Beyond the Ariel

[phase_accurate]

What's your opinion about using linearisation FIR filters for music playback?

If you belong to the camp that dislikes stored-energy then you'd rather not use FIR filters.
If you are afraid of the post-ringing caused by stored energy the use of FIR filters would introduce something that is even worse: PRE-RINGING.

Regarding the Manger transducer: Yes the decay shows effects of stored energy mostly around the two infamous holes at 0.8 and 1.6 kHz.
Above that it is very good and its initial rise-time is very good at just 13 us.


Regards

Charles

 

[jzagaja]

The problem with ringing is that it is created everywhere in recording productions now - when antialiasing digital filters are used. What can we do? Back to the analog sources or use AD/DA without AA filter (NOS-ADC).

If Manger is so good then we can easily evaluate the codecs artefact and this ringing. If the ringing is there then why measured CSD after DSP applied is so clean (much better than Manger MSW, and 4 times less price)? Is it out of band ringing?

Jordans after linearisation have impulse response very close to the Dirac pulse. My ears aren't trained maybe that's why I don't hear the cone breakups (but I'm attending to jazz clubs few times per month).

Do you have Manger's square wave response? Mine looks like this:

http://audiostereo.lukarnet.com/gfx/700000/708561_3.gif
(1kHz)

And step:
http://audiostereo.lukarnet.com/gfx/700000/708561_1.gif
http://audiostereo.lukarnet.com/gfx/700000/708561_2.gif

 

[phase_accurate]

You can see the step response of the bare Manger driver on their website and it is for sure better than the one from your link.
I bet it would even improve significantly if the response errors are EQed out.
It wouldn't be difficult to guess the square-wave response from this.

The step response of a Manger + Woofer combination using a prototype active minimum-phase crossover of mine can be found here:

http://www.diyaudio.com/forums/showthread.php?postid=1139711#post1139711

Regards

Charles

 

[jzagaja]

Step response of Zerobox in Gelsenkirchen is smoother than from metal cone - see appended picture. Does it explain the subjective impression that Jordan JXR6 and Manger MSW are completely different drivers? Jordan is electrostatic archetype with breathing clarity and MSW is extremely smooth but without this electrostatic clarity.

Or maybe we should look in the time domain in other way - mentioned power cepstrum used by Holland and Newell for direct radiators and horns comparisons.

 

[chrismercurio]

Jumping in...Mangers and Hemp

I had Mangers in a diy system for about 3 years. It was extremely difficult to mate them to anything else and have the sonic picture sound complete. On their own, despite the limited dynamic range and top octave extension, they are quite special. It was a little like having a JX-92 that sounds MUCH better for a LOT more money.

After reading the whole thread this morning I think it is worth mentioning that Hemp Acoustics has an 8", 12" and 15" coaxial drivers available. The concentric tweeter is a compression driver, and the cones are obviously hemp. They also have a 15" designed specifically for open baffle use. Lynn mentioned a desire for hemp, they industrial type of course, earlier in the thread.

C

 

[Lynn Olson]

Personal Preferences

I had Mangers in a diy system for about 3 years. It was extremely difficult to mate them to anything else and have the sonic picture sound complete. On their own, despite the limited dynamic range and top octave extension, they are quite special. It was a little like having a JX-92 that sounds MUCH better for a LOT more money.

After reading the whole thread this morning I think it is worth mentioning that Hemp Acoustics has an 8", 12" and 15" coaxial drivers available. The concentric tweeter is a compression driver, and the cones are obviously hemp. They also have a 15" designed specifically for open baffle use. Lynn mentioned a desire for hemp, they industrial type of course, earlier in the thread.

C

The Hemp Acoustics 12 and 15 drivers are in fact my first choices for the new project - I am awaiting specs for the horn and bass unit to evaluate their suitability. Hemp-composite cones have much less breakup and "cone cry" coloration than anything else I've heard so far - I surmise the hemp is tougher, more elastic, and more resistant to failure modes than paper, carbon-fiber, or other synthetics. I've always been impressed that ropes strong enough to hold ocean liners were made of hemp. That isn't a light-duty application!

The G load forces where the voice-coil former meets the cone are extremely high, and many materials audibly fail at higher SPLs. In addition, when the speed of sound in the VC former and cone are different, that creates a reflection that propagates up and down the cone, and sets it up for bending modes. This is why the choice of glues joining the VC former to the cone are absolutely critical to the sound of the speaker - it's both a mechanical weak point and a region where two materials have different speeds of sound.

As for the Manger, eh, not my cup of tea. Not a fan of PHY either - it's a popular type of "hifi" sound I don't care for. The audiophile "exotics" have a rather artificial and unnatural sound to my ear, and the waterfall typically has a chaotic character in the critical 1 to 5 kHz region.

This is the frequency band where hemp excels - even the Tone Tubby 12" guitar speaker has smoother characteristics than the fancy exotics. See the measurements I've made of drivers used in very expensive and well-reviewed audiophile speakers, and compare the smoothness of the rolloff region (a figure of merit that I consider very important).

 

[IslandPink]

Baffle Design etc

Hi Lynn,
Nice to see you thinking about a very timely and relevant subject ( for me ! ) .

It would be really good, as Scottmoose has said , to get JamesD onto this thread, because he has done some serious modelling and practical work with OB design, driver choice and series crossovers . His 'Quasar' I design which is buried somewhere in Bert Doppenberg's BD-Design forum, is a very capable design . The ( 20mm Acrylic ) baffle transitions from Open at the top to shallow U-baffle at the bottom and the bass radiation I believe goes to Cardoid at the lower end as a result . I can vouch for the performance , and the quality of the drivers ( AER and Supravox ). Bass tone and speed is glorious .

James has been very tied up with work projects and family stuff in recent weeks . He may be back in audio circulation in the next week or two , so I'll alert him to this discussion .

Good luck with the recovery, Lynn . I'm sure Thom will have you on the treadmill soon !

Mark

 

[BudP]

Indeed. just what is needed, folks with working experience who have solved some of the mounting plate / driver / room problems successfully and also not successfully. Those errors will teach us as much, or more, than their successes.

We need to get Gary Pimm to discuss his system, choices, and a few pictures too. He has solved the low frequency issues and has gone a long way towards solving some of the other issues.

Bud

 

[Lynn Olson]

Hi there, IslandPink!

Hi Lynn,
Nice to see you thinking about a very timely and relevant subject ( for me ! ) .

It would be really good, as Scottmoose has said , to get JamesD onto this thread, because he has done some serious modelling and practical work with OB design, driver choice and series crossovers . His 'Quasar' I design which is buried somewhere in Bert Doppenberg's BD-Design forum, is a very capable design . The ( 20mm Acrylic ) baffle transitions from Open at the top to shallow U-baffle at the bottom and the bass radiation I believe goes to Cardoid at the lower end as a result . I can vouch for the performance , and the quality of the drivers ( AER and Supravox ). Bass tone and speed is glorious .

James has been very tied up with work projects and family stuff in recent weeks . He may be back in audio circulation in the next week or two , so I'll alert him to this discussion .

Mark

Well, that's along the lines of what I've been thinking of since last summer - a baffle that is flat for the upper driver, and has (one) side panel that extends maybe 12" deep at floor level. This extends the path-length for the bass driver to maybe 24" or so, yet leaves the diffraction-sensitive wideband driver alone.

One difference between using an AER/Lowther and a potent 12 or 15-inch coax is not needing to use a highpass electrical filter for the mid/hi driver - I can run the 400-watt prosound driver full-range, with no cabinet loading, and not fear the driver will be destroyed. I'd never dare try that with an AER/Lowther with its 1mm of linear excursion and very soft suspension.

But - if you can just find a big cap of high quality (not easy) - an AER/Lowther would most certainly be a candidate for the mid/hi driver, since the efficiency from 500 Hz on up is pretty high, and the big boy down below can handle the bass power. After listening to the Ariel for 12 years, though, I kind of hunger for the stupendous 120dB peaks of horns and prosound gear, and that would be certain death for the AER/Lowther, at least as a direct radiator.

The real challenge is going to be domesticating the prosound stuff - this is the part where I'm jumping into the wild blue yonder. (Well, once I get walking again. First things first. The next X-ray on May 1st will show whether I can start putting any weight on the left leg - so far, it's toe-touch only, and walkers, crutches, and wheelchairs to get around. If the X-ray and the surgeon give the green light, then I can start up physical therapy again.)

 

[jzagaja]

As for the Manger, eh, not my cup of tea. Not a fan of PHY either - it's a popular type of "hifi" sound I don't care for. The audiophile "exotics" have a rather artificial and unnatural sound to my ear, and the waterfall typically has a chaotic character in the critical 1 to 5 kHz region.

Don't understand. "Exotic" drivers have clean and robust either CSD and impulse response. Don't you like ESL57 now? For example Bandor 50mm unit is popular among musicians in UK (Doreen Bance words) - are they stone deaf?

I understand that dynamic range is important for either speakers and photographic cameras but do you really need a rock concert in your house?

I understand also that we cannot describe the driver subjectively just looking on the graphs.

 

[ScottG]

120 db!..

..then you are almost exclusively looking at large diameter pro coax units, high pass limited, with a compression driver that will be horn loaded by the cone for a good bit of their bandwidth. (..line arrays are the other option, but they typically don't have the dynamic "feeling" in comparison.)

My biggest problem with such a design is:

1. Compression drivers and dampening - they are WAY over dampened, and even with a high output impedance amp are going to sound quite "pin-point" and less "3D".

2. The substantial portion of sub-optimal cone-horn loading will be smack in the midrange.

I can't help but think you would be FAR better off with something like the Azurahorn (Le Cleac'h profile) extending down to almost 500 Hz. Then the trick would be to obtain a compression driver with out as much physical dampening.. Perhaps a Cryo'ed Radian 950?

For the dipole midbass I'd almost certainly recommend the Jensen Neo 15-150 in the 16 ohm version x2 wired in parallel. It has low mass for its sd, decent linear excursion, high fs., moderate "q", and a freq. response that is near tailor made for an open baffle dipole in the <300 Hz range when considering gain from amplifier impedance interaction on typical high output impedance amps. With such an amplifier you would be looking at a fairly flat response to 50 Hz, and considering its sd, high fs, and using two drivers per channel - high spl's should be possible (..and of course low compression at sane levels).

Though expensive.. that (with a wide dispersion super tweeter and a very good subwoofer) should make an excellent system.

 

[Lynn Olson]

Don't understand. "Exotic" drivers have clean and robust either CSD and impulse response. Don't you like ESL57 now? For example Bandor 50mm unit is popular among musicians in UK (Doreen Bance words) - are they stone deaf?

I understand that dynamic range is important for either speakers and photographic cameras but do you really need a rock concert in your house?

I understand also that we cannot describe the driver subjectively just looking on the graphs.

One of the nicest things about being an old fart is I no longer feel a need to defend my tastes in music and hifi. During the time I was doing reviews for Positive Feedback magazine, I discovered my tastes align with only about 2 to 5% of audiophiles, and I was most certainly out of sync with all the mainstream magazine reviews. So be it.

I can rant and rail from the sidelines, and best of all, don't have to satisfy nitwits in the Marketing Department, or suck up to the High and Mighty Big Magazine Reviewer. Similarly, THX, NRDC and other name-dropping doesn't impress me that much - I've met them, and eh, so what. I'm more a fan of obscure folks most people have never heard of - from my perspective, these folks are the real innovators, not the marketing bozos with the four-colors pictures in the magazines.

Most mainstream high-end stuff sounds pretty unpleasant to me - mechanical, filled with annoying resonaces, rather poor tone colors, and very deficient in spatial qualities. At the last RMAF, I was astonished how similar the high-end transistor amps sounded - and I don't mean neutral or accurate - they sounded noticeably hard, forward, and metallic, enough so that it was clearly audible before you walked into the room. At least it saved time - if I heard a chrome-plated metallic coloration in the hallway, well, that was one room I could skip.

My hifi pal John Atwood noticed the same thing - and both of us have enough transistor design experience to be a bit surprised at the uniformity of the "high-end" sound of the transistor brigade. Well-designed transistor amps can sound quite different from each other, some neutral, some soft and gentle, and some aggressive and punchy, depending on design philosophy. Our mutual guess was the Big Two Magazines were enforcing a certain type of sound in order for products to get a good review, and over a period of time, this had an effect on the mainstream manufacturers. Once again, I was grateful I no longer have to bow and scrape to these clowns, and can design things for my own tastes, not the boss down the hall.

Moving on to the other comments, you can only sort of guess what a driver sounds like from the CSD. Huge peaks, though, are plenty audible to a skilled listener (one with design experience, not a magazine reviewer). With more experience, you start to discern the sound of an "overcooked" crossover - I've certainly done my share of those!

A cluttered display, as opposed to massive un-equalized peaks, sounds different - to me, it sounds kind of hashy, like poor-quality digital or low-quality transistor amplification. I'm not surprised there's been a fad for tube preamps to be used with CD players and transistor power amps - the tube preamp is basically a 2nd-harmonic generator to "cover up" the cluttered and disordered spectra of the other components.

This "hash" type of distortion (and partially-correlated noise) has an interesting subjective effect as the percentage goes down.

1) At high levels, it sounds like mass-fi junk audio - a portable MP3 player, car stereo, home theater, etc.

2) As it drops towards "audiophile" levels, the sound becomes more open but quite often kind of sterile as well, with lots of detail and "transparency" but muted or absent tone colors. (Most non-DIY audiophiles never experience sound beyond this level.)

3) As the designs in the electronics and speakers start to show some refinement (designed by engineers with background in other areas of electronics or perceptual psychology), tone colors become more vivid, and the impression of in-the-room realism grows. This is fairly rare in commercial high-end equipment, and is a big part of the incentive of the DIY movement of the last fifteen years.

4) At the highest levels (in research labs and the most advanced DIY systems), the "hifi" or "canned" quality disappears quite suddenly and the emotional impact becomes much greater. I've had other designers describe this as the "right-angle turn" as a perceptual threshold is crossed - you're no longer even talking about transparency, but so caught up in the tone colors and emotions that words no longer describe what you're hearing.

I've heard all of these levels, and find it very difficult to sustain anything close to Level 4. The equipment at this level is annoyingly tempermental, and you find you're working with clearly audible effects that are not well documented in the literature. Anyone that claims everything they do is at Level 4 is a b*******er, not a serious designer. I am not aware of any commercial equipment at any price point that is even close to Level 4, and I've heard a lot of stuff over the years.

Moving on to ScottG's insightful comment, an undistorted 120 dB is probably asking for too much. (110 or 115 dB, maybe? Not all that uncommon for pro-quality direct-radiator studio monitors.)

I'm trying to stay away from horn sound - as good as the Azura horn is - and I've heard it at some length - and it's not quite what I'm looking for. The Azura is probably at the top of the horn mountain, but it still sounds like a horn in spatial and tonal terms.

There's an undefinable something about horn loading that sounds "projected" into the room, instead of the float-in-space quality of the most advanced omnis (MBL or the Walsh project in this forum) or the best dipoles. We're talking best-of-breed here, and all of these technologies have immediately distinctive spatial and tonal characters. I want the dipole sound, not the horn sound - to me at least, they are on different ends of the spectrum, and each has their fans and adherents.

The wave expands into the room in a quite different way in the critical midrange region, and this naturally affects the perception of the sound. To me, horns sound more immediate and concentrated, but also more in your lap, as opposed to impressions of far-away distance and space. This is a function of the proportion of energy radiated onto the side and back walls, or put another way, the ratio of direct-arrival sound versus total room energy sound. Along with different energy-storage properties, this is the area of greatest difference between direct-radiator dipoles and horns.

The coax part of the dipole is the most unpredictable - and most expendible - part of the design. I could just as easily go with a large-diaphragm bass-mid and a pro-quality ribbon tweeter, if the coaxes turn out to be kind of sucky (and that could easily happen, based on the indifferent performance graphs I see for the woofer part of coaxes). That would avoid horn loading entirely, giving semi-180 degree dispersion up to fairly high frequencies. At least dipole baffles are simple to build, unlike the scary transmission lines I'm used to.

 

[ScottG]

There's an undefinable something about horn loading that sounds "projected" into the room, instead of the float-in-space quality of the most advanced omnis (MBL or the Walsh project in this forum) or the best dipoles. We're talking best-of-breed here, and all of these technologies have immediately distinctive spatial and tonal characters. I want the dipole sound, not the horn sound - to me at least, they are on different ends of the spectrum, and each has their fans and adherents.

The wave expands into the room in a quite different way in the critical midrange region, and this naturally affects the perception of the sound. To me, horns sound more immediate and concentrated, but also more in your lap, as opposed to impressions of far-away distance and space. This is a function of the proportion of energy radiated onto the side and back walls, or put another way, the ratio of direct-arrival sound versus total room energy sound. Along with different energy-storage properties, this is the area of greatest difference between direct-radiator dipoles and horns.

The coax part of the dipole is the most unpredictable - and most expendible - part of the design. I could just as easily go with a large-diaphragm bass-mid and a pro-quality ribbon tweeter, if the coaxes turn out to be kind of sucky (and that could easily happen, based on the indifferent performance graphs I see for the woofer part of coaxes). That would avoid horn loading entirely, giving semi-180 degree dispersion up to fairly high frequencies. At least dipole baffles are simple to build, unlike the scary transmission lines I'm used to.

Exactly! Not only is everything "projected" forward, but also "sound stage"/reverberant reflections don't sound "right". Even an omni like the Duvel's has something "wrong", but it isn't the same kind of "wrong". (..obviously I'm no descriptive laden reviewer )

You can alter this perspective somewhat for the better with a really good Le Cleac'h horn the way Le Cleac'h himself does - i.e. crossover well above the horn cut-off. Still, (to me) it just isn't "right". In a heartbeat I'd sacrifice spl's for something like that midbass driver open baffle and a fostex 126e radial (upward firing no waveguide) to about 1.8 kHz, and then look for a good ribbon above that.

With that walsh thread in mind here is a photo that might interest you.. (under "customer systems" and then "Sven Boenicke")

http://www.audio-consulting.ch/

Makes me wonder what such a driver, (in such a configuration), might be like with these for tweeters:

http://www.warrengregoire.com/hifi-stereo-ikonoklast-tweeters.htm

 

[BudP]

ScottG,

I did apply that EnABL pattern to a set of horns once..... To some Radio Shack systems with a high efficiency 15 inch woofer, a mid and a tweeter horn, not time aligned in any sense. The untreated speakers were not much fun to listen to.

The cone, front plate and horns were all treated. The results were odd.

The result was a VERY forceful speaker with razor sharp detail and near perfect time alignment, as the EnABL pattern appeared to enforce this condition onto the overall sound, even though all of the drivers were mounted flush with the front cabinet surface.

Did not make them any more pleasant to listen to though, but they were a lot more efficient and would play quite a bit louder than before so their new owner was very happy.

Did this for some teen owned boom boxes too, with the same results, but was always worried I would have to face the irate townspeople with their pitch forks and torches so I stopped doing that. Then the cars full of woofers and amps began to show up and I knew I was again safe.

Bud

 

[mpmarino]

Perfect thread!

Since the demise of my tannoy drivers I have been banging my head against the wall as what to do, those are big shoes to fill. I have a set of Eminence Beta 12cx and compression drivers doing nothing in the garage. I have yet to hear a 'regular' speaker that 'takes me away' like those tannoy CD drivers did. I have been leaning towards trying the eminence in an open baffle (Qts >.4).

This thread is on target with what I want and I have a feeling the thread starter is on a quest for a sound I would enjoy

>following closely<

 

[Lynn Olson]

Exactly! Not only is everything "projected" forward, but also "sound stage"/reverberant reflections don't sound "right". Even an omni like the Duvel's has something "wrong", but it isn't the same kind of "wrong". (..obviously I'm no descriptive laden reviewer )

You can alter this perspective somewhat for the better with a really good Le Cleac'h horn the way Le Cleac'h himself does - i.e. crossover well above the horn cut-off. Still, (to me) it just isn't "right". In a heartbeat I'd sacrifice spl's for something like that midbass driver open baffle and a fostex 126e radial (upward firing no waveguide) to about 1.8 kHz, and then look for a good ribbon above that.

With that walsh thread in mind here is a photo that might interest you.. (under "customer systems" and then "Sven Boenicke")

http://www.audio-consulting.ch/

Makes me wonder what such a driver, (in such a configuration), might be like with these for tweeters:

http://www.warrengregoire.com/hifi-stereo-ikonoklast-tweeters.htm

My my my! Ain't Sven's system something! Lookit that field-coil 40 cm driver - what a sight to see! Freaky tweeters as well!

The Le Cleac'h implementation makes sense - as well documented by Geddes, conventional horns - yes, that includes the Azura - go into "high order modes", or as Geddes abbreviates it, HOM. Dunno why he calls the modes high-order - they start right at the bottom of the horn's usable range and are troublesome for at least two octaves above nominal cutoff. In essence, the horn behaves like a pipe at the lowest frequencies, with the associated pipe modes, which appear as ripples in the impedance curves. By crossing higher, Le Cleac'h is avoiding the worst of the HOM. One gotcha is horns have fairly limited bandwidth to start with, and the Le Cleac'h crossover throws away another two octaves.

The origin of the highorder modes is simple enough - there's a sharp edge on one end of the horn, and for compression drivers, a hard reflecting surface on the phase plug on the other end. This is acoustically similar to a pipe with one end closed and the other end open.

To "de-Q" a highly resonant system like this, it helps to remove energy from each reflecting surface (and surface edges reflect just as much as hard surfaces). Removing the phase plug helps *if* the diaphragm is acoustically transparent and not backed up by a small compression chamber. No free lunch though - removing the phase plug invites self-cancellation effects at higher frequencies as the diaphragm size approaches the radiated wavelength. The non-coherent launch into the horn has the net effect of narrowing the HF dispersion at the top of the horns frequency range.

On the other end of the horn, EnABL treatments, Mamboni felt triangles, or cutting a pattern of holes in the horn would provide a gentler termination at the horn mouth than just a sharp edge looking into the room.

All of these techniques can be confirmed without expensive MLS measurements - the ripples in the impedance curve should diminish as the treatments become more effective, making the horn more like an infinitely long horn and less like a pipe of finite dimensions. As the HOM diminish, the "projected" sensation should diminish as well.

As for the success of the Geddes implementation, I dunno. I've never seen before-and-after impedance curves or the impulse-response of the horn portion of the Geddes Summa system, so there's no way to objectively evaluate the real-world success of the foam treatment techique.

If I follow the coax path, I'm definitely going to have to do something about the HOM of the compression driver - however, the techniques to improve the HOM should also improve the behaviour of the woofer cone as well (felt strips, EnABL, etc.)

 

[Variac]

Lynn, Have you ever heard the Supravox field coil drivers? (like in Sven's Speakers). I think they come in a couple of diameters. They have been discussed in some threads here and the imressions were good.

Just curious. I think I prefer the way you are starting out with cheaper drivers and some mods. If that pays off, it is a much greater benefit to us all.

 

[phase_accurate]

The disadvantage of the "oscillating throat" which is inherent to those coaxials that use the woofer cone as continuation of the horn can be solved by the construction shown under the following link. Due to the shape of the horn it acts more like a phase-plug for the woofer compared the horns of the Altec-style coaxes which act as obstacles. It is of course still a compromise (like any speaker) but an interesting one admittedly.

http://www.avantgarde-acoustic.de/hoerner/imgbrowser.php?produkt=solo&bildnr=1

Regards

Charles

 

[planet10]

Re: Hiya Bud, Good Questions

4. The thread mentioned earlier about the ISIRIS shows around Page 5 or so some sort of modelling program that simulates the frequency response for the given baffle shape. This looks interesting, and the freq response shown is entirely consistent with what I've measured on MLSSA for open baffles.

Svante's The Edge...

http://www.tolvan.com/edge/

dave

 

[jzagaja]

Lynn - I don't understand people who looks for a reproduced sound that involves emotionally. It's a typical audiophile mania to chase the rabbit. For me life is simple - I'm doing my personal recordings of live concerts and looking for close reproduction. For me Jordan drivers do the job perfectly for the money. If a person is not interested in live music then it is a good idea of using headphones to hear how a perfect speaker without resonances should work. Surprisingly most people won't chose this reproductors - they prefer colored sexy sound. Why? because they don't know how the musical instruments sounds in nature!

Have you seen Summa impulse and step?
Have you seen impulses before and after HOM reduction in Geddes OS waveguide?

I can send them private (they are too big for the forum).