Opinions on passive vs. active

[Mr.Ed]

OK guy (and gals), I want opinions on the SQ differences between passive and active crossovers.
Budget aside, amp qty aside, I'm only interested in the differences you hear in sound qualty.
I'm trying to decide which way to go. It would be really great if anyone has done a side by side comparison. But definitely looking for opinions. TIA.
-Ed

[BobEllis]

OK, I'll bite. Properly implemented, with similar quality parts and components in the signal path I think that both will sound about the same. I've done a comparison with the same drivers/cabinet in a 2 way, but it wasn't truly valid since the filters didn't have the same transfer function and it wasn't A/B. There seemed to be a bit better control of the bottom end when active, but I was expecting that and may have heard it whether it was there or not. Fixing phase alignment of the passive XO made a huge difference - my son (generally disinterested in audio quality) asked what was different when I added the phase correction. If you haven't seen it, check out http://linkwitzlab.com/ for lots of active XO design information.

For me, it is easier to get good results with active. Yes, capacitor quality matters there, too. I have plenty of decent to high quality Pass clones and other amplifiers, so that part of the budget doesn't matter to me. My frugal nature makes it much easier to swap out a couple resistors in an active setup to keep tweaking than to even consider changing out a Clarity Cap. When I have done passive designs, I reach the "that's good enough" point a lot sooner, even if there is still performance to be gained. Using MKP caps (I prefer stacked film type, Wima FKP) in the signal path makes a huge difference compared to MKT types. I've also found using MKP caps in the opamp bypass positions sounds better. Of course you don't have to use opamps, see http://www.htguide.com/forum/showthr...ole-Line-Array for an example.

Where active really excels is rapid prototyping with a digital crossover. The miniDSP has a good following, but I haven't used it. I use Allocator and Arbitrator from http://thuneau.com/products.htm and an old M-Audio Firewire 410 interface. It's really easy to change filters and EQ to get where you need to be. Not always fun trying to create an analog equivalent filter, but that's life. I'm seriously considering a couple of pro audio interfaces so that I can do a 5.1 or 7.1 system all active.

[JonP]

Pretty much what Bob said.

If you take the amps and DAC out of the equation, well, that's removing nearly all of what would affect the SQ, of the part of SQ that is affected by the crossover anyway... Then it comes back down to how well the crossover design is doing the job of handling the drivers, and that's a whole other area of science, art and skill...

I did read a whitepage someone did on active crossovers once upon a time, and while being quite technically correct, they claimed that a passive crossover had high distortion, (??) even had graphs to demonstrate it, and that was one of the claimed improvements of the active system... I never understood how he was getting that kind of distortion, (up to 2-3%!) where it was coming from... Always wanted to try to figure it out or try test measurements. I think he must have done something wrong in his measuring, to create the illusion of high distortion, and got fooled. Unless you have an electrolytic, or a ceramic of the high K type you shouldn't be using, and they have a large voltage swing, or a lossy cored inductor being worked to its limits, I can't imagine (the right kind of) passive components adding measurable distortion.

And yes, even though I'm still a padwan when it comes to speaker building and crossover design, I would underline Bob's comment on how the digital systems are SO MUCH handier in being able to change things quickly! Even popping resistors in and out of pin sockets beats soldering and unsoldering big caps and inductors!! (not to mention having to have $$ stock on hand to do so, vs 1% resistors costing .02 ea)

[Carl V]

Once upon a time Paradigm offered a Passive Studio 20 and active 20.
A passive Studio 40 and an Active Studio 40 as well as an active LCR or Passive LCR.
Most reviewers all stated that the actives played louder with less distortion than the passives.
I beleive Stereophile amongst others reviewed them.

A College buddy Set-up his recording studio with Active M&K monitors. He switched from passive M&K
monitors. His next step up was to Powered M&K Monitors with a side firing Powered Bass section.
he ran a 5.1 Recording/monitoring studio. He had 3 powered M&K MX200 subs. He stated that he liked
the sound of the powered M&K monitors, over his previous passives. He rec'ved a nice discount for outfitting his studio with them.
He also stated that clients liked to see his set-up with the well respected gear & components. He rec'vd referral jobs
based solely on his gear & the active aspect.

I heard a similar story from a guy who went form passive WESTLAKE Monitors to Active Monitors. He was a clsoe friend of the M&K
guy. Maybe they both suffered from group think mentality...but I doubt it. They spent a fair amount of green for their Monitors.

As has previously been stated quality of parts & quality of execution is critical. But the nod was given to high quality active monitors.

I mentioned all those IC, OP amps, PCBs etc., the answer was if I was worried about that....don't even look at the signal path that went
into the "A" chain....gawd help you if you look behind the curtain at what happens after the tracks are laid down. The Active XO will be
the least of your worries. Sorta liike Sausage...don't ask what went into it.

[5th element]

It depends entirely on what you want or need to do to make a system work. I've not got time to read all of the posts above me, but it's probably been mentioned that the hardware for an active system is a lot trickier to get right. This has a lot to do with it, imo, especially as there are large pitfalls in most active, be they DSP based or analogue, systems that really don't come into it with a standard passive one. In other words it is very easy to buy a world class two channel DAC with vanishingly low distortion, super low jitter and class leading SnR and then pair it up with a power amp of similar design and get very good technical results from the input to the output. It is another thing entirely to be able to replicate that kind of performance with a DSP based or analogue active crossover. If however you are comparing apples with apples, in that all of the system design is equal with whatever approach you take, then as far as I see it, it is simple.

If a passive crossover allows you to do everything that the drive units within the system require you to do, to make them work flawlessly together, then there is no overbearing reason (there are a few small ones but they are small enough to be ignored imo) why the active system should sound any better than the passive one.

Now this is a rather big thing to say and I've put two important words in italics so that they aren't missed.

I am sure you've heard it said, or read it before, that loudspeaker design is a large exercise in compromise. This is usually directed towards how the system is put together from a fundamental point of view. Ie are you using a 1" tweeter with a 7" woofer, or are you going three way, are you using a wave-guide, going for a high efficiency design. Do you need something compact, but nevertheless full range..etc etc. That is getting besides the point here, however in my experience there is nothing more compromising than the passive crossover when it comes to actually getting any of the fundamental system arrangements to work properly.

This again falls down to what you define as working properly calls back to the word flawlessly above and then basically requires you to look at whether or not the passive xover can actually do everything that's require to do that.

In my opinion if a loudspeaker has to use asymmetric acoustic slopes, or use odd xover topologies to properly phase align the drivers then it isn't anywhere close to working 'properly'. Typically this is something that is very difficult to do with your standard passive design unless you're lucky, or you've gone specific routes (that usually compromise the system in some other way) to enable it to happen.

Now I've designed a fair few systems, both passive and active and I almost always get annoyed with the passive systems. There is usually always 'something' that is a downright pain to address that you wish wasn't there, because it's something that makes the design less than perfect. With the DSP system I don't care, you simply whack in the required filters to make it work, usually it's something that requires a bit of gain, something you cannot do with a passive xover, that gets in the way. Sure it's not going to be a huge issue, but nevertheless, it is a problem.

Then there's the question of whether or not flawless includes the filter complexity. When you go DSP this has no bearing on the eventual quality or cost of the design. Once your processor is there, if you need to add in more filter blocks then that is what you do, there's no downside to it, you are simple linearising the system at the cost of nothing. With a passive system this is about as far from the truth as you can get. Every passive element you add in costs money, creates insertion losses, has the ability to interact with its neighbours (inductors), will diverge from the value you actually want (parts are usually +-5%) and takes up precious cabinet space. There is a lot of pressure on you as a designer (even for your own designs) to keep complexity to a minimum for all of the above reasons.

For me passive crossovers are an evil in most cases. They impose limitations at every corner and prevent me from doing my role in designing good loudspeakers, I am 99% of the time making compromises when I do a passive design. If you happen to have the perfect drivers, in the perfect cabinet, that end up only needing say 2nd order electrical passive networks, resistive attenuation, some impedance compensation and a high frequency notch or two, to land on a system that is perfectly phase aligned on the listening axis, with perfect acoustic target slopes (ie NO asymmetric slopes etc). Then yeah passive is fine. Otherwise, imo, no.

[Mr.Ed]

Thank you guys for your opinions, and thanks Bob for the links. I'll take a look at them. It sounds like active are preferred maybe by a small margin for a number of reasons. Good info.
If going active, is there a preference for analog vs. digital.
Just trying to understand enough to make an informed decision.
-Ed

[CraigJ]

Hi Ed,

You may find more information here in Dr. Linkwitz's DSP challenge; http://www.linkwitzlab.com/LX521/DSP_challenge.htm IMHO, Davey has done an excellent job putting together a package that is fairly straight forward.

Craig

[BobEllis]

Rod Elliot has an article about going active that may explain why some report lower distortion with active crossovers. Think of the waveform resulting from a LF peak with a HF peak superimposed. It can be quite demanding, but the two peaks sent to separate amplifiers may be easily handled. Rod notes that the energy in typical music seems to be fairly evenly divided above and below 300 Hz, suggesting that may be a crossover point to consider.

http://sound.westhost.com/bi-amp.htm

http://sound.westhost.com/bi-amp2.htm

[JonMarsh]

Seems almost redundant for me to chime in... but then one of my favorite sayings is "No generality is worth a damn, including this one!"


The guys above have definitely made thoughtful posts about the pros and cons. regarding distortion in passive crossovers, this is most usually the case if using cheap cored crossover coils with inferior core materials, and electrolytic caps on the top end, which have their issues with delectric absorption and voltage nonlinearity. Just earlier this week in the Wavecor Ardent thread we hashed out a bit about the relative merits of different film caps.

I think Matt made a very cogent argument for the DSP school of thought, and at the right price and performance points (upper mid?) there's a lot of flexibility and performance to be had. OTOH, there's no way to integrate sensibly a DSP solution with my TotalDAC-D1 Dual or the Berkeley Alpha DAC RS that I have my eye on- I'd be much more sanguine about using an ASP type of format for the final product, though rapid prototyping with a DSP, such as the miniDSP (have one, haven't tried using it a system proto, though). For a nice three way setup, an ASP and the corresponding channels of amplification does get a skosh pricey, and believe it or not, you can start to have failure points just due to connectors and cables, that can be very annoying. (BT, DT).

In the end, it's about using your experience to make the right choice for what you want to accomplish. There are two three way designs on my drawing board which are planned to use a passive mid to high and an active low to mid/high. One uses the DSP built in to a Metric Halo ULN8 eight channel DAC, so that's kind of cheating and raising the bar for the DSP output conversion to a fairly high level- one high enough for perhaps 98% of users? Cheating, I suppose, but you know what they say, if you aren't cheating, you aren't trying hard enough! :T

And with that, I should remind myself of another favorite saying of mine that I too often forget- it's sometimes better to be silent and thought a fool, than to speak forth and remove all doubt! Alas, I have a habit of speaking forth too much! :B

[Hank]

"Men of few words are the best men"
- Shakespeare, "King Henry the Fifth", Act III, Scene II
"Better to remain silent and be thought a fool than to speak out and remove all doubt"
- Abraham Lincoln
"Silence is a friend who will never betray"
- Confucius
Those quotes sit on my desk and Abe's words are why I rarely post here (except for some 3M shilling :W)

[5th element]

Yes, I suppose I should add to this that analogue passive crossovers also require considerable thought into minimising system losses and thoughts in keeping distortion and noise low. DSP based systems are quite nice in this regard and are generally quite 'easy' to put together if you're doing it yourself. They don't necessarily need to cost the earth either, especially if you compare this to how much a complex multiway passive xover can cost. Of course, if you're the kind of person who's going to buy all the DSP system as boxes off the shelf then this is likely to get very costly.

From my point of view there are quite a few of the newer ICs out there that allow you to add DSP functionality to a project without too much hassle, or additional cost. Usually when building things I tend to look at systems, rather than just speakers and if done in that way, going DSP can often be the most pragmatic approach, but it all depends on your end goals for the system.

The passive approach can lead to very satisfactory results and is certainly the only real option when designing commercial loudspeakers, or coming up with DIY designs that are for others to build.

[Mr.Ed]

Thank you again guys. More good info. I've been reading on the Linkwitz and ESP sites. Lots of info to digest. And Jon, I don't think anybody here would consider you a fool and besides, I asked for opinions so all is good.
Sometimes I find it easier to understand with a hands on approach so I may just pick up a Minidsp to play around with.
It's not a huge investment and I already have a number of amps, so why not. Thanks again.
-Ed