PDA

View Full Version : Time Alignment



alskinner
03-12-2005, 05:26 AM
I know this subject has been covered in other threads, but I found this interesting. I recently aquired a pair of 2360H horns with 2445J drivers. My thoughts were to start another project using my Altec A7 bottoms with the 2360 setup. Anyway, just to test the 2360s out I hooked them up in place of the 2397/2441 combo on my hybrid 4345s. The sound was amazing, the speakers dissapeared and the whole room seemed to become the speaker with sound coming in 3D. Of course it is somewhat impractical to have a horn the size of the 2360H mounted 30" in front of the rest of the drivers but it is a great sound. I am going to experiment with these a few days or until my wife gives me the "evil eye" to move them. I guess one of my questions is is there an analog way of time alignment and if so is there any links on how to accomplish this?
Thanks for any help

AL

RacerXtreme
03-12-2005, 07:18 AM
Did you just win those on EBAY? How would you describe their sound as compared to the 2441's ? I'm leaning towards buying the exact same setup to run on top of eight JBL 2225's in 4 Altec 210 VOTT cabinets. They do eat up a lot of room don't they..... I know some people feel the 24 db slope in my Rane AC23 electronic x-over makes the sound seem "squashed" (is that the right word???) but the time alignment feature is really helpfull. Are you familiar with Rane gear?

Nice horns. Let me know if you're wife ever convinces you to sell them. I've been watching Ebay and there has been a few sets of them up there lately. Some guy in Florida has a bunch of them.

c-ya

Guy

alskinner
03-12-2005, 08:18 AM
Thanks for the reply,

I did win these on e-bay from ABrown in Pensacola, Florida. Fortunately he lives about 100 miles from me or I don't know if I would of bought them. They looked a little dingy but when I go them home and cleaned them up they looked great. The drivers were also in great shape. He has 2 pairs up for auction right now. I don't know what the shipping would run due to their size.

As far as the sound against the 2397/2441. They are two completely different sounds. 2441 has a smooth sound but is rather pinpointed far as listening posistion. If you move from a seated position to a standing position you can tell a difference in the sound because of the horizontal dispersion.

On the other hand the 2360H/2445 the sound is also very smooth. I was at first concerned with the sound of titanium diaphrams due to the expression by some of their harshness. On the 2360H they are extremely smooth and because of their dispersion characteristics will fill the room without a paticular sweet spot in the mid-range. When I close my eyes I cannot tell where the speaker is as it appears to be coming from everywhere, yet it is still coherent with the instuments and vocals seeming to come from different places.

Of course these are first impressions and I have them very crudely set up at the moment, but if I had the real estate to put these horns in I would certainly give them strong consideration. One thing I have pondered is hanging them from the ceiling in order to get the physical placement right without taking up a lot of floor space.

As for Rane, I am familiar with their products and personally don't care for the sound. I know their are a lot of people that do and I respect that. I use an Ashly 4001 4 way crossover and absolutely love it. That's why I am trying to see if there is an analog time-delay that will work for the midrange only.

Earl K
03-12-2005, 08:50 AM
Al

Re; Time Alignment

Just use a physical alignment topology, ( ie start by aligning the voice coils of the horn driver with the woofer below it ).

Horns like this are not going to get "flush-mounted" into a box ( to look nice & vacilitate portability in the SR business ) . Since they exist "outside of the box" you have the flexibility of shifting it around in depth until you obtain the best "summing" at the crossover point .

A simple tone generator ( sine wave ) & Radio Shack SPL meter are all that is needed to accomplish this alignment task .

I think that there's a paper on the procedure , over at Todd Whites' Unofficial Altec site - buried somewhere in the archives.

:cheers:

alskinner
03-12-2005, 09:32 AM
I am about to come to the same conclusion. Physically aligning the drivers until the SPL is the same is a good basic approach and I know that will work. I am working on getting an RTA, O'scope and SPL meter as far as signal generator I have one based on software that has worked well for me aligning Compression driver diaphrams. Since aestetics are not as much an issue with me than the sound I want, I don't mind putting the horns any where they need to be. I do have the driver on the Horn almost lined up with the bass driver on my 4345 Hybrids but as I said it is going to stick out anywhere from 26 to 30 inches from the rest of the drivers, this will probably not be as much an issue with the A7 cabs as the driver magnet on the 416 woofer is already about 27" back from the mouth of the bass horn opening. Again thanks for the input. I still have a lot of homework to do before I can draw some final conclusions and I do need the equipment where I can somewhat validate what my ears are telling me.

Regards
AL

Mr. Widget
03-12-2005, 10:28 AM
I am working on getting an RTA, O'scope and SPL meter as far as signal generator I have one based on software that has worked well for me aligning Compression driver diaphrams.

I don't know what you were planning of paying for all of that, but the Clio Lite version will do everything you need and more and is $700. This includes a calibrated mic... not just a test mic that is sorta calibrated like the Behringer, Rane etc. but a mic that has calibration data files for it.

I really haven't scoured the internet, but it is possible that Soundeasy or one of the many other even lower cost options might do what you need. For time alignment using Clio you can shoot an impulse response of each driver and then physically move them or digitally move them until they are all at the same position in time. Here we see that this driver is 3.6 milliseconds from the mic.

Widget

Earl K
03-12-2005, 10:30 AM
Al

- I'd recommend you buy an old Behringer CX 2300 crossover . Pay as little on eBay as possible.

- Just use the LF output on it to drive the woofers. This is a "tunable" crossover.

- The analog delay section on the woofer section allows you to electronically shift the woofers back approx. 23 in. So, that's your maximum allowed physical horizontal spread between the voice coils. Also, make sure the Horn Driver is behind the woofer for this woofer delay to work.

- Drive your horn amp with something else that is more transparent sonically.

- I use a Passive Line Level HiPass ( 2 caps per balanced XLR input on an amp ). This gives a 6 db slope.

- Get more slope ( say 6 or 12 db more ) by using some passive parts after the amp - before the horn driver.

- Since you really want to protect any compression driver from DC ( offsets from the amp ) - this caps' inclusion represents a 6 db "opportunity" for a crossover. Pick a value that is actually working as part of the crossover. Choose a value that is small enough so its' Fc point that is dead-on to your desired crossover area.

- Follow this DC blocking cap/ crossover by creating an ad hoc High Frequency Lift section/CD Horn EQ that really just consists of an 8 ohm resistor ( in series to a 8 ohm JBL driver ). A cap ( paralleled to that resistor ) allows High Frequencies to "bipass" the resistor . At high frequencies ,this caps uf value is "in series with the previous cap" so you need to adjust its size ( upwards ) to account for this fact.

- Follow this "EQ" setup with an approprite sized coil and you've got a CD compensated 3 pole crossover ( that is as "transparent sounding as your amp is ). Keep the coil "after" the EQ section . Coil size in this position is minimized by it's proximity to the low impedance of the driver ( or being situated after the impedance of the resistor ). The coil is quite effectively isolated from the DC blocking cap by the resistor - allowing greater freedom of choice in value selection ). Translation ; coil size that is based on a low impedance is smaller than a coil working into a higher impedance ( for the same crossover frequency ) The smaller coil size saves money .

- With all the electrical "spreads" occuring between the 3 poles of this sort of crossover topology - you're now essentially listening to a 3 pole Bessel arrangement. The maintenance of the Bessels' superior transient response is worth the (minimal) effort of deviating from a more classical arrangement.

- Do all this and I think you'll soon forget the inclusion of that lowly "Behringer" part into your system .

Oh , buy the older CX 2300, not the 2310 or 3400. The newer Behringer stuff is "Surface Mount, Wave Soldered" stuff that is all quite goofy on the inside and virtually impossible to fix since SMD stuff is ,"just so teeny-tiny" .


:cheers: ( a few thoughts )

alskinner
03-12-2005, 11:27 AM
Thanks Mr. Widget

It will take me a little while to digest your points as old age, but what you all are saying makes sense. As far as the test equipment I think I will look into CLIO Lite since the only need I have for the equipment is strictly home audio. $700.00 does seem like a good deal for what I need and a decent RTA is going to be $500.00 or more as a standalone. I was concerned about the mike being calibrated for the unit I bought but since CLIO Lite comes with a calibrated mike that I am sure is calibrated with the software and hardware that should alleviate that concern.

Also thanks to Earl K. I understand what you're saying but will have to scratch my head and butt to put it all together.

To me JBL drivers are like a great blank canvas, great quality drivers for the most part that can either become a masterpiece through using the right electronics, room acoustics and signal processing or they can be just great initial driver that squak and honk

Again guys thanks for the help, I have been out of the electronics arena for some years and it's gonna take me some time to get back up to speed. Shared knowledge and oppions is to me what makes this such a great site.

AL

Ian Mackenzie
03-12-2005, 12:46 PM
Al,

Cool idea.

Your original idea to used within a A7 enclosure makes sense or perhaps look at an Edgar mid or midbass horn. Steve Schell has had great success with horns, may be worth sending him a pm.

Ian

djrobertc
03-12-2005, 03:10 PM
To me JBL drivers are like a great blank canvas, great quality drivers for the most part that can either become a masterpiece through using the right electronics, room acoustics and signal processing or they can be just great initial driver that squak and honk

Wow - thats an excellent analogy!!

:applaud:

Al Klappenberge
03-12-2005, 03:22 PM
Guys,

I have done some research into the time alignment thing and I say it's a red herring! All time alignment buys you is a smooth crossover area. Your brain can not detect it uness it is so far out that a single sound can be heard twice. What is a better fix is what I am doing for the Klipsch community. That is extreme-slope crossover networks that keep the drivers from interacting by seeing to it that only one is playing any give sound. This stops comb effects as well. Experiments with 3 seperate sine wave singnal generators generating the fundamental, 3rd and 5th harmonics of a simulated square wave shows that you can not hear the phase relationships at all! You can't even hear the phase changes while watching the waveform on an oscilloscope for clues! It sounds like a continuous raspy tone. That would be the effect of drivers being out of time except that the phase doesn't continuopusly change. It would just be fixed at some wrong phase. In short, don't bother with it!

You will find a paper I wrote on the subject on my web site http://www.alkeng.com/klipsch.html look for the download dump-truck.

Al K.

alskinner
03-12-2005, 04:04 PM
Thanks for the information. I could see the benefits of the design, but I wonder if the transition is still smooth between drivers at a 120 decibal slope/Octive? Also does this eliminate the physical alignment of the drivers for the most part. in other words if the midrange horn was say 10 inches back or forward from the Woofer, the horn that I am trying to inigrate is 35" from the driver to the mouth of the horn. Am I correct in that it wouldn't be an issue with this type crossover. I typically use active filters that eliminate a lot of the phase shift but am willing to try a good passive design. As I said in my earlier posts I am catching back up on my electronic knowledge and appreciate all the help on this issue.

Regards
AL

Al Klappenberge
03-12-2005, 05:18 PM
Al,

Smoothness through the crossover is strictly a matter of phasing. Since your
ear is deaf to it, all that matters is that the two drivers do not interact
with each other. If they are exactly out of phase, you will get a dip. If
they are exactly in phase, they will add slightly or sum flat of both sources
are 6 dB down. If they are some odd phase angle apart the transition will be
smooth having far less frequency response error then the speaker drivers
themselves. If they cause a dip, chances are all you need to do is invert the
polarity of one driver and everything is fine. The advantage of the extreme-
slope is that the frequency band where they can interfere with each other is
reduced in a big way. The big advantage though is the improvement in stereo
image you get when only one driver is playing at a time. The classic "sweet
spot" virtually goes away! Another point to realize is that the phase
nonlinearity introduced by even a 120 dB / octave filter at the crossover is
minuscule compared to the phase shift introduced by the time delay from the
driver to your ear, which is a constant phase shift / Hz (flat group delay).
The phase shift from the networks is totally overwhelmed by the propagation
delay.

The plots below show the point. The top curves are the "transfer function" of
the propagation from amp input to test mike output. The top plot is phase and
the bottom is frequency response. The black and yellow phase curves are with
the mike in different locations just a few inches apart (the amplitude plot
reference was moved intentionally). The bottom set of curves is the transfer
function of an 8500 Hz extreme-slope network alone. The speaker was using and
identical network at 7500 Hz (where the red marker is). Note how the speaker
phase looks like a sawtooth going +-180 degrees many many times more than the
network alone!

Note too that there is another extreme-slope network in the speaker at 700
Hz. If you look closely, you can see a tightening of the "saw teeth" below
700 Hz. This is the additional propagation time through the horn-loaded
woofer of the Belle Klipsch.

Al K.

alskinner
03-12-2005, 06:26 PM
Thanks Al

That clears up a lot being able to see the graphs. With the A7 bass cab I am shooting for crossing it over at 800 hz and 9-10K for the 2405 tweeter. I would like to try the crossover in those ranges. I know that you are probably extremely busy but would appreciate any help you could give me computing the valuse when I finally get ready to put the system together.

Thanks
AL

Al Klappenberge
03-12-2005, 06:52 PM
Al,

I will be glad to help. I want to get into the ALtec stuff some more. All I do these days is Klipsch!

Al K.

Al Klappenberge
03-13-2005, 04:24 AM
Alskinner,

I am going to modify what I said earlier. After some private message discussion with Mr Widget, I got to thinking. The time alignment is necessary if more than one driver is making a sound at the same time. Unless you have extreme-slope networks to see to it that only one at a time is sounding, you will need to time align with such a long horn! My thinking is getting biased by the extreme-slope idea and I forget about the vast majority of systems using networks with only 6, 12 or 18 dB / ocate slopes!

Al K.

alskinner
03-13-2005, 05:44 AM
I am grateful for all input into the system I am proposing to put together. It becomes a lot more difficult to put drivers together in a configuration that has probably not existed before. The good thing is there is a good bit of information from JBL on the 2445J/2360 combination and the Altec A7 bass horn. Getting them to play together nicely takes a lot (at least on my part) of research and planning. Having said that right now I am not sure if I will keep the 416 drivers in the A7 or opt for one of the JBL woofers such as a 2225 which will bring about another variable. The time alignment issue just popped in my mind as one of the first things to work out because of the midrange horn and the other drivers.

It was somewhat easier to modify the 4345 using the 2397/2441 based on the known cabinet dimensions and the excellent posts on building the 4345 from Ian Mackenzie and others. Also using a good quaility electronic crossover like the ASHLY 4001 helped eliminate the construction of a passive crossover. Even with that it took a lot of amp juggling to get in place the best amp/driver combination out of what I could afford and already had.

Al
I am glad to see people like you working with the classic Klipsch designs. I have heard some great Kliipsch speakers in my time. I think a lot of Klipsch bashing comes from people not being able to set them up properly and some of the design characteristics that need updating. One of my fondest memories of Paul Klipsch is a picture of him creating some of the first designs out of plywood in the garage of his home. Don't have the picture now, but I remember saying to myself, this guy has got the guts to follow what he believes in.

As far as the Altec A7

I great theater speaker in it's day. However in a home setting most would agree the sound of the upper end is anything but HiFi. I know several people that have attempted to modify the existing system through crossovers and equalization and a lot of them ended up throwing their hands up. Even though the Bass Cab is only good down to 40-50Hz, it has unique sound that is hard to replicate in anything but a bass horn.

Thanks for letting me rant on. One thing I do know is a need to get some test equipment together to be more accurate in the questions that I am asking
:blah:
Again thanks to all of you.

Al Klappenberge
03-13-2005, 06:34 AM
Al,

I have been around the L-C filter design world for quite a few years and I
have never seen a passive filter that will give time delay on the order
necessary to time align even to a very small degree. I can easily design
group delay equalizers (all-pass filters) that can yield a big group delay
peak someplace, but that is useless! Group delay and time (propagation) delay
are two different things. All-pass networks will NOT do it! To really time
align you need a digital delay device and multiple amps. All this costs money
and you still can't get two drivers in alignment except at one spot in the
room. This cuts down the multipath "smearing" but can't do much for the
"lobeing". An extreme-slope passive network, though expensive because of its
parts count, will make the time alignment problem a moot point and will allow
you to do away with the active crossover, delay circuits and multiple amps.
You would need a calculator to add up all the equipment costs to figure which way would cost less!

I would like to do a set of extreme-slope networks that could be used in a
classic Altec speaker, like the A7, just to illustrate the point. I would be
happy to do the design just for the fun of it if I can get together with
someone who has a set he wants to modify. If your project is using drivers
that are common to a lot of the Altec speakers, it might be a good candidate.
The most important consideration is the woofer driver. I like to design the
woofer / squawker crossover to incorporate the woofer voice coil inductance
right into the woofer filter. That makes the network woofer-specific to some
degree.

Al K.

alskinner
03-13-2005, 09:41 AM
Found a pretty good site for the layperson or those just beginning in electroacoustical theory.

http://www.lenardaudio.com/education/06_x-over.html

Although not real technical it does give a pretty good explanation of basic crossover theory and time alignment. as pointed out on the site one of the original uses for time alignment was for auditoriums where alignment would have an effect on the overall sound.

Thanks
AL

Al Klappenberge
03-13-2005, 10:34 AM
Alskinner,

I looked at your link briefly. Needless to say, I don't agree with their warnings against complexity. This is what I call "conventional wisdom". I believe it's derived from years of experinece by people who know speakers but have had no experience with complex filters. I believe Paul W. Klipsch is an example. A bad complex filter will yield bad results. This conventional wisdom comes from the days of "image parameter", "constant-K" and "m-derived" techniques. You can't make a good high order filter that way! I want to prove my point by actually building extreme-slope crossover netwroks and have people listen to them. So far every set I have built listeners have confirmed what I have been saying. Here's an example: http://forums.klipsch.com/idealbb/view.asp?topicID=62518&sessionID={6661738C-2BFB-4F23-AC94-622FBEB3365F}

AL K.

alskinner
03-13-2005, 11:46 AM
Al

I agree with you on the part about the site being filled with "convential techniques" as I meant it as a link for just basic understanding. After all the creator of the site is also a speaker system designer and I am sure some of his bias is because of the products he designs and sells. People like yourself, Mr Widget, Earl K, Giskard and Ian Mackezie and a host of others on the forum are way beyond the basic theories , just thought it would be a good starting point with people with no or very little experiance. Thinking outside the norm is what makes for new innovation and techniques. After all crossovers wether passive or active qre one of the main things make or break a great speaker system. Complexity shouldn't be an issue if the results desired are reached. I can't remember the guys name but he built a lot of the crossovers for places like Studio 54 and people will pay thousands of dollars for a set of them, comming as far away as Japan. I am sure he was another guy that went against the theories of the day


Cheers
AL

Mr. Widget
03-13-2005, 11:58 AM
I can't remember the guys name but he built a lot of the crossovers for places like Studio 54 and people will pay thousands of dollars for a set of them, comming as far away as Japan. I am sure he was another guy that went against the theories of the day

That would be Ed Long. He definitely broke new ground. His ideas about time alignment while ground breaking were more conventional than his "Live End Dead End" studio control room designs.

I agree with Al Klappenberge, that the use of extreme slope filters does lessen the need for time alignment significantly, but with some of the new technologies becoming available we will be able to have our cake and eat it too.

Widget

Ian Mackenzie
03-13-2005, 12:00 PM
Wasn't it one of the Sheffield Lab owners that offer'd new crossovers for the 604-8G Altec Duplex?

Ian

Mr. Widget
03-13-2005, 12:02 PM
Doug Sax, founder of Sheffield, started the ball rolling.

sfogg
03-13-2005, 12:10 PM
"I agree with Al Klappenberge, that the use of extreme slope filters does lessen the need for time alignment significantly, but with some of the new technologies becoming available we will be able to have our cake and eat it too. "

Sure, the DEQX box would let someone cross over as steep as 300 db/octave AND allow time alignment between drivers along with other processing but that is a $3k box and of course you need more amps with that too. And for those of us with three speakers up front we would need two of them. Gets very expensive.

I'd love to use the DEQX box but Al's crossovers worked out very well in my setup using an Altec 902/511B on a LaScala bass horn.

Shawn

Mr. Widget
03-13-2005, 12:47 PM
....but Al's crossovers worked out very well in my setup using an Altec 902/511B on a LaScala bass horn.


Hard to argue with that.:)

I intentionally didn't mention DEQX in my post because I expect there will be many other products in the coming years that will also take care of these issues. In the mean time, going passively is certainly one solution.

Widget

Al Klappenberge
03-13-2005, 12:50 PM
Guys,

I never thought about speaker system problems for applications like nite clubs and such. I would have to assume there is a long list of "gotchas" involved there! Shawn is one of my first users of extreme-slope networks. He seems to be happy with them.

I suppose if you had a way to time align along with extreme slope crossovers you would be having the best of everyhting. There is a small band of frequencies right at the crossover where adjacent drivers can still interfere with each other and aligning them might help with that. It's a small problem compared to the overlap you get with conventional crossovers though.

The link Al gave is a good overview of the job a crossover does and following its advise certainly wouldn't get you into any trouble. I guess I'm just a rebel!

Al K.

sfogg
03-13-2005, 12:57 PM
I am sure you are correct that there will be more options as time goes by.

This could do it too:

http://dias.umist.ac.uk/PAG/signalwizard.htm

Though you would one board per channel. Probably will be more options when/if Tact or Behringer implement much higher slopes into some of their products too. A very high slope crossover option in the DCX2496 would be fantastic IMO.

Shawn

sfogg
03-13-2005, 01:03 PM
Al,

"Shawn is one of my first users of extreme-slope netwroks. He seems to be happy with them. "

Yup, I have been very happy with them. They are what really made the speakers sing.

Now I just wish Lexicon would implement very high slope crossovers in my pre-pro to cross from the mains to my Sub1500 subwoofers. Fourth order is just so slow. :)

Shawn

Al Klappenberge
03-13-2005, 01:18 PM
Shawn,

The idea of an extreme-slope active crossover for subwoofers has crossed my mind. The other filter software I sell (Dr George Szentirmai's S/FILSYN) will design pole-placed or elliptic active filters that could be used for a crossover at subwoofer frequencies. It is someting I may consider doing someday. I just hate getting into PC boards! I am getting stuck with that now anyhow though. The switched tweeter L-pads I am putting in my ES5800 squawker / tweeter networks is becoming a pain to wire and the switch I have sellected is not what I had hoped. Good switches cost in the range of $25 each! I am working on replacing it with a PC board having a movable plug to replace the switch and a small transformer to replace the resistors. A PC board for an active extreme-slope subwoofer crossover would be a MAJOR undertaking! All I can say is: SOMEDAY!

Al K.

boputnam
03-13-2005, 02:04 PM
...I got to thinking. The time alignment is necessary if more than one driver is making a sound at the same time. This is a most interesting discussion...

Per the above, I cannot think of a time when that is NOT occurring. Even a single note will have harmonics octaves above the primary tone, and those harmonics will almost certainly emanate from an other element in a multi-element cabinet.

sfogg
03-13-2005, 02:32 PM
Al,

Thats cool that it can be done actively too.

The sub crossovers in my pre-pro are done via DSP. So literally for Lexicon to implement a very high order crossover would be for them to write the code to do that assuming they have the DSP cycles and such needed.

Shawn

sfogg
03-13-2005, 02:51 PM
"Per the above, I cannot think of a time when that is NOT occurring."

I think Al meant when two drivers were reproducing the same frequency sound. With Al's crossovers the harmonics would still be nearly entirely reproduced by one of the drivers unless the sound was right at the crossover point.

My ES crossover cross at about 600hz. There is only about a 50hz band where the drivers aren't more then 10dB down relative to each other.

Shawn

Al Klappenberge
03-13-2005, 03:01 PM
Bo,

Shawn is right. The harmonics are simply the componenets of a complex waveform. The crossover will simply direct them to the appropriate driver and ONLY that driver.

Al K.

Mr. Widget
03-13-2005, 03:30 PM
I think we all agree that having all drivers reproducing part of a note (a note being a group of frequencies made up of a tone and it's harmonics) will be best if time aligned. Just how audible this may be is certainly up for debate.

Here is an excellent example of what Al is talking about. It is from recent work we have been doing with Project May. In this example we are changing the configuration below the crossover point. The crossover is a low order passive design at 800Hz. As you can see changes in the woofer configuration cause changes in the frequency response as high as 3200Hz! The significance of this is that for every frequency on either side of 800Hz (for at least 2 octaves on either side of the crossover frequency) both drivers are reproducing the same frequencies.


Blue is MTM.
Red upper woofer disconnected and replaced with resistor.
Green lower woofer disconnected and replaced with resistor.

sfogg
03-13-2005, 03:47 PM
"The significance of this is that for every frequency on either side of 800Hz (for at least 2 octaves on either side of the crossover frequency) both drivers are reproducing the same frequencies."

Another test a person can do to hear this is to play pink noise on their system.

Then standing fairly close to a vertically aligned speaker stand up and squat down. As your position changes you will hear the interaction between the drivers as changes in how the pink noise sounds.

When you do the same thing with Al's ES network the sound hardly changes at all until you get out of the dispersion of the horns.

This is signifigant as the sound the listener hears at their normal listening position is a composite of the on axis and off axis sound from reflections within the room.

Shawn

alskinner
03-13-2005, 04:04 PM
Thanks Guys

I think I finally understand what Al is saying, although there is some interaction between drivers the 120db/octave keeps the interaction between the same frequencies going to different drivers are kept to a minimum resulting in a a lot more coherent system without the smearing and phase shifts associated with lower order crossovers. Am I close?

Also it is certainly nice to have a thread where differing viewpoints can be discussed with civil discourse.

Cheers
AL

Mr. Widget
03-13-2005, 04:09 PM
I think I finally understand what Al is saying, although there is some interaction between drivers the 120db/octave keeps the interaction between the same frequencies going to different drivers are kept to a minimum resulting in a a lot more coherent system without the smearing and phase shifts associated with lower order crossovers. Am I close?

By George, I think he's got it.:)


... nice to have a thread where differing viewpoints can be discussed with civil discourse.

Don't push it. :applaud:

Al Klappenberge
03-13-2005, 05:04 PM
YEP! He's got it! :applaud:

Al K

Al Klappenberge
03-13-2005, 05:23 PM
Guys,

Here's an accoustic plot of the actual network that Shawn is using (600 Hz) being tested in my speaker. It's a Belle Klipsch woofer, JBL 2426h driving an Altec 811b horn. The two plots show the driver wired in and out of phase. My "ES" networks acts just like 2nd order Butterworth networks in that the outputs of the two channels are exactly 180 degrees out of phase at the crossover.

Al K.

Tim Rinkerman
03-13-2005, 05:24 PM
Man, I was really getting into this thread,thinking I was really learning something out of the box, and indeed I agree with all that was said...and then I went to view your home system,Al,....and please tell me that those Rane EQ's are a mis-print,or you modified(gutted) them to not make that "RAIN" sound when you insert them into the signal path!

Al Klappenberge
03-13-2005, 05:31 PM
Tim,

It seems that equalizers are like cats. You either love 'em or you hate 'em! Believe me, I use them very sparingly! They are both set exactly alike and are only being used to remove a room resonance that was making the system sound boomy. They are set to mostly cut and only to boost a db or two when all else fails! I think I have two of the five sections bypassed. I learned the hard way what an incorrectly adjusted equalizer can do to you. They can drive you up the wall ! :blink:

Al K.

scott fitlin
03-13-2005, 05:45 PM
That would be Ed Long. He definitely broke new ground. His ideas about time alignment while ground breaking were more conventional than his "Live End Dead End" studio control room designs.

I agree with Al Klappenberge, that the use of extreme slope filters does lessen the need for time alignment significantly, but with some of the new technologies becoming available we will be able to have our cake and eat it too.

WidgetThe crossovers that were used at Studio 54 were built and designed by Alan Fierstein for Richard Long in the 70,s! They did state in their brochures they had implemented " linear phase " technology in their crossovers and that it was the first time this technology had ever been used for disco playback sound systems!

I have and use these crossovers, and I can tell you his units have the clearest imaging I have heard on just about anything!

They had indeed broke new ground and succesfully outdid everytrhing else available at the time, and it still outdoes many things more modern, made today!

Of course, I might be willing to try the DEQX based on what some I respect have been telling me of the unit, and its linear phase filters!

Al Klappenberge
03-13-2005, 06:01 PM
Scott,

As soon as somebody starts talking linear phase a big red flag goes up in my head! Linear phase is very easy to do in an all-pole lowpass, somewhat difficult to do in a narrow bandwidth bandpass, VERY difficult in a wide bandwidth bandpass and totally impossible to do in a highpass filter! Many desingers simply "invert" and scale out a linear phase (Bessel) lowpass into a highpass. This is what the text book Bessel crossover are. The results retain the very slopy attenuation of the lowpass but none of it linear phase properties. They are a waste of parts! The easiest wideband linear phase bandpass is called a "Blinchnicoff / Zverev" filter. I have the software to synthesize it singly terminated (which you need for a crossover) but I really doubt may other people have it! Besides, like I said, the human ear is deaf to phase relationships of components of a complex wavefore anyhow. Like I showed earlier, the linearity of the netwrok is overwhelmed by the "excess phase" contributed by the propagation delay. So I wouldn't believe any of the claims unless they were proven to me!

Al K.

Tim Rinkerman
03-13-2005, 06:01 PM
OK,Al..I'll let you off this time:) I was a live sound guy for a long,long time and any time I came into a place that had Rane equipment in the rack,I knew I was in for a long day....I guess I mostly ran into their 1/3 octaves and eek!ametric paralizers, but when I pushed the insert button on any of them,even with no boost or cut at unity gain,it was like somebody leaked in a pink noise generator.

And I do understand what you're saying about x-over points and slopes and couldn't agree more!

Mr. Widget
03-13-2005, 11:25 PM
AL K.,

I am sorry but I have to agree with Tim on this one. Take your Ranes and set all of the faders to flat, and compare Rane flat and bypassed... those damn Ranes really muck up the sound. I've read their catalogs and sales hype. They say all the right things, but in practice they just plain sound bad. I have a pile of their gear in my HT rig... it was all booted out of my music system. Someday when I have a few extra bucks it will go the eBay way and be replaced.

Widget

Mr. Widget
03-13-2005, 11:33 PM
The crossovers that were used at Studio 54 were built and designed by Alan Fierstein for Richard Long in the 70,s! They did state in their brochures they had implemented " linear phase " technology in their crossovers and that it was the first time this technology had ever been used for disco playback sound systems!

Did I get my Longs mixed up? Oops!

Ed Long was the time alignment guy who ultimately got the 604s to work out for Urei, along with his LEDE studios. I remember hearing the Urei 813s in one of his studios and it was pretty damn good... not a Westlake though.

Widget

Al Klappenberge
03-14-2005, 05:04 AM
Guys,

For some stupid reason I am unable to see page 3 of this thread. It acts like the server is down! 1,2 and 4 ar ok. Finger that out!

As to the Rane equalizer, I have tedsted it in and out set flat and the difference is very small to my hearing. I find taking the boom out of my listening room is an acceptable trade off! I woulder if you have listend to the parameteric equalizer? I don't have much resptect for the 1/3 octave veriety.

Al K.

Al Klappenberge
03-14-2005, 05:10 AM
Guys,


NEAT.. The post I made above cleared the problem with page 3 of the thread. Computers! go figure!

I suspect that the problem with the Rane stuff was the setup. The more equipment you have in a chain the more chance that something is not set right. Even a simple thing like gain distribution can kill you! That's another good argument for a passive crossover rather then muit-amping with active crossovers. Passive stuff is simple and the entire thing is in one place and under the complete control of the designer. Nobody else can change anyhting but maybe to add attenuation to each driver if needed. It's idiot proof!

There's another factor that could be at work here too. Rane equalizers use op-amps in their filters. Early op-amps had slow slew-rates which added a dullness to the sound. New, faster ones solved that problem. You may have been testing an eq with the old slower opamps.

Al K.

Robh3606
03-14-2005, 05:46 AM
'That's another good argument for a passive crossover rather then muit-amping with active crossovers. Passive stuff is simple and the entire thing is in one place and under the complete control of the designer. Nobody else can change anyhting but maybe to add attenuation to each driver if needed. It's idiot proof!'

Hello Al

I disagree on that one. I think it's very difficult to do a passive crossover and get it right. Seperates the men from the boys. I still have my training wheels on with the passives;) . Actives are comparatively very easy to set-up. Your right about the gain especially with the noise floor, have to get it all right. But once you get the mix worked out you can shift your frequencies on the fly and it's much easier changing out to a different driver. Also like that snap in the transient response you get with no passives in line. I quess it's all about trade offs. I would rather have the flexabillity with an all active set-up which once set-up basically stays there asside from either a crossover frequency shift or a driver gain setting.

As far as time alignment well as long as it is below the Blauert an Laws Criteria for group delay is it really a problem?? I look at time alignment and phase as the same animal. I don't worry about it too much. Just try to use common sense and not get into an area where interactions are destructive.

Rob:)

Al Klappenberge
03-14-2005, 08:31 AM
Robh---,

I agree that it is difficult to make a good extreme-slope crossover. It does seperate the men form the boys. Once it is done properly though, it's done! The customer has no more control and it will be right no matter what he does. I have been doing multiplexers like this at frequencys up to 2000 MAGAHertz (microwave) with as many as 5 contiguous ports for many years. I actually sell the software to design them to filter companies. I am 99% sure they are right. In any event, making them right is the desingers problem. Once it's done, the users is off the hook. He can't screw it up unless he starts mdifying the filters somehow.

As to the Somebody & Laws group dealy standard (I forgot the name!), look carefully at the article. It is about Group delay from all-pass filters, not time delay! They are very different things and many people confuse the two. Group delay is a measure of phase linearity, and THAT'S ALL! Your ear is deaf to phase! They also say that the effects can only be heard though headphones by people trained as to what to listen for! You can't hear it through a speaker palying music. It is really an over rated thing. Nevertheless, I have designed the Klipschorn woofer filter so that the group delayy peak is under the limits give in that article. You do it by limiting the filter stopband slope.

AL K.

Al Klappenberge
03-14-2005, 08:46 AM
Guys,

Here's the computed performance of the 400 Hz filter for the Klipschorn. I take it back! The group dealy peak right at the crossover is a bit over the limit if you extend the B&L curves. I don't think that would be a significant flaw even if you could hear it. I don't beleive you can though.

Al K.

Robh3606
03-14-2005, 09:01 AM
Hello Al

"I agree that it is difficult to make a good extreme-slope crossover."

I meant all not just the extreme but also the 12dB networks commonly used.


"As to the Somebody & Laws group dealy standard (I forgot the name!), look carefully at the article. It is about Group delay from all-pass filters, not time delay! They are very different things and many people confuse the two. Group delay is a measure of phase linearity, and THAT'S ALL! Your ear is deaf to phase!"

I have seen a couple of references that lump the entire system delay from the crossover and driver offset such as horn length into that number. If I can I will post the reference.

Rob:)

Robh3606
03-14-2005, 09:04 AM
LOL :applaud:


Guess we have read the same stuff.

Rob:)

Al Klappenberge
03-14-2005, 09:28 AM
Rob,

Yep.. That B&L article seems to be an industry wide standard. It never seases to amuse me though how widespread the confusion is between true time dealy or propagation delay and group delay. So may very smart and educated people seem to think they are the same thing. I have even read papers by Paul Klipsch, who was without question a very smart guy, which seem to use the two terms interchangably.

Al K.

boputnam
03-14-2005, 11:15 AM
I think we all agree that having all drivers reproducing part of a note (a note being a group of frequencies made up of a tone and it's harmonics) will be best if time aligned. That was my point. Now, back to lurking, and learning... :yes:

Al Klappenberge
03-14-2005, 11:48 AM
Guys,

Common sense says that all components of a waveforem reconstruct that waveform properly when each component is in the proper phase, BUT: an oscillocope can add up all those compoents accurately to reproduce that waveform. I do not believe the human BRAIN is capable of that. Remeber that the brain's main concern through history was just to tell the person if what he hears is food or if he is to become food for what he hears! Details like phase just isn't important! All I know for sure is that I can't hear it. I have actually done experiments with multiple signal generators to test this.

Al K.

boputnam
03-14-2005, 12:09 PM
...tell the person if what he hears is food Sounds like Hearing-Gustation synaesthesia to me, a troublingly interesting affliction...

But I'm WAY :offtopic:

:slink:

Al Klappenberge
03-14-2005, 12:48 PM
I think that means one who tastes what he should hear. I have heard of that, believe it or not. It actually a real condition! Maybe there's some cosmic significance there! :biting:

Al K.

Mr. Widget
03-14-2005, 01:39 PM
"Details like phase just isn't important!"

Little sonic cues like phase help blind people navigate their sightless world. Time shift, phase shift, doppler shift, all of these and many other acoustic subtleties are used by the brain to construct an aural image. I think we all hear them and we choose to pay attention or not.

Look, we can listen to a boombox and know that it is trying to reproduce a kettle drum. We also know that it isn't an actual kettle drum. As you get more and more of the pieces right the illusion becomes better and better. Frequency response is the first step, time and phase along with lowering distortion etc. are refinements.

Widget

pangea
03-14-2005, 02:15 PM
I've read this thread with great interest, but I must admit, ... a lot of it I didn't quite understand.

So please, could anyone explain to me (in simple terms), if what has come to light in this great thread, will change or add anything in regard to my MTM hybrids, which use 3'rd order B/W and overlapping frequencies on the M-B from 100Hz to 897Hz through my Behringer UltraDrives?
http://audioheritage.org/vbulletin/showthread.php?t=4157


BR
Roland

Al Klappenberge
03-14-2005, 02:18 PM
Widget,

Yes, but we are talking about the phase between two different things! You are comparing apples to oranges! The phase relationships between harmonics is what is involved in a single loudspeakers reproducing a complex waveform. The mechanics of stereo image, or location of a sound is the phase and amplitude relationship between to sensors (your ears) listening to same sound! Now your brain has a reference to make a comparison. It can't do that when the two things being compared are two, three or more times each other in frequency! This is why loudspeakers should be identical. In fact, I make my networks in matched set osf 2 or three using matched pairs of capacitors! My extreme-slope networks have typically been within 2 degrees of each other in insertion phase.

AL K.

Al Klappenberge
03-14-2005, 02:23 PM
Pangea,

My simple answer to your questions is a resounding "I have no idea!". Could you explain more about what they are. I'm new here and all I know is the Klipsch world lingo!

:blink:
Al K.

Mr. Widget
03-14-2005, 02:34 PM
"Yes, but we are talking about the phase between two different things!"

I would submit this is where the "it is open for debate" part comes into play. I have no opinion here. There are those who would argue bitterly that they can hear the phase anomalies created by multi-way systems and others, I believe you are firmly in this camp, saying that humans cannot perceive the phase errors typical of multi-way speakers.

I do not know. It is hard to verify either side of the debate. Maybe some humans can perceive these anomalies and others can not. I do know that there are things I can measure that I can not hear, and conversely things I can not measure that I do hear.

Widget

Mr. Widget
03-14-2005, 02:38 PM
So please, could anyone explain to me (in simple terms), if what has come to light in this great thread, will change or add anything in regard to my MTM hybrids...

There is no easy answer. In all likelihood steeper slopes would make your speakers sound better.

Widget

pangea
03-14-2005, 02:53 PM
Pangea,

My simple answer to your questions is a resounding "I have no idea!". Could you explain more about what they are. I'm new here and all I know is the Klipsch world lingo!

:blink:
Al K.

I will gladly try to explain what I've built.

My question was more or less directed to Earl K., Mr. Widget and others who kindly have been sharing their know how on my thread, but of course I would be honored if you would like to add some of your obvious expertise on the subject.

I became intrigued by the D'Appolito MTM configuration, when in fact I had already set my mind on building a pair of Imperials, but then I suddenly knew instantly I had to build what you can see, in my avatar, as well as in my thread.
I had to take the stuff I already had, which was some JBL 2215H, 2123H, 2441J on a pair of Selenium horns and finally the 075 on top.

To make a long story short, they sound extremely homogenous, transparent 3-D and holographic, although the "sweet-spot" is quite narrow in the horizontal plane, but there it's very stable though.

I've been reading your posts with great interest and from what I understand, I can see that what you're saying, doesn't quite cover the MTM configuration, where two drivers are over-lapping each other to create the sonic "glue", which is needed in an MTM configuration.

BR
Roland

Ian Mackenzie
03-14-2005, 03:00 PM
Phase shift, time delay and multi point sound sources are all intermingled issues.

The Dunlavy web pages have great coverage of this. Don and I talked about this in another thread recently. The Vanderstein system is also a great example of time and phase corrected design.

The actual transient performance and step response function are probably equally if not more important to subjective taste that phase shift.

There have been a lot of studies on this with varied findings as the human ear/mind is perceptually more sensitive to some fequency bands than others regards audible perception of group delay (or phase shift).

Another variation of perceptual awareness is amplitude variations. The ear is more alert to overall balance and broad response variations than random small variations.This is why 1/3 octave smoothing is more relevant in actual measurements. (If steep or otherwise low order slopes produce the best overall response then so be it, and this will be dictated by the drivers at hand and overall layout.)

For example the ear is sensitive to amplitude variations below 150 hz and group delay, also in the 2 khertz region.

About MTM, they are an interesting genre of loudspeaker but the popularity for better imaging is over rated and best left for small bookself models where tiny drivers can be used to maximum effect

Some may consider this all theory, but its based on science and practical research.

My 2 cents worth:)

Ian

pangea
03-14-2005, 03:04 PM
There is no easy answer. In all likelihood steeper slopes would make your speakers sound better.

Widget

Hi!

You mean higher slopes on all x-over points and all drivers?

Right now I'm using 3'rd order B/W on all points. Tomorrow I'll try the 48dB B/W on all x-over points.
Now it's a bit late here.

BR
Roland

Mr. Widget
03-14-2005, 03:21 PM
To make a long story short, they sound extremely homogenous, transparent 3-D and holographic, although the "sweet-spot" is quite narrow in the horizontal plane, but there it's very stable though.

Then why do you want to mess them up!

Just enjoy them.

Widget

Mr. Widget
03-14-2005, 03:30 PM
The Dunlavy web pages have great coverage of this. Don and I talked about this in another thread recently. The Vanderstein system is also a great example of time and phase corrected design.

Two interesting examples. They both have taken on phase and time alignment as primary objectives and they both do remarkable jobs in the imaging arena. I have issues with the Vanderstein's dynamics and other areas, but in the imaging department they are really top drawer. The Dunlavys too are great speakers in terms of imaging. Actually they are pretty damn good all around. I believe that both of these designs use very low order networks with massive overlap between drivers.

I guess that goes to show there are many ways to get there. If you have excellent drivers that will work well together you can let them overlap. If you have drivers that get a bit complex out of their pass band or are using drivers with inconvenient acoustic centers, steeper slopes are a better solution.

Widget

Al Klappenberge
03-14-2005, 04:43 PM
Guys,

All of this stuff is open to debate! The undelying thing here is human perception. There is no balsted way to hook an oscilloscope up to the human brain to see what it's hearing through all the different little hairs inside the inner ear, all the nerves that run to the brain or how the brain "proceses" all of it! That makes it impossible to definitively say what is correct here and what is not! So I am not going to make and definite statement beyond what seems logical to me. I think this point has to be extended to the methods used to research the problem too. Very often conclusions are reached through experiments that are not valid for one reason or another. People also have a very GREAT tendency to hear what they want to hear. I am no exception to that either. It is not an easy subject! I think the min thing to consider is that very few people have heard extreme-slope crossovers completely and properly implemented into an audio system. This means most people can't talk about them except in theory as we are doing here. I only have seen three examples of them, but I suppose there are more. Mine, Joseph audio and one guy in a magazine article that scaled a set from normalized elliptic function lowpass tables. Those looked very similar to mine. Anyhow, to my ear, the difference when I installed the extreme-slope networks was far more obvious than any other single change I have ever made in my equipment.

Al K.

Mr. Widget
03-14-2005, 05:20 PM
I agree with everything you said!!!!

Two items I would like to point out. From what you have said your experience is primarily with Klipsch designs which does make your practical experience limited when thinking about all aspects of audio. I don't mean to discount the importance of what you are talking about or what you have accomplished, but only to point out that what may be the salvation for a horn based design originally created in the 50's-60's may not be the same for say an electrostatic panel or contemporary ribbon.

The other item, you have obviously invested quite a bit of time and effort designing these networks... it is hard not to become biased after that experience. For me to apply a 100-300dB/octave filter is a matter of typing in the number and even though it really was that easy I have come to a similar conclusion.

With those two thoughts in mind, I certainly agree with you that for systems such as yours, mine, and many others the extreme slope is a very real improvement. In the case of the Joseph Audio products... I was aware of the SEAS drivers that they are using long before I was aware of extreme slope networks. When I looked at those drivers I doubted I could get the most out of them as they are far less than ideal out of their passbands.. for anyone who has heard the Joseph Audio Pearls, those drivers are really remarkable when used with extreme slope networks!

Widget

alskinner
03-14-2005, 05:33 PM
Guys

I certainly appreciate all the feedback and thoughts on the subject. Although not as tecnically proficient as a lot of you all, this thread has opened up a world of possibilities. To quote Robert Heinlein in "Stanger in a Stange Land", "I am but an egg" with much to learn and experiment with. I am definitely interested in AL's extreme slope concept and as soon as I get the drivers together to use, I am going to ask for his help in constructing a pair. I still think I will have to play around with the physical alignment of the 2360 horn to get the basic timing in place.

As pointed out there are countless variables in reproducing complex music. Every thing from human physical condition ( Right now I have a terrible head cold and everthing sounds like it's coming from a tin can) to Room acoustics, space/time perception, electo/acoustical physics, pyscho acoustics, quality of drivers, electrical components, amplifers, preamps.

I think in the Wal Mart world of today the word "audiophile" has been diluted to a nonsensial word. I don't beleive it is something you can buy off the shelf without the passion to make audio the best it can be.

Like Widget said earlier there may come a day when we can have our cake and eat it to. Would be nice to have all audio problems whisked away with a magic box. On the other hand, sure would take a lot of the fun out of trial and error until we know a speaker system inside out.

There is something about the creation process that for me is extremely rewarding.

Anyways thank you all and as soon as the A7/2360 monstrosity starts coming together I will keep all posted

In the meantime its time for my medication

AL

Al Klappenberge
03-14-2005, 05:48 PM
Widget,

I agree that it is not a proven fact that the extreme-slope idea will apply equaly to horn systems and direct radiators, but the theory seem to point to that. What difference would it seem to make what is making two sounds, two horns, two direct radiators or a driver with a reflected wave from a wall generating a standing wave someplace in the room. The idea is the same. Consider the lobing from two drivers making the same sound as standing waves. What's the difference?

Al K.

Mr. Widget
03-14-2005, 07:09 PM
No argument here... I just have issues with absolutes. Let's see if in ten years everyone is using them.

Consider it a personality quirk. :hmm:

Widget

pangea
03-15-2005, 02:51 AM
Then why do you want to mess them up!

Just enjoy them.

Widget

You're absolutely right, but I was thinking of two things.

One, being that I thought I could have a somewhat wider "sweet-spot", if the discussion here could lead to any concrete improvements on my hybrids. :applaud:

The other thing I was thinking of, was if phase and time alignment doesn't have that much impact on the result, as the steepness of the filters, then what is it that causes the soundstage to collapse so easily?
I mean I have only to move one of the speakers a little bit or one of the drivers even, to totally loose the soundstage, but after having the UltraDrive do it's thing for a few seconds, it's all there again like magic.

What is it I'm missing?

BR
Roland

Al Klappenberge
03-15-2005, 05:01 AM
Widget,

With only me making them, I doubt they make much of an impact on the world in 10 years, or 50 for that matter!

Al K.