DSP vs. active analog crossover vs. passive analog crossover. What is your take?

@mikelavigne wrote:

no simple answer to this question.

the crossover should compliment the media, and the degree of room acoustical work that has been done. dsp fixes room<->speaker integration issues. but it’s not the only way to do that. at lower price points and integrated systems dsp does result in a more listenable performance. but past a certain price point it starts to get in the way of signal path purity for 2 channel.

so no absolutes in this question.

it depends....

We may be addressing different aspects here. Digital Signal Processing can be many things, but in my use of the Xilica DSP unit it serves only one purpose: acting as a digital cross-over, and nothing else. No "room<->speaker integration," surround processing or other; 2 channels only (i.e.: 3 outputs pr. channel), with 4 outputs dedicated for the 2-way main speakers, and 2 outputs for the subs. It’s worth noticing that as a digital XO in my case there’s no passive ditto in the signal path, so each amp output channel sees its respective driver terminal directly.

Is this how your experience with DSP, acting actively as a digital XO solely, has also been formed - apart from being used as a room correction unit over already passively configured speakers?

i own the ultimate dsp processor, the Trinnov Altitude 16. i use it with my 9.6.3 surround sound system for my home theater. it uses ’object based’ dsp to create soundscapes that do make movies more real. horses for courses, and it’s the ideal solution for those type movie or concert recordings.

I looked up the Trinnov - an impressive beast of a multiple-feature, high quality apparatus. For home theater use in particular I gather it’s a godsend, and with digital inputs and a digital source one would avoid the A/D conversion, not that I find the added A/D conversion to make much of a difference sonically, if any.

for traditional 2 channel i prefer an analog crossover since i’m a big analog recording guy, as well as have no intention of ’double’ converting my digital files with dsp for my 2 channel listening. that would suck the life out of the music.

The A/D to D/A conversion necessitated with a DSP in your case with an analogue source - if the DSP were to act only as a digital XO sans any passive XO - to my mind would be the lesser evil compared to the influence of passive cross-overs on the output side of the amp, but again, that’s just me.

but i do have a separate dedicated 2 channel room. my 2 channel room is purpose built and highly tuned to avoid the need for any dsp. i’ve fixed the room, and retained the purity of my analog signal path.

With my former passive configured all-horn speakers I at one point had 3 processing stages involved: the passive cross-overs in the speakers as an analogue "processing" stage, the Xilica DSP to high-pass them, and JRiver Convolution hosting a software for room correction in both the amplitude and time domain. It was a capable setup overall, but I prefer my current fully active setup with only one processing stage: the Xilica DSP. Room acoustics have been optimized with both diffusion and light absorption, so no room correction.

my 2 channel room speakers have 2 towers per side; each tower is 7 foot tall and 750 pounds (3000 pounds total). one tower per side is passive, from 35hz and up, the other tower is active and powered for under 40hz, (on paper) -3db at 7hz and -6db at 3hz. so serious bass capability. the passive tower rolls off at the bottom and cannot be used as a stand alone speaker. the bass tower uses analog adjustments and gets it’s signal from the passive tower speaker terminal so it’s signal mimics the sound of the main amplifier. the crossover can be adjusted from 50hz down to 20hz for best room integration.

i get awesome performance, considering the years of work i’ve put into the room.

Indeed sounds like an awesome speaker setup - kudos. Have you tried running the whole system fully actively with no passive XO’s involved?

i think many situations can benefit from dsp, but at the tip top of the 2 channel music reproduction food chain it’s a penalty.

I differ here, certainly as outlined above using a DSP unit as a digital XO - sans any passive ditto - only.

Apart from listening to a variety of bundled-package active speakers I’ve heard passively configured speakers that I knew very well being converted into active ditto, and it has certainly convinced me of the merits of properly implemented active configuration.

Most relate to "active" as bundled speaker packages - calling them simply, and rightly, active speakers - with built-in amps, DSP/electronic cross-over and sometimes DAC(s). However ’active’ per definition simply means the filtration is done prior to amplification on signal level, and not (passively) on the output side of the amp between that and the drivers.

That is to say: active configuration isn’t defined solely or even strictly as a bundled package (that one usually calls an active speaker), but can as well act - true to the definition of active config. - as a separate/discrete component solution; with a free-of-choice DSP/electronic XO, the number of amps necessitated to drive the respective driver segments, and DAC and cables - essentially as you would operate with separate components and cables with a passive speaker setup (sans DSP/electronic XO), the vital difference though being - to reiterate - that it would be actively configured.

With that out of the way my reference above to the passive speakers being converted into active is as a separate component solution - i.e.: fully and truly actively configured. Running them actively meant a more resolved, transparent, smooth/easy on the ears, transiently snappy, dynamic and tonally more authentic presentation. More accurate, in my understanding of the term, and yet more musical. The passive iteration by comparison felt sluggish, coarse and veiled (which to some may equate into more "musical"), and it wasn’t due to a badly implemented passive XO with so-so component quality.

Setting up a separate component, actively configured speaker setup without preset filter settings from a manufacturer comes with a caveat: you’re left to your own devices dialing in those filter values, and to a newbie this may come off as intimidating and even off-putting. It’s a learning curve and a process that takes patience to wring out the full(er) potential, for sure, but it’s very rewarding - not only eventually.

In my own case with my separate component, actively configured speaker setup, it came as a surprise how fast I was able to dial in filter values on my Xilica DSP unit for an initially quite satisfying sonic result, but as time went on I was able to refine the results much further with measurements, extended listening and gaining more knowledge. Having a horn that amplifies linearly (as opposed to some waveguides) no doubt helped with the pleasing initial results.

To me active configuration ultimately holds the bigger sonic potential, insofar one is able to harness it either finding the right preassembled and -configured active speakers, or going about it on a self-taught/collaborative separate component basis. The latter scenario has the advantage of a carte blanche plate; every parameter, physical as well (and not least), can be scaled and chosen according to desire. There are no limitations other than what you impose on yourself or am willing to learn in the process.

@mikelavigne wrote:

interesting, i actually own a "new-in-box" un-openned Xilica DSP XP-2040 i had purchased to use with my Trinnov for my 3 subwoofers (happy to sell it cheap). it turned out that my 3 Funk Audio 18.0 subs in my Home Theater have an even better internal crossover, the ALLDSP module that is able to be ethernet networked and tuned remotely.

My Xilica XP-3060 is exactly that: remote controlled via ethernet for on-the-fly filter parameter tuning from the listening position. Doesn’t get much easier. Delay, Q, slope types and steepness, etc. - vastly more elaborate settings options than any built-in DSP solution I’ve tried. It’s built for the pro environment and looks that as well, but who cares, or should care about segment and aesthetics when it’s sonically rather transparent and fairly priced? Oh, well - the latter (i.e.: fair price) may rub some the wrong way as well..

Actually considered a pair of Funk Audio 18.0’s for my setup some years back, but went the tapped horn route instead.

but back to the subject of a dsp crossover and whether it’s suitable for the top level 2 channel music reproduction?

i think we see it mostly in high performance 2-channel with DIY active horn systems where otherwise it’s just not very doable. so in those situations it’s simply the only choice, not that it’s inherently better than analog.

A DSP cross-over for active config. comes in especially handy with bandwidth limited horns in the need of steep slopes, not to mention that horns are very revealing in exposing the sonic advantages a DSP XO offers while not least getting rid of the passive XO. Indeed I’d say they’re instrumental in getting the most from horns, but a successful DSP pairing is not limited to horns; low efficiency direct radiators can suffer from heat build-up not only in the voice coils but passive XO’s as well (that cause changing filter values), and so running them actively - also for other reasons here - would as well be a great boon.

for instance Magico brought out their $600k Ultimate Horn System some years ago, but it used a dsp crossover for the best performance and they only sold a few. the marketplace did not warm to the dsp idea as most high level users want all analog signal paths. we can argue about performance, but 2 channel at the top has it’s perspectives.

Well, kudos to Magico for sticking to their guns with a DSP solution "for the best performance," as you put it yourself. Why users don’t comply with that would seem to be more about habits and conservatism than sound per se, or a particular sound perhaps they’d expect to be a product of an analogue (i.e.: passively configured) approach.

it is interesting that the new G3 version Avant Garde Trio uses a completely analog crossover.

Maybe not much more interesting than what’s possibly illuminated with above Magico example already, the difference though being that Avantgarde may simply have complied with the costumers here. Not saying passive XO's in conjunction with horns can't provide for very good results, they most certainly can.

@panzrwagn wrote:

Active crossovers remove coils, capacitors and resistors from the amp-speaker interface with huge benefits in overall system performance: dynamic range, 'speed' and control of drivers. The benefit is due to lacking those energy storing or energy dissipating components, the amps have an easier load, less phase shift and less time smearing/latency/hysteresis and more control over the driver.

Exactly. 

@tannoy56 wrote:

... if I had learn anything from the past, there is no substitute for large drivers.

Indeed - there’s no replacement for displacement, as they say.

Cool speaker setup of yours, btw. Wasn’t aware Tannoy made this line/type of speakers. Should be both potent and very well sounding. Down the line I’d definitely go with a (Lake-based) DSP cross-over, as has been suggested.

@kingharold wrote:

I use a DEQX DSP which uses PCM coding, not DSD. DEQX is coming out with a new line of DSPs, supposedly this quarter. I wish they would use the DSD, but I doubt it. Whatever they use I am looking forward to upgrading.

Drool inducing new upcoming line a products. Wouldn’t mind a Pre-8 lying under the Christmas tree in the near future..

I use Bill Fitzmaurice designed HT Tuba folded corner horn bass bins driven by 15" woofers which at 200 Hz cross over to Oris 150 horns with AER BD3 drivers. The Oris AER combo plays the range from 200 Hz up to 8 kHz where Fostex t900a bullet tweeters take over. Thanks to DSP the system is remarkably flat from 25 HZ where the output is identical to that of the 1 kHz reference output up to a little over 20kHz. The roll off below 25 Hz is typical for a horn loaded woofer.

I place great value on the system being fully horn loaded. Horn bass is the icing on the cake. It lends a smooth continuity and rightness to the SQ that I don’t feel is achievable by hybrid systems mixing horns with other bass alignments.

Kudos, absolutely agree on your assessment of horn bass and its contribution to the overall sound. I use a pair of tapped horns myself (i.e.: MicroWrecker). How many THT’s have you implemented?

Considering that Mobile Fidelity got away with sneaking a A/D conversion and a D/A conversion into their Original Master LPs for fifteen years perhaps analog purist aren't as able to hear digital processing as they have believed themselves to be.

DSP uage in its different varieties of functions and the assessments of it has a tendency to be very unspecific - i.e.: the context of its implementation is unclear, and typically the statements are rather heavily theorized and not coming as much from personal experience.

In this particular context the big win of omitting the passive cross-over (with a DSP acting instead, actively, as a XO) is something it would seem few have experienced and can actually comment on. 

@kingharold wrote:

Hey, phusis. I only use two HT Tuba folded corner horns. Considering that the enclosures are eighteen cubic feet each even that occupies a fair amount of floor real estate. Besides that my room only has two suitable corners.

Two of them certainly gets the job done properly. A single quarter wavelength horn-loaded 15" with a tune around 25Hz (or just below) will do stuff not only in quantity that far exceeds a 15" direct radiating driver. My MW’s are 20cf. each, also 15"-loaded and placed in their respective corner, so in the same ballpark as your HT Tuba’s. Powerful stuff, as it should be and with headroom to spare.

@erik_squires wrote:

There’s no clear winner among all of the choices. Convenience matters a lot. I run passive crossovers with my main speakers, but active to my sub which right now is only for home theater.

Why wouldn’t there be a winner? Not trying to turn this into a contest or throw about absolutes, but to those willing to go the distance and forego convenience there very well could be a winner - and by wide margin.

My take is that this is a hobby and you should focus on what you want to learn and how much of your system do you want to build vs. buy. How important is it for you to have separates? After a lifetime of buying into the all separates mentality I’ve given up. Separates are not actually better.

With an active setup using a DSP acting as a digital XO only and sans any kind of passive XO in the mix, a separates solution isn’t only or as much about importance than it is necessity - unless you’re buying a pre-developed and -assembled finished product that is already bundled.

A bundled active speaker as such can be great while potentially expensive, but as a "DIY" option - at least as it pertains to just using separately housed amps and DSP and setting up filter values by yourself - a separates solution IMO is just easier and more straight forward to install instead of building everything into the speakers (pre-assembled or not), while offering the choice of components and quality as well as their wider range more easily.

Where importance really enters the stage to me is the uninhibited nature of putting together an active-as-separates DSP-based setup and the unrestricted physics and principles it offers with speakers in particular. Unless we’re talking larger ATC models like the SCM 150 or 300A’s, actively configured speakers only really come in bundled, smaller direct radiating packages, and they just don’t float my boat. The larger mentioned ATC models are great, but even so a carefully implemented large horn-based setup simply takes it to another level in several respects. It’s just physics, and no they’re not in vain nor overkill in domestic environments.

I got to hear the original B&W Nautilus driven by a ridiculous number of Krell amplifiers and crossovers. It was not all that.

Actively driven? Well, it’s an expensive setup (which is not saying much), and only one of many.

Enjoy the hobby, but don’t obsess or think any particular way here is THE way.

Again, why not? Why not obsess and go bonkers with what one finds is the way? It’s not preaching the gospel; to most it’s just a single-minded adventure into realizing sonic goals, and with no self-imposed convenience restrictions or other and a dedicated space to go sound galore - well, let’s go explore and party :)

@erik_squires wrote:

First, I only answer a few points because it’s clear to me you are now using circular logic. You point left and then right, much like Kenjit. Are you kenjit?

There is a nice approach to this. Pulling the "kenjit-lite" card (irrespective of the man’s doings, whatever they are; it’s your belittling intention referencing to him that’s a bit coarse) when things get unclear to you and you form your funny conclusions, is less instructive, so let’s go over your remarks one by one:

(your reply to my "why there wouldn’t be a winner?")

Because of:

• personal values
• The impossibility of trying to define best.
• The diversity in implementations.

For a consumer, you can no more define best type of crossover than you can best type of amplifier. An absolute inviolate hieararchy is impossible.

To reiterate and (hopefully) clarify: I’m not trying to steer this in the direction of a contest or state with certainty that there is, let alone what is the best for all. What I meant is that there can be an obvious winner for the individual who embarks on this journey and has chosen an actively configured DSP-path, or otherwise for that matter. To him or her it might be clear as rain.

And (again) as I said: why wouldn’t there be a winner? It’s not saying there is a winner, but it’s entertaining the thought challenging your opposite notion that there isn’t one.

(Actively driven? Well, it’s an expensive setup (which is not saying much), and only one of many.)

"And this is where you go 100% kenjit. You argue in the same piece there must be a best, and then that one example doesn’t prove anything. You can’t have it both ways, Kenjit-lite."

First: you didn’t answer my question. Were the B&W’s actively configured? The Nautilus’s are expensive, that’s a fact, as are a bunch of Krell amps, but please enlighten me as to why that guarantees great sound?

And what’s your point with "it was not not all that" as it applies to separates here - as that one example alone? How is that representative of anything other than a specific context confined to that very demo?

Lastly: can’t see how you’re masterminding my claimed "there must be a best" from above quote. You’re creative for sure.

Correction: right as rain.

@russbutton wrote:

Now when you add in mid and bass drivers, with high and low pass filters there... It's a real mess. But we're not done there yet. Nope. Many of your extreme hi-end loudspeakers add in equalization to their crossover designs, which makes that impedance curve even worse. This is very hard for an amp to properly manage. That's why people drop many, many thousands of dollars on things like Krell, Threshhold, Bryston, or Rowland Research solid state power amps.

Now when you use an active crossover, an amp channel only has to manage a single driver. There's no passive, reactive component in between the amp and the loudspeaker driver. Then you don't need a megabuck amp to deal with it.

...

Another major benefit is that you can use much, much lower powered amps when you use active crossovers. A lot of power is wasted having to push through a passive crossover. You really don't need to push many watts into a tweeter or mid-range driver to get a lot of level out. You could even run a single ended tube amp on your tweeter, and a mid-level tube power amp on your mid-range driver, and a solid state amp for the bass driver. You have a lot of options.

So instead of dropping $7,000 on that Threshold Stasis 8.0 power amp. You could spend much less on an active crossover and the various much more modestly priced amps of your choice.

Unapologetically concise, to-the-point - one that actually gets it. Thank you, @russbutton.