Round Tractrix

[Al Klappenberger]

Here's another way to look at the directivity. Real egg-head stuff!!

Lee sent me the raw dat and the software to organize it all sorts of ways. It will display the polar plots one frequency at a time. Doing that shows the beanwidth just starts to narrow up a t 4000 Hz and only drops ab it to 6000 Hz. That says the 4000 - 6000 Hz Crossovers we have been using is perfect for a 3-way system. I'll post those next.

Al K.

[jwc]

Al...not to thow a monkey wrench into this but....I wonder if the upper frequency coverage would be better with a non coaxial 2 way driver? Like say a JBL 2446 or something else?

jc

[Al Klappenberger]

Here's the mid-range polar plots separated.

Al k.

[Al Klappenberger]

Jc,

"I wonder if the
upper frequency coverage would be better with a non coaxial 2 way
driver? Like say a JBL 2446 or something else?"

I am no expert, but I think the performance of the horn is separate from the driver once the anomalies involved with the crossover are factored out. I would like to see test to be sure of that, but a driver that will go from 400 Hz to tweeter frequencies is rather hard to come by!

Al k.

BTW: I have the JBL 2426h (a 1-inch driver) and have tested them on several horns. It makes it flat to 6KHz but drops quickly above that. If the 2446 is similar it would require major EQing to use it seriously for a 2-way system. I have done that with the 2426 and it sounded fine though. It is a possibility.

[robi]

Happy to see your development!!!

Thank you from Italy!!!!!

[mikebse2a3]

I am no expert on this sort of thing, but I think this is typical of using a big horn to go wide range. It's going to beam the highs.

Hey Al

First, I'm so glad to see test like these being performed..! [Y]

Large Horns can do excellent polars over a very wide bandwidth if shaped properly(Klipsch's K510 and K402 are excellent in this regard in the Horizontal and of course the K402 is also superb in the Vertical while the K510 does start to exhibit Vertical Polar broadening like all horns with small vertical mouth dimensions.

mike tn

[mikebse2a3]

Jc,

"I wonder if the upper frequency coverage would be better with a non coaxial 2 way driver? Like say a JBL 2446 or something else?"

I am no expert, but I think the performance of the horn is separate from the driver once the anomalies involved with the crossover are factored out. I would like to see test to be sure of that, but a driver that will go from 400 Hz to tweeter frequencies is rather hard to come by!

Al k.

Roy has shared with us in the past that once the frequency wavelengths start to get close to the horn's throat diameter (ie: 2" in this case = approximately 7000 Hz ) the phase plug and other internal driver design structures begin to determine the polar response.

Note: I'm assuming no narrowing, diffraction slots or other obsticals in the horns throat area in the above comment.

So a different coaxial or drivers like the K69-A might improve the upper polars somewhat above approx. 7000Hz but it does look to be narrowing even below this as evidenced by the change from 4000 Hz to 8000Hz so I suspect like others have said in that this horn would probably be best used as a midrange in a 3-way system.

mike tn

[Al Klappenberger]

Mike,

I tend to agree. The driver used in the test was a coaxial B&C DCX50. It's the same type driver as l istend to a week or so ago, just a different one. The polar response above the crossover (about 9 - 10 Khz) acted squirley. The bottom line for me is that nothing beats a good 3-way system with extremely steep crossovers between them. The killer is overlap, not delay or phase errors!

Lee set me vertical polar plots a while ago too. They show as smooth transition to a narrowing pattern as frequency went up. Again, this went squirly as the tweeter section cut in.

Al K

[mikebse2a3]

The killer is overlap, not delay or phase errors!

Having gone with active crossovers and being able to compare shallow versus steep crossovers I like steep crossovers also becauses of the improvement in clarity heard through the shared crossover regions when two-drivers are covering the same frequencies. Being able to do time alignment is also audible as an increase in clarity through the shared frequency region(at least in the region of 500Hz crossover points in my experience).

mike tn

[mark1101]

Just another point..........an additional advantage of time alignment is that the entire frequency spectrum gets aligned. We always seem to dwell on the "overlap" area, when in reality the entire driver (actually all drivers) are out of alignment which skews the entire guitar or any instrument or vocal spanning a crossover point. With passive crossovers, probably the best you can do is steep slope networks to remove the smear in the "overlap" region.........but in reality the drivers are still out of alignment everywhere else. This can only be fixed by aligning the voice coils or by digital delay. Just wanted to toss that in the discussion since alignment had come up.

[Al Klappenberger]

Mark,

I think what most people miss on the time alignment thing is that "time" and "phase" are related. If only a single driver is making each "harmonic", than the only relationship that is misaligned is phase. I believe you ear is totally deaf to the phase relationships of harmonics to the fundamental. I experimented with this by using three different sine wave generators to simulate the harmonics of a square wave (1 Khz, 3 Khz and 5 Khz sumed together with levels 1/1 1/3 and 1/5.). This simulates each harmonic reproduced by a different driver. Only in this case the phase is continually changing, not just knocked out of phase to a fixed degree. I listened to the sum through headphones while looking at the signal on an oscilloscope. The waveform crawled like a worm through all sorts of shapes. To spite looking at the scope as a clue, the sound was a continuous raspy tone, just like the crude square wave it would represent if the phases were fixed and aligned properly. I conclude that YOU CAN NOT HEAR THE PHASE ERRORS BETWEEN COMPONENTS OF A COMPLEX WAVEFORM, even if they are continuously changing let alone if they are fixed! What you CAN hear is if each component comes to you twice or if they are so far out they become separate sounds resolvable by your brain. Your brain can't sum up unrelated tones in real time. It's just not fast enough. Only an oscilloscope can do that! It can compare phases of the same tone if they come to both ears at once. That's the stereo image. The only question becomes: how far out of time do the components have to be before the brain can resolve them as separate sounds. I don't know!

Al k.

[mark1101]

I've heard the delays are close enough when within 1 Msec. But I haven't experimented with this. You may be right with what you are saying above. What I can say I noticed was that using your ESNs was a huge improvement in clarity over shallow sloped networks. But when I went to the active system which included settings for EQ, channel gain, and channel delay.........that was another step forward and an even clearer more coherent presentation. But again, I can't say WHY that is........just that I noticed it. I alwayas assumed that it was the driver time alignment.

[Al Klappenberger]

Mark,

That 1 mSec number would have to be explained more as to what it really means. It's out of context.

It could be that the difference you heard with the time aligned system was the difference in damping factor gained by removing all the losses involved with the passive crossover. This is what bi-amping is all about. Like with the 1 Msec number. It's hard to pin down what the actual change you heard is really the result of.

Testing for the delay time that would be required for the different components of a waveform to be perceived as separate sounds is difficult considering a short "pip" type of sound is constructed of may components itself, so it's becomes difficult to identify what "harmonic" is what! I think such a sound would be required just to identify it over another sound. That's one reason I say the 1 mSec number is out of context. That stuff gets sticky! Remember the classic tap dancer thing from years ago. Two taps were clearly being heard in their entirety twice. The complaint was not a thud followed by a snap assuming a "tap" consisted of thud+snap at the same instant. Using the same analogy, what was heard was thud+snap followed by another thud+snap, not thud, then, the snap an instant later. Confusing isn't it? Taking it even further, What delay is requited between the thud and the snap before you hear each separately? That is what I don't know!! [:S] I think THAT is time alignment. The snap twice is driver overlap.

[Chris A]

From http://forums.klipsch.com/forums/t/22533.aspx?PageIndex=2:

Around 1978 there was a company called UlraPhase, Don McDonald was the CEO.

They made a line of time aligned speakers. Good drivers, good cabinets, good crossovers, good design, etc.

They
had a showroom display that was quite interesting. It was an actual
crossover plus the midrange and tweeter out of one of their speakers,
all wired up. The point was to play this at a reasonable volume level
and move the midrange and tweeter backwards and forwards in your hands.
There was a point where the sound just came into 'focus'. The range of
motion to hear the sound pop in and out of focus was on the order of ±
1/2".

[Al Klappenberger]

It would seem that this question should be resolvable, but moving drivers back and fourth by hand seems a bit crude. The "1/2 inch" thing seems a bit to simplistic, You can change the path length from two things to you that are sitting 6 inches apart horizontally by a half inch just by taking a single step to one side! I think there has to be more to it that that.

One thing I think is for sure though. Perfect time alignment is obviously better, if you can do it without compromising other factors. Digital delay equipment is the way to do it. That means bi-amping or tri-amping. That can be a bag of worms. I think it's the way to go if your are up to handling the worms! Most people are not.

Al K.

[djk]

" It would seem that this question should be resolvable, but moving drivers back and fourth by hand seems a bit crude. The "1/2 inch" thing seems a bit to simplistic, "

Why don't you try it before commenting. Think about what the crossover frequency of interest might be, the ±1/2" is about 1/8W at the crossover frequency.

[Al Klappenberger]

Dennis,

"Why don't you try it before commenting"

Very simple: I'm not about to start removing drivers from my speakers and running back and forth moving each driver 1/2 inch at a time on both speaker while resetting the test program each time. Even with help from another person. It's silly! A study like that should be double-blind anyhow. In short, I don't beleive the test that came up with the 1/2 in number was conclusive.

Al K.

[Marvel]

Since I have more time than equipment, I just moved the tweeter. I have an Eminence APT150 (same dirver as Bob's, but on the Eminence horn). Since it wouldn't fit the baffle in my LaScalas, I mounted it on a small baffle to set on top of the cabinets. At first I just had it along the front edge of the cabinet, and it sounded good.

I decided to move it to the back, to align it with the K55. BOOM!!! The imaging became cleaner, tighter, with the soundstage now HUGE.

I thought I would have a lot of trouble withreflections off the top of the cabinets, but it hasn't been a problem for me. I even tried different materials on the top of the cab, to see if I could hear a difference. No change.

Bruce

[Al Klappenberger]

Marvel,

What networks in in your LaScalas?

AL K.

[Chris A]

decided to move it to the back, to align it with the K55. BOOM!!! The imaging became cleaner, tighter, with the soundstage now HUGE.

Bruce,

You know that this makes you the new poster child for the "time-aligned" crowd... [] This is very interesting that it works well just as you have implemented it.

If I were you, I'd make patent application ASAP...

Chris [:#]