Active vrs passive

[dewardh]

Originally Posted by CopperTop

but they were rubbish with the original passive crossovers

I also have heard examples like that . . . decent driver selection, well regarded (DIY) passive crossover, dramatic improvement when switched to active. Not active that tries to mimic the passive design, but active that does what active does best.

Most of the "art" of passive crossover design seems to be endless tweeking to hide, or at least disguise, the design's inherent flaws. Downright silly stuff, like asymetric (often too shallow) slopes adopted because it's the only way to "passively" correct for driver offset. And then the inevitable assertion: "If you match that transfer function with an active crossover it will sound the same". Well maybe yes, but why would one? Why not design it right in the first place? And no, you won't be able to match that transfer function in a passive design . . .

[gk7]

Originally Posted by CopperTop

... three way Tannoy speakers...

Which Tannoy model is three way ?

[5th element]

Originally Posted by dewardh

I also have heard examples like that . . . decent driver selection, well regarded (DIY) passive crossover, dramatic improvement when switched to active. Not active that tries to mimic the passive design, but active that does what active does best.

Most of the "art" of passive crossover design seems to be endless tweeking to hide, or at least disguise, the design's inherent flaws. Downright silly stuff, like asymetric (often too shallow) slopes adopted because it's the only way to "passively" correct for driver offset. And then the inevitable assertion: "If you match that transfer function with an active crossover it will sound the same". Well maybe yes, but why would one? Why not design it right in the first place? And no, you won't be able to match that transfer function in a passive design . . .

Exactly.

[Pano]

Originally Posted by dewardh

Most of the "art" of passive crossover design seems to be endless tweeking to hide, or at least disguise, the design's inherent flaws.

Perhaps that's one school of passive crossover design, but there are others. The nearest of which might be to work with the driver's flaws, which they all have. Another is to find the acoustic slopes that work the best for the drivers, boxes, baffles and room and achieve those via passive means. Nothing wrong with that.

Downright silly stuff, like asymetric (often too shallow) slopes adopted because it's the only way to "passively" correct for driver offset.

Some of that exists, but I wouldn't call it great design (unless it works ) Silly ideas and bad designs are not limited to passive or active. Both can be just as bad. With DSP we now have great, very flexible tools at our disposal. That does not mean everyone knows how to use them or how to make speakers sound good with them. The new tools may be faster to learn and implement than the old, but that doesn't necessarily make for better designs.

[Pano]

Originally Posted by CopperTop

I can't say that it's passive vs. active, as such, that accounts for this difference, but getting the actives to this state of rightness didn't involve any black arts or tweaking in response to listening, nor a long apprenticeship; merely measurements and automatic software

True, you can do it that way, but perhaps a brief apprenticeship in the black arts of crossover design would do you good. These days anyone with modest outlay of money and material can make a computer perfect crossover - and they'll sound like it. I don't exempt myself from that, I believed for years that the only way to a good crossover was active means and paper perfect slopes, points and FR. Ultimately it was a road to frustration.

Audio shows are full of computer designed crossovers, you can tell when you walk in the room. Mostly correct, but lifeless and artificial sounding. That's true for active or passive versions.

I may be a lone voice in this thread crying out against the mutual love-fest for actives crossovers saying "Beware! There is more to learn than you think, don't close off future paths of knowledge."

I like active crossovers and use them, but working with passive crossovers taught me a lot, and I'm still learning. Passive crossovers taught me much of what's important.

[wintermute]

Just on the coppertop quote in pano's post above. When I made the new passive crossover for my MTM's There was NO listening, tweaing, listening cycle at all.

I took accurate measurements of the drivers, imported them into speakerworkshop simulated various passive networks until I had what I wanted (4th order bessel with excellent phase tracking through the crossover region, and quite a way either side).

I then ordered the caps, wound the coils and built the final crossover. After building it I measured the response and compared to the sim and they were VERY close. Any differences will have been due to the fact that the new measurement was not done in exactly the same place as the original measurements used for designing the crossover.

The only "black art" involved in this process was the careful selection of capacitor and inductor values and electrical order to obtain the desired phase matching and acoustic slope (which was done completely in software). Something that would be no different should you decide to go active if you want full control over what the drivers are doing, rather than simply pressing a button and letting the software do what IT thinks is best.

A year later and I have not altered the crossover in any way, even after moving house..

Tony.

[Pano]

Measurements? Software? Simulations? Phase tracking?? Tony, really. Obviously you folks down under don't know how to properly design a passive crossover. Where was the chicken blood? The burning of sacred herbs? The full moon incantations? The guilding of ears? How can you possibly design without that?

[mondogenerator]

Pano and Wintermute. I hear you both, ive been spin quoted enough to serve the pro actives.

When I first started DIY I textbook table designed a 2 way, 1st order. Without experience I vehemently believed that if I went 2nd order, then Id hear some malady, some ear bleedingly horrendous colouration. I soon found how wrong I was!

Im neither for or against actives. Id like to make one, but I just dont feel DSP at a price I could afford is good enough. Its just getting accessible. Id go analogue but thats just me.

Zobels do a lot for any amplifier, whether it drives a passive or active speaker. Just the same as Bi or Tri amping, often criticised, actually has a marked improvement in aspects of reproduction if done property. The question to me is whether DSP, op amp, line level passive, or high level passive. The most elegant active solution in my minds eye, is a set of class A bjt amps, each band optimised for gain, if required then input or output RC networks to dial in what remains of the slopes. More power per device in class A due to a reduced bandwidth, simplest way to integrate filtering. With modern spice software, just like with PCD Boxsim and others simulation is quite reliable and it should be quite possible. I must try it.

[Charles Darwin]

Originally Posted by gk7

Which Tannoy model is three way ?

From the top of my head and not counting models with super-tweeters I can think of the Buckingham Monitor (10"DC + 2x 12"), FSM, 215DMT (both 15"DC + 15" woofer) and the Dreadnought (15"DC + 2x15").
Rather large and fairly uncommon all of them. FSM and 215DMT can be run either 2 or 3 way.

[Barleywater]

Originally Posted by CopperTop

.....My own experiments (gradually heading towards this ideal) suggest to me that the automatically corrected and crossed-over speaker, far from sounding flat, neutral (in a bad way) or 'grey' as I might have imagined, sounds magnificently colourful (in a good way) and fills the room with sound.......

Which software are you using for IR measurement and correction generation?

Originally Posted by mondogenerator

......Id like to make one, but I just dont feel DSP at a price I could afford is good enough. Its just getting accessible. Id go analogue but thats just me......

DSP may be performed with computer/laptop, multichannel sound card, and freely available software; workable crossover filters may be made using Audacity, PEQ filters easy to work with and export using Room EQ Wizard, which also has excellent swept sine measurement. Of course there is DRC and associated GUI. HOLMImpulse is also excellent measurement software.

Audacity with Aurora plugins has excellent convolution engine for testing, and Kirkeby inverse filter, also available for Audition/Cool Edit. I can do all analysis and correction generation, simulation in Cool Edit Pro 2.1 with Aurora plugins. Copies are floating around cyberspace.

I have for grins simulated two way Linkwitz-Riley 24dB/Oct crossover with LTSpice, fed sim a swept sine, used resultant IR to correct the all pass phase smear, and rerun sim with corrected sweep that returns beautiful IR. Cool Edit is much easier.

ECM8000 or EMM-6 with calibration <$100 dollars this side of pond; suitable preamp/soundcard start in same range. This is in league with big fat microphonicly plagued crossover caps.

I've got a PIII laptop with no sale value, and got my first DSP system up on it with Creative's old Extigy USB and Souceforge Convolver plugin for Windows Media player; two way stereo + sub.

Very good two channel DSP EQ/correction may be worked with any decent speaker.

Regards,

Andrew