deepthoughts

Congrats on getting the new version launched. The ability to present the data in so many different ways is quite impressive. The only tool I could see as highly useful would be a way to more easily capture the displayed graph to post somewhere else. Fortunately screen clipping is much easier these days, but just a thought for all of those cases someone wants to share a graph or comparison.

Pennynike1 took a first step of stacking the Othorns in the double stacks (I wasn't there). Of course any stack over 6' warrants a picture: After some tinkering he currently spun 2 of the Terraform D18s around to slide into the outer cavities firing toward the front wall. Remember this all has a screen hung in front of it. Here was the last in process picture I got: I'll be curious to see how things are measuring when I visit next and see if we might be able to use the 2 remaining Othorns to good effect at the sides or rear of the room.

It is rather narrow at just under 13.5' wide, with a small divider middle depth and a depth of I think >30 feet.. There is a lot of gain, but the hurdle is really getting the sound to the seats. Just firing a 36" deep subwoofer with rear SpeakOn puts the front of the sub ~40" from the wall, which immediately sets up some potential issues around 85Hz. He was also seeing some peculiar behavior when the subs were aimed at the right wall vs left wall I expect due to the big bass trap along the top of the right wall. I'm hopeful that closing off the ceiling in the front of the room will dramatically improve that peculiarity. In my experience, getting the best coupling to the room and delivery of the frequency range to the listening position can easily make for more gains than 6 vs 8 boxes, and I think we can put the last pair to use at the rear of the room.

I getting a serious lack of creativity in implementation here. If you look at the front right corner, he is seriously limited in placement options there with the sump pump & well, so nothing can directly go in the front left corner. What's also not obvious is that the right cinder block wall is a mid-wall of the basement with the same size space on the opposite side. Complicating further, there is currently an opening the height of the floor joists at the top of the wall on the right running the full length of the room. He is going to close off the ceiling in front first, and then the rest as that will go a long way to make things more symmetrical. Yes, I already directed him to install high grade plywood so it might survive the high SPL. I was hoping someone had a model they could quickly adjust to model the horn extension and boundary loading. Fortunately Josh has the DXF files posted with the side profile. Some quick measurements and re-approximating the path with just 1 segment from front to back of driver gave me the 4th segment to experiment with as the extension/loading. The impedance and general shape are correct, with a little less energy modeled in the 40-60Hz range, but not much. I do understand how the tapped horns function vs conventional horn, as I was working with Tom Danley when he did the first tapped horn design which was to solve an application I was looking to fill. What started as a dual 12" with a high-flow port, quickly had the driver slide back in the port, and the port was lengthened into what we now know as a tapped horn. Early on Tom measured a few of the tapped horns in singles vs stacks at Sound Physics Labs and then the early models from Danley Sound Labs which showed that the performance generally tracked the models where added spacial loading often damps the upper bandwidth harmonics. Take most of the tapped horns and compare them with 4.0 Pi or 2.0 Pi loading vs say 0.50 Pi loading, and you will see the same trend. Even before we create any sort of extended pathway in front of the horn, stacking the subs in a large block, or creating the same spacial load with boundaries confines the expansion past the mouth to a 90 deg opening that forms a radius out to the edge of the box face. Modeling this to an added length of 24-30" beyond the effective length of the single box with an effective area of 6000-10,000 cm2 shows very significant changes to the horn's response, especially in smoothness and broader output gain. The effect is most beneficial at smaller spacial angles. Looking at firing the sub into the other stack of subs or a corner boundary, we end up with a 24-36" extension, depending on how it's figured and sized. Modeling that range with different exit angles shows you get a bit of a saddle in the response, but you can push the lowest excursion minimum down from ~28Hz to below 25Hz. With 6 units, we're not too worried about the increased excursion around 30Hz, which I'm still seeing to be below ~15mm with 80V input (ouch that'll be loud). From past experience with BDeaps, LAB Horns, and a few others in confined spaces, planning for room interaction becomes tricky, as the boxes become so big as to almost create new walls or boundaries that other interactions can get minimized or react a bit differently than expected. For example, if we just stack the subs with the mouths to the center in a 2 wide x 3 tall stack, we now have a 6' square front baffle that is ~3' from the front wall. Setting the stack of 3 subs with the mouths firing left or right along the front wall, we now have a 3' deep x 6' tall boundary and the rest of the box for the sound to wrap around. Now it starts making sense to talk about just walling off the front of the room at 3' deep so the subs radiate from flush with an effective wall, but with box faces so large, how much difference will that actually make? It will be different, but knowing how much gets much fuzzier. In the end real world measurements and lots of experimentation tells all.

I was visiting one of the crazier bass enthusiasts out there this weekend, the infamous pennynike. Anyone who's been in a few listening sessions with him know he has a serious appetite for bass. He basically bought a home so he could add 4 of my dual 18", 11Hz tuned, Terraform D18s. He had a lust for more, and found Josh's Othorn design. He grabbed 2 from Josh, and was loving the intensity they could add to the VLF extension the Terraforms provide in his basement bunker. Of course if 2 are good, 8 should be even better... and some enabler built him up 6 more cabinets. That's how he got to this with a pair of SP2-1200 amps powering the 8 Othorns. ? He had everything set in place before he got around to covering the walls, where this picture shows a lot more of the craziness going on here: I also want to point out the pink rigid foam wedged into the window well openings on the front wall... Those were added after he literally BROKE the windows. After visiting and seeing what was roughly going on, but not yet breaking out my microphone and just going on what he has measured, it would appear things would be much happier if we change the 2x2 stacks deep, to 3 units set on their side for a 6' tall stack, with the mouth against the front foundation wall. As the front center location did measure rather well in his room, I'm thinking to put the 2 stacks of 3 firing at each other against the front wall. This brings me to the question... for this thread. I've done some creative horn extensions with some of Tom Danley's past designs, with very useful results. With the 2 stacks firing at each other, lets assume for a moment we confined the height to the 6' tall stack. This would provide a much tighter spacial load, along with extending the horn length by most of the 36" height of the cabinet. We can vary the distance between the two stacks to adjust the area of the added segment, and I'm wondering if anyone has a model already in Hornresp who could check what range/type of effect an extension might have? I'm thinking I would start with about 36" between the two, which gives an extension area for each of about 18" x the 24" height of each for about 36" figuring the 90 deg turn and mouth/end effects. If anyone had time to model a 12" x 24" (1860 cm2), 18" x 24" (2790 cm2), and 24" x 24" (3720 cm2) extension that is 36" long, it would be appreciated and interesting. -Mark

I saw this in a local Dolby Cinema and was very impressed with the Atmos mix and overall sound, along with the movie as a whole. Are there specific scenes that really throttle the 20Hz range creating the spike we see in the graph?

The part some might skip over is investigating ways to slice & dice the cabinets into pieces with fractional performance. Obviously the image above would be easy to slice in 1/2, but what about 1/3rd or 1/4 or even less? As you mention, sometimes drivers aren't so hard to find, and you increase the number of voice coils to split the power between. Separating boxes also gets around the issues of spreading the loading point along the horn path, although I suspect 2 with some careful spacing might even be beneficial to resonances, but that would require some testing... That gets to the other very useful reason to look at fractional designs. Proof of concept. If you have a design that looks very exciting, but you don't know how bends, real losses and Q's of resonances will be in reality, it can be very worthwhile and easier to test a 10-15" driver in a smaller package with the same acoustic design. The ultimate test is if a pack of the smaller subs is comparable to a single large sub, which it theoretically should be.

One way to up the challenge a bit is to dive deeper into these huge cabinet designs that are fun to model and every once in a while build. One great property about acoustics is the ability to scale things. Maybe burn some late nights working out what a sliced fraction of such a design would look like, what parameters are needed for a 15", 12", or even 10" driver to maintain the same response? Might any of those parameters be something you can find? The suitable smaller drivers might not have the same excursion, but things might get interesting if you can actually use multiple units. No matter the type of alignment I'd say that being able to scale most all the performance qualities to be an interesting challenge. Ricci mentioned the Bose boxes earlier, and ironically the 6th order bandpass was a now expired Bose patent. Sometimes looking at designs in a different scale or with different technologies we stumble on some fun possibilities. If you can make a $1 driver sound impressive in something the size of a briefcase, what can you do if that's the size of a refrigerator with high excursion parts? Similarly, maybe when scaled down there are parts from a different market or application that you wouldn't have bothered modeling, but in fact the math suggests it scales nicely in some interesting design. Personally I still would love to have the time to create my concept of a subwoofer system using 3 or more bandpass designs intended to blend together into a multi-way subwoofer with the ability to place each bandwidth in efficient locations. The upper 1/2 of the bandwidth really gets interesting when you look at small modules and bandpass designs with significant gain. Of course getting better results than just a pile of compact sealed subs always comes back to the execution. Excuse the diversion, back to big obscene subs.

LspCAD 5 includes an optional attempt at modeling port compression (can check a box to see with and without model). Unfortunately it doesn't really allow a simple model of a commonly flared port. Instead they have a model more akin to a flared vent that defines more of an hourglass shape with an effective, assumed radius greater than 1/2 the length of the port. Comparing straight ports and more significantly flared ports gives some interesting points of consideration. One thing overlooked in your assumption above is that as the port compresses, the driver excursion increases. This is seen very easily in the LspCAD models and can be estimated/correlated by comparing high level impedance sweeps around tuning, and comparing driver vs port near field measurements. With high excursion woofers you will often see the driver picking up a good bit of the output load as excursion increases in an exponential manner due to port compression. This actually reduces the observed output compression, so remember the port itself is even more non-linear than the total SPL compression suggests. As with most things, the models are overly conservative and compression isn't quite as severe as suggested, but it's a lot more than none. Long ago I recall Deon Bearden relaying on that in his testing and research most port designs start compressing anywhere past 10m/s. That's not to say the ports aren't useful past that point, but we should understand that behavior isn't linear, just as with real drivers well before the rated Xmax. The complicating factor is always the wide bandwidth, complex signal consideration. If a component of a complex signal pushes a port into severe compression, how does this impact the rest of the complex signal being produced at the same time?