the ancient urge to build a dual-woofer three-way

Yeah, before we start throwing money at a new three way idea, let's back up a minute.

There are two advantages to going with DSP/local amps that have been driving me forward, the second one is much more compelling than the first:

1) Modest designs like Kite can deliver more performance in a smaller package and at a lower total system cost.

2) The frontier of the possible can be advanced, and by a big measure.

I hadn't even considered a third possibility that the technology could be used to build a better bedroom speaker, probably because I don't do that sort of listening, haven't had a stereo in my home in decades; have never had a reason to "dial in down".

To be sure, Kite's programming aggressively sets out to prove point #1 without much thought, if any, about low level work.

Perhaps the smart thing to do first is to make a program that sounds nice and pretty at 75dB and not care about how it does at high amplitude, maybe even prevent it from playing loudly because it'll likely extend the tweeter's band, making it more vulnerable. You do have four Presets to play with.

I can write the program, email it, and you can try it and see if you like it. This is one of the best things about DSP.

Meanwhile, I'm working on challenge #2. I do have customers who want to go all out.

If we think there is a market for a fully tricked out bedroom speaker we can do that too.

I guess it depends on what all out means. The system I was referring to earlier was the better part of $20k, for a large two way and a 15 watt Tube integrated amp. I'm guessing not many users will ever get over 90db with it, and will spend thousands of glorious hours listening to small jazz quartets and vocalists. Is a 3d, holographic representation of that performance any less All out then orchestral music at realistic levels? Or recreating the Who playing live?

Not an answer, perhaps. But just a different perspective.
 
Last edited:
Yeah, before we start throwing money at a new three way idea, let's back up a minute.

There are two advantages to going with DSP/local amps that have been driving me forward, the second one is much more compelling than the first:

1) Modest designs like Kite can deliver more performance in a smaller package and at a lower total system cost.

2) The frontier of the possible can be advanced, and by a big measure.

I hadn't even considered a third possibility that the technology could be used to build a better bedroom speaker, probably because I don't do that sort of listening, haven't had a stereo in my home in decades; have never had a reason to "dial in down".

To be sure, Kite's programming aggressively sets out to prove point #1 without much thought, if any, about low level work.

Perhaps the smart thing to do first is to make a program that sounds nice and pretty at 75dB and not care about how it does at high amplitude, maybe even prevent it from playing loudly because it'll likely extend the tweeter's band, making it more vulnerable. You do have four Presets to play with.

I can write the program, email it, and you can try it and see if you like it. This is one of the best things about DSP.

Meanwhile, I'm working on challenge #2. I do have customers who want to go all out.

If we think there is a market for a fully tricked out bedroom speaker we can do that too.

To answer the second part of what you were suggesting, I think a low level program would be a fantastic idea. I'd love to mess with that.

I do wonder what this approach would do for a really brilliant, lower level setup. It's the kind of thing a lot of the high efficiency guys like to do, but done in a totally different way. And no, I have no clue what the market would be for a speaker that can do some of that magic, but in a totally different way. Do I see the high efficiency/SET guys beating a path to your door? Nope. However, might there be guys who would appreciate that kind of performance, but in a way they never thought possible? Maybe. Be interesting to find out.

One thing I know your new approach does is that it totally messes with the scale we used to think was possible . Listening to the Kites, you'd swear they were a dramatically bigger speaker then they are. I'd love to know how this tech could cheat the common wisdom in different ways.
 
Raw naked protos are now running. Getting the chambers to seal from each other properly was a bit of a pita, held things up.

The stance and stature seem appealing. Next for some filter work, and see how things go. At this level of resolution there's precious little forgiveness for getting the filters right, so it'll take some time. But the first arbitrary crack at it sounds pretty decent - close enough to "useful" that I can hear a little timing issue between the tweeter and mid. There's no sense in fussing over the high and low passes until we set that right - you don't want to make up with amplitude what you're losing by timing mismatch induced cancellations. The passive guys are stuck with that screw-up - their hands are tied unless they physically offset the tweeter.

The teeny, tiny, little Predators with a 1x500W will stand in for the real thing below 30Hz while we build the big boys. We'll just have to put up with a modest limit on the very bottom maximum while we work to free ourselves from it. Also we'll make another set-up for just running the mains, alone.

That's a Black Bird atop the Predator. She's looking for a home.


20190918_111639.jpg
 
Very nice.
The design you are using scales very well.

Man, I would love to hear that. My not fully broken in Kites with the Predator sounded amazing. I'd love to hear what these do.

Btw, listening to a pair of speakers with Pat is a very interesting experience. He just clues into so much of what is going on. It's like watching a baseball game with Babe Ruth.
 
Wait! Did I miss something? When did you get the predator? Pics? Update your thread?

- Woody
Hi Woody -

We built the Predator as a "same footprint" sub/stand for Kite. The trick to arriving at a narrow enough baffle is an eight incher with stooopid big excursion. The idea is to be a "speaker stand" that can do bottom octave rather than merely taking up space. ((I'd imagine that, if one would take the time to learn the processor app, one could use them to augment any other bookshelf.))


While I'm building the big stack of ten inchers, the Predators are standing in (pun intended), doing the below 30Hz chores. Of course, when the chips are down the mains can outpace them, but I just played Roger Waters The Wall 2010-2013 at a belly-massaging level and found all of the extremely low and FAST stuff came thru quite pleasantly. Helicopter blade slap. There's a lot of hard hitting violence on that album that would otherwise not be evident. Shook the building; its' an old converted horse barn.....but still. Banged the stops a time or two but you can sure hear Water's intent to impress in the nether regions.

And we've just confirmed Hypex's claim of N-Core's ability to drive a 2 Ohm load. Nice for line arrays; use fewer amps.

The idea for the 'big boys' is that no driver will even come close to working hard. One ten per side with 500W is OK if you want "good". I'm aiming way above good - so four tens, 1KW. (2KW if you prefer.) FWIW - the N-Core is acquitting itself quite nicely in the bottom octave for tight grip. You know, most systems make nothing like fast bass in the bottom octave - more like a loud blur. Perhaps beauty at the bottom is the last thing you'd hand your money over for but, just me, it's totally worth it - you've gone the rest of the way, why not finish the job?

So far the mains have pleasantly surprised by being more than willing down to 30. So they're great on their own, most of us would call that "full range". It's kinda funny to just leave the need to do the bottom of the bottom. So, yeah, big boys to do 5-30. Tee, hee.

Here are the drivers in the mains:

20190916_114620.jpg

20190916_114702.jpg
 
Never said I got one. Just that I had listened to the combo. When I picked mine up, they were sitting atop the Predators.

Ahh... that makes sense. I love the idea of that combo, so if you had, I wanted impressions. Just thinking on the future, ya know?

- Woody
 
I've long been a fan of odd order Butterworths, stuck with single poles for passive work. Just me, it gets ya the most natural and realistic vocal presentation, cleanest transients, least electronic sounding.

We've been finding happy sound with five poles. To execute that for this midwoofer, with a small shelf to flatten the naturally rising response, in a passive we'd need this:

Five Pole Butterworth Bandpass Passive.jpg

Ouch my wallet! Who wants to pay for that thing? And who-the-hell would want to listen to it? Whaddya nuts?


Happy Days, now we hook up directly to the amp and just do this in the processor:

Five Pole Butterworth Bandpass DSP.jpg


Each little circle represents a biquad. A five pole Butterworth is a cascade of a single pole Q 0.5, and two two poles of Q 0.62 and Q 1.62. You do need to know the correct Q values to write the filter properly, but after that it's a snap.

Here, the issue of the day is the relationship of the midwoofer and tweeter as it effects female vocal. That's always the starting point. If you don't get that right nothing else you do will matter a whit.
 
I wish every Havenite could hear the magic that this all creates. How different it is from what they have heard before.

Should I load my speakers in the trunk and do a road show? :)
 
Rainy Saturday morning filter work, I can count on being left alone. Mic at listening postion. Looks like when you goose it, maybe a little rich in the midbass, eh? Sounds nice, anyway, pretty close to our target response. Solid bottom octave. It takes about 6dB at 15 to get there, so the demand for juice is really gettin' up there. Am running the 500W units on the subs, while waiting for the 1KW ones to arrive. Presently, with that much boost, we're clipping the peaks on rough program. You can only really notice it with the mains muted. But anyway, yeah, 1KW will buy 3dB, which oughta be sufficient.

We're settling in on a set of midwoof/mid/tweeter relationships. The process starts at female vocal, if good then male vocal, then piano, then sax, then every possible thing, all the while returning to the beginning following each stage, lest you've undone earlier work. It's a bit like squeezing a balloon, you fix one issue and another problem can pop up somewhere else. Our std set of torture tracks is getting flying colors and then some, especially a 24/192 female vocal we just added to the list. The idea is to look for trouble, so our various tracks each poke at a different set of problems. For instance, we use a female vocal that's right on the edge of shrieking; if you're hot in the upper mids or edging into cone cry, it'll jump right out. In another, the voice can easily be swamped, but should not. And so on. You have to pay attention to variations in focus or wander because they're clues to underlying issues.

It's easy work, really, with this hand full of Aces, parts wise. Usually you find yourself 'splitting a few babies', but this time not so much - we're pretty much getting what we want.


IBIS092919B.jpg
 
Rainy Saturday morning filter work, I can count on being left alone. Mic at listening postion. Looks like when you goose it, maybe a little rich in the midbass, eh? Sounds nice, anyway, pretty close to our target response. Solid bottom octave. It takes about 6dB at 15 to get there, so the demand for juice is really gettin' up there. Am running the 500W units on the subs, while waiting for the 1KW ones to arrive. Presently, with that much boost, we're clipping the peaks on rough program. You can only really notice it with the mains muted. But anyway, yeah, 1KW will buy 3dB, which oughta be sufficient.

We're settling in on a set of midwoof/mid/tweeter relationships. The process starts at female vocal, if good then male vocal, then piano, then sax, then every possible thing, all the while returning to the beginning following each stage, lest you've undone earlier work. It's a bit like squeezing a balloon, you fix one issue and another problem can pop up somewhere else. Our std set of torture tracks is getting flying colors and then some, especially a 24/192 female vocal we just added to the list. The idea is to look for trouble, so our various tracks each poke at a different set of problems. For instance, we use a female vocal that's right on the edge of shrieking; if you're hot in the upper mids or edging into cone cry, it'll jump right out. In another, the voice can easily be swamped, but should not. And so on. You have to pay attention to variations in focus or wander because they're clues to underlying issues.

It's easy work, really, with this hand full of Aces, parts wise. Usually you find yourself 'splitting a few babies', but this time not so much - we're pretty much getting what we want.


View attachment 16705

A little rich in the midbass sounds like a nice problem to have. Bet it makes for some lovely listening.

Man, I can't wait to hear these!
 
Yes! I challenge all comers to attempt to drain my beer fridge. Not possible. Bring your sleeping bag and give it a try. While your at it try to blow up my stereo. Ha!
Do we have a 2020 Haven annual gathering in the making? Three of the four main ingredients are present; fridge full of beer, great music/system and a good host! (We can always order the food in). 🍻🔉🛏
 
Back
Top