The new speakers (well, the prototypes I heard) and the Gedlee Summa's share the property of good integration at close listening distances. 1 meter is way too close (and too close to measure accurately), but 3 meters and further works just fine. They're not near-field speakers ... way too large for that ... but the crossover and driver integration are very good at a 3+ meter listening distance. In subjective terms, integration is better than most commercially available 3 and 4-way systems, which have a kind of a disjointed sound when I hear them at hifi shows.
A very common problem with (commercially available) 3 and 4-way systems are phase angles between drivers that exceed 90 degrees. In addition to making measurements difficult (response changes dramatically with small changes in microphone position), the large inter-driver phase angles are directly audible as a quality of "phasiness" or lack of coherence. Part of the reason for the fullrange driver movement is a desire to get away from the phasiness that multiway systems can exhibit.
I don't know why commercially available high-end speakers have such poor crossover integration, with nulls and polar problems that are obvious as soon as you walk into the demo room. This is a solved problem in the DIY community, whether people are using miniDSP or passive crossovers.
Gary or Lynn I'm curious if you've ever heard Vivd Giya with classical? I was a bit surprised by these. Not being a fan of the magazine reviewer favorites speakers plus the almost alien looks of the Vivid I expected them to be cold as ice. I was pleasantly surprised about how even handed they were with complex symphonies. I thought they sounded like an active ATC except with better imaging and soundstage.
A brief visit to the webpage for Vivid Giya shows aluminum-cone drivers with Linkwitz-Riley 4th-order crossovers and a claimed efficiency of 91 dB. I suspect it might be just a bit lower than that, since metal diaphragms tend to be a little heavier than paper equivalents, and since aluminum is prone to very high-Q resonances, maybe a little bit of damping is sprayed on the back side of the driver (or not). My own guess for efficiency of the midbass driver (which probably sets the efficiency of the whole speaker) is in the 88~90 dB range, which would be very typical for aluminum-coned drivers.
No guesses about the sonics. It would all depend on how well the inherent modes of the aluminum are controlled ... in the crossover with notch filters, with sprayed-on damping material, or some other method of resonance control. The really exotic approach would be beryllium diaphragms all around, like the justly famous Yamaha NS-1000 studio monitor of the late Seventies.
The Giya models remind me of the B&W Nautilus speakers, with similar tapered-tube rear chambers and a low-diffraction cabinet profile using molded materials. B&W used to have LR4 crossovers back then, but the recent models have gone to low-order crossovers instead, while retaining the Kevlar-cone midrange.
The goal for the "Beyond ..." project was flat response, low diffraction, and efficiency in the 97+ dB/meter range, with subjective sonics biased towards classical music reproduction. For my own personal use, I'd like to push it towards 100 dB/meter, and that looks like a pair of 15" drivers are required, or if the space budget is relaxed, a midbass horn.
Gary's system is up and running, so I'd recommend sending him a PM if you want a personal audition. There are no direct commercial equivalents for what he's built.
Last edited:
For ~ 95db/watt efficiency.....
...there are many positive reviews for a 12" midbass crossed ~1200Hz to a large horn with 1" CD up to 18Khz.
-Dr. Geddes gets impressive directivity control from a 12" midbass under a large OS waveguide with 1" CD.
-PMA! gets impressive results from the JMLC 350.
-Put them together with a pair of 18" woofers.
...there are many positive reviews for a 12" midbass crossed ~1200Hz to a large horn with 1" CD up to 18Khz.
-Dr. Geddes gets impressive directivity control from a 12" midbass under a large OS waveguide with 1" CD.
-PMA! gets impressive results from the JMLC 350.
-Put them together with a pair of 18" woofers.
Attachments
Interested in this proposition : maybe a Radian 636 or 745 Be could be a good match.
For the 15 driver in MLTL : do you think Altec (416) can do the job ?
Can you advice how to MLTL time align the 15 ?
Recently, I tested many types of DSP crossovers with 2 way large horn monitors (Auguspurger / Kinoshita style) in an acoustically controlled room. Phase issue is surprisingly obvious with all of the non time coherent (=no linear phase) crossover.
I have never realized this issue until I actually hear it myself.
The easiest way to check this is listening to the mono, solo instrument recording, such as Casales's Bach, with one speaker only. If the step response is time coherent, the image is very small, like tiny full range speaker. But if step response is non time coherent like most of the commercial speakers, the image is vertically very high oval shape. I have heard many audiophiles praise the speaker with this pseudo vertical high image, which is actually nothing but the effect due to phase shift of the drives. It is probably a good sound effect when one want to mimic the concert hall experience, though.
The problem is, a time coherent "huge" speaker sounds like a bookshelf speaker with great bass, and I think the market is very small for that kind of thing. Many think a huge speaker should sound like a huge speaker.
PS: And this is why Dunlavy had closed their business. Only "Golden Ear" mastering studio guys bough them.
Last edited:
My point is everyone is looking for a small speaker sounds like a big speaker, not a big speaker sounds like a small speaker, therefore all commercial high end speakers are all out of phase, I guess...
I am not an expert, but I think LineSource's suggestion is a complete package, and changing one part throws the rest off.
That is: the BMS 4592Nd is a 2" coaxial.
Changing from that to a single diaphragm 1.4" driver will mean less low end capability, which then suggests either going to a higher crossover frequency or using a bigger (longer) horn, which will then give a narrower pattern.
For time aligning, the BTA (Lynn Olson) version uses physical adjustment - shifting the horn back and forward. But he is (or was) a picky electronics purist who rejected DSP for most purposes.
Systems using DSP can do the same thing electronically. This can allow for slightly better physical layouts. That is, the horn mouth and bass driver can be on the same baffle (like in the Summa speaker), rather than using physically separate parts which can reflect off each other. Layouts like the Summa are also a bit more space efficient, in that the space behind the horn is not "wasted".
There are MLTL designs available on diyAudio for the Altec 416 and 515 which model well for ~600Hz crossovers. Duke uses this "corner friendly" cabinet shape.
============
If you bi-amp with one (tube?)amp directly connected to the coaxial BMS 4592Nd or BMS 4594Nd with its passive crossover, you can include an analog or digital delay on the second amp to the ~600Hz woofer to compensate for the physically long throat (delay) on the horn.
For a 600Hz crossover, both the 2" BMS 4592Nd and 1.4" BMS 4594Nd would sound great.... the BMS 4594Nd looks like a better choice because it will have slightly superior high frequency extension from its 1.4" throat.
If you want an all passive crossover, you can select one of the "approved" "quasi-optimal" crossover circuits, and make minor adjustments to the crossover frequency to achieve the closest phase-shift to the mimic the equivalent physical alignment delay:
-------------------------------------------------
3rd order Butterworth Low Pass, -3dB @ Fx*0.87 (+)
3rd order Butterworth High Pass, -3dB @ Fx*1.15 (-)
Offset = 0.22*c/Fx
4.97 inches @600Hz
4th order L-R Low Pass, -6dB @ Fx (+)
3rd order Bessel High Pass, -3dB @ Fx*1.4 (-)
Offset = 0.29*c/Fx
6.55 inches @600Hz
4th ordrer Butt Low Pass, -3dB @ Fx* 0,93 (+)
4th order L-R High Pass, -6dB @ Fx (-)
Offset = 0.31*c/Fx
7.00 inches @600Hz
6th order Bessel Low Pass, -6dB @ Fx*1.25 (+)
2nd order Butterworth High Pass, -3dB @ Fx*1.3 (-)
Offset = 0.40*c/Fx
9.03 inches @600Hz
6th order L-R Low Pass, -6dB @ Fx * 1.06 (+)
3rd order Butterworth High Pass, -3dB @ Fx* 1,13 (-)
Offset = 0.445 c/Fx
10.05 inches @600Hz
6th order L-R Low Pass, -6dB @ Fx * 1.07 (+)
4th order L-R High Pass, -6dB @ Fx * 0.92 (-)
Offset = 0.465 c/Fx
10.50 inches @600Hz
Last edited:
Pretty good guesses Lynn, but the efficiency is lower: Vivid Audio Giya G3 loudspeaker Measurements | Stereophile.com
Very nice vertical and horizontal off axis responses in all honesty, but the inefficiency is a no go for me. If I was into boat anchor amplifiers and wanted great directivity control, I would keep a watchful eye on Andrew Jones and his upcoming designs with Elac. My only quibble is the inefficiency with his designs once again. But I sure appreciate the directivity control he brings to all his designs, and really liked his TAD's he had designed previously.
Best,
Anand.
Well spotted, they are designed by Laurence Dickie who did the B&W Nautilus. I thought their midrange (along with ATC) were closest I have heard to matching my ESL57 (maybe it is their dome midrange? I'm not sure). Superb dynamics and no distortion that my ears could detect even when playing the Decca Solti Ring Cycle at loud volume. I'm of the DHT low power fan camp as well. The G3 were something that came along and defied all my preconceived expectations of low efficiency speakers. So much so that I'm now thinking about Class A SS amps (Pass amps have always been dear to my heart since my first Aleph 30) with a DHT preamp.
I'd love to hear Beyond the Ariel project. Of the dozens of horn speakers I've heard I'm always hyper aware that I'm listening to horns, I imagine JMLC's horns would break this mold.
Last edited:
Same designer for the Giya & the flagship Nautilus. The Nautilus was practise.
dave
Don't think JMLC horns are a magical solution to negative associations you have with horns. It is all in the application not just in the horns. It is not like they do not honk! I take advantage of JMLC horns in my system, they have pros and cons...
Horns can not be thought without corresponding drivers, their relative positioning and the lobing they are going to create and addressing of these in the crossover.
Listening to classical, I think you would like to take advantage of a 1st order crossover if you go the multiway horns route. The impedance correction is also very important if you first order. Otherwise you will not get the filtered result you are after from each channel. So all in all, even though it will be a first order crossover it will be a complicated one and has to be designed by a competent person. Filter design is a scientific artform.
I also listen to classical and was not happy with the systems that can be had, then I built my own. You can see my system and crossovers in here: http://www.diyaudio.com/forums/multi-way/266064-full-range-5-way-horn-project-25.html
Which is also very astonishing : this a 4 ways... ?! If it is, this guy is a very talentuous designer...
So if the subjective dynamic gap with a "low" efficienty driver could give the same feeling than high efficienty drivers, one can ask if an OB 4 ways* could be even better ??? Or any hybrid ESL !
*: to use the drivers in their "linear curve" (before the first null) in relation to their sizes. One can ask if a large flat panel as Troels G. uses again can be also an alternative to putt less filtering and have a little better efficienty with low-mid & upper-mid (170 Hz to 2500 Hzz iirc with a 220 cm² Sd driver)
Last edited:
To go higher in efficiency, maybe a sealed Altec 416 instead of a 12' can go 100dB down to 80 or 100Hz. Crossover between SEOS and Altec would be say around 700Hz passive.
Under, as you propose a vented 15' or 18' for under 100Hz, being active. If active maybe it could also be sealed with high power amp.
About SEOS waveguide, I did read Gedlee papers about HOM, constant directivity and phase matching. He is speaking about Oblate Spherical horn : SEOS is, but not circular as he advices : any problem ?
For SEOS driver, a Be 1.4" Radian is mentioned up to 18kHz, could be a good match.
By the way I don't understand the need to go high in frequency above 15 kHz as if you make an audiogram, you will see that the majority of people have a fall of 20-30dB over 15kHz.
