If you believe that an improved anechoic response leads to an improved listening position response it is interesting. Otherwise it probably means nothing.
Believe is one thing, reality is another thing! I do realise that directivity matters, but aren't you'll taking it a bit anal;o)
Believe is one thing, reality is another thing! I do realise that directivity matters, but aren't you'll taking it a bit anal;o)
I just skimmed through that thread and it's obvious you've spent a lot of time optimizing the simulations and I appreciate and try to learn from that what I can. In the end the big question is how much of all this is mostly academic and how much of it really has substantial effects. I find that really hard to estimate without having actual measurements. Now there is of course value in the little details regardsless of this question, but then it's more about basic science rather than creating a speaker for real world situations.
Sorry, I think I was unclear. I meant the forward backward distance. If you move the free standing WG back far enough it becomes part of the baffle. Now moved the other way, i.e. forward, how far does it have to protrude from the baffle to have this decoupling effect?
I have seen enough real measurements that correspond with properly constructed BEM simulations to accept that getting the best simulated response before building is a worthwhile option. As to how much it matters, I'll know when I've built something based on it.
Putting anything in front doesn't allow the termination to have the same effect.
I don't think so, my aim is to make the best anechoically optimized waveguide speaker I can.
well nobody has designed horns this way... why poo poo an effort to answer this very question? its all speculation until we have data
I believe the fundamental research on the secondary flare concept (intentional or accidental as it may apply here) was conducted somewhere in the early '70s.
I agree, we do not always know the actual benefits of what we are doing - perception wise - but experience has suggested that the objective best is found to be the preferred by the majority of people. Hence, doing anything but trying to find the very best with simulations is just being lazy. Building and testing is what takes all the time.
You lost me here - and I was around in the 70's - I don't know anything about what you are saying - it's new to me.
I think Allen is referring to Don Keele's work using a secondary flare at a steeper angle.
https://www.xlrtechs.com/dbkeele.com/PDF/Keele%20(1975-05%20AES%20Preprint)%20-%20Whats%20So%20Sacred%20Exp%20Horns.pdf
https://www.xlrtechs.com/dbkeele.com/PDF/Keele%20(1975-05%20AES%20Preprint)%20-%20Whats%20So%20Sacred%20Exp%20Horns.pdf
Yes I couldn't remember. Is that the one that mentions the 2/3 transition point or is there more than one paper on this?
It's a diffraction device.
It's a diffraction device.
I stand by what I said. we are just starting to do the first testing with these profiles. nobody has built and listened to these exact profiles before. so its impossible to say whether listeners will feel its 'worth it' until we have some real solid research. who knows, maybe these will become the most preferred designs out there in a few years.
You can change that manually in the file lib\scripts\report2.gpl.
Change these lines -
set xtics add ("0.2" 0.2, "0.5" 0.5, "5" 5, "" 15, "20" 20)
set logscale x
set format y "%.0f°"
set xrange [0.2:20]
to something like this:
set xtics add ("0.1" 0.1, "0.2" 0.2, "0.5" 0.5, "5" 5, "" 15, "20" 20)
set logscale x
set format y "%.0f°"
set xrange [0.02:20]
I think the horn is superb.
It was not abandoned, the Tritonia waveguide is still waiting in my workshop to be properly measured and the whole loudspeaker finished.
The freestanding waveguides (as presented so far) are somewhat smaller becaue they can be smaller. This is what I noticed as one of the first things and it makes them, among other things, so attractive to me now. That they can be made quite easily at home because they are axisymmetric (no pattern flip) is also not an insignificant advantage to a hobbyist, IMO. That they have virtually no noticeable diffraction seems to be taken as granted now but this was not a matter of course for years.
The rest is just quibbling.
Last edited:
Sorry, I think I still didn't get my question right. 🙂
In the free standing simulation there is something like 1cm of air between the woofer baffle and the waveguide. What I was wondering is whether it's that tiny gap that makes the difference or if not what feature of that situation makes it behave that way. What happens if you move the waveguide back 1cm and have it touch the woofer baffle? It then basically becomes part of the baffle with a protruding lip. Same effect or does that kill it?
The reason I ask is that it's not very intuitive to me (which might well be a case against my intuition) that this 1cm of space makes the waves coming around the mouth of the WG simply don't see the woofer baffle or act as if it wasn't there. If it's not that tiny space that makes the difference, it's probably wiser to look at the WG as part of the baffle but a baffle that is shaped a certain way to acoustically hide elements on it. In some sense like the Neumann line of studio monitors where the woofers have this ridge around them to 'hide' them.
There is a small gap for two reasons, the first is to do with the practicality of the simulation. If I join the surfaces I have to join the meshes and make sure that they are correct. This can be a pain and if you don't get it right you get garbage results.
The effect of the rollback termination is best seen through observation fields, this allows you to visualize what is happening at different frequencies. mabat posted quite a few throughout the thread I don't have them bookmarked to link directly. If you join the waveguide to the woofer cabinet that will be disrupted. I don't intend to simulate it as it is a lot of work for something I cannot see any benefit in.
What I have found through doing a lot of these simulations is that human intuition and acoustics do not always go well together which is one of the reasons I find them so fascinating 🙂
The effect of the rollback termination is best seen through observation fields, this allows you to visualize what is happening at different frequencies. mabat posted quite a few throughout the thread I don't have them bookmarked to link directly. If you join the waveguide to the woofer cabinet that will be disrupted. I don't intend to simulate it as it is a lot of work for something I cannot see any benefit in.
What I have found through doing a lot of these simulations is that human intuition and acoustics do not always go well together which is one of the reasons I find them so fascinating 🙂
Fair enough, to be as small as possible for a given acoustic effect certainly is a good argument. And please don't get me wrong, I'm not trying to be overy critical, I'm absolutely baffled (pun intended) about what you (and this thread collectively) have come up with. I just didn't really understand some of the forks in the path. And I'm not seeing how an axisymmetric waveguide is any easier to build? Especially since all the attention has turned to printing. Pattern flip of non-axi WG's certainly is an issue, but it's a compromise that has its benefits (not the pattern flip, but the non-axi in general) depending on the situation and how one deals with it.
I guess I was just so excited about the Tritonia as that quite literally ticks all my boxes, that I'm a little bummed it's just sitting in your workshop. 🙂
There are always so many forks in the path... It's even succcess I still have come up with a waveguide after all 🙂
Yeah, Tritonia has been resting in peace for a while. It will come to life as soon as this pandemic madness ceases, which I still hope will happen some day.
Yeah, Tritonia has been resting in peace for a while. It will come to life as soon as this pandemic madness ceases, which I still hope will happen some day.
Last edited:
I remember these rollback termination posts. I just fail to see how a flat surface just behind the rollback does not influence or disrupt the termination. I get the simulation issues of connecting the two and just in general the amount of work it takes, wasn't trying to suggest you should do those. My reasoning is just that there's too many variables involved to come to any real conclusions and/or their relevance.
There's already a pretty strong evidence that the presence of an enclosure behind it doesn't really spoil the pristine performace of the device itself. This was measured e.g. by Pet007 quice convincingly. I can't see why is this still so hard to accept.