Active Crossovers

[DrWho]

ben: I don't understand why the driver
alignment is good only for one point in space. In my mind's eye, once
the sound leaves the speaker with all drivers 'time aligned' the sound
should propagate in unison (barring room interference, of course). So,
I would think that the sound would be 'aligned' for any location in the
room. I guess that is not correct.

It's actually good
for more than just a single point in space when all your drivers are
stacked such that they're centered above each other. The problem is
that when you move your head up and down, the relative difference in
distance between the HF and LF changes. If you move side to side, then the relative distance doesn't change...so perfect time-alignment only holds true for a
plane level with where you choose to time-align.

However, if you choose
a good xover frequency between the HF and LF, then you can
theoretically take advantage of the phase shift (time delay) that
happens when you move vertically in order to gain better control over
the vertical polar response at the xover frequency. You'll only get a
dip if the HF and LF are physically too far apart. Another thing that
can happen is that the vertical polars of the LF will be tight enough
at the xover frequency such that there isn't enough energy off-axis to
create any noticeable phase cancellation from the extra delay.

In other words,
you can get effective time-alignment over a much larger 3D area if you
consider the use of delay an acceptable means for polar steering. If
you stay within one wavelength, then I'd argue the extra delay off-axis is not
creating an audible compromise.

[ben.]

ben: I don't understand why the driver alignment is good only for one point in space. In my mind's eye, once the sound leaves the speaker with all drivers 'time aligned' the sound should propagate in unison (barring room interference, of course). So, I would think that the sound would be 'aligned' for any location in the room. I guess that is not correct.

It's actually good for more than just a single point in space when all your drivers are stacked such that they're centered above each other. The problem is that when you move your head up and down, the relative difference in distance between the HF and LF changes. If you move side to side, then the relative distance doesn't change...so perfect time-alignment only holds true for a plane level with where you choose to time-align. (snip)

Yes it does. Try it. The real world is out there, Mike. Put down the books!!

[ben.]

Ok, my point here is that time alignment in a home hifi environment using the better Klipsch speakers will provide extremely small improvements, and if you want to start splitting those hairs, the razor blade's going to cut both ways.

[DrWho]

The amount of variable in this issue are huge! Frankly, I'm amazed speakers work as well as they do.

Me too []

I think an important thing to keep in mind is the relative magnitude of everything. Obviously time-alignment itself isn't the most important variable.

Just another thought to throw out there...the lack of phase (time) alignment at the xover frequency will require one of the drivers to work harder if you desire to fill in the gap on-axis. You will also get a resultant shift in the polar response through the xover as well. The creates a narrow region of increased distortion and then offsets the tonal balance of the off-axis sound. Filling in the gap usually increases the amount of overlap between the two drivers, which then makes the time delay much more audible; and moreso as the region of overlap increases. I bring this up to imply that the effects of time-alignment (as imperfect as it may be) provides benefits to other behaviors that I would argue have a larger effect on the sound. The thing is, you can't get those benefits without phase-alignment...

Rane has a really cool article that describes the polar shifting behavior:
http://www.rane.com/note160.html

[DrWho]

Yes it does. Try it. The real world is out there, Mike. Put down the books!!

There is always a plane where the pathlength difference is zero. Not all points fall on that plane, but that doesn't mean the plane doesn't exist...

Before you add time-alignment, the plane is centered and normal to the vector between the two points. When you add delay to time-align, the plane tilts so that it cuts through the point where you chose to do the time-alignment. Describing it as being the same when you move horizontally was just an oversimplified way of describing the "real world" [][] In real life, the plane of zero pathlength difference usually ends up tilted a bit...

[ben.]

Yes it does. Try it. The real world is out there, Mike. Put down the books!!

There is always a plane where the pathlength difference is zero. Not all points fall on that plane, but that doesn't mean the plane doesn't exist...

Before you add time-alignment, the plane is centered and normal to the vector between the two points. When you add delay to time-align, the plane tilts so that it cuts through the point where you chose to do the time-alignment. Describing it as being the same when you move horizontally was just an oversimplified way of describing the "real world" In real life, the plane of zero pathlength difference usually ends up tilted a bit...

While you'd certainly be able to draw a line where the path lengths remain equal, performing my experiment would demonstrate just how impractical the majority of those locations would be for a listening position relative to a single speaker assembly.

Then what happens when you add a second speaker system on the opposite side of the room? Then how many points in space are now time aligned for both assemblies? Disregard the interaction between left and right sides, which time alignment of drivers can't correct. Just consider that curved line of points you're describing and think about how many points of intersection are possible with two speaker assemblies.

[ben.]

oh, then do the same thing with a 3-way...

[Coytee]

This is all giving me a headache... I'm selling everything and getting a pair of 901's that need refoamed.

[]

[Rudy81]

Ah, the 901's. Now there's are speaker! [<)]

[Coytee]

Yes, perfect signal alignment [A]

[Rudy81]

Many years ago I owned, and was very proud of, a set of 901's. I though I had a great speaker at the time. CD's were coming out and I was starting to build up a collection when I ran into a CD, Overture of 1812, that had a WARNING label on it. It indicated the CD could damage your speakers if played too loud due to the 'digital cannons'. Well, the 901's couldn't reproduce a pop gun, no matter how much power I pumped into them....it's that physics thing. I decided to get a real set of speakers, and got to audition a set of La Scalas, with the CD, and it blew me out of the room. I have never owned any other speakers since. I love my Klipsch Heritage.

[Coytee]

time alignment in a home hifi environment using the better Klipsch speakers will provide extremely small improvements

If that is in fact, accurate then there is something else to mention (for my experiences)

Although I'll admit that listening to a pair of Heresy's and then switchout out to a pair of Khorns/Jubilees is a much larger "OMG" experience, I none the less, noticed a very real difference from going from Khorns to Jubilees.

Various factors are/were:

1. ES slopes in the Khorns (my opinion, those required more time therefore distance for the sound to meld together into a coherent single wavefront)

2. Jubilees are signal aligned so all the sound is starting at the same bell

3. Jubilees are 2-way rather than 3 way so there is one less component to mix into the wall of sound and this alone might also allow for a quicker melding of the wave front.

I do not know if it was the change from a 3 way to a 2 way that did it, the change from a steep slope passive to a signal aligned active that did it BUT... with those said, I can assure you that the qualitive difference between these two experiences in the same corner of the same room with the same electronics was noticable. It was not as dramatic of a change as I said above, from going from Heresy's to Khorns but it was a definate increase in creating more of a "one voice" speaker. Now that I've had the difference, I can see how I was able to discern the height of the Khorn by noticing the sound from the different drivers. This seperation of sound is one thing that went away after 30' (I actually measured). The Khorn became a 'single voice' at this distance.

Today, I can be 10' away from the Jubilee's, staring at them, focusing on the sound TRYING to ascertain which half is creating part of the spectrum. All I can usuall come up with is "gee.... it really sounds like all the sound is coming from the center of the speaker" It's darn near impossible for my tin ears to differentiate the different drivers. The "one voice" seems to start at the face of the speaker and unlike my Khorns, doesn't take 30' to gell.

I personally attribute the Khorns sound in part, to the ES crossovers which I deduce would require more time (therefore distance) for the sound to mix because they separate the drivers more. I would think a Klipsch "A" crossover would not have the same effect however, it's going to have more over lap so again, we're into tradeoffs.

All the above to suggest to anyone reading.... my personal experience was very positive from going to a steep slope passive to a signal aligned active. If there is anyone nearby, we could probably duplicate that with my LaScalas since they still have Al's ES networks in them. Be a chance to hear it for yourself if you wanted.

[ClaudeJ1]

Many years ago I owned, and was very proud of, a set of 901's. I though I had a great speaker at the time. CD's were coming out and I was starting to build up a collection when I ran into a CD, Overture of 1812, that had a WARNING label on it. It indicated the CD could damage your speakers if played too loud due to the 'digital cannons'. Well, the 901's couldn't reproduce a pop gun, no matter how much power I pumped into them....it's that physics thing. I decided to get a real set of speakers, and got to audition a set of La Scalas, with the CD, and it blew me out of the room. I have never owned any other speakers since. I love my Klipsch Heritage.

Great story. All you needed was an M&K subwwofer to get the missing 2 octaves of bass. I remember, that recording was one of the early CD's, back when myself and about 3 other members of the Audio Engineering Society even had a CD player.........circa 1983, when you either got a Sony or nothing.......because taht is all there was at over a grand apiece (a lotta money for a kid in those days).

Anyhow, I had Khorns and those did a pretty good job on the cannons, but not as good as Danley Super Spud, DTS-10 would do today, to bend your walls and windows. HAH!

[ben.]

time alignment in a home hifi environment using the better Klipsch speakers will provide extremely small improvements

If that is in fact, accurate then there is something else to mention (for my experiences)

Although I'll admit that listening to a pair of Heresy's and then switchout out to a pair of Khorns/Jubilees is a much larger "OMG" experience, I none the less, noticed a very real difference from going from Khorns to Jubilees.

Various factors are/were:

1. ES slopes in the Khorns (my opinion, those required more time therefore distance for the sound to meld together into a coherent single wavefront)

2. Jubilees are signal aligned so all the sound is starting at the same bell

3. Jubilees are 2-way rather than 3 way so there is one less component to mix into the wall of sound and this alone might also allow for a quicker melding of the wave front.

I do not know if it was the change from a 3 way to a 2 way that did it, the change from a steep slope passive to a signal aligned active that did it BUT... with those said, I can assure you that the qualitive difference between these two experiences in the same corner of the same room with the same electronics was noticable. It was not as dramatic of a change as I said above, from going from Heresy's to Khorns but it was a definate increase in creating more of a "one voice" speaker. Now that I've had the difference, I can see how I was able to discern the height of the Khorn by noticing the sound from the different drivers. This seperation of sound is one thing that went away after 30' (I actually measured). The Khorn became a 'single voice' at this distance.

Today, I can be 10' away from the Jubilee's, staring at them, focusing on the sound TRYING to ascertain which half is creating part of the spectrum. All I can usuall come up with is "gee.... it really sounds like all the sound is coming from the center of the speaker" It's darn near impossible for my tin ears to differentiate the different drivers. The "one voice" seems to start at the face of the speaker and unlike my Khorns, doesn't take 30' to gell.

I personally attribute the Khorns sound in part, to the ES crossovers which I deduce would require more time (therefore distance) for the sound to mix because they separate the drivers more. I would think a Klipsch "A" crossover would not have the same effect however, it's going to have more over lap so again, we're into tradeoffs.

All the above to suggest to anyone reading.... my personal experience was very positive from going to a steep slope passive to a signal aligned active. If there is anyone nearby, we could probably duplicate that with my LaScalas since they still have Al's ES networks in them. Be a chance to hear it for yourself if you wanted.

There you go. If you really want to hear the effect of the crossover on all this, do it with your LS. Then the results would be meaningful.

Attributing the difference between Khorns and Jubilees to the type of crossover is like eating an apple and an orange, then saying they taste different because one was red and the other was orange colored.

#3 on your list of guesses should be number one. Using one driver from what - 800Hz? on up is a big difference. Plus they actually use a high end driver right? The B&C DE-72 or 75? Those are a couple of the best compression drivers available and are in a completely different league than what was used in the Heritage products I'm familiar with.

AND... the wave front does not MELD. It's still two wave fronts and they can be manipulated to arrive simultaneously at a given point, but they do not "join together".

[AltmanEars]

One can activate Khorns but in practice I think you'll find the Jubilee active solution to be ultimately better and cheaper.

For optimal performance you tend to need a very good crossover and 3 way xovers are typically twice as expensive as 2 way.

To blend the drivers it is significantly better to have the same amps meaning you will be buying 4 really good single ended monoblocks.

I know this is different than what you have read in the audio press but trust me it is absolutely true and it really speeds your design time.

[Coytee]

Using one driver from what - 800Hz? on up is a big difference. Plus they actually use a high end driver right? The B&C DE-72 or 75?

Actually, I think we cross electronically on the bass bin about 450 and pickup on the HF driver about 380. I don't know how that translates into acoustic crossover. (looking at my RACE file and low out is Linkwitz 24/450hz and high out is LR24 380hz)

We're using the K69, not the B&C drivers. The only person that I'm aware of that has the B&C driver is Bob Riff. Everyone else has (or started with) the K69 and a handful of guys have now gone to the TAD 4002. Someday when Mr. Budget and I are smiling with each other, I'll fork out for a pair of TAD's as well. A definate increase in sound quality/clarity.

AND... the wave front does not MELD. It's still two wave fronts and they can be manipulated to arrive simultaneously at a given point, but they do not "join together".

Perhaps so, but tell my ears that what they are hearing isn't a coherent "bubble" of sound []

[Rudy81]

What is the accepted 'length' of the Khorn bass bin from driver to mouth? I figure I'd ask in case I ever get to play with this time alignment thing. Somewhere I thought I saw 8'???

[ben.]

Using one driver from what - 800Hz? on up is a big difference. Plus they actually use a high end driver right? The B&C DE-72 or 75?

Actually, I think we cross electronically on the bass bin about 450 and pickup on the HF driver about 380. I don't know how that translates into acoustic crossover. (looking at my RACE file and low out is Linkwitz 24/450hz and high out is LR24 380hz)

We're using the K69, not the B&C drivers. The only person that I'm aware of that has the B&C driver is Bob Riff. Everyone else has (or started with) the K69 and a handful of guys have now gone to the TAD 4002. Someday when Mr. Budget and I are smiling with each other, I'll fork out for a pair of TAD's as well. A definate increase in sound quality/clarity.

AND... the wave front does not MELD. It's still two wave fronts and they can be manipulated to arrive simultaneously at a given point, but they do not "join together".

Perhaps so, but tell my ears that what they are hearing isn't a coherent "bubble" of sound

The crossover that low is even better, then. Wonder what the K69 actually is... My point stands that it must be light years ahead of the poor old drivers PWK used to such great effect.

As far as what you're hearing, what I've been trying to do is help you understand what it is you're hearing and why, so when you give advice it's rooted in reality and not conjecture so as not to confuse matters further. I know you're trying to help and it's really cool that you are so generous with others to expose them to new gear and help them improve their systems, but some of the conclusions you come to are not logical.

One of the things I'm trying to make clear is that the coherent bubble of sound you're hearing has more to do with the fact that you're hearing the bulk of the sound from one source, not the fact that you are time aligned around 400Hz. Try changing your delay settings a few milliseconds - see if it makes it all fall apart or if it still pretty much hangs together. Then you're actully learning things. You've got everything you need to stop guessing.

[DrWho]

While you'd certainly be able to draw a line
where the path lengths remain equal, performing my experiment would
demonstrate just how impractical the majority of those locations would
be for a listening position relative to a single speaker assembly.

And then when you time-align along an axis that
intersects the listening position, now you've got a lot of seats
covered within that plane - or at least very close to it.

Then what happens when you add a second speaker system on the opposite side of the room?

I don't think that is at all related to time-alignment through a crossover...

==============================

The worst case scenario for a system time-aligned along a
forward firing axis is the spot directly above and below the speaker,
since that will be the point of greatest time-differential. If you
xover at say 190Hz with two omnidirectional devices, and the acoustic
centers are ~3ft apart, then you've got 180 degrees of phase rotation
due to the propagation difference. If the design goal is a controlled
polar response, then having acoustic rejection directly above and below
the speaker is actually beneficial to keeping the wavefront firing only
forward. If you xover at a higher frequency (say 400Hz), then the 180
degree rejection node moves forward (at 400Hz, I believe it would be 60 degrees away from the forward firing axis).

In
the real world, the HF is generally getting wider on the polars and the
LF is generally getting narrower. If the wavefront of the LF never
overlaps with the HF, then there is no time-arrival difference because
only one drive unit is propogating off into that direction. On a
slightly different note, let's say your HF loses vertical pattern
control before it loses it in the horizontal (which is often the case).
If you can xover low enough to where the vertical polars of the LF are
wide enough, then you can use the vertical propogation difference to
create phase cancellation and tuck the vertical polars back to where
you want them to be. I could see this being much more advantageous if
the LF bin was starting to clover-leaf in the vertical (since that
allows the forward wavefront to stay tight). If the polar response of
both the HF and LF are identical, then the system polar response
doesn't change.

Btw, the same physics behind the polar steering
I'm describing also apply to systems that aren't time-aligned. The
problem there is that the polars at the xover frequency get kicked off
in crazy directions. And if the delay is beyond just one cycle, then
the polars change dynamically with the input signal...

The point I'm trying to make is that it involves 3D behavior and the time arrival differences behave differently based on the size of the wavelengths in question. The nice thing about a crossover between two drive units is that the region of overlap is relatively small and easy to deal with. Full-range line arrays with several different units gets a lot more complicated...yet they still steer the polars quite well (provided there's no air turbulence).

[seti]

I'm using the B&C DE-82-TN. That is the driver Roy & PWK had picked for the 403. I really like it so far. I'm crossing them at 600.