Active crossovers measurably better?

[stunta]

In theory, active crossovers are better, all else being equal. I am yet to come across measurements and/or double blind A/B tests that prove this in practice. Has this been done?

If we take the best known passive crossover implementation and compare it to the best known active crossover implementation for the same speakers, is there proven audible difference?

 

[dallasjustice]

In theory, active crossovers are better, all else being equal. I am yet to come across measurements and/or double blind A/B tests that prove this in practice. Has this been done?

If we take the best known passive crossover implementation and compare it to the best known active crossover implementation for the same speakers, is there proven audible difference?

You are really asking 2 different questions:
1. Is there a measurable difference between active and passive crossovers?
2. IF there is a difference, is it audible?

To answer the first question, you need to specify which type of active you are asking about:
1. minimum phase or linear phase?
2. Time aligned or time coherent?

IMO, the best type of active crossover is always linear phase and time aligned. Why? Because it most accurately reproduces the incoming signal. IOW, it changes the music the least of any crossover. Is that measurable? Yes. There are many folks who have posted their step responses using this type of crossover. A good guide would be Mitch’s book on DSP. There’s many examples showing what I’m talking about.
https://www.amazon.com/dp/B01FURPS40/ref=cm_sw_r_sms_awdb_t1_HYNAAbFM1SSRR


All passive crossovers are minimum phase. Almost none of them are time aligned (exceptions are Thiel and Vandersteen). The good ones are time coherent. IOW, the step response of each driver fully cycles into the next driver. A good example of this would be Vivid or Revel loudspeakers.

IMO, the best question is:
Is the best passive crossover indistinguishable from the best time aligned and linear phase active crossover where best is defined as best overall polar plots and step response. I don’t know if anyone has ever done that experiment. It would be interesting. IMO, if listeners were trained, the digital solution would win 100% of the time.

 

[stunta]

Thanks for the technical clarifications, but the question effectively remains the same. I am treating components as black boxes and the question in layman terms because that is exactly who I am in this context

 

[dallasjustice]

Thanks for the technical clarifications, but the question effectively remains the same. I am treating components as black boxes and the question in layman terms because that is exactly who I am in this context

Because your question asked about “measurements”, I assumed you were curious to discover why one type of crossover may be better than another. I’m afraid there’s no avoiding the underlying technical questions if you want a real answer to your question.

 

[stunta]

Because your question asked about “measurements”, I assumed you were curious to discover why one type of crossover may be better than another. I’m afraid there’s no avoiding the underlying technical questions if you want a real answer to your question.

For this question, I am not interested in sub-types if one sub-type is already established as superior within it's type.

I was asking about measurements across passive and active, not their sub-types.

 

[dallasjustice]

I was asking about measurements across passive and active, not their sub-types.

There’s all sorts of active and passive crossovers. In fact, one could make an active crossover behave exactly the same as a passive crossover. So you’ll need to be more specific.

 

[RayDunzl]

In theory, active crossovers are better, all else being equal. I am yet to come across measurements and/or double blind A/B tests that prove this in practice. Has this been done?

If we take the best known passive crossover implementation and compare it to the best known active crossover implementation for the same speakers, is there proven audible difference?

I currently have a pair of active speakers, JBL LSR 308, and a pair of passives, martinLogan reQuest, in the same room, fed with the same source/dsp/DAC/preamp. I wouldn't call either one of them "the best". Good enough, maybe.

I use a miniDSP to apply a "filter" to signal going to the speakers (a different customized filter for each), to "correct" the measurement obtainedat the listening position.

As an example, here are the uncorrected (red) and "corrected" (black) "step response" measurements for each.

I'm not sure I could identify which traces belong to the active and which belong to the passive.

In either case, the correction gets them both close to looking like what "the best" would shoot for.

For casual listening, I find them rather indistinguishable. I can easily forget which is playing, and have to look at the equipment LEDs for hints. For critical listening, I prefer the MartinLogans imaging and lower distortion at higher SPL.

So, I use "Active Signal" and am currently satisfied. Have I coined a new phrase?

As a bonus, it can be applied to any speaker I bring in to play with, active, or passive.

(I'm not a "play music from my computer" type - you could perform the same magic from the PC)

(Cosmic will argue that correcting the speaker itself is better, because he can)

 

[March Audio]

Hi Stunta,

As dallas suggests there is no way to avoid the technical details of this question as you should compare apples to apples, however my experience is this.

I decided to design my own 3 way speakers. I initially designed a passive crossover. It is a LR 4th order at 600Hz and 5000Hz. After tweaking it sounded very good. I will try and dig out the measurements.

I then proceeded to implement a fully active DSP crossover using Acourate software. This used linear phase and time aligned using 2nd order NT filters. Driver response was corrected with nearfield measurements. There was a significant improvent to all measurements and subjective quality. Much better clarity, detail, neutrality and imaging. Not to mention how much easier it was to tweak and finalise xo spec. Changed it to 700Hz and 4700Hz.

As far as I am concerned active dsp XO is the way forward. Hope that helps

 

[Wombat]

Some people use active crossovers to tune their systems and then build a passive substitute - note I didn't say equivalent.

I am not aware of measured comparisons but modern active crossovers offer far more flexibility than passives for DIYers. I use SS amps and I like my drivers connected as directly and safely as possible to them.

Some pros and cons: http://parallelhomeaudio.net/ActivePassive.html

I believe active crossovers are capable of measuring better through a transparent amplifier, and digital crossovers even better, than passive.

FWIW I have read more comment on the www. of DIYers preferring active to passive but these are usually from SS amp users. Active can be expensive with pricier tube amps.

 

[oivavoi]

I welcome this thread. I have also searched for objective comparisons of apples to apples when it comes to active and passive crossovers, and not found it. My subjective impression is that active crossovers are vastly superior at lower price points, but that passive crossovers can get acceptable when they become really elaborate and expensive.

 

[DonH56]

There are many but I do not have any at hand. DuckDuckGo found me this one, by a local professor even:

AES E-LIBRARY
Active and Passive Filters as Loudspeaker Crossover Networks

This tutorial paper defines the function of a crossover network and then explores methods of meeting this function. For moderately priced two-way loudspeakers, a passive network at about 800 to 1600 Hz. will continue to dominate the designs of the future. However, the use of active filters (electronic crossover networks) and buffer amplifiers offers the most significant means of loudspeaker improvement in the next decade. As one typical factor, crossover frequencies need to be lowered and crossover slopes increased and the active filter is the only economical method of doing this.

Authors: Ashley, J. Robert; Kaminsky, Allan L.
Affiliation: University of Colorado, Colorauo Springs, CO
AES Convention:39 (October 1970) Paper Number:772
Publication Date:October 1, 1970 Import into BibTeX
Permalink: http://www.aes.org/e-lib/browse.cfm?elib=1240​

Gotta' love the prediction made in 1970 (predictig active would take over by the 1980's...) I am also still awaiting my flying car.

I thought of writing an article comparing them, as it is easy to simulate, but don't have a way to do the experimental work so gave it up. That and Work and Life is sucking way too many hours... And I am no speaker guru.

 

[Thomas savage]

Thanks for the technical clarifications, but the question effectively remains the same. I am treating components as black boxes and the question in layman terms because that is exactly who I am in this context

This happens to me here often

Using CAD it’s amazing how you can achieve very good result when building passive crossovers so many passive crossovers are not what there were years ago. ( so I’m told)

I only know of one guy personally who has done extensive ‘listening’ in this area, dic who designed my speakers (Vivid) He seemed more satisfied with a passive CAD designed crossover than going active but I think in the last couple of years that tide has turned.

In 2018 I can’t think of a benefit a passive cross over can hold over active but it boils down to implementation, the differences assuming the ‘best’ implementation might be tighter than a few think.

If I bought another pair of speakers I would hold out until I found a active pair I loved the sound of, I think everyone here would go in that direction if they have not already.

Iv always wondered what my vivids would be like with a well implemented active xover, I don’t find myself wondering what my new JBL 705p’s would sound like with a passive network..

Nothing technical , no measurements here soz

you can make a passive measure better within software like I think @dallasjustice does, then you have to ask is that passive network in the speakers he uses ‘ transparent ‘ .. I’m sure he has fantasies of ripping it out though.

 

[Purité Audio]

As BE (Alan) says DSP allows you to implement using steeper slopes for the crossover, but DSP also allows you to correct delay,phase, correct driver anomalies Kii use the crossover to create a cardioid response.
I have found that companies don’t use DSP either because they don’t have the skill set to implement or their ‘core’ customers wouldn’t like it.
Keith

 

[Cosmik]

I currently have a pair of active speakers, JBL LSR 308, and a pair of passives, martinLogan reQuest, in the same room, fed with the same source/dsp/DAC/preamp. I wouldn't call either one of them "the best". Good enough, maybe.

I use a miniDSP to apply a "filter" to signal going to the speakers (a different customized filter for each), to "correct" the measurement obtainedat the listening position.

As an example, here are the uncorrected (red) and "corrected" (black) "step response" measurements for each.

View attachment 10166

I'm not sure I could identify which traces belong to the active and which belong to the passive.

In either case, the correction gets them both close to looking like what "the best" would shoot for.

For casual listening, I find them rather indistinguishable. I can easily forget which is playing, and have to look at the equipment LEDs for hints. For critical listening, I prefer the MartinLogans imaging and lower distortion at higher SPL.

So, I use "Active Signal" and am currently satisfied. Have I coined a new phrase?

As a bonus, it can be applied to any speaker I bring in to play with, active, or passive.

(I'm not a "play music from my computer" type - you could perform the same magic from the PC)

(Cosmic will argue that correcting the speaker itself is better, because he can)

As far as I can work out, the thing that DSP correction can't do for a passive speaker, is to eliminate factors such as suck-outs or level mismatches between drivers. It can correct the measurement for a single point in space, but as soon as you move away from it, the artefact will begin to show up in comparison to a better crossover. So if you had a choice between pre-correcting a passive speaker, or applying an active filter to each driver individually, the active version could be different (better) even if conventional on-axis measurements didn't show up a huge difference.

And then if you can correct each driver individually and render it linear phase using DSP, and add time alignment delays, the on- and off-axis behaviour becomes different between the two types, again. (And time-aligned linear phase is the way to reproduce the waveform without linear distortion, of course).

Passive crossovers reputedly change character with power level. They are more difficult to match accurately between stereo channels.

Once you start down the active road, with the crossover sounding 'better' (more transparent?) you are suddenly in the realms of there being no downsides to crossovers per se. If you jump from two-way to three-way without any penalty from the hitherto always odd-sounding crossovers, everything gets so much better including:

• Evenness of dispersion versus frequency (lack of 'beaming')
• Drivers in their ideal frequency ranges (less distortion)
• Less power into each driver
• Less intermodulation and doppler distortion (mid range has its own dedicated driver)
• Etc.

The demands on each amplifier are very low.

Active EQ control means you aren't dependent on ports or other passive methods to bolster the bass, so you can use sealed woofers which work better on many levels (no smearing in the time domain, control of cone all the way down, roll-off inherently complements room gain - but you can choose whatever you want).

And does the better cone damping of an active speaker result in something more than just a flatter frequency and phase response? I don't know the answer, but I suspect there is something better about it on more levels than just the behaviour with steady state sine waves.(.?)

If you put it all together, it may be different in some obvious ways, but also subtle ways that don't necessarily show up clearly in the measurements. Reviewers of active speakers often talk about 'control' or 'grip' and I think that is basically a sense of solidity, rather than a thing that wafts about as the signal changes.

 

[Rodney Gold]

Dic recently set up an active (trinnov based crossover/correction) quad amped version of the Giya G1 spirits..I havent spoken to him about it but the Trinnov local agent has heard both passives and actives in the same environment and they recon the active version is stunning
Im thinking of going that route..however cost is an issue .. I would need 2 extra amps to supplement my exisiting 2 at eu 27000 and a trinnov altitude 16 at a further eu 20000.
I would also most likely have to ask dic and the trinnov guys to come here and set it all up

Apart from that , I dont see how one could set up a truly objective AB/DBT of a passive vs active crossover version of the same pair of speakers without major interruptions and shuffling boxes around

 

[Purité Audio]

Why on earth would you need such expensive amps, Benchmark at one ninth of the price, instead of Trinnov ask Christof from Illusonic to do it for you.
I wasn’t hugely impressed with Trinnov not so much the product but the support from the company which was weak at best.
Keith

 

[dallasjustice]

Dic recently set up an active (trinnov based crossover/correction) quad amped version of the Giya G1 spirits..I havent spoken to him about it but the Trinnov local agent has heard both passives and actives in the same environment and they recon the active version is stunning
Im thinking of going that route..however cost is an issue .. I would need 2 extra amps to supplement my exisiting 2 at eu 27000 and a trinnov altitude 16 at a further eu 20000.
I would also most likely have to ask dic and the trinnov guys to come here and set it all up

Apart from that , I dont see how one could set up a truly objective AB/DBT of a passive vs active crossover version of the same pair of speakers without major interruptions and shuffling boxes around

There’s certainly a less expensive and probably more effective way to do the same thing.

I was wondering about why Dic built the spirit with an external crossover. It’s a good thing he did that; to give his customers the option to tryout digitally active crossovers and sort future proof things. I think all passive loudspeakers should have external crossovers. It makes things more flexible.

 

[Rodney Gold]

The local Trinnov agents and their international sales manager are exceptional..helped me out with setup and tips and tricks even when knowing I bought the unit 2nd hand .
As to a cheaper method of making my spirits active , I cant get away from the cost of a trinnov as it is the tool needed..I could probably get away with cheaper amps than my Devialets , but they sound so good on my speakers and I already have 2 .
I did try a 650w class D abeltec power amp I have stashed away just in case , its based on the anaview ALC1000 module , using the Trinnov as a pre/dac/correction and it didnt sound wonderful..

I tried the abeltec 2 ways , driving the speakers via the trinnov as well as using the Devialets crossover with the abeltec only driving the bass drivers and the devialets driving the rest. Neither way sounded great

At any rate , I don't have the experience to get a quad amped G1 spirit system sounding as good as Dic would and would probably make a dogs breakfast of it as a DIY project
The vivid factory will not offer any guarantee on their speakers if you do a DIY fully active job.

 

[Thomas savage]

If I could do what @BE718 has done and build some fantastic actives I would. Whenever I see him posts the measurements and pictures of the speakers I get tinges of jealously

 

[Cosmik]

I would need 2 extra amps to supplement my exisiting 2 at eu 27000

I use a Sony STR-DA1200ES for one pair of three-way speakers. Going price on eBay seems to be about 30 - 150 euros.