48db/oct crossovers causing listening fatigue?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Linkwitz-Riley Crossovers: A Primer
I am posting this again. Towards the end of this article(LR-8 transient response), it brings up the question of audibility. What is or could be sub-clinical in a laboratory environment could very easily be a cause listener fatigue in long term listening situations(And then again, maybe not). Some things(that are sub-clinical) just can't be tested reliably in the lab. This is when we have to trust our(gulp!) individual instincts.
 
I have tried lr 2,4 and 8 at variuos frequencies with my minidsp-controlled 4-ways. Responses eq'd and distortion kept low I prefer LR2 for every stage. Low order sounds less fatiguing and more natural. My friens agree. For insanely high spl higher order is preferred.

Low order is more critical of timing and symmetry up to rwo octaves past xo to soud really good, but is worth the effort. Marsh is very pedant with his analog crossovers!

I raise my hat for J Marsh and wish him all well and blessing.
 
I've recently gone from LR4 to LR8 at 130Hz on my OB WAW set up, it's an improvement mostly because it deals better with the resonance of the U frames. Who is Jon Marsh?
Also read the Rane note linked in above post, very interesting


Did you implement with DSP or op-amps? I'm assuming it's LP's on the woofers only and the wide range are left unfettered so you minimize the power dip.
 
I have tried lr 2,4 and 8 at variuos frequencies with my minidsp-controlled 4-ways. Responses eq'd and distortion kept low I prefer LR2 for every stage. Low order sounds less fatiguing and more natural. My friens agree. For insanely high spl higher order is preferred.

Low order is more critical of timing and symmetry up to rwo octaves past xo to soud really good, but is worth the effort. Marsh is very pedant with his analog crossovers!

I raise my hat for J Marsh and wish him all well and blessing.

High order crossovers in the midrange always yield a 'signature' sound to consonants. I use audiobooks, of all things, to test for this. With lower order crossovers the sound is more articulate. My reference speakers are Vandersteens, with 1st order xovers, so I may be accustomed to that.

Having said all that, LeCleach has a great paper about how to achieve very good phase response with 3rd order crossovers and some manipulation of the driver setback.
 
So Butterworth is better?

Have a look at the graphs again. Lower orders have less power radiated by tweeter. In many rooms, this will sound as more laid back, analog, less bright etc. I bet allot of people will like that

Butterworth radiates overall more power through xover. The smaller/livelier the room, the more audible the difference will be from LR. Sometimes, the power dip due to LR is helpful if the tweeter has very high dispersion through xover.

Old school British designs often used huge woofers crossed to tweeters up high, and the woofer would have a big deficit off axis. Maybe a Butterwork sounds better in that config? Perhaps its why many of these old systems used Butterworth xovers.

Look at the drivers' dispersion and power handling, pick xover frequencies and orders that doesn't let the drivers distort, then guess at the phase matching that makes the power off axis smoothest but by far first optimizing the reflection off the side wall over total power. Then tweak by ear/measure, rinse and repeat. With measurements and sims, LR vs Butterworth distinction are immaterial now, the design method is all about phase matching on and off axis and trading the two off.
 
I understand.
In my case this will be active/dsp.
Right now I use LR24db to keep my fullrange alive outdoor when pushing hard.

When I use lower order it will need to play deeper and maybe run into xmax or thermal problems I think.

From a short test just before I can say LR sounds brighter/harsh/more direct than BW at the same frequency.
BW will sound smoother, not as bright.

From my understanding LR will fit seamlessly and flat.
Now when I use BW which I intend there might be dips or peaks.
How I am going to solve the problem?
Leave a space or overlap?
And then by how much?
I want to crossover at min. 80Hz max. 110Hz.
 
Would need measurements to know whats going on in your system, but a textbook Butterworth will sum with a 3dB peak on axis. A nice trick to come close to making it sum flat again on axis (but of course putting a hole back in the power response) is to make the high and low pass of the Butterworths intersect at -6 dB. Multiply low pass frequency by 0.77 and high pass by 1.3.

These are all estimations, again you need real measures, but this is the gist of it.
 
I was speaking about acoustic slopes, and with my speaker off-axis response (power response) remains the same with lr 2/4/8! That is why I believe that the difference that I hear comes from phase shift difference/gd.

To globalplayer, I have found lr2 (12) acoustic best for my HT subwoofer too! I use minidsp 2x4HD between pre and power amps. Equalize room responses first at least one octave past xo (don't eg room modes) Then place one main speaker next to the sub, play both with xo and set delay for the main speaker by looking at summed response and step response (I use 0,7ms, remember to check polarity as per acoustic measurement - don't trust markings on cables and connectors!). Then put speakers at normal places and adjust levels. Low order has tighter and snappier sound to my ears!

I have found it very educative to take measurements and to listen to music while making changes to a speaker/dsp. It takes just a couple of mouse clicks!
 
Last edited:
This sounds like a really good plan.
Thanks for the tips.

In the meanwhile I changed all crossover points by ear to BW24 and leaving gaps between bands. Sounds much more smooth and relaxed but looses some punch in the lows.

Maybe LR for kickbins and sub and BW for tops might be best.
But then again the smooth sound is nice also.

Only thing I measured yet was chassis distance from the next chassis for rough delay figures.

I will measure FR etc. also in the near future.
 
No suggestions, just curious about your solution. I'm suspecting the power dips are more severe as the XO filter order increases. It would require measurement to prove it.
I've been reading and experimenting with this S. Harsch XO
However, due to the low xover and U-frame resonance already mentioned, and the fact that the Eikona is already rolling off acoustically, there appears to be few downsides and more upsides to the LR8 and......it sounds better:)
 
After using the minidsp for a while I try to use Bessel filters in favour. They sound more dynamic and "lively" to my ears than the usual Linkwitz-Riley oder Butterworth filters. The measured step response is showing that too. When you combine suitable loudspeakers with Bessel filters, the result should be good enough even vor critical ears.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.