A transient response simulator

[forr]

Forr,
Thanks for highlighting the alternative quasi transient perfect xo topologies. Have you tried the Harsch XO? I find that it has an almost flat phase response with just a 55deg mild rise above the XO point. It makes pretty decent square waves too.

I did it a long time ago and again these last days.
My system is based on ScanSpeak drivers :
Revelator 18 cm 18W85351W,
Discovery 10 cm 10F8414G10,
Illuminator 25 mm D3004 660000,
crossover frequencies at 344 Hz and 3.44 kHz.

As I have built a command at distance for my digital processor, I could change its setup within one second or so just by pressing a button. So I could compare the crossovers BUT18, L-R24, Lebbolo and Harsch almosts instantly.

I did not find great differences between them when listening right in the axis.
However I like to move in my listening room and not to have too great changes in the general sound balance.

At the end of my recent tests, my prefered configuration is still the Lebbolo because it gives, I think, a largest ambiance.

 

[xrk971]

Link for Lebbolo xo please?

 

[bwaslo]

Keep in mind that the drivers themselves also have a phase response that must be included in any delay consideration. Do not just look at the response of the crossover itself.

That's a point that gets overlooked a lot in crossover discussions. With speakers, we don't start with flat responses and just filter them to fit together -- the speakers start with their own built-in filters that we have to work with (or work around).

In light of the fact that you can now get a stand-alone FIR based equalizer for about $200, or do it with nearly any PC and soundcard, what's the rationale (even assuming that tranisent perfect were a major benefit)? it would seem that going nuts to get transient perfect with a passive crossover is like fixing hail damage to a car's paintjob with a marker pen -- just spray paint the damn thing!

 

[forr]

Link for Lebbolo xo please?

I gave the configuration here :

http://www.diyaudio.com/forums/multi-way/293175-transient-response-simulator.html#post4754119

For a crossover frequency at 1 kHz (-6.1 dB) :
hi-pass: order 1 at 384 Hz, order 2 at 1.170 kHz Q=0.633, delay 180 µs, inverted.
lo-pass: order 1 at 2.604 kHz, order 2 at 855 Hz, Q=0.633.

 

[Grasso789]

No. You need flat group delay, which is the same as linear phase response.

No. Flat is linear. A line may be horizontal, rising or falling, essentially constant or changing with constant pace. The opposite of linear is bent, changing with changing pace.

It's helpful to remember this difference by thinking like this: you want all frequencies to arrive at the same time, so they should have the same amount of delay. Phase lag, in degrees or radians, is like -delay*frequency, so the phase will be linearly increasing with frequency for a fixed delay.

Rite.

I believe, that a loudspeaker should have the more phase turn the louder it is. The average for home HiFi should be Pi phase turn between 150 and 7 KHz. I do not believe too much in science and technics.

A few months ago i built a small left/subwoofer/rite combination, in which around XO point satellites have same phase as subwoofer, just a second-order hipass for the sats and same-order lopass for the sub. A bass-enabled center and two first-order hi-passed sats would have been closer to the truth. Now i believe, that loudspeakers need not have phase turn, see there.