Driver delay/time alignment in Heritage speakers?

[Islander]

Many speaker designs try to compensate for time discrepancies between drivers by placing the tweeters back from the front of the enclosures, in some cases even having a separate mid/high cabinet that can be slid back and forth or tilted very precisely, depending on the listening distance, to ensure that all the musical notes arrive at your ears at the same time.

Then I look at the La Scalas, with the tweeter driver about three inches back from the front of the enclosure, the squawker about two feet back, and the woofer about three-and-a-half feet back, and don't hear any "time smearing". Could it be that time alignment just isn't a problem in typical listening situations?

Then I looked at the specs for the Klipsch theater speakers and noticed that most of them have built in delay between the drivers. In the case of the 3-way Jubilee 535, it's 3.5ms for the midrange and 4.6ms for the tweeter.

My question is this: is there any built-in delay in the Heritage speakers between the drivers, perhaps in the crossovers? Or does something in the cabinet design correct for horn dimensions that look like they should cause a large time misalignment?

[sfogg]

" the squawker about two feet back, and the woofer about three-and-a-half feet back, and don't hear any "time smearing"."

You might not hear it as distinct smearing of notes, you might just notice the comb filtering and altering of FR that results from the drivers in effect being out of phase with each other. Or you might not notice it at all.

"Could it be that time alignment just isn't a problem in typical listening situations?"

Some people appear to be more sensitive to it then others. Or perhaps some people know what to listen for (and what it sounds like) and others don't.

" is there any built-in delay in the Heritage speakers between the drivers, perhaps in the crossovers?"

Nope. The drivers are out of alignment. Like most things it is a tradeoff. The horns in that layout with the passives basically prevent you from being time aligned while on the flip side you get the distortion/dynamic benefits of the horns. Some are willing to make that tradeoff, some aren't.

Shawn

[TBrennan]

It's a controversey.

This entire thing goes back to the Eleanor Powell "double tap" incident at MGM studios in the early 1930s. When monitoring playback of Powell tapdancing an echo was noticed on the taps. John Hilliard, subsequently the greatest of all horn designers, was a young sound engineer at MGM and traced the problem to the 8 foot difference in the path lengths of the horns in the 2-way Western Electric monitor speaker.

Hilliard did experiments that showed that the effect was time and frequency dependent and that a delay of less than 3 milleseconds (about 3 feet) was inaudible using crossvers between 350 to 800hz.

A result of this incident and WE's refusal to market their improved Fletcher loudspeaker was the decision of MGM sound honcho Douglas Shearer for MGM to design their own improved system which would among other things minmize time delay. Hilliard was put in charge of the project which then developed the famous Shearer Horn. An entire galaxy of talent worked on this system including Hilliard, James Lansing, Bob Stephens and even RCA's Harry Olson.

Note that though the folded basshorn of the Shearer cut time delay compared to the WE snail horns it didn't eliminate it totally but Hilliard's later Altec Voice of the Theatre did. So it seems that even though Hilliard thought some delay was inaudible the existence of any at all still nagged at him.

Which brings us back to the present day. Is the time offset of the big Ks a problem? Theoretically the distance falls under the distance Hilliard thinks was undetectable. Might it still be a problem? Perhaps. But keep in mind that all loudspeaker designs make compromises, given a choice, as PWK was with his design, between lower distortion and time delay which would you choose? I'd take the lower distortion of the horn system.

Photo of Shearer Horn from Lansing Heritage site

[sfogg]

Tom,

Knowing how big that multicell is that is a *big* bass horn!

"Hilliard did experiments that showed that the effect was time and frequency dependent and that a delay of less than 3 milleseconds (about 3 feet) was inaudible using crossvers between 350 to 800hz."

Do you know where Hilliard did the experiments? In other words was it done in a theater or in the home. I would guess it was likely in a theater where the importance of off axis power response (other then just strict coverage angles) is different then in a much smaller listening room. I'm not sure the understanding of that importance was really all that well known back then. I believe Acoustic Research was one of the companies that started stressing the importance of flat power response back in the day.

As far as being inaudible the FR changes from comb filtering is easy enough to demonstrate, in a home setting.. harder in a theater when you are 30+ feet away from the speaker, to someone that wants to hear it.

"But keep in mind that all loudspeaker designs make compromises, given a choice, as PWK was with his design, between lower distortion and time delay which would you choose?"

Absolutely... engineering is knowing what compromises to make.

" I'd take the lower distortion of the horn system."

It is possible to have both.

Shawn

[TBrennan]

Shawn---I don't know where Hilliard did his experiments, I'm sure he did an AES paper. Hilliard published an article in Audio magazine back in the 70s that covered the subject but I only remember the main points. And Altec honchos Badmaieff and Davis talk of it too in "How to Build Speaker Enclosures".

(Note that Badmaieff and Davis, though Altec engineers, said of the Klipschhorn "For it's size this type of horn has greater dynamic range and lower distortion at maximun spl than any other type". High praise indeed.)

"It is possible to have both."

Are you using a digital crossover?

[sfogg]

Tom,

"Are you using a digital crossover? "

Yup, three DCX2496s. Interestingly (or not) the automatic time alignment on the Behringer is putting 7.35ms of delay on the K510/44xt compared to the LaScala bass bin. I have run that numerous times and it is consistent.

When it was run three way I think it was putting about 4.5ms on the 805b/288 and about 7.8ms of delay on the 2404H.

Those times are considerably higher then I would have expected so it could be a measurement error/bug in the Behringer or maybe something else is going on which is throwing it off. Or maybe that is the delay needed in that setup. I want to try and verify those numbers though.

Shawn

[PrestonTom]

Tom,

"Are you using a digital crossover? "

Yup, three DCX2496s. Interestingly (or not) the automatic time alignment on the Behringer is putting 7.35ms of delay on the K510/44xt compared to the LaScala bass bin. I have run that numerous times and it is consistent.

When it was run three way I think it was putting about 4.5ms on the 805b/288 and about 7.8ms of delay on the 2404H.

Those times are considerably higher then I would have expected so it could be a measurement error/bug in the Behringer or maybe something else is going on which is throwing it off. Or maybe that is the delay needed in that setup. I want to try and verify those numbers though.

Shawn

Shaw,

The delays do seem a bit large.

As I recall you were mimicing some steepish filters by cascading a couple of filters. I assume that equates to a large number of taps and correspondingly a convolution that will could have its own considerable latency (but you already know this.....). Without knowing the sampling rate and number of taps etc, I can only guess, but I would not imagine it would be more than a msec or two. Is this done to a comparable degree for each driver or would one have more DSP-delay than the other (hence the exagerated driver mis-alignment).

Incidentally do you also toss in a cap to protect the drivers in case the Behringer etc has a hicup?

Also incidentally, you folks always bring up the "Tap dancing" anecdotes. There actually has been other research on this topic. Some of it was recently reported by Moore et al and published in JAES. Two or three msec is the ballpark critical value for audibility. But there are some caveats...

Good Luck,

-Tom

[TBrennan]

"so incidentally, you folks always bring up the "Tap dancing" anecdotes."

Well it's a good story with a great cast of characters. Hollywood's Golden Age and all that, Thalberg was Shearer's brother in law you know. And MGM did have the best sound while Shearer was in charge. (MGM had the best picture too, Shearer was involved in the development of the "Camera 65" anomorphic 65mm process used for Ben-Hur and Mutiny on the Bounty)

And it's a story that puts the problem in terms people can immediately understand; "double-tap, I get it".

[sfogg]

Hi Tom,

Yup, that is a possibility. But I am pretty sure I also tried very low order filters (first or second order) with no other EQ and the Behringer came back with the same numbers. Which at least suggests the Behringer keeps latency consistent through all the outputs.

Does the speed of sound get slower in a horn? The medium is somewhat different (different impedance then 'open' air) so that might not be as crazy as it sounds. I want to try and verify those numbers some other way. A 500hz tone (crossover point) on the scope and analyzer summed pretty close to as high as it would with the delay number the Behringer came up with.

If my laptops external sound card wasn't misbehaving I could see this easier with ETF. I need to spend some time working on that.

"Incidentally do you also toss in a cap to protect the drivers in case the Behringer etc has a hicup?"

The amps are AC coupled so that protects against Behringer problems. The DCX2496 also has good muting to prevent against turn on/off transients, unlike the BFDs. From the Teacs I run straight to the amps. The Teacs outputs are relay switched for on/off (and I believe DC protection) so they are dead quite turning them on and off as well.

Shawn

[Islander]

Thanks for the responses. There are some very knowledgeable dudes (and dudettes, I'm sure) on this forum.

So it appears that any time misaligment in Heritage speakers is so minimal as to be inaudible to most untrained ears, and is less noticeable the further the listener is from the speaker. And yet the pro cinema Klipsches have built-in delay. In a typical cinema setting, most listeners are probably 20-200 feet away, so why would the design engineers feel the need to address this "problem", unless they were expecting an audience of sound engineers, each with his favourite sound meters in his pocket?

[sfogg]

"So it appears that any time misaligment in Heritage speakers is so minimal as to be inaudible to most untrained ears"

Or it is audible and some don't recognize it for what it is. Not many have heard the difference between a time aligned K'Horn and a 'regular' K'Horn for example.

" and is less noticeable the further the listener is from the speaker. "

That wasn't exactly what I meant. I meant the demonstration of one of the ways to hear the effects of comb filtering is harder to do the further away from the speaker you get.

"And yet the pro cinema Klipsches have built-in delay."

They don't have built in delay. They have available delay settings to be using in the digital crossovers that are paired up with the cinema speakers which are bi/tri/quad amped in many situations. There is no advantage on its own to having drivers out of time alignment. Digital crossover allow the drivers to be time aligned. As such that ability within the digital crossovers is used as it takes the problems of misaligned drivers out of the equation.

Shawn

[Islander]

"So it appears that any time misaligment in Heritage speakers is so minimal as to be inaudible to most untrained ears"

Or it is audible and some don't recognize it for what it is. Not many have heard the difference between a time aligned K'Horn and a 'regular' K'Horn for example.

" and is less noticeable the further the listener is from the speaker. "

That wasn't exactly what I meant. I meant the demonstration of one of the ways to hear the effects of comb filtering is harder to do the further away from the speaker you get.

"And yet the pro cinema Klipsches have built-in delay."

They don't have built in delay. They have available delay settings to be using in the digital crossovers that are paired up with the cinema speakers which are bi/tri/quad amped in many situations. There is no advantage on its own to having drivers out of time alignment. Digital crossover allow the drivers to be time aligned. As such that ability within the digital crossovers is used as it takes the problems of misaligned drivers out of the equation.

Shawn

Thanks for clearing that up, Shawn. I guess Klipsch assumes that anyone reading the specs of pro cinema speakers knows what they're looking at, unlike a newbie like me...[*-)]

[DrWho]

This topic comes up from time to time, so let me start off by saying
that I'm adamently a supporter for signal-alignment. I would be more
than willing to demonstrate the effects for anyone that happens to be
visiting Champaign, IL.

Two of the issues that come up are the "tap dancers" and the
"comb-filtering" - both of which are real concerns, but I don't feel
they get to the heart of the issue.

The tap dancing is just
determining the threshold at which our ears are able to distinguish
between two seperate sounds. In other words, two sounds that happen
within a certain window of time are going to be perceived as a single
sound. Usually signal arrival differences around 10-20ms are considered
the threshold - though I'm sure in very controlled condititions you can
distinguish over much shorter periods.

Comb-filtering is what you see when you study the steady state
signal of the response...it's basically the average frequency response
over a long period of time.

You could minimize the effects of
both of these concerns, yet you will still be able to hear a difference
between two aligned and two unaligned signals - even if they don't
share the same frequency bandwidths.

Let's say you have driver
component A and B. Component A is the HF driver crossed over with an
infinite slope at 200Hz to B, the LF driver that arrives 5ms late. You
won't have any comb-filtering and you won't have the double tap issues
(or at the very least, let's ignore them). A simple analogy to use is
the sound from a kick drum...

A classical kick drum has two parts to the sound - the high
frequency attack of the hammer hitting the head, and the low frequency
boom. 600-3kHz is pretty much the attack and the low frequency boom is
pretty much 30-200Hz depending on the drum. But using this crude
working model, it should be apparent that the attack is going to happen
5ms earlier than the boom part of the sound (assuming that the attack
and boom happen at the same time....they actually don't, so it's going
to increase the distance by 5ms). One of the tricks I've picked up in
the studio is to seperate the components of the kick and move them
around in time to change the timbre of the kick drum. There is no
frequency response change or double-tapping happening, but the timbre
is changing. The reason the timbre changes is because the impulse
response changes and our ears hear a sort of average of the impulse. If
you change the frequency content at any given time, then the sound over
that short period is going to sound different. By delaying different
frequencies, you are changing the tonal balance at all time because
nothing is arriving at the same time.

I know I'm butchering this up (I'm in a rush to go visit me grandpa
in the hospital), but I plan to return to this later (probably late
tonite). Anyways, my point is that the issues of tap dancing and
comb-filtering really don't matter...you can still change the timbre of
a sound by moving bandwidths around in time.

In other words, the signal-alignment issues in the Heritage speakers
are indeed an issue - but as mentioned, the compromise in timbre is
really going to be small in comparison to the distortions it reduces.
Heck, I would argue that it's probably small in comparison to other
room acoustics issues - probably more on par with the differences heard
between different amplifiers.

Anyways, gotta run....I'll have my flame suit on when I come back []

[bodcaw boy]

imagine a klipschorn with a digital processor to provide delay and crossover?

roy delgado

[Champagne taste beer budget]

If my wife understood somewhat of that being discussed in this thread, and I could get her to read all the way through it, I might have a chance of her understanding my disease.

A very interesting read, no doubt, keep up the good work.

[BigStewMan]

know I'm butchering this up (I'm in a rush to go visit me grandpa in the hospital),

doc--i hope your grandpa's doing better.

best regards, steve

[DrWho]

Thanks for clearing that up, Shawn. I guess Klipsch assumes that anyone reading the specs of pro cinema speakers knows what they're looking at, unlike a newbie like me...[*-)]

I wouldn't say it has anything to do with newbies directly...pro cinemas are already going to have active crossovers with dedicated amps for each of the drivers, so throwing in the delay compensation is essentially a free improvement. It's rare to see crazy people like Shawn with active crossovers in the home because it means spending two to three times more on the amplifiers. It also requires the end-user to have some sort of clue what they're doing since they will be required to set up the active crossover. There is a lot of potential for making things worse (which doesn't help the reputation of the idea).

It's also bad marketing...providing the arrival offset between the drivers implies that there is a problem []

[DrWho]

Thanks Steve - he's making good progress and it's nothing super serious (or I wouldn't have mentioned it so casually).

[Klipsched with Yamahas]

Ah......Timbre: The combination of qualities of a sound that distinguishes it from other sounds of the same pitch and volume ( a piano, clarinet, oboe, and violin can hit the same tone but not the same timbre)

Just in case []

Man, we ought to have a sticky on DEFINITIONS! Maybe I'll post up and ask!

Interesting thread, and now I think I'll stay w/passive's! [:|]

[ben.]

Ah......Timbre: The combination of qualities of a sound that distinguishes it from other sounds of the same pitch and volume ( a piano, clarinet, oboe, and violin can hit the same tone but not the same timbre)

That wouldn't exclude two kick drums that are the same except the arrival time of a particular frequency band. If you can hear the difference, that's a timbre shift by definition. He used the term correctly for the working environment as well.

I showed my wife what a 1ms delay sounds like when applied to one channel of a 2-ch live recording. She swore at me to change it back. She doesn't usually swear.

BTW, a good rule of thumb to help some folks conceptualize the matter -- 1 foot of path difference is roughly 1 ms of time.

Mike, try this little trick next time you've got a few minutes to play. Use the kick or a trigger to trip the key of a gate (use a decent gate with some features). Feed something else through the audio in of the gate and process seperately. You'll find it quite useful I think. Rather than slamming compressors or crashing mic pres, you can add crunch to a nice kick drum simply by adding pink noise filtered to taste with EQ or a crossover. Or use a sine wave or synth tone to add focus, punch or resonance depending on the length of the hold or release setting on the gate. Throw some oddball effect on if you want. If it's too much, you can turn it down or shorten the hold... There's more tricks in there but I ain't givin it all away... This may all be mundane in the world of plug-ins, but it was a serious tool mixing down from 2".