View Full Version : Help with 4312's

06-17-2006, 09:54 PM
Hello! Anyone here with intimate knowledge of early '80's 4312's?
I have a strange low freq. response problem.
My ears have heard it from the beginning, but I finally have the tools to really see it in great detail.
Both speakers now have a sharp 20db dip at 125Hz, and a sharp peak at about 270Hz.I've religiously checked for air leaks and loose joints-none found.The 125/270Hz relationship made me suspicious of this.
For years, I wrote this off as room boundary issues, yet every room I put them in, regardless of room dimensions,I ended up with the same result. I recently tested these 4312's outdoors, one at a time,flush against a large brick wall, a reference mic 3.3 ft. away, the closest north/south hard surface at least 250 ft. away (the only way that I can mimic the original JBL test conditions), and got approximately the same results. SPL level was 80.
The room they live in is treated, with massive bass traps in the rear to reduce room boundary effects.
I found a test graph I made in the '80's with a Goldline RTA in an untreated room, and the dip at 125Hz was 10db down.
Today, in a treated room or outdoors, it's 20 db down.
The testing software I'm using is ETF 5.XXX (demo version).
This software also reveals the 250-270 Hz box resonance that's always bothered me from day one.
Has anyone tested these monitors this extensively?
Has anyone seen this before?
Thanks for everybody's help!

06-18-2006, 08:59 AM
All your "issues" appear to lie within the range of the 2213H. There is no Lo-pass filter. The 2213H runs wide-open, allowing it's natural roll-off.

The MF has a 1.5KHz Hi-pass - everything you mention is far below that.

I wonder if a recone would be in-order...?

Tom Loizeaux
06-18-2006, 12:51 PM
Your attention to your 4312s is impressive. I don't mean to question your sense that the irregularities are getting worse, because you may actually be experiencing a problem ... but it's possible you are simply hearing more clearly the limitations of that design.
I like my 4312As, but I suspect the 4412s, with their true 3-way crossover and foam surround woofer would prove to be flatter and be a more accurate performer.
Please keep us posted as you learn more about this.



06-19-2006, 12:25 AM
Hello everyone!
Thank you for sharing your ideas.
Since I first noticed this problem to a minor degree right out of the box, I'm not sure if a recone would help,unless of course the magnets need to be recharged...that was JBL's last suggestion. All previous suggestions by the JBL crew have not solved this problem. The recharge/recone would be an expensive gamble to take. I've looked at the 25 watt 40Hz output from the speakers on a scope, to check for major suspension problems,and the waveforms are clean and symetrical.
I wish I had tested these speakers outdoors back around '88 or '89 when I first tested them indoors with an RTA.
Like Mr. Loizeaux, I've wondered if this is just a design flaw.
I was hoping for a response curve like the one Mr. Loizeaux posted, but I've never measured a low freq. response like that. If you start with that curve,then draw a dip starting at about 92Hz,with a maximum attenuation of 20 db at 125Hz, then back up at about 164Hz, and a peak of 8-10 db at approximately 270Hz, then you have something like what I measured in a treated room and outside.
Since the original literature regarding freq. response testing raises lots of questions, I was just wondering if anyone else had tested their 4312's to see how their set really measured up.
Mr. Loizeaux's statement about the low end response of 4412's possibly being more accurate than the 4312's is true in my case. There is quite a low end difference between my one of my 4312's and a 4412 side by side.
The cabs of those models are very close in specs, and I've been tempted to try and find a good used 128H-1 and drop it in and see what happens.
Unless I'm mistaken, it's either the cabinets or the 2213H's.
If it's the 2213's then the big question is are both actually defective?
If I ever figure it out, I'll post my findings.
All insights and suggestions are appreciated!
Thanks everyone!

06-19-2006, 09:03 AM
...I've been tempted to try and find a good used 128H-1 and drop it in and see what happens.I've done that, myself. The 2213's went over the transom...

I don't advocate swapping elements, but I ran some models in WinISD and BoxPro comparing the 2213H vs the 128H in the 4312 cabinet - the swap was reasonable and the resulting low frequency response superior (acoustically - I never measured it). Ideally, I suppose the network should be re-worked to include a Low pass filter.

06-19-2006, 01:28 PM
Thank you Mr. Putnam!
I believe I see some light at the end of this tunnel....
Did you by chance,happen to notice a whopper of a dip in modeled response around 125 Hz with the 2213H's?
I would guess modeling programs would allow you to tweak the flux density to see what field variations do to the response?
That might explain a few things.
I'll Google these programs to see if there's a consumer demo version available.
I'd like to experiment with the numbers on a program like that.
I might actually learn something!
Thank you Mr. Putnam for mentioning these!
With everyone's help, I feel like I'm finally getting somewhere with this.
Thanks everybody!

06-19-2006, 01:45 PM
WinISD is shareware, although they appreciate donations!


The models do not show a dip at 125Hz - they show a gentle curve and in general are used to predict the (extreme) LF response expected from the input box dimensions and Thiele-Small parameters.

I regret I never measured the 4312's - before and after. The acoustic difference was quite dramatic (as I recall - the 4312's are boxed away right now) and so I was not inclined to do any measuring. Also, at that period, I did not use an EQ in the signal path so any measuring would have only served as observation not as guidance...

06-19-2006, 09:06 PM
The 2213 is a High Qts speaker and as a rule should be in a sealed box. Venting a speaker with a Qts above .38 almost always results in a peak in the response. Some like the peak and tolerate it, I don't like the transient response that it gives, too punchy for me.

The 128H is an excellent 12, probably one of the best JBL made, and it should go ruler flat to 40 Hz in 1.7 cu ft without much trouble. I know mine did.

If you run an impedance curve on the 2213 in the box you will see how different the 128H is, provided you had a 128h to swap out with.

06-21-2006, 12:20 AM
Hello everyone!
I think I've learned 2 things already:
JBL should have used the 128H's in the 4312's, and obviously I have some kind of funky problem with my pair of 4312's, that is making mine even worse.
I think it comes down to the 2213H's or the cabs having a bizarre resonance.
Since it's nowhere near as expensive to build a MDF test cab, compared to having a 2213 overhauled, I think I'll try that first. I can always build a test speaker out of it later....
I'll try to build one using WinISD as a guide, and see if I get better results.
I'm also going to rerun all previous tests on one of the old girls sometime this weekend,with the port plugged, to see what that changes in the real world.
(Thanks for the tip Duaneage!).
If I find something interesting, I'll let you know.
Thanks everyone!

06-21-2006, 11:39 AM
So, the first (and easiest) step toward "fixing" L100 bass would be blocking the port?


[We KNOW how to do that.... :yes:]

That gives Qtc 0.874, F3 43.92 (black).

Then upgrading the damping to "heavy" yields Qtc 0.779, F3 45.72 Hz (red).

[Not sure of the net internal box volume, so "first approximation"....]

06-21-2006, 12:03 PM
[We know how to do that.... :yes:]Gollum...? Hello??

06-21-2006, 12:08 PM
Gollum...? Hello??This forum knows the local hardware plumbing department well by now.

Well, me and CLIO do, anyways.

Need I illustrate (again)?

If I do that, I'll have to take some measurements.

[And NOBODY wants to see that happen, certainly.... :p ]

06-22-2006, 06:05 PM
Here are some response measurements from the old girls.
Since there's little difference in response 3.3 ft away inside or outside, I decided not to scare the neighbors anymore than I have to.
The measurements were taken inside the treated room.
The jpg's are in pairs, one pair is 20HZ-200Hz, the other pair are 100Hz-1000Hz.
Two are with the port open, the other two with the port sealed.
The green line is the calibrated response curve of the sweep tones I'm sending to the amp/speaker, and the blue line is what I'm getting back.:blink:
Let me know what you think.
The limited demo version of the software I'm using for this is free. If anyone wants info on it, just let me know.
Thanks everybody!

06-22-2006, 06:25 PM

Explain (again...?) the test set-up?

What is the mic (brand and model)?
Where is the mic in relation to the woofer? You say 3.3-ft away - why?
Describe the "treated room", again?

06-22-2006, 09:10 PM
Hello Mr. Putnam!
Here's the skinny: The mic is a Groove Tubes MD5SM fet microphone, that has a razor flat response from 22HZ to about 5Khz. From there to 18Khz, it's tested and guaranteed not to deviate more than 1.5 db.
The 48 volt phantom p.s. is a Newmann.
The preamp is a custom built ultra low noise unit,built around a Burr-Brown INA103 op-amp.The soundcard is a 20 bit Echo Layla, the power amp is a Yamaha P2250, and the speaker cables are 12 gauge Monster Cables.
I mention that because JBL once told me the problem was occuring because I was using 15 gauge standard cables, which could cause excessive capacitance and excessive resistance, reducing the damping factor of the amp.
A valid argument I guess, but resulted in no change in my situation.
The mic is placed dead center of the speaker's x/y axis, 3.3 ft away.
The reason for this, is because this is the intial measurement point JBL uses for their freq. response measurements.
According to the supplied literature and techs at the factory, they test the speaker flush against a large baffle centered in a near anechoic enviroment, with the mic centered on the speaker 1 meter away(3.3 ft).
If necessary, they repeat these tests, swinging the mic in an arc, to publish response variations due to listening angles.
I'm just trying to stick with the program.:)
Concerned room boundaries was still wearing me out, I tested them outside, flush against a brick wall, to simulate JBL's setup--with no real difference below 300Hz. Above 300Hz, it smoothed out some...
The room, well that's a novel....but here goes!
Structural dimensions are: 15' 3.6" long, 11' 3" wide, and 8' tall.
The north wall (where the 4312's live), is modified into a octagonal shape to reduce standing waves and head distortion.50% of this wall is covered with Auralex--couldn't afford to put wide band traps everywhere.The front of the east and west walls are covered with pyramid Auralex from 2ft. up.Only the first 9ft. are covered.
In the center of the rear wall there's a open 5 ft wide x 30" deep closet that's packed to the ceiling with bats of R13, with somewhere between 7-10 inches of dead space behind all that.
On each side of the closest are home built (and not too pretty) blue cloth covered open frame bass absorbers that's comprised of 8 layers of plastic wrapped R-13 each, packed down to 10" deep units.These are 5ft tall, and are on stands that lift them to the ceiling. According to my calculations, they yield .99 absortion down to about 60Hz, falling to about .50 at 40Hz in their coverage area. I may be wrong about that, but if I walk to the back of the room, I don't hear bass booming anywhere.
I think that about covers it.Hope this post wasn't too long in length!
Thanks everyone!

Tom Loizeaux
06-23-2006, 05:58 AM
At first glance it looks as if the low frequency response doesn't change that much when you closed the port! I'm surprised at this.

I'll be watching for other comments and developments.


06-23-2006, 08:22 AM
With that test gear set-up, why are you working on 4312's...? :rotfl:

Seriously, the LF curve looks very strange. That extreme roll-off starting at ~200Hz is just wrong. And, as Tom sez, there is no material difference between plugged and ported.

What is the source data - Pink Noise?

Can you feed your Pink Noise signal output directly into your measurement set-up, bypassing the mic, and the 4312? I'd like to see that trace...

06-23-2006, 11:11 AM
As Tom observes, the fact that there's no difference in measured response vented vs. closed is telling. Something's up.

My take on it: you can't measure LF response using the method described. Nearfield, ground plane, and in-box are options. Nearfield, you have to sum with the port output when measuring vented.

Also, you've got to get the woofer up off the floor to measure it in that range, even nearfield; 1/2 the ceiling height's optimum.

What's "nearfield" in this instance? Mic within one driver diameter. Some advocate at like 1/4" out from the center of the dome. Try that, and see if you don't get more rational results....

06-23-2006, 11:52 AM
Some advocate at like 1/4" out from the center of the dome.Yea, when Widget is using CLIO he gets his measurement mic right up close.

06-23-2006, 04:24 PM
Hello everyone!
Mr. Putnam, the input signals I'm using are software calibrated sweep tones.The green lines in the graphs are the actual freq. response curves of the tones that are sent to the power amp. The small deviations in the green line are sound card freq. response deviations.
I forgot to mention that the speakers are 3 ft. off the floor-sorry about that!
Included here are more curves.
Zilch is on to something, something's definately up,I just don't know what to make of the results.:blink:
These tests were run in the room, all speakers tested in the same location for comparison. The up close mic'ed positions are centered on the 2213H and E120 ,rather than the 4312's x/y axis. The mic is 2 inches from the dust caps.
Since the original published freq. response curves were taken 1 meter away, on axis, I don't know how to interpret these results.
I will repeat these tests outdoors again this weekend, and I'll capture the results and post them here. I'll capture 1 meter x/y axis, and 2 inches away centered on 2213H's. Due to mic/mic holder dimensions, 2 inches is as close as I can get to the dust cap.
The last graph is mislabled. It should read:JBL E120 utility cab(approx.1.75 cubic ft.),mic'ed 2 inches away from dust cap,centered on dust cap,3 inch port hole, no duct, well padded.
Here they are!
Thanks everyone!

06-23-2006, 04:39 PM
Well, there you go. :thmbsup:

You can now even see the difference between closed and vented behavior.

There's nothing wrong with the woofers, it doesn't look like.

E120 in 1.75 cuft. w/ 3" dia x 0.75" port:

06-23-2006, 04:50 PM
Hello Zilch!
If I get similar results outside, what will that mean?
I definately remember seeing that dip in response outside as well.
There is an overhang above the speaker where I tested it. Could this cause the same problem? I can move it over to be flush with the chimney next time.
More confused than ever!
I would like to thank everyone for everybody's help.
The lightbulb's got to light up eventually!:)

Tom Loizeaux
06-23-2006, 05:04 PM
Isn't it true that if you position your mic close to the woofer, you will get a low end roll-off reading when the signal starts to reach cabinet tuning. At that point the port is contributing most of the SPL and the cone is doing the least.
To get a true performance curve for this speaker, wouldn't you need to make the mic placement closer to equal-distant between the cone and port?

Just my thoughts...


06-23-2006, 05:46 PM
Isn't it true....Yes. You're supposed to measure the driver and port separately, and add them, weighted for the differential diameters.

Nearfield between the port and driver would be a useful approximation, I would think. It's easy to watch what happens on an RTA as you move the mic closer and further.

There's also a high frequency limit to summing the nearfield port measurement, 1.6 X the tuning frequency, according to Dickason, due to crosstalk between the vent and driver.

I forget what D'appolito, Keele, and Small each say on this subject; they've all been there.

If I want to see the real VLF response, I gotta get up close, is what I know.... :p

06-24-2006, 10:22 AM
Hello everyone!
Don't know what it means, but here's woofer+port 2 inches away.
More curves comin' sometime tommorrow.

06-25-2006, 04:41 PM
Hello! I wonder if JBL may have left an important piece of information out about their testing procedures.After I saw the same dip in low freq. response from my E120 loaded utility cab, I started having more questions about what they are actually doing.
After getting the same deep dip outside @ 1meter, I'm really curious.
There's no room boundaries out there.
For all these tests, the speaker was mounted flush against a brick wall(large baffle) for low end support, and the mic on it's x/y axis for the 1 meter measurements.The attachment limit is 5, so we'll skip the dust cap measurement for now...
The testing conditions are different this time, to eliminate overhang boundary questions.
The closest n/s boundary is 65 ft. (compared to over 200 ft. previously)
This is the best I can do to re-create the original testing conditions.
"The accompanying graph and specifications were compiled from measurements made under standard laboratory test conditions.The loudspeaker system was mounted flush in the center of a large,flat baffle in an anechoic environment;a calibrated condenser microphone was suspended at a known distance from the sound source,sufficiently far to be safely out of the near field;and all electronic equipment was checked and calibrated before tests were run."
As you can see from the graph that Mr. Loizeaux posted, the "known distance" is 1 meter.
The only difference I can see, is that I'm using calibrated sweep and MLS signals.I'm certaintly not getting curves like this!:(
Something's missing in this equation.
In a couple of weeks,I'll be able to run 1/3 octave pink noise measurements. If this reveals a different story, I'll post the results.
For now, here's what I measured outside.
All thoughts and comments are appreciated!
Thanks everyone!

06-25-2006, 05:18 PM
Your nearfield measurements seem to be a reasonable approximation of the factory curve; they capture the L100 et. seq. LF "bump" quite nicely.

As soon as you move away, I think reflection from the ground (floor) is what's causing you difficulty. Lay the speaker on its back, away from walls, etc., and suspend your mic 1M above it; I suspect the "dip" will disappear.

See P.3, here:


06-26-2006, 08:38 PM
Holy cow,thats it!!!:thmbsup: :applaud:
I've been so consumed with wall boundaries, I never stopped to think about floor or ceiling boundaries! I really pulled a :homer: on that one!
That's the one constant that has remained unchanged in over 22 years.
Now, if I can just fix the problem.....
I would like to thank everyone for their thoughts and opinions on this, and especially Zilch, for helping me understand where I've been going wrong all this time.
It's been a real learning experience for me.
If I solve the problem to a significant degree, I'll post what was done, and the resulting curves afterward.
Thanks again, everybody!

06-26-2006, 08:47 PM
Good one, Zilch!

I was going to suggest you forgo the esoteric woofer-specific measurements and take some in/at the listening position and see what the overall cabinet reproduction response is. I still think that would be worthwhile - room effects being what they are...

06-29-2006, 12:48 AM
Mr. Putnam, you are right, and I'll be doing a lot of that for a while.
I'm hoping a listening position/speaker position compromise will fix the problem.
I don't even want to think about building more bass traps! Especially traps for the floor?!:blink:
If I find something interesting and helpful for the whole group, I'll post curves and all.
It may take awhile....
Thanks again,everyone!

06-29-2006, 11:36 AM
I'd certainly be interested to see your results with the box outside aimed skyward.... :thmbsup:

06-29-2006, 06:41 PM
Hello Zilch!
Weather and lower back permitting, I'll try and tackle that this weekend.
Lowering and lifting a 4312 to and from the ground ain't as easy as it was
20+ years ago...:eek:
It's definately worth the effort,though!
Hopefully, curves will be a comin' shortly...
Later everybody!

07-02-2006, 09:42 AM
I just recently became a registered member here and I also have some questions regarding the JBL 4312's. I have a small recording studio and have a pair of 4312's I am considering recapping. Yesterday I managed to get one crossover out of the cabinet, which by the way is a pain because the screws that hold the crossover are under a thing sheet of aluminum bearing name and model. Very difficult to remove without bending and permanently damaging the nameplate. Who was the genius who came up with this?
In any event I replace the orginal caps with some Solens. I am not yet sure if there is any notable improvement. An additonal bummer was that I either blew out the LE25-2 diaphragm or it was blown so some time unbeknownst to my knowledge until yesterday.
My question is regarding the original 3uf and 8uf caps and the .01K caps. What type of caps are they, oil in paper, electolytic or polypropylene. I am not a tech wizard although I do have some training in electronics. Also why type of caps are the .01K and why did they notate their rating this way rather than simply in microfarads.
I will have to send in the LE25-2 for a relacement diaphragm and am also looking at the possibility of substituting the 2213 woofers with some 128H's as I have been hearing the bass response on those woofers is a bit more appealing than the 2213's which seem nice and tight but like the other poster mentions seem to have a peculiar response most likely in the same frequencies he mentioned.
Any help will surely be appreciated. Thanks

07-02-2006, 10:19 AM
The key to removing the foil labels is a heat gun.

I believe the original caps were Mylar.

If you're using metalized polypropylenes to replace them, current practice is not to use 0.01 uF bypass caps with Solens, but if you feel you must, AudioCap Theta film and foil polypropylenes would be my choice.

0.01 uF is same as 0.01K nF, no? :D

Regarding swapping drivers, 4312s are "Control" monitors. If I wanted true studio monitor voicing, I'd probably move up to 4412s:


07-02-2006, 05:16 PM
I concur with Zilch, the 4412 is much superior unit. And you can get used pair for less than the upgrades mentioned above. I got mine for $333.

07-02-2006, 07:34 PM

Thanks a heap for answering the questions and pretty quickly I might add. Yes you are right about the bypass caps regarding how they are labeled. I will look into what you suggested regarding the 4412's in the future. Right now I have so many things going on as far as tweaking gear in the studio I want to get past that before I get knee deep in any thing else. But thanks so much for the help.

07-05-2006, 06:32 PM
A few related questions and if Zilch can make some recommendations that would be great. I replaced the original mylar caps in one of my 4312's with Solens. I have not removed the original bypass .01K caps yet as he mentioned it is current practice to do.
My question is what will be the sonic difference if I remove the bypass caps?
I have only done one cabinet so far and therefore it is easy to make some comparisons. What I notice is the response difference coming off the dome of the woofer which I assume is primarily lower mid frequencies. The cab with the original caps seems to project a bit more but is also somewhat harsh compared to the recapped cab.
I would also ask if anyone is familiar with Jensen caps and would appreciate some suggestions for alternatives to the Solens without getting into the high price caps. Yes I assume Jensens would be costly and possibly overhill for these old 4312's. I just won an auction for a pair of 128H's and when I receive them with be trying them out in the cabinets. I got the 128H's pretty cheap. they might however, need reconing.
I have followed the discussion regarding the peculiar bass response and would say that mine also seem to have a somewhat overly tight bass response and at times it seems the woofer is not handling the load of the lower mids and the lows. As mentioned they are running full out and that might be part of the problem. Might be time for a recone as well.
Any further insights are appreciated.

07-05-2006, 07:59 PM
While I did promise to try Solens with and without bypass caps, I have not done it yet.

You're in a good position to A/B evaluate the difference, actually. Recap the other crossover and remove the bypass caps. Then compare.

These differences, like from various capacitor types and brands, are not subject to measurement; they're more subjective listening preference kinds of variations.

There are several threads here describing the performance of different capacitors. I just try to follow JBL spec practice. Right now, I'm playing biased (but not bypassed) Solens in my Minis.

I believe there are different grades of Solens, as well. I can't begin to sort it all out. If I were manufacturing speakers, I might be better able to make the requisite comparisons. That's why I rely on the manufacturers; they presumably do all that stuff....

07-06-2006, 02:59 AM
Hello all!
A family medical emergency has grounded testing for awhile.
I will perform skyward on-axis tests and post results ASAP.
It now may be up to a month before I have the time.
I suspect the floor boundary is a big problem in sonicmerit's listening arrangement, as it has been mine for all these 20+ years.
The PDF Zilch posted, perfectly explains this problem, and the curve is almost identical to my measurements. All of mine are here, so compare mine to the floor boundary curves in the PDF for reference..
A floor boundary can blow your low end response, regardless of the model of speaker.
I didn't capture them, but super quick testing revealed speaker height to listener/mic, and speaker distance to listener/mic, both affected floor boundary altered response curves.
The only thing that I noticed not changing was the "Q" of the response dip.:blink: Not sure what that means yet...
It either shifted in frequency, and/or amplitude.
More on all this when I have time.
Later everyone, and thanks again Zilch, for solving a two decade mystery!:thmbsup:

07-06-2006, 05:57 PM
I have one more question for Zilch. I tried posting this last nite but must have done something wrong as it did not post. Anyway here is the question.
zilch you mentioned that it is current practice to remove the .01 bypass caps when using solens as I have done on one cabinet so far. I would like to ask what I can expect in the difference in sound versus leaving them in. Also are you familiar with the Jensen caps and how would you rate them. Thanks

07-06-2006, 06:48 PM
So many questions, so little time. I have yet a few more questions directed at anyone who will answer them. I realize the question is off thread as it regards JBL L19's. I also have a pair of L19's I found in the junk some months ago. The woofers are pretty much trashed, tweeters seem to be ok.
I planned at that time to get the woofers reconed. Would like to know what the model# for both the woofers and the tweeters are and any other relevant info anyone can provide. Are recone kits available for these woofers? The cabinets are in near perfect condition, no dings, no scratches. I tested them when I found them and everything worked but as I mentioned the woofers surrounds look like they were torn off and whoever had them had placed some duct tape around the cones. Some repair job eh?

07-06-2006, 06:49 PM
I believe it's in one of the threads here that using bypass with Solens (or, at least, "Fast Caps,") makes the VHF sound "edgy," or some such.

Giskard reported that JBL is using unbypassed Solens in its latest crossover designs, and slapped me silly when I suggested I was gonna bypass them anyway in "Keeper" crossovers. :p

I didn't; Thetas are hella expensive. That's where I was gonna see if I could hear a difference, so it could still happen, but I can build with Solens only (no bypass) cheaper than using Daytons w/Thetas, my prior standard approach.

Jean (B&K Man,) who lives near Solen in Canada may be able to find out (or, perhaps, already knows,) about the different cap lines Solen offers.

I have no experience with Jensen caps.

Look up L19 in the Consumer Tech Sheet links section at the bottom of the main forum page....

07-07-2006, 10:40 AM

I'll try your suggestions and thanks for taking the time to answer my questions.

07-08-2006, 02:01 PM
Old 4312,

I recieved this letter via e-mail a few days ago. I just recently purchased a pair of 128H's on e-bay auction. Apparently this gentleman felt inclined to write me after watching the auction himself. I think you will find some of his comments useful regarding the problems you noticed with your 4312's bass behavior.
From Dan Marshall,

The 128Hs are much better suited for use in small bookshelf enclosures; say
a bit less than 2 cu ft. The 2213s and the similar 123-x variants require
about six cu ft enclosures for smooth, extended bass response. Loading them
in smaller (less than 2 cu ft) enclosures results in a hump in the response
w/o much bass extension. I am not sure why JBL did this in the first place;
they ought to have known better. You might have to tweak the enclosure vent
tuning a little to optimize the bass response.

If you aren't already familiar with the on-line (free) speaker design
software from Linear Team, you can easily find it with a Google search. It
allows you to plug in the speaker Thiele-Small parameters and it will give
you the optimum enclosure size and vent frequency. The 4312 enclosures you
have will be about the right size for 128Hs. If you have an audio
oscillator it is easy to check and optimize the vent tuning frequency. You
can also plug in the T-S parameters and your box volume and then experiment
with optimizing the vent tuning frequency by noting the effect on the
frequency response.

In the final analysis though, somewhat better results can oft times be had
by tuning and measuring (or listening), as most speaker design software
doesn't take Ql into account (nor the room boost), both of which can have a
considerable effect on low end extension. I made a spreadsheet using the
T-S design equations back in the mid nineties that took Ql into account and
plotted the freq resp for eight integral values of Ql (4 through 11). The
effect of Ql on bass extension is quite significant, though it is ignored
(assumed to be about 7) in most programs. Perhaps you are already familiar
with all this, so I won't pontificate in too great detail.

My eldest kid has a studio and we are launching a record company, though I
am mostly a bystander supplying the financing. He uses L100s for mix down,
which are like the 4311s. We just bought a lot of new gear, the latest and
greatest stuff including a new dual core AMD computer and the newer version
of Cubase and a bunch of other just-released equipment. He can do something
like 100 tracks and keep everything in a format that is easily manipulated
for the final mix down before going to digital audio and can send these
files directly to the record producer, or some such. He could tell you a
lot more about it than I can. He has made a couple CDs and intends to make
more right away and is set up to go live in the San Francisco area within a
few weeks. They are apparently in an 80s revival there now.

He is pretty big on low bass (a chip off the old block). We bought a
horn-loaded sub by Tom Danley that is quite something. It handles a true
kilowatt rms at an efficiency of 103 dB/watt on down into the low thirties
at full tilt (133 dB Max SPL). It is being driven by a QSC 3402 amp which
is capable of 3,400 watts bridged into 8 ohms. Alone, it makes a pair of
the big dual 18 inch JBL subs sound totally sick by comparison. All they do
is thump with little on the bottom end. He is now bugging me for a new pair
of mains. I have gathered together some deluxe components, now all I need
is the energy to construct some enclosures. Better go for now.


This may or may not answer some of your concerns. This gentlman was pretty cool in sending me an e-mail just trying to be helpful.

07-08-2006, 02:12 PM

Just wanted to let you know I followed your advice. Yesterday I pulled the crossovers out of the right side 4312. I proceeded to remove the bypass caps and replaced the larger 3uf and 8uf caps with Solens.
Here is what I found out. When A/B ing the two cabinets, one with bypass caps and one without, there is a noticeable difference, particularly in the performance of the midrange frequencies. As you mentioned it is strictly a matter of preferences. It seemed to me the cab without the bypass caps was a bit more edgy in the mids, and that is not necessarily a bad thing, it also seemed to create a bit more depth in the mix. My next move is to take out the bypass caps in the left cabinet so both are without bypass caps. I will run them that way for a while to get used to them and make a determination at that point whether to put Theta's in or not. Thanks for the tip

07-08-2006, 02:27 PM
The 128Hs are much better suited for use in small bookshelf enclosures; say a bit less than 2 cu ft. The 2213s and the similar 123-x variants require about six cu ft enclosures for smooth, extended bass response. Loading them in smaller (less than 2 cu ft) enclosures results in a hump in the response w/o much bass extension.

I am not sure why JBL did this in the first place;
they ought to have known better. You might have to tweak the enclosure vent tuning a little to optimize the bass response. To make the most saleable and successful loudspeaker of all time, of course. Buyers were impressed by the boomy bass.

[And STILL are, apparently.... ;) ]

07-09-2006, 02:11 PM
Another question for anyone who can answer it. I guess I must have caught recap fever as I now intend to recap an old set of Cerwin-Vega V-33's. Fortunately I have the schematic for the crossovers for these that I obtained a few years ago from their tech via the internet.I know it is not a JBL question but just thought someone here might have some idea of this symbol denotes.
The question is this. There is one symbol on the schematic I cannot decipher. It looks like an hour glass. Actually like two triangles point to point and has a line through it with each end bent over in opposite directions. A friend thought it might denote two zener diodes which would make it a dual diode. As the zener diode symbol is very similar except it only has one triangle not two.
Does anyone here have any idea of what I am referring to? And if so what the heck is it?

Earl K
07-09-2006, 02:38 PM

- From your description,these sound like 2 zener diodes facing each other .

- In the schematic, are they wired in parallel across the horn driver circuit ?

- If so, then this is an old Klipsch, SR trick . ( If I recall correctly ) this was used in their industrial usage, La Scalas . The "zeners" open up past a certain voltage, dumping voltage to ground. A crude form of hard limiting with a bunch of sonic "chatter" as the ziodes close and recover . Not elegant sounding at all. Personally, ( if this is what they are ) I would disable them for HiFi usage .

<> regards

07-09-2006, 02:56 PM
Yup. Transient (or surge) suppressor, clamps (clips, passes) everything above the design voltage. Consider it a bidirectional Zener:

Earl K
07-09-2006, 04:14 PM
Thanks for the viusal, Zilch ! :)

07-09-2006, 08:08 PM
"To make the most saleable and successful loudspeaker of all time, of course. Buyers were impressed by the boomy bass.

[And STILL are, apparently.... ;) ]"

I bought mine because back in '83-84, they were the only speakers I could find that didn't sound like they had a bedsheet over them, and didn't have a nasty resonance when playing "Guitar Etude No.3" from "Twin Sons of Different Mothers".
What can I say, I was still a teenager...:rotfl:
To sonichermit: thanks for the info.

If possible, I'm going to attempt the skyward test next weekend.
This should be the definitive test to verify ceiling/floor boundary interference.
Later, everyone!

07-15-2006, 05:22 PM

Yup, that is the symbol alright. I did a bit of research on it as well and asked an electronics tech that has worked on my studio gear. He called it a transorb, is that essentially what we are talking about?
Would you recommend removing it? I have completed replacing the caps in one of the V-33's with Solens. It did make an audible difference. Granted these are merely PA cabinets from way back when but I wanted to do a little experimenting just for the fun of it. They do after all have "horns" for midrange which will never be smooth and are of an old design. Much has changed since they were manufactured. But in any case recapping has made them sound a bit more tolerable.
If you think removing the zener diode will made any type of audible difference it would be a simple matter to merely snip off the leads and pull it out. Thanks for answering my incessant questions.

07-15-2006, 08:52 PM
If it's wired as a protection device, as Earl suggests, disconnect one end of it and see if you hear a difference.

Then, if you resell them, you can reconnect.

Depending upon how they have been used in the past, the device could become leaky and audible.

07-19-2006, 04:26 PM

I'll try that and see if there is any difference. So far putting the Solens in the old Vegas (one that is done so far) it does sound audibly better than the other that has the old caps in it.I'm waiting on one part before I can recap the second one. The bottom end is deeper, mids sound a bit less harsh and more defined. The cabs have 18" woofers and they can really put out some bottom end and for PA cabs the bass is pretty tight as well. I have owned them since 1977 so I guess I got my money out of them and then some.
By the way I removed the bypass caps out of the remaining 4312 so they are both unbypassed right now. Frankly I think they sound better now then they ever did before. I am very pleased. Thanks for all the help.

07-23-2006, 08:35 AM
Hello everyone!
Sorry about being a week late.
As promised, here's the skyward test results.
As you can see, the dip only moved upward in frequency because of boundary removal. It's still there, but it's now at 186 Hz rather than it's usual 126Hz. The low end looks even worse than usual.:(
All test parameters are the same, except that the speaker is on the ground facing upward. As usual, the straight line at the top of the graph, is the frequency response of the input signal to the amp.
Since the distortion traces looked great, I'm assuming this is either a cabinet resonance or magnet issue or design flaw.
Thoughts and opinions are appreciated.
Thanks everyone!

07-23-2006, 10:15 AM
The fact that the frequency changed indicates you're on the right track. I wouldn't be suspecting the speakers.

Wavelength at 186 Hz is ~6 Ft., so 1/2 wave is 36". What differential path lengths to the microphone equate to that?

Was it on its back on the ground?

If so, and you elevate it, like on a crate or small table, keep the same mic distance from the driver, and the notch frequency goes down, then it's bounce from the rear boundary, now.

The literature is replete with the difficulties of measuring low-frequency response due to such effects. There's four ways I know of to get it, and they each produce a different result:

1) Anechoic

2) Ground Plane

3) Nearfield

4) In-box

Sounds like you're in a good situation to experiment with ground-plane versus nearfield measurements....

07-24-2006, 08:02 AM
Was it on its back on the ground?:yes:

"As promised here's (sic) the skyward results..."

07-25-2006, 07:47 AM
The answer to many acoustic testing questions was already here and I just didn't know where to find it!
That's how they're doin' it.
That's how they remove the rear boundary problems as well.
Once again, thanks Zilch for the advice on that one!
So,if I understand this right,you have to build a 30' x 40' wall...:hmm:
I think it would be less expensive to soffit mount the 4312's in the room.
There's less lumber that way.:D
If I had access to scaffolding, I could raise the speaker up 16 feet outside (facing upward), and see what I get. 16 feet is approx. half wavelength of 35 Hz.If the dip moved down to that range, I would consider this resolved.
Zilch, could you please give me more info on ground plane testing?
I guess this is the next step for me to follow.
A JBL tech once mentioned this to me, but didn't go into details about how this works.
I've already done the 2 inch to cone tests, so I guess that qualifies as nearfield testing.
It appears that I'm reaching the limits of what I can verify, yet again..
I've been working on and trying to resolve this problem off and on for over 20 years now!
Thanks to all of you, I'm finally grasping the reality of the situation:
It takes a lot of effort and even more cash to get solid low end response in a listening room.
Where there's a will,there's a way.
Till then,thank you Sony, for 7506 headphones.
Thanks to all of you as well, for the advice.
I'm listening and learning a great deal not only about loudspeakers, but also the physics of sound.
Hopefully, other newbies like me are as well...

07-25-2006, 12:38 PM
Zilch, could you please give me more info on ground plane testing?At this point, I'd say you need to spring for a copy of D'Appolito's Testing Loudspeakers. There's a section describing the theory and method, as well a low-frequency response plots very similar to yours. ;)

Interestingly, the "Hole-in-the-ground" approach to LF testing got JBL into some mid-bass problems when used for speaker design. You'll find that documented Re: L212 here on the LH main site.

Nearfield is also described by D'Appolito. Variously, in the literature, the requisite distance is 1/4", 1", or anywhere closer than one driver diameter. :p

It works best with closed boxes; otherwise you have to separately measure the port output and sum that with the driver's, corrected for the size differential. That's too much work for the Zilchster, so I'm looking into in-box measurement per Small for vented boxes. Mic wire conveniently comes out via the port. :)

Once the speaker is in a listening room, all bets are off, as you have discovered, and there's a HUGE literature dealing with room effects. Toole has a paper on subwoofers in rooms on the JBL Consumer website, and his article in the June AES Journal covers the whole spectrum well. Recommended reading, with lotsa good references for further inquiry, too. :thmbsup:

Tom Loizeaux
07-25-2006, 03:00 PM
I find this "searching for the truth about 4312s" quite interesting, as I have a pair of these and would like to know what they do technically. If the "problem" is, in fact, a design flaw, then we should all appreciate being informed.
I do want to say that I just spent over an hour sitting in front of my 4312As. They sit on top of my 4343s, so they're slightly above my head when seated. As I've explained before, I add a very slight bit of help from a small subwoofer (JBL PB12) to fatten up the very low end. I don't know how techically perfect these 4312As are, but they sure sound great! If I didn't have my 4343s, these would make me completely happy. Great low end, warm, rich lower-mids, clear upper-mids and a nice, crisp top end.
When I switch over to the 4343s (which are bi-amped) and match the listening levels, I get more a little more beef in the low end, maybe a touch more smoothness in the low-mids, and a bit more clarity in voices and upper-mid instruments.
I don't notice and dips or areas lacking in the 4312s. To me it's a matter of refinement and the ease at which the 4343s play that separates these two designs.


07-25-2006, 03:05 PM
Hello Zilch!
Thank you very much for the tip!
I'm off to Amazon, then on to JBL consumer.
It's nice to know I'm not the only one who has wrestled with this!
I sure wish I had conversed with you about all this 20 years ago!
I'd probably be blasting instead of typing..:D
Thank you for all the help!

07-25-2006, 03:54 PM
It's nice to know I'm not the only one who has wrestled with this!Reviewing the thread, your nearfield LF measurement between woofer and port (Post #26) seems to provide the best characterization thus far. Next time you're set up measuring, repeat that, but then close the port with a plumbers' test plug and measure again. I believe you'll find the difference to be, well, "dramatic."

When you work with groundplane, please take pics for the forum of your test setup. I don't believe anyone has ever show how to do it here. It's not very difficult; outside away from walls or structures will give the best results, as long as there's no railroads, freeways, garbage trucks, or earthquakes to interfere.... :p

07-26-2006, 04:58 AM
Is this in the neighborhood of what we are talking about?
I can handle that.

Originally posted by Zilch:
It's not very difficult; outside away from walls or structures will give the best results, as long as there's no railroads, freeways, garbage trucks, or earthquakes to interfere....:p
Naw..Maybe the sound of a few freaked out squirrels hitting the ground--but I can always retest around the thuds.:rotfl:
To do it on a fairly flat surface, I'm gonna need more cables.
Also, I don't have a camera. I don't know if those disposable cameras would work for these kind of photos.
I'll look into it...
Later everyone!

07-26-2006, 06:30 AM
G'day! I may be taking this sideways... but I something I saw in an earlier post resonated(:p) with me. With my 4410s, I went through about 4 amplifiers until I found one that would drive them "properly". By properly, I mean controlling bass that was audibly unpleasant (I never did measure the response). I mean really unpleasant - resonating and floppy.

I finally found a build-it-yourself kit MOSFET power amp (185W per channel) that sounds awesome. It's defining characteristic is that it has a HUGE damping factor (over 1000 as I recall). After checking the other amplifiers that didn't sound good, their damping factors were in the low 100s. This seems to be a characteristic of JBLs - great bass but only if you use the right amp.

Ever since this experience, I've used amplifiers that have very high transient current output devices which give them high damping factors - this is the only way I've found to control the bass response. I've been using this approach in my live sound production as well - I drive an enclosure with two 2241H drivers with a bridged Crown CE2000 amp and it controls the bass quite tightly (the damping factor is rated at 400 but seems to be higher to my ear).

Hope this hasn't been too far off topic...

07-26-2006, 09:20 AM
...it has a HUGE damping factor (over 1000 as I recall). After checking the other amplifiers that didn't sound good, their damping factors were in the low 100s. This seems to be a characteristic of JBLs - great bass but only if you use the right amp.Not really off-topic, at all. It is certainly worth discussing.

I too, long-ago, awoke to the importance of output impedance / damping factor in bass response. Ironically, the cabinets I most struggled with this were my 4312's... :hmm: That is, before I swapped-in the 128H's and got a real amp into the signal path.

If the amp is struggling with "back EMF", it is woefully less efficient. JBL certainly knows this as evidenced by their new line of differential-drive woofer motors which improve the amp's chance of efficient operation.

Hey "old_4312" - have you tried different amps, and what is the damping factor of the amp you are currently using?

07-26-2006, 01:51 PM
I finally found a build-it-yourself kit MOSFET power amp (185W per channel) that sounds awesome. It's defining characteristic is that it has a HUGE damping factor (over 1000 as I recall).Please start a thread on this amp over in "Lansing Related Gear."

It's likely members want to know more about it.... :yes:

Is this in the neighborhood of what we are talking about?Yup.

I don't know how much "flat" is actually required outside of the immediate measurement area.

I'm thinkin' the driveway may be good enough for this, as long as no air-conditioning compressors are running nearby. You'll be able to read the ambient noise baseline immediately.

I doubt turf is gonna reflect much in the area of interest....

07-27-2006, 01:19 AM
Hello everyone!

Originally posted by BOPUTNAM:
Hey "old_4312" - have you tried different amps, and what is the damping factor of the amp you are currently using?
The current amp is a Yamaha P2250.It has a damping factor of >110 up to 1kHz.
I've tried listening evaluations with a Crown Microtech,UREI 6290 and a Mackie M1400 as well.
The better amps with higher DF, made the low end sound cleaner and more punchy, but thinner.
If I ever fix this problem, I suspect I will do another A/B test with a Mackie to see what the end result is. Without the canyon in bass response, I will most likely prefer the Mackie. There is some resonance I hear with the Yamaha, that I don't hear with the Mackie.There is also a resonance I hear around 270Hz that doesn't go away regardless of the amp.
It seems to be more pronounced on lower register female vocals than anything else. For anyone who's curious, listen to track 7 on Trisha Yearwood's "Where Your Road Leads". When she sings the words, "you mean to me", in the first line, that's the range where my 4312's howl a little bit.It doesn't seem as noticeable with male vocals or music beds.
I'll deal with that after I resolve the bigger problem!:)
If the groundplane tests reveal the 4312's are ok, and hopefully they will,then I'm really gonna have a tiger by the tail.
The whole front end of my listening room will have to be overhauled!
I've never built bass traps for the floor.:blink:
I'll post the groundplane test results as soon as I can run them.
If I can find a camera that will work, I'll include photos as well.
Question for Zilch:Should I run the nearfield test at the time of the groundplane test for comparison, or should I elevate or position the speaker skyward for the nearfield test?
Thanks everyone!

07-27-2006, 01:54 PM
Question for Zilch:Should I run the nearfield test at the time of the groundplane test for comparison, or should I elevate or position the speaker skyward for the nearfield test?Just elevate it off the ground. That's the virtue of nearfield: the mic's so close that no reflections from anything else have opportunity to appear, so skyward is not necessary. Just get it up off the ground.

Groundplane actually USES the reflection to create a virtual second driver below the ground and acoustically sums the two.

FUN, huh? :D

07-27-2006, 09:50 PM

Originally posted by Zilch:
Groundplane actually USES the reflection to create a virtual second driver below the ground and acoustically sums the two.
Thanks Zilch for explaining that!
I was wondering why the mic had to placed 2X the original free field mic distance.
If I understand this right,the virtual driver like a real driver, doubles the SPL, so the mic must be placed back twice the free field reference position to compensate for the 3db rise in pressure.
Just for grins, in addition to the nearfield measurements, I'm also going to repeat the elevated 1 meter away test with the ground boundary only--with no rear wall interference.That will show exactly what the floor/ground boundary alone is doing to the acoustic output. The on the ground skyward test is showing the effects of the rear boundary.
I hope the next set of tests finally solves this mystery--and my neighbors won't stop talking to me!:D
I'm ordering the book Zilch mentioned in a few days.
For anyone who's interested, it's available at Amazon.
Thanks everyone!

07-30-2006, 08:16 AM
Hello everyone!
Ran tests this morning. Per Zilch's request, I purchased a disposable zoom camera and took photos of the tests.
Groundplane results weren't the same as the JBL specs, but they looked much better.
As soon as I can get these developed, I'll post pics and all.
I'm very interested in getting the group's opinions on the test results.
Later everyone!

07-30-2006, 12:04 PM
Groundplane results weren't the same as the JBL specs, but they looked much better.
Yup, each method gives different results. See D'Appolito Fig. 4.23.

Ground plane should approximate the rising response (above 100 Hz) of anechoic. Halfspace per JBL would be flatter, according to that. Most importantly, the boundary effects should be gone.

Thank you for doing these experiments. I look forwared to seeing your groundplane results, and what happened when you closed the port(s). :thmbsup:

08-02-2006, 12:45 AM
Originally posted by Zilch:

Yup, each method gives different results. See D'Appolito Fig. 4.23.

Ground plane should approximate the rising response (above 100 Hz) of anechoic. Halfspace per JBL would be flatter, according to that. Most importantly, the boundary effects should be gone.

Thank you for doing these experiments. I look forwared to seeing your groundplane results, and what happened when you closed the port(s). :thmbsup:
No problem! I'd like to thank everyone here for their help in diagnosing this problem.A special thanks to Zilch for explaining groundplane testing and for the help on isolating the different effects of boundary interference.
It's been a real learning experience for me!
The response graphs of each isolated boundary makes it all pretty clear.
I hope others will find this information useful as well.
Providing the photo's are useable,would the group prefer the photos and tests in pairs, or posted seperately?
I should be posting them late Wednesday night.
Later everyone!

08-03-2006, 02:37 AM
Well the photos are not great, but that's the cost of using a disposable camera,I guess.
I've tweaked them as best I could.
There should be enough image on each one,to see what I was doing.
I'll post the photos and tests in groups.
I believe the mystery is solved, finally after 20+ years,but I would like to hear any comments anyone might like to add.
Thanks everyone!:)

08-03-2006, 02:51 AM
Acoustic tests continued..

08-03-2006, 03:12 AM
Last page.
Two reposts:
4312 flush on ground, mic'ed @ 1 meter, skyward measurement, rear boundary interference only,reposted for comparison.
4312,elevated 29", mic'ed @ 1 meter, flush against brick wall. Rear and ground boundaries effecting acoustic output, reposted for comparison.

08-03-2006, 10:39 AM
If I'm reading your results correctly, groundplane max and min are 89.5 and 83.5, i.e., +/- 3 dB, better than any results I've ever obtained measuring LF conventionally. The technique is credited to Mark Gander at JBL in Stereophile 28:1, January, 2005, P. 68.

I'd call it most excellent work, Old_4312. May we have a summary of your findings, then, regarding LF measurements? Some readers may not get what's going on otherwise.

Also, tell us about the measurement software you're using, please.

Tip: See D'Appolito Chapter 4 regarding loudspeaker measurement microphones.

08-04-2006, 12:10 PM
Hello everyone!
Originally posted by Zilch:

May we have a summary of your findings, then, regarding LF measurements? Some readers may not get what's going on otherwise.

I learned so much from these tests, I'm not sure where to start with a summary.
I'll do the best I can-forgive me if I get longwinded.
Lessons learned:
1. Unless you perform nearfield testing, it is impractical, if not sometimes impossible to do meaningful LF testing of loudspeakers in a less than perfect acoustic space, unless your objective is to see the result of the marriage of those speakers to that space.:blink:
2. All boundaries have an impact on LF response of a loudspeaker.
3. When 2 or more boundaries are acting on a loudspeaker system,
their effect is cumilative. If these boundaries are far enough apart, they can introduce more than one large dip in frequency response.
If they are identical, they will create a significant dip in response, as much as a 20 db loss. If the half wave length frequencies of the two are close in frequency, they actually can widen the "Q" of the dip, effecting even a wider number of bass frequencies.
Check out the "4312 outside" response graph, and compare it to the skyward test and the 1 meter floor/ground boundary tests, which are both one boundary tests only. The "4312 outside" test shows the effect of two similar, but not identical boundaries workin' my poor 4312 over. That's one whopper of a dip there! This is pretty much what I'm seeing in my kinda sortof treated room.:(
*Note* Be sure and observe the difference in db's per line. In some tests, the difference is 3db per section, others it is 6db.
4. Groundplane testing in a wide open environment, with no vertical structures closer than 40', yields acceptable LF response measurements.
The closest boundary in my case was 43 feet.
5. The null points and depth of bass loss in db's, caused by room boundaries, are not fixed parameters. They actually vary in frequency and depth depending upon the distance and position of the microphone/listener from the loudspeaker. Refer to the 2 floor/ground boundary tests in my last posts.
The only difference in these 2 tests, is the distance of the mic in relationship to the speaker. Everything else is the same.
Changing microphone/listening positions can have a significant effect on measured/perceived bass response.
6. No one room boundary appears to be more or less important than another.
Addressing one or two acoustic problems in a room, does not a accurate listening/monitoring room make!
7.It takes much planning, considerable construction, and a great big wad of cash to even get close to published frequency response curves or "grade A" studio quality monitoring to happen in an enclosed room.
To quote Seinfeld: "Good luck with ALL THAT!!"
Unless you're able to address and correct every boundary problem in a room, including floor and ceiling, don't expect perfect bass response.
Carefully placed bass traps, the removal of right angles,etc.. is not enough to get the "Holy Grail".
Whew!! I think I may have rambled too much there.
I did apologize for getting long winded, right?
I think that's the gist of it.
If anyone has questions, I'll do my best to answer them.
Zilch or others here in the forum can probably do that better than I can...
One last lesson: If you have a loudspeaker or acoustics problem, this is a great place to come for help!!!
It sure helped me to see the light of day!
In the next post, I'll describe the software, "ETF5. " I've been using for the tests. It's a great, affordable piece of software. It does have limitations and certain hardware requirements, though.More on that later.
Thanks everyone!

08-06-2006, 05:32 AM
Hello everyone!
Per Zilch's request, here's some info on the software, ETF 5.XXX, I've been using for these tests.
If you want to download it, it's available as a demo or full version from www.etfacoustic.com (http://www.etfacoustic.com)
The demo is the same as the full version except:1. You can't save files with it, and 2. It will only perform 2 analysis per launch.
If you need more than 2, all you have to do is re-launch it and start over
---not reboot.
It does however, have a sequential test function (by clicking the red arrow at the right of the analysis screen's toolbar),once an analysis has been run.(This also works in the demo.)
You can then click the "play" button for manual one shot updates, where the program will repeat the noise burst, then display the result, or you can set it to loop mode, and it will consecutively repeat the noise bursts and display any changes made in the measurement as you change speaker/mic positioning, eq changes etc..., just like a real RTA.
What I like most about the program, is that it calibrates itself, so you don't have to burn up brain cells calculating the differences in decibels between two noise signals across the freq. range, or go through the task of trying to calibrate a hardware RTA to its own signal, dealing with crest factor and all that crap.
To use the program, you must have a full duplex soundcard.
Also, if the soundcard has a DSP mixer, and direct monitoring capability,
you have to mute the monitoring of inputs one and two, before you make the soundcard connections,or you will get a loop, which might damage the soundcard. It doesn't mention this in the manual-so beware!http://audioheritage.org/vbulletin/images/icons/icon4.gif
When you read how the soundcard is wired up for this thing, you'll understand why;Channel one output is connected to channel one input, and ditto for channel two, for initial soundcard tests.
Then, the channel 2 output is disconnected from the #2 input and connects to the signal path feeding the speaker, and the channel 2 input receives a line level microphone signal measuring the speaker.
This allows the program to calibrate itself to it's signal, and any variances in soundcard freq.response. If you look at the graphs I posted, that line at the top, is the reference noise signal + innerchannel soundcard freq. deviations.
For me, the trickiest part of the program was learning to always set the gating function up right. Too short a gate time, and low frequences aren't displayed, too long a gate time will add too much room noise. Around 49 ms seem to work the best for low freq.response measurements.
The manual is pretty thorough, and explains everything fairly well.
The software also offers impulse testing, with either graph display or waterfall displays for identifying speaker or room resonances.
*The soundcard must pass the initial soundcard impulse tests with flying colors to utilize the impulse testing feature.*
The more spurious noise and ringing your soundcard generates, the harder it is to discern resonance from soundcard artifacts. The manual explains this quite well. My 20 bit Layla barely squeaks by, but that's possibly due to driver issues, since it seems to works best with DirectX7 and it's own ASIO drivers.
Newer, high end soundcards will not likely have any problems.
I don't know what type of drivers this software uses.
Acoustisoft may have soundcard recommendations.
The only thing I don't like about the program, is that you can't custom select the frequency co-ordinates. You have to choose from a predefined selection from a drop down dialog box. If you're trying to zero in on a specific freq. range, that can get a little annoying--especially if the db range changes per range selection.
Once you do the initial soundcard testing, it's not necessary to repeat that test everytime you launch the program. If you're using the demo version, this soundcard test will eat up one your two allowed analysis.
If your soundcard will allow you to save sessions or setups, it's good idea to save 1 setup just for ETF5, after the soundcard is wired up and all the levels adjusted.
If you're using the demo, don't forget about the sequential RTA option!
If you need to make constant adjustments while observing the result, that's the way to go.
I believe the cost of one license is 40 bucks...that info is available at the website.
The software can also perform even more advanced testing such as psychological testing(for speaker placement), PSD testing(for noisy environments), and IMD and HD distortion tests as well--but these are available only as individually purchased plug-ins.
One little trick: I've been capturing snapshots of tests by using the print screen key, and then pasting the clipboard in a jpeg editor/viewer.
That's not the same as saving a test, then reloading it into the program and continuing to work with it, but it's a simple way to store a specific one time result.;)
That's about it for my ramblings on the subject.:D
The manual explains all the rest.
Hope this has been helpful.
I've received so much help here, I hope I've given a little back!
Later everyone!

08-06-2006, 07:37 AM
I think the site is actually www.acoustisoft.com - 'looks interesting!


08-06-2006, 12:00 PM
Hello everyone!
I goofed on the url, originally.:o:
I corrected the link in the original post.
The url works now.Actually, both listed url's will work.
Thanks,Johnaec for bringing this to my attention!
Hope this didn't cause too many problems...
I'll watch my fingers a little closer next time.
Later everyone!

08-06-2006, 01:13 PM
You did all this with FREE software?

[How embarassing.... :thmbsup: ]

08-08-2006, 05:56 AM
Except for the one limitation, it's a great piece of software.
After I buy the "Testing Loudspeakers" book, I plan on buying a license.
There are advantages to saving and reloading analysis files for further scrutiny.
One license is cheaper than a one day Goldline 31 band RTA/AKG ref. mic rental--and I'd rather use the ETF5.:)
Later everyone!