man the depth was breathtaking even in those days. Then a friend let me listen to his klipsch La Scalas. Cut open in his living room with a 90 Watt Valve Amp (Futterman OTL-1) I believe it was. Man that thing was down right scary and had such an open natural sound. Best thing was you could turn around and run your guitar or sing through it and it still sounded great!

Listen to a good horn first and you most likely be a horny convert!

[eweitzman]

Sydney:

The sub is about 5' to the right of the right speaker. The mains and sub are all about 8' from the listener.

I'll try to answer your question "at what frequency do front and back wave cancel?"

The front and back waves of a dipole cancel at all frequencies in the plane of the baffle. There's a key frequency that determines the useful upper limit of a dipole, so maybe that is what you are looking for?

Below this frequency, the response rolls off at 6db/octave and above it, the response is more or less flat, but with severe comb filtering. When 6db/octave EQ is applied, the response gets tilted clockwise so the previously-sloped lower frequencies become flat and the useless upper range gets rolled off. This frequency depends on the spacing between the two sources, the so-called dipole distance D.

Depending on the baffle implementation (flat baffle, H-frame box, etc) the distance D will correspond to some dimension of the baffle. In the Orions, D is about 15" for the woofers and 4" for the midrange IIRC. The baffle is 12" wide with maybe an inch between drivers and the edges of the H-frame, but the woofer's D is large because it is mounted in an H-shaped frame whose sides are about 15" deep. The crossover frequencies are ~140Hz and ~1400Hz, so the upper usable frequency cutoff must be somewhere above these two crossover points for the woofer and midrange.

Art:

I'm not sure we need to agree to disagree. You can get much higher SPL in the deep bass with other dipoles than you can with the Orions, but you'd have to accept a much larger speaker. Living-room friendliness was one of Linkwitz' design goals, so perhaps you can't improve on them while respecting this constraint. Maybe we can agree on that?

Incidentally, a friend has a set of Scanspeak drivers as used in Krekovsky's Not-an-Orion and hopefully he'll build that dipole someday and we can put them head to head (or pole to pole, or whatever.)

Ron K:

A couple of decades ago I figured out that the high end emperor had no clothes, so I learned some electronics and got into DIY tube amps and high efficiency speakers including horns. Never been a speaker builder, though. It's way too much work.

My last pair of horns were the eXemplar horns designed by John Tucker and Gary Markwort (?). These were inspired by A7's after these guys spent some considerable effort optimizing A7', coming to a dead end, and writing a couple of articles. Unfortunately, the leXemplar's ow bass wasn't that good. They use Altec 515-8G drivers with a tractrix (vertical) and CD (horizontal) horn and four extremely large sonotube ports for extension below what you can get from a ~2'x3' mouth (130Hz IIRC?). Apparently, some former Altec designer told them that the 515 was meant for this type of application. So I was never happy with the bass, but in love with the snap and dynamics and clarity in spite of a bit of coloration and brittleness in the mid highs. This was a two way system crossed at around 800Hz with remarkable integration at close distance, say under 10'. The highs were reproduced by a 909C compression driver in a similar but smaller horn.

I thought the strength of this and other horn systems lay in the low powered amps that drove them because they were usually so much sweeter to listen to than much more expensive systems with big amps and box speakers. Eventually I figured out that it was the low driver distortion that made it happen and tubes got in the way, but not in a bad way. Once I heard the Orions, I knew that you could have it all, that Linkwitz achieved that same magic with dynamic drivers and cheap amps without the drawbacks of horns. And with unbelievably realistic soundstaging and scale. But compromising deep bass SPL. Let's say I'm nearly ready to look for a new hobby after a couple of decades seeking audio nirvana. The last hurdle is to recreate, in the living room, the SPLs that I measured at a Bjork concert last month. (Where, incidentally, my wife and I were probably in the oldest one percentile out of 25,000 people.)

- Eric

[gdougherty]

Never been a speaker builder, though. It's way too much work.
- Eric

LOL, and I look at circuit design and say I'd rather build a speaker than an amp any day.