May even begin to approach true full range electrostatic panels

If your electrostatic speakers (Quads 57 or 63, B&W DM70 or Martin Logan) require repairs, you can get DIY kits here: http://www.eraudio.com.au/ESL_Repair...pair_kits.html

Hi Tom, stereo imaging seems more focused on cone driver dipoles that I have compared with electrostatics. Familiar with ESL63, and Apogee electrostatics, and quite a few cone driver dipoles including Kyron Gaia which have the best image definition I have ever heard.
https://kyronaudio.com.au/gaia.html

Thanks Studer! They're the five Soundlabs I pictured in Post #17 above. Early on, prior to the current construction design, they were fragile, and needed to be returned to the factory for repair if arced. These new 7 1/2 foot 845PX (sans the wood trim) have been super reliable, even driven to above real orchestra SPL's.

Tom

20+ years after hearing Peters' stacked Quads i listened to the first Eggleston Alondra (i think) which i was told used 6db crossovers..sounded amazing.
I believe Bob Ludwig uses a high power ( more drivers) version for mastering.

Yes, the PX Teflon coated mylar is a big benefit to the Soundlabs reliability and I think sound.

Lets examine this closer. 4th order crossovers using active filters, opamps or many transistors, passive resistors and passive capacitors all add their little errors to the stew.

That is a lot of extra stuff in the audio path. Then there are the filter designs themselves. Are they designed with Bessel linear phase or the far more common steeper slope Butterworth curves with their inherent extra phase shift and non-linear phase curves?

What does a square wave show on a scope? Ringing? What do the residual THD and noise specs show?

Well designed passive 2nd order crossovers add no noise and THD. They don't add extra phase shift. They don't use sand power resistors, electrolytic caps or even wire inductors. They use wirewound resistors, very high quality film caps and air foil inductors. These parts are far more refined and expensive than the low cost surface mount pcb's used in most active crossovers today.

The high end speaker market is almost exclusively passive now. There is good reason for this, the active processor sound does not reveal what those passive speakers can.

The pro end is all active
What does this tell us ?
Roger

This is true but... the distortion and coloration in the crossover electronics is -nothing- compared with the massive distortion and coloration in even the best speaker drivers.

The key is what does it take to get the coloration in the speaker down as low as possible, and if doing so happens to require having more coloration in the crossover, that is still a win overall.

Mind you I am not in either one of these camps... from my perspective I have heard systems taking both approaches that sounded good.

Also, of course, you COULD make a fourth-order minimum-phase passive filter. It would be fearfully inefficient but that would probably be fine. If your goal is narrow filters and you want to keep it passive, there are no reasons other than cost and convenience not to do so.
--scott

Convenience reigns supreme.
I spent a couple of decades playing with Quad ESL63s and understand the allure of the transient and step response and character of these speakers. I have a set of them sitting about 8' from where I type this...
I, however, have never looked back since switching to the Dunlavy SC-V that I have now. It took me many years to figure out that the impulse and step response of these large speakers with first order x-over was the thing that let me hear the true advantages of DSD and it's absence of time based artifacts. Ditto for hearing the time base artifacts of low sample rate PCM.
As always, YMMV.
-mark

Not a single electronic component in my crossovers. Just beautiful mathematics. They will be perfect forever. The drivers on the other hand ...

Infinitesimal compared to an analogue mixing desk.

Bessel isn't linear phase, either. Bessel can also be as steep (or not) as Butterworth or any other topology. A lot of designers favour Linkwitz-Riley these days, since they sum flat on-axis.

All IIR crossovers (ie, anything implemented in the analogue domain - passive or active) will screw up a square wave, as will the frequency response of any given driver. We haven't even got to room acoustics yet.

THD+N of a well-implemented analogue active crossover will be very low.

In theory, all passive crossovers (not just 2nd order) will add no noise or THD. Your argument here seems to be that expensive components must be better, but consider this:
Passive crossovers must be capable of withstanding considerable voltage inputs and current flow, and as a result they must be electrically rugged. That's what drives the cost up.


A 12dB/octave slope is insufficient for a highpass filter, too. That slope results in constant excursion as frequency decreases, which is bad news for a tweeter!


We've also avoided the fact that inductors are a really crappy component to have to use: additional DCR, interaction with other inductors (or really, anything metal nearby), self-resonance, etc.
Forget using super-low-resistance speaker cables if your passive crossover is adding 0.5ohm in series with your bass driver. And yes, that will have a knock-on effect on the LF response.

The high-end market also seems to think vinyl is the most accurate recording medium, that spending £1000+ on a couple of 1m XLR cables is justified, and that little wooden stands will decouple those cables from the Earth's magnetic field.
I'm much more interested in what the likes of Genelec or ATC are up to with their high-end models.

Passive crossovers decouple the speakers from the amplifiers, and even using really high-quality components won't change that.

https://sound-au.com/biamp-vs-passive.htm

I've heard of people using caps to "protect" compression drivers in PA systems, to find the diaphragms have shredded because the midrange and bass drivers have applied so much pressure.
Removing the capacitors dropped the failure rate from half the HF drivers per gig to zero, because the amplifiers then could short out the back EMF, dramatically reducing the excursion of the diaphragms from the external pressure.


Again, the way to do crossovers properly is to use FIR processing and separate amps per passband. The advantages are undeniable, to a point where I don't understand why we're still talking about this:

- Square-wave capability (yes, even with real-world drivers)
- Proper protection for the drivers from out-of-band signals
- No dependence on the driver's impedance curve, particularly relevant at higher power levels where the DCR may shift due to thermal build-up in the driver
- Drivers "see" high damping factor across the audio range


Chris

maybe cause some folks dont like the way it sounds.
while in theory you are obviously correct...in practice not enough attention to detail is spent on chosing amplification that does not exaggerate the artifices from both digititization and FIR processing.
I find taking the easy way out with ICE Power...Hypex...Powersoft is not the proper solution.

Let's take the degenerate cases on the far sides of the continuum.

Altec A-5. Several feet between drivers, so when you're sitting in the center of the auditorium there is good phase alignment between drivers, but if you are seated up front or up in the loge, you are several inches closer to the acoustic center of one driver than the other. Comb filtering in the crossover region? You bet! And the crossover is at 500 Hz so that comb filtering is right in the vocal region where it's most offensive! Going to a high order crossover (analogue or digital, your choice) is a staggering improvement because you now have a narrower frequency range where comb filtering is taking place.

On the other end... Tannoy dual-concentrics. The acoustic centers of the two drivers are very close to one another, so the distance between them is very close no matter where in the room you're sitting. The driver impedances over the crossover region are pretty uniform, so there isn't even any need for zobel networks to flatten that out. A passive second order crossover works just fine and there's no need for additional conversion stages or boxes of op-amps. It's all very clean and simple and works.

It's horses for courses.
--scott

these folks have a very loyal following.

https://www.me-geithain.de/en/aktiv-lautsprecher.html

The result is so much more precise and predictable if digital crossover is used.

It is incredibly easy to implement a digital crossover in just about any DAW, either going with sub group, or foldback. There are millions of different EQ plugins you can experiment with, let alone infinite adjustability, in your listening room. You can get perfect phase response, frequency response, no added distortion, or noise. If you worry about putting too much power into a driver, you can easily limit the power output with a limiter in your crossover design. By not having passive crossover between the amp and driver, the amp can control the driver movement much better. Lastly, you get far greater efficiency by going active.

I really enjoyed their coaxial monitors when I heard them in Europe a decade ago, but nobody else seems to have ever heard of them. They have no US presence at all. These are folks who deserve more attention.
--scott

I hear this argument a lot, and I don't necessarily buy it because the tolerances on the drivers is usually far wider than the crossovers anyway. What good is having a fraction of a percent precision in the crossover when the Fs of the woofer varies by 10% from unit to unit anyway?

Now, the argument that makes real sense is that a measurement system can compensate for driver imperfections and changes in the drivers as they age, and that IS something that has been integrated into a few (but far too few) digital crossovers.
--scott

It is assumed that you take your own raw measurement of each and every driver in and out of the box before you start doing digital crossover design. It is pre-requisite.

You assume the role of a speaker designer if you decide to implement your own digital crossover in the DAW. You have to bypass the in-built crossover in your current speaker, find a solution for multi-channel amplifier to drive each driver individually, as well as finding a multi-channel DAC. You may also want to tweak the gain of your amplifier to compensate for the speaker drivers’ sensitivity.

My recording rig monitoring speakers are also digitally crossed over so I usually do a quick measure/alignment wherever my in-field control room happens to be before I start to record. Granted, not a lot can be done in the field to improve the sound but at least I have some wriggle room.

In principle, this is pretty much straight forward. In reality, you do have to open some boxes and heat up your soldering iron. But, I promise you, the reward can be great.

I like my Geithain RL906, a lot. Imaging of a good Blumlein pair though them is smile inducing! Also, small / highly portable, which I like.

However, I don't have a broad enough experience with other, similar speakers to give a helpful comparison.

Nice setup . Just curious is that Jardiniere and pedestal made by Bretby?

Glanta by WW Worthington, thats all I know
We had a couple in an old Victorian house