4 dual opposed SSD 15s - Page 5

Bosso,

You may want to think about how you put your e-mail address up there so you do not get tons of spam hitting it.

James

To add to the joy of your teasers, bosso, might I also inquire as to why you made no mention of a 15" driver, as that's the one you seem to prefer for your builds?

Once .mac disappears (end of June), I'll have a new e-mail account, web sites, etc.


The 15s are all mine. All new basket, surround shape, surround material, triple shorting rings, just enough motor and just enough throw. I figured I may as well have a driver that suits my design goals vs using the latest whatever and adapting the system to it. So, no sharing.

OTOH, my design philosophy is the simplest way from point A to point B. I'm not gonna blather on about "proprietary" this and "exclusive" that. The basic platform is a standard overhung with some aluminum strategically placed to lower 2HD and inductance (flux mod), enough manageable throw, well built with good mats for the $$. Since I don't use 18s and the DIYers seem to prefer them, I will probably offer the idea and see if there's any interest after I'm finished testing the platform and parts combination.

There's always the chance that my design will be an epic fail rather than a magic potion.

I actually have guys like Husker in mind. I've been there myself, waiting for an eternity for the latest DIY driver-of-choice. Being the guinea pig and waiting, waiting, waiting. They finally arrive, you build, you like, they disappear forever. It's always been that way in DIY and it's always been a royal PITA.

Blueprint, Adire, Exodus, Stryke, AE, TC Sounds, CSS (SDX), Avalanche, Destijl (Fi), et al. Bummer.


Yes, I had the same reaction when I first plugged the SSDs into the system, using just 4 of them, then scaling the system up to 16, then settling on 8. It's a very compact system with serious firepower for relatively little cash.

I agree (and so has everyone who's heard my systems over the years) that the SSD driver adds a certain percussive nuance that is unique in my experience and certainly seems to fly in the face of all the "Inductance! Oh noes!!" talk.

I've kicked, pushed, prodded, dangled and dumped, throwing everything modern digital formats have at them. And, I haven't nailed down exactly why, they just sound good. I'm glad you agree, but I was pretty darned sure you would (or I wouldn't ever have suggested the idea to anyone).

BTW, the elegant solution to stacking is outstanding and your spacing is just about perfect, (IMHO, YMMV, etc). Awesome with the ABS plumbing fittings. I love dual-use apps. Thumbs up.

Wait til you get a load of the system with the signal shaper in line. It truly is like have several subs to choose from without the hassle of having several subs. You can select a curve that brings the low end of that scene in Hulk to parity, or choose a curve that de-emphasizes that part of the BW and lets you push the system to its max above 20 Hz, which is impressive by any standard, and everything in between.

Anyway, glad to see you finally got up and running, and very happy to hear that your results are making it all worth the effort.

Bosso

Bosso,

That's a very convincing plot! Well done! The test engineer in me sees one more piece of information that might be very telling and should even be pretty easy to get. If you've still got them stacked up that way, could you run a sweep with the middle two units disconnected? All of the benefit for that 73Hz null happened with the addition of the last unit, so I wonder if it was a combination effect of all 4 units together or if the magic lies in having point sources at those two corners. If that works, the only other thing to check would be having ONLY the top unit playing, just to ensure that it's not in some kind of miracle spot all on its own.

Have you attempted a test like that? What are your thoughts?

matt

Actually, you can see an improvement going from 1 to 2 modules in the dip, which goes from -30dB to -22dB.

The final improvement is from 2 modules to 4 modules. If you disconnect the center 2 modules, the passive drivers would skew the results, so I didn't consider placing just 2 modules, one at the floor, one at the ceiling, as you suggest, because it would be mechanically difficult.

This is where others might lend a hand and experiment a bit, but I've found that that generally does not happen. Most people post the Harman papers 2D approach or the Geddes stuff and the majority follows.

Imagine if the Harman papers had combined a stack in one corner and a horizontal spacing across one wall. That would address the major problems in most every room. But then, Harman doesn't have a stackable system to sell.

In my sub experiences, spreading out 'x' number of subs across the floor space doesn't deal with floor-to-ceiling induced peaks/dips and is not doable in many rooms where floor space is dominated by typical furniture.

Columns can be designed to completely hide the fact that they're subwoofers and take up less floor space than the average high performance ID sub, so it seemed like a good idea to investigate to me. It will be cool if gpmbc gets to the measurements stage and we can compare results.

Bosso

Thanks for that. Very useful graph. It does look like the horizontal corner stack takes care of one big mode, but doesn't otherwise do much. Perhaps your idea of a corner stack and subs on the back wall would be an all-around great solution, comparable or even better than a Geddes-style placement.

One critique, if you don't mind, is that your scale over-emphasizes the first-mode region, where, as theory predicts and your graph shows, the subs pretty much just sum in phase. Likewise, it under-emphasizes the modal region, where multiple subwoofers have response-smoothing effect.

A combination of the full vertical stack with a proper Geddes style multi-sub system might be the ticket!

It might not be needed at all, depends on the room. I can get a flat response with minimal filters with all subs going across the front. It gets easier stacking in the corners in my room(front corners).

This is exactly why I've been sitting on my hands gents; if Dave is guiding the design of a raw driver with DIY opportunity, I'm all-in!!!

Dave,

Sign me up my friend!

Larry

Correct. Not all rooms and system configs are the same or will measure the same.

Just thought that it would be a cool combo though and work on all axis.

It might help to match the 1 module vs 4 modules traces:



I don't see them summing in phase. 10dB is too big a difference to make that generalization. Part of the exercise was not only investigating the elimination of the floor-to-ceiling mode, but the addition of a 4th full BW virtual point source vs the usual 3 virtual point source phenomenon from a floor placed corner placement. I see evidence of that in the comparison. Yes, it should be 'pretty much' an in-phase summing, but the influence of the added point source should not be discounted in the overall results.

Just to clarify, I picked the 'worst' corner in my room to emphasize the improvements of the configuration, not to show the best placement and delay/phase tweaks before any smoothing EQ might be employed (which is the SOP for most).

Here's a stack in the left front corner, which is a much better starting place in my room with the chosen seating positions:



I would love to see the differences with the stack as is in that graph + 3 modules spread across the front wall, but my 8' long console that houses my players/amps/AVR, etc, is in the way.

I've tried using a stack in one corner vs the other, both corners at once and breaking the stacks down of all possible separate placements in the front corners and actually, the front right corner stack is the best placement in my room with the seats located as they are.

Since this response is so outstanding as is with no post smoothing EQ, alternate placements would only be for experimentation in my case, and I've just had too many other things to investigate besides LP FR.

The whole point here is that rooms are all 3D spaces. We've experimented and posted about all the standard tweaks over the years, but all of them have been 2D options. I just wanted to break the ice with a 3D approach.

Bosso

Mostly, they do sum in phase in the first mode region, from my read of the graph. You have a pretty big room, if memory serves, so the modal region extends down a bit lower than most.


You mean "another 3D approach."

The Geddes methodology very much considers all three dimensions of the room. After all, the recommended starting placement is, "one in a corner, one far away from that one, and a third far away from both and above the room centerline.

Indeed, one of his two main criticisms of Welti/Devantier stem from their 2D orientation. (The other is that asymmetric placements were not investigated.)

hey bosso,

interesting news on the driver. i look forward to seeing what you have cooked up there. have you contacted the other guys who you reference in order to see why they failed? hopefully, your wheels won't get banged up by the same potholes as the other guys.

the closed form solution for the wave equation in a 3-d space is in the harman paper. their choice of 2d testing was one of praticality, not of ignorance.

while i agree 100% that there are benefits to vertical placement, 3d thinking is nothing new. geddes championed this idea a few years ago and it never took hold. the dearth of solutions is a matter of the market. the infinity irs had a stack of subs, but the market just wasn't there, so it died. if you are doing a custom system here or there, you may have something and folks would be well advised to listen.

even more advanced strategies such as using rear subs that are time aligned to absorb the front bass wave are also possible. with proper placement (1/4 up, 1/4 wide) and the right digital compensation, room effects will be almost completely eliminated, but that typically requires an un-home-friendly, expensive, and complex system.

if you really want to get this into folks homes, i would suggest dropping the 15's and use something like 8's or 10's in a slender vertical stack in each corner. that will accomplish what you are shooting for while providing a more friendly form factor.

I'm not a Geddes fan for several reasons, but the 'above the centerline' and '1/4 wide, 1/4 up' ideas both predate Geddes.

I know you've found value (in reading anyway, haven't seen your real-world results) the Harman 'papers'. I haven't. Nor will I be convinced otherwise by having it trotted out once every few months for another decade.

My thrust is not the mass market. That market is covered logarithmically.

Most every person who can afford a world class system goes to a dealer/installer and takes their recommendation. That means they invariably end up with a system that's well beneath their means and ends in an ignorance is bliss scenario.

If I had a nickel for every time I heard a person of means say something like "I just bought the best one they had.", I'd be retired. And, that applies to every product, not just audio.

Of course, lots of systems on the drawing board, that's part of the draw for me, but as it's become a cliche to say "It's all about compromise.", I say that's just a veiled admission to the holes in a system that's been designed by conventional wisdom with the primary goal of making money. I don't believe in the concept, nor do I see any compromises in my own system.

I think having options is the driver here. Stack, place separately, dial in full BW, dial in flat-to-20 Hz, wide F6-shallow roll off, narrow F6-steep roll off, tune to 30 Hz, tune to 10 Hz, scale to reference level in 6k cubes, scale to moderate level in 2k cubes, hidden in a dedicated space, prominent in a living space, etc.

There doesn't exist such a system. That's either because there is no demand, or because it's never been an option.

Bosso

I dig your style...Dude

http://www.madronadigital.com/Librar...imization.html

Give Yates the money, let the computers crunch the millions of numbers, and put your subs where he tells you. In an ideal world.....

I wonder what it costs to perform this service? My guess is $10,000 and higher.





Forin

nice data.



that doesn't look too shabby across six seats.

"If I had a nickel for every time I heard a person of means say something like "I just bought the best one they had.", I'd be retired. And, that applies to every product, not just audio."

i know where you are coming from and agree 100%.

It tickles me how you lap up the JBL stuff with no scrutiny whatsoever. You're looking at 2 octaves. Think there's any smoothing applied to those graphs?

This is how that little snapshot contrasts with what I usually look at, which is the full spectrum of the low freqs and an octave above to note the 'transition':



I applied smoothing to the 20-80 Hz slice and overlaid it onto the graph (the dashed trace) with no room correction applied.

If someone is willing to pay for that result, then I'm in the wrong business.

Bosso

What? Just go to the commercial threads. Hell, a single F-20 can get you those results for $250 and high spl's! The whole point to all these multiple sealed subs is not for headroom, it is needed for the low octaves to be produced correctly. I know they don't think the low end is needed but I keep telling them they have not heard it properly yet. To be honest, I have felt it many times but up until recently I really have not experienced it and it only happened when I boosted the low end. I used to run 10 dBs hot and still do when I don't boost the low end but when you add a LT or boost to the low end 5 dBs hot seems too much now, it really is incredible how that works out!

That Harman graph is just the difference between 6 seats. An overlay of your one graph is comparing apples to pork chops.

You know what we call a graph like the Harman one posted above in the engineering world? Useless.

Their data covered less than 10 units on the vertical axis, but they displayed 50 units worth of axis. You only display data like that if you don't know what you're doing or you're trying to deceive someone. Either way, it doesn't do a good job of conveying the pertinent information. And that was exactly bosso's point.

...or maybe it just illustrates the small variation among six seats to regular people that don't drive trains. Pretty easy to read to me, but again, I don't drive a train.

Hey, thanks for the heads up and all, as if it's my 1st day off the turnip truck, wit the fruits and meats.

I was using the graph LTD posted to make points you apparently missed. Let's scale the graph to a scale I use and am much more familiar with:



Now, let's go into the article and see the results of adding 3 subwoofers to the non-specific room in question, for a total of 4 subs and run the magic JBL room correction gadget (and, of course, the magic smoothing button on our measurement software), you know, so it'll be pork chop to pork chop:



So, buy 4 subs, place them according to some BS suggestion, run some auto EQ, so everyone can equally have a BS FR? The tight variance is a) BS and b) Irrelevant to what each listener will hear. Right, so what means is used to actually tame that terrible response? PEQ, JBL brand, of course.

So after all that Harman jargon and hardware, post the final seat variance and average FR for me, will ya?

This is just another Harman advertisement.

My graph has no smoothing, no auto EQ, no post EQ and is a single sub location. The variance from seat to seat requires no Harman magic wands and fees. And, my room is not a magic room, nor do I have a magic wand.

Take 20-80 Hz, be non-specific of 'seat' and 'room', set a nice tight scale, smooth the traces, place 4 subs in a non-descriptive placement relative to the non-descriptive 'seats', run auto-EQ, then apply post PEQ and then wrap it all up by just saying all is well, just buy a bunch of Harman stuff and pay a lot of money to people who buy a lot of Harman stuff and bingo, 15 or 20 dB of EQ cut across half the bandwidth in focus is, thank goodness, a simple thing and will definitely sound great.

BTW, just for the heck of it, post the sq ft that this article claims a seat fits into. Hopefully it isn't according to the goofy diagram used in the article from Welti's comic book version of a room, where 16 people (those are little heads with ears sticking out, yeah?) are crammed into a space that would barely fit 2 recliners.

Pork chops, yeah, got it.

Bosso

bosso,

i just assumed that they were applying the same smoothing as they did with the first plot (shown below). perhaps they are cheating by applying additional smoothing to the graphs, but i kind of doubt it.

as for jbl/harman, one reason that i tend to like the jbl/harman findings is that they often present their findings in papers or conferences where peers will shoot them down if they are wrong and/or cheating. their science seems pretty good and i can't think of a single other company that has contributed as much research into the public domain as they have.

"So, buy 4 subs, place them according to some BS suggestion, run some auto EQ, so everyone can equally have a BS FR? Right, so what means is used to actually tame that terrible response? PEQ, JBL brand, of course."

where are you going with that one man? ricci has independently verified the benefits of the suggested placement.

http://www.data-bass.com/data?page=content&id=80
[graphic near the bottom of the page]

Bosso that is a single point measurement from the main listening position right?

I would like to see some other people try a vertical corner stack and measure their results like you did with combinations of only the bottom operating, the whole stack, just top and bottom, middle etc. Hopefully Gmbc will. I would be elated with a response like that from a single corner placement but have never been able to get anything even remotely that good. I have never tried any subs higher than waist level though.