drawbacks w/ constant directivity horns

[mini]

Hi,

So called constant directivity horns for example narrow its radiation between 500Hz and 1kHz, but widen it again above 1kHz(1). With the 1.5" "optimized aperture" design off axis level exceeds the on axis level by several dBs between about 3kHz and 12kHz(2). The latter combined with the narrower radiation below causes a serious inbalance.
As most people wouldn't listen on reference axis this may be an issue. At home and especially with horns the reverberant field can't cure that. To tow the speaker in isn't practical due to sheer size of common (to us) cabinets.

Any idea?

Thanks a lot


(1) This is true with the groundbreaking EV HR90 up to now with JBLs PTWaveguides. Not to forget E.Geddes - his famous design shows the effect too, which makes "higher order modes" a/k/a HOM apparent at a glance to the educated reader 8-]

(2) http://www.jblpro.com/pub/obsolete/23525354.pdf , "Horizontal Off Axis Response"

[Zilch]

Any idea?Crossover filter "fixes" the dip at the low end.

EQ the high end where you listen.

I'm not seeing the actual on-axis frequency response published anywhere there....

[mini]

I'm not seeing the actual on-axis frequency response published anywhere there...


Hi,

I've posted a reference to JBLs datasheet regarding the 235x series.

(2) http://www.jblpro.com/pub/obsolete/23525354.pdf (http://www.jblpro.com/pub/obsolete/23525354.pdf) , "Horizontal Off Axis Response"

With the 90°x50° waveguide off axis level exceeds on axis level by several dBs between 3kHz and 10kHz. Second to that level below 3kHz goes down with listening angle as it should with a CD design. This effect comes up with off axis angles as low as 15°.

You don't see the inbalance e/g using this device as intendet from 600Hz up to ~20kHz? Equalization to flat response at some special angle would in any case alter linearity near besides that angle quite a lot. That in consequence gives a definite "sweet spot". Ain't it a contradiction to what "CD" is meant to do?

I wonder why this hasn't been discussed yet regarding home applications of medium to full size CD devices. May be in this respect the older horn/lens 2..3-way combinations do better?!

Juergen

[Zilch]

The on-axis response is "normalized" to perfectly flat there, so we don't know the actual frequency response. Do we suppose the response above 0 dB is coming from spontaneous generation?

It doesn't tell us much more than how the horn behaves with the particular driver and mouinting alignment employed for the testing. Until it is used in a final system design and measured, the specs provide only guidelines for comparison to alternatives.

Once you EQ the system to the preferred listening angle, the minor deviations shown become factors only in the reflected soundfield, in which context a variation of 1 - 2 dB would be swamped by other variables such as whether the drapes are closed or not or even if and where the dog is in the room.

Compare these specs to the measured polar response of the old exponential horn/lens combinations, ignoring the fact that they have long since been abandoned for other reasons as well.

You want "perfect" CD, use the Geddes oblate speroid waveguide; he tells you to EQ that on the listening axis as well as NOT to listen on-axis, for reasons most obvious. His polar measurements of the $9.90 JBL PT waveguides I sent him, using the cheapest compression drivers I could find, are posted on his website:

http://audioheritage.org/vbulletin/attachment.php?attachmentid=34325&d=1222640963

Finally, recognize and appreciate that you're looking at pro SR horns and waveguides here, which we "adapt" to home use DIY. For JBL's best efforts in horns intended for home use, you must look to their TOTL consumer product offerings and clone those.


I wonder why this hasn't been discussed yet regarding home applications of medium to full size CD devices. May be in this respect the older horn/lens 2..3-way combinations do better?There are plenty of other forums rife with endless techno-babble about the comparative virtues of one approach versus another. Sure, do Tractrix, if you want, or Le C'leach, or quadratic whatever. Bottom line, at some point, the theoretical speculation must end and the rubber hit the road, or it's all just so much wank and blather.

DO it, measure it, and listen to it.... :yes:


http://audioheritage.org/vbulletin/attachment.php?attachmentid=12960&stc=1&d=1139807745

[Robh3606]

Hello Zilch

Good post reminds me of "Just do it"

Rob:)

[Ian Mackenzie]

Hi,

I've posted a reference to JBLs datasheet regarding the 235x series.

(2) http://www.jblpro.com/pub/obsolete/23525354.pdf (http://www.jblpro.com/pub/obsolete/23525354.pdf) , "Horizontal Off Axis Response"

With the 90°x50° waveguide off axis level exceeds on axis level by several dBs between 3kHz and 10kHz. Second to that level below 3kHz goes down with listening angle as it should with a CD design. This effect comes up with off axis angles as low as 15°.

You don't see the inbalance e/g using this device as intendet from 600Hz up to ~20kHz? Equalization to flat response at some special angle would in any case alter linearity near besides that angle quite a lot. That in consequence gives a definite "sweet spot". Ain't it a contradiction to what "CD" is meant to do?

I wonder why this hasn't been discussed yet regarding home applications of medium to full size CD devices. May be in this respect the older horn/lens 2..3-way combinations do better?!

Juergen
I think the questiom must be related to the end use application.

For a stadium or concert hall where to want sound to each everyone in the audience yes there are benefits.

But for home use the term CD is too much of a generalization for the way the current crop of JBL Consumer horns are designed .

[johnaec]

DO it, measure it, and listen to it.... :yes:Sheesh, Zilch - is there *any* JBL 2-way combo you *haven't* tried yet?? :p

John

[Zilch]

Seems I'll be waitin' a VERY long time for my 476BEs to arrive here.... ;)

[mini]

The on-axis response is "normalized" to perfectly flat there, so we don't know the actual frequency response. Do we suppose the response above 0 dB is coming from spontaneous generation?

Sorry, I won't get this. The + does come up cause of "higher order modes" that alter the (intended) zero-order spherical radiation pattern.



Once you EQ the system to the preferred listening angle, the minor deviations ... and where the dog is in the room.

If equalized to flat on axis at 15° off axis You have to face a step of +6dB @ 3kHz that extends up to 10kHz. That assumeably would be apparent even to the least demanding listener.


You want "perfect" CD, use the Geddes oblate speroid waveguide;

These too show serious "HOM" - HOM may be definded simply as deviation from the intended radiation pattern (sperical, plane ... whatsowever), expressed in orthogonal functions, alike Fourier analysis. T'm convinced that the ondulations of Geddes' OSW in frequency response over angle ain't intended. They do no that much better than conventional devices. That shouldn't suppress the merits he earned with theortical findings - great work!


the $9.90 JBL PT waveguides

O/k, The curves I saw didn't look that good as Yours here. It's a pitty they aren't useable with lower cutoff. In Germany they can't be bought too.

BTW: all above didn't "bring the rubber to the runway" :D What to do if one owns the "optimized aperture" somethings already? To equalize to flat on listening axis would affect nearby listening positions. That is my practical finding with these. Remember that I swaped the diaphragms of the 2447s already to the plane ones of my 2445s.

Thanks a lot anyhow

[Zilch]

That's certainly not how Geddes defines HOM, and like "Waveguide," I would presume since he believes he coined the term he is therefore entitled to define it. Go over to the forum he frequents and suggest that his waveguides are a veritable festival of HOMage, and see how he responds. :p

********

Those are not my curves, they are HIS curves. Go back and look at his report again, but this time, scroll down to the EQ'd comparative data; that's where those I posted above may be found.

Have you asked Guido if he can obtain these cheap waveguides for you? They're also available with 1.5" throats, but not so cheap. You may be surprised to discover how low those might actually be used with certain drivers. There's actual curves posted for them here on LHF, and not just by me. :yes:

********

What's that in the upper left of my pic above? Have you ever heard one? :)

[Ian Mackenzie]

I hope you don't plan on bragging on about them or posting something retarded with respect to them.

.:bs::wasnt-me:

[mini]

That's certainly not how Geddes defines HOM, ...
Those are not my curves, they are HIS curves.

Great,

You've sent him the w-guides, that he subsequently measured. On the other hand I would insist on what HOM has to be meant with. He describes his finding of OSW as that the solution of the wave propagation equation in special coordinates can be found in 1 dimension. In spite of solving the 3-dim equation for complicated shapes of the horn he solves the easier equation for a specific geometry. The result should be a perfectly spherical waveform at the mouth with all frequencies.

That isn't true with his own designs. Not only is the frequency response contaminated with ondulations from internal interferences but this too depends on angle. That's what "HOM" makes, isn't it? O/k, what about perturbation analysis? Why not describe deviations from a perfect waveguide/in-wave/out-wave with a - example given - Fourier-series? Find the transformations! That is what I would expect when someone claims to deal with Higher Order Modes.

Yes, I've got the 235-2 WG with the intended driver myself. It's not bad but I wondered if anybody had made the same observations as me.

[Zilch]

I can only suggest taking the HOM issue up with Geddes himself.

I run the #3 and #4 drivers on 4313B's above list on 2352s, and find them to be "Transcendent" using either of those.

[Zilch]

You like them better than those super expensive sonoglass Consumer horns that no one can get even if they wanted them!? :blink:Nope.

Those are "uber-transcendent." ;)


Why don't you hook toddalin up with some of those 2352's and the 2452H-SL's I have for sale and save him from those nasty old fog horns with the plastic lenses?You have 2452H-SLs?

Would you run them on PT-F95HF (the combination I'm mostly building with) and post the curves in an appropriate thread, please, in comparison to 476Be and 435Be?

I forget where you posted those, but you suggested that I send you 2452H-SL to try the same.

*******

I don't know, but I certainly would have dumped the potato mashers as well, long ago. Todd and I have had this discussion, actually, with respect to potentially switching to PT waveguides.

I BELIEVE 2352s would generate WAF issues.... ;)

[Zilch]

I haven't opened the boxes yet. The person who ordered them lost his job and I can't return them for credit if I open them. If I get socked with a restocking fee, which is fairly likely, I'll probably keep them and run your curves.I can still send you a pair to measure, if necessary.

Let me know.... :yes:

[Zilch]

Does anyone have any money to buy these things from you? Or are you guys getting them second hand?I bought one pair factory new as baseline, but all of the others have come from the lowball SoCal vendor of pro drivers with indeterminant provenance.... ;)

[Zilch]

Wow! And you've not suffered a single pinched voice coil? That's amazing. Who is this SoCal vendor? This is where all those 2435's came from too right? What is he charging for these?Yes, the same one who provided the flood of 2435HPLs members bought last year. Supposedly "new," but not, apparently.... :dont-know

[pos]

Would you run them on PT-F95HF (the combination I'm mostly building with) and post the curves in an appropriate thread, please, in comparison to 476Be and 435Be?I think you already measured them:
http://audioheritage.org/vbulletin/showpost.php?p=124188&postcount=84

And I posted the comparison here, with the same scale:
http://audioheritage.org/vbulletin/showpost.php?p=219998&postcount=32

[Zilch]

I think you already measured them:
http://audioheritage.org/vbulletin/showpost.php?p=124188&postcount=84

And I posted the comparison here, with the same scale:
http://audioheritage.org/vbulletin/showpost.php?p=219998&postcount=32
We've moved substantially further ahead since then, actually:

http://audioheritage.org/vbulletin/showthread.php?p=204970#post204970

[pos]

I thought you were looking for raw measurement on the PT-F95HF waveguide?

[Zilch]

I thought you were looking for raw measurement on the PT-F95HF waveguide?Well, yes, but how well we can get them to play is always a matter of considerable interest, too.

4313B uses different measurement methodology from what I do, and under better conditions. Whatever he is willing to do will be valuable.

This is largely uncharted territory, but also a very nice place, in my experience.... :yes:

[johnaec]

Who is this SoCal vendor? This is where all those 2435's came from too right? What is he charging for these?I'm sure Zilch is talking about this guy: http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&rd=1&item=370115980411&ssPageName=STRK:MEWA:IT&ih=024

I've emailed him before to ask, but he's really dodgy about his sources, though he certainly seems to have them... List his stuff - sometimes he's really got a bunch of decent drivers. Listing his completed items shows what he's been getting, and the deals some people get! I think I've bought some dias. from him before - can't remember...

John

[boputnam]

...he's really dodgy about his sources, though he certainly seems to have them... We'll find him. We're on it...

[johnaec]

JBL seems quite interested to discover his source... Evidently he is "just up the street"...I'd hate to blow a somewhat hidden but legitimate source - if I remember right, he mentioned he'd built up a bunch of sources, many of which were connections to SoCal sound companies with extra things like those service packs JBL sells for stuff like the Vertecs, where you buy a pack that has a whole bunch of different parts in it, drivers, dias, handles, hardware, etc., (if you dig down in the documentation enough you usually find stuff like that listed where you get a bunch of parts at once), and he was picking up stuff that was mainly left over.

If he ain't doing anything legally wrong, at least he's been a decent source at times for many... JBL has to be aware of him - he's been around for years.

John

[mini]

I can only suggest taking the HOM issue up with Geddes himself.

I run the #3 and #4 drivers on 4313B's above list on 2352s, and find them to be "Transcendent" using either of those.

Hi,

Sure these horns sound slick and acute. In spite of being "HOM"-ies as can be. In my understanding there is nothing to argue about with Geddes himself. The HOM-issue should be seen as a somehow missleading speculation - until the claimed effects are made available to common engineering or at least - direct - investigation. Everywhere in HiFi a bunch of wishfull thinking in front of imaginary amazements hinders clear sight, people are to much used to it.

If I liked to stick with JBL, I take the el cheapos. The off axis performance seems to be great. No HOM but plain spherical radiation. There's an issue with loading though. The raise of THD that You measured with lower frequencies is a remarkable finding. I didn't see such with other smaller horns - e/g RCF H100 and appropiate drivers. My medium format horns - EVs and JBLs - did the other way round: high THD at high frequency and lowering THD way down to some distinct lower cut/off. Below the latter a sudden rise of THD with queasy noises.

I understand that the vintage CD approach has to be taken as it is. I was curious on what people - if any - do with the leakage of lower hights off axis. It seems that the alternative is to accept or split up with it.

Thanks a lot

[Zilch]

I wouldn't consider my distortion measurement definitive of anything. You can see me asking how to do it there. ;)

Geddes's analysis of his measurements of the cheapo JBL waveguide is at variance with my own, of course. I tried to elicit some discussion of the results on that forum without success. It seemed as though most everyone was more interested in using the thread as a platform to forward their own off-topic agendas.

Nobody ever seems to get down to the head-to-head competition data comparing the EQ'd waveguides' performances. At worst, I don't see how anyone can conclude that the JBL "horn" did anything less than remarkably well.... :thmbsup:

[mini]

I wouldn't consider my distortion measurement definitive ... head-to-head competition ...

Zilch,

It ain't easy to purchase JBL parts at reasonable prices in Germany. I'm near to ask Guido for some el cheapos. As soon as I get a pair I'll offend You with my own measurements ('hope You get the irony).

btw: http://www.fesb.hr/~mateljan/arta/

Measure distorsions via a method found by Mr. Farina, that is deconvolution of a log swept sine, transforms the HD components in time domain to seperate impulse responses, in the end gives HD f.response. You should give it a try. Acurate HD2 and HD3 within seconds for the whole spectrum.

Jurgen

[Ian Mackenzie]

Mini,

You raise an interesting point about distortion and low end directivity.

Most of the el cheapo CR horns of the PT type were made for integration into some of JBLs SR systems where they were crossed over at 2K hertz and above from a matching PT midrange wave-guide. ie they are not intended for 800 hertz crossover. The increase in distortion is the result of the driver dias being unloaded and increase in excursion.

As far as domestic home audio goes according to Geddes, there is little point in controlled directivity with horn loading below 1000 hertz for home use but he does recommend matching the dispersion of the woofer and the horn at the crossover point. Typically this is 1000-1200 hertz for a 15 inch woofer. To properly load the driver this requires a large wave-guide of 15 inches or more.

This is based Geddes overall approach to understanding small room acoustics in relation to what is important to the human ear as a result of his scientific study. (less than 25 sq metres)

The problem is people shy off from large horns and wave-guides and try to use smaller devices and they wonder why the $299 Ipod dock sounds better than the diy speaker.

My recommendation is leave it to the experienced loudspeaker engineer and build a proven diy system than base your project on a lot of ad hoc data posted on numerous forums.

Geddes paper on his Suma system is quite a valuable read if an understanding of what a loudspeaker does in a room is desired. To appreciate what Geddes is saying you need to read the paper several times because the complexity and the number of dimensions to the discussion.

Geddes also markets a wave-guide to his design.

Many of todays domestic loudspeakers are designed for not what looks good on paper on and off axis but how they perform in a domestic home listening room in so far as the response and dispersion in the reverberant field is concerned. Some of the more recent JBL Consumer 2 way systems are designed along these lines as opposed to the older 43XX blue baffle series (JBL Japan range)

CD or no CD ie a direct radiator it really comes down to specifics and your particular needs..assuming you understand what they are.. But people still like to make a labour of love out of doodling with all sorts of drivers and horns like they were coming of a conveyor belt.

If your are interested Acoustic Elegance are working on a large wave -guide for the Diy market to mate with their 15 inch drivers. I have no affiliation but their products appear to be very well made and are value for money for the diy market

[Zilch]

Geddes teaches that horn loading is a concept derivative of old-school horn theory which does not apply to waveguides.

He also teaches that distortion as we know and measure it is irrelevant to sound quality.

JBL uses some mighty small horns in its current TOTL monitors, down to 1 kHz:

http://manuals.harman.com/JBL/HOM/Technical%20Sheet/4348%20ts.pdf

They ALSO use some rather larger ones at 750 Hz:

http://manuals.harman.com/JBL/HOM/Technical%20Sheet/4338%20ts.pdf

Go figure.... ;)

[Ian Mackenzie]

I said earlier it comes down to specfic of the design.

JBL know what they are doing

Specifically the PT devices you have been rooting around with in retard mode are not but you keep trying to make it work.

[Zilch]

Specifically the PT devices you have been rooting around with in retard mode are not but you keep trying to make it work.Yeah, well, seems quite a few more have joined the PT Corps of late.

When are you going to actually DO something other than blog and blather here, Ian? :dont-know

[mini]

Geddes teaches that horn loading is a concept derivative of old-school horn theory which does not apply to waveguides.

He also teaches that distortion as we know and measure it is irrelevant to sound quality.

JBL uses some mighty small horns in its current TOTL monitors, down to 1 kHz:

http://manuals.harman.com/JBL/HOM/Technical%20Sheet/4348%20ts.pdf

They ALSO use some rather larger ones at 750 Hz:

http://manuals.harman.com/JBL/HOM/Technical%20Sheet/4338%20ts.pdf

Go figure.... ;)

Hi,

I'm very pleased with Your attentivness (irony). Yes, I do appreciate Geddes' work on horn/waveguide theory. Not to say it was the most delightful reading regarding audio transducers of the last years (Distributed Modes Loudspeaker, better known as NXT flat panels was an other one). It had been real fun all the way. When Geddes uses HOM as an issue to encourage his own designs it becomes slightly uncomfortable though. But in the end his kits/boxes may be the best ones available by now especially for the prices called!

As Zilch remembers I own a pair of impaired EV horns type HP940. This "loads" down to --400Hz as can be seen all to clear from distortion and intermodulation figures. I'm pretty sure Geddes didn't mean it that way when he claims THD wouldn't matter. High THD from overstressed drive and/or compliance is of course a well known issue. For my taste it sounds very disgusting. What could be of less importance is the adiabatic non linearity with compression drivers at higher frequencies. But again, the scientific evidence regarding that is as flimsy as with some other claims in audio. I would really like to see some additional tests of that hypothesis - 4 eyes principle just 2 be sure.

Do You have any suggestion what PT Waveguide to take with 1" or 2" drivers? I won't let the fancy 235-2 alone using the appropriate 1.5"er with a brand new capture.

Jurgen

[Zilch]

Do You have any suggestion what PT Waveguide to take with 1" or 2" drivers? I won't let the fancy 235-2 alone using the appropriate 1.5"er with a brand new capture.I have built several systems using BMS 4555 on the 1.5" PT-F95HF. JBL 2452H-SL mates very nicely with it as well.

RobH has used the PT-H1010HF with 2435HPL for over a year now, and reports he is quite pleased with its performance, sufficiently so as to retire the 2344As in his DIY 4430s.

For that same 100° x 100° pattern, I also like the rectangular PT-F1010HF. The PT Waveguide white paper says the difference is that the "compact" rectangular variants ("F") have lower distortion, but less vertical pattern control. The square "rotatable" ones ("H") appear to play a bit lower.

Several others have built using the PT-F64HF, and I like that one, too, but their 60° x 40° pattern is a bit narrow for my taste. My preference is biased by having 4430s as my mains for over 20 years now.

If you want to get an idea of how these perform, the thread-on 1" PT-F95HF is only $9.90. JBL Part# 338800-001, it's the one Geddes measured for us. I particularly like it with BMS 4552nd using the P-Audio PC-35 throat adapter.

Various members on another forum have now completed 20 systems using this waveguide and inexpensive entry-level Selenium drivers, with another several dozen presently being built by others:

http://audiokarma.org/forums/showthread.php?t=150939

[Robh3606]

Hello Jurgen

You can also go with an adaptor for the 1 inch drivers for the 1.5" horns. Attached are photo's of one on a 2416. So any thread on driver 2426, 2425HS or 2418 will all work as well as any other manufacturers drivers as well. Advantage with the 2426 and 2425HS is you can use aquaplased diaphrams in them.


RobH has used the PT-H1010HF with 2435HPL for over a year now

I have 2 pair of these in 2 DIY set-ups both use aquaplased 2435's. My goodness how time files it's actually 2 years now!! If you can get them they can sound quite good. The only real problem with them is they are 1K horns. Too bad they don't go lower than that. I have done sims where they look preaty good using the rolloff to get them lower but have yet to have a listen to see what's going on.

Rob:)