Ah, now I understand. Simulating axisymmetric waveguides and adjusting power response/DI is one of the very few things I actually know how to do so far. However, I'd fall on my face if I tried to design an asymmetric waveguide. Zero chance I'd be able to design one with different coverage angle patterns that complemented each other inside a room. Neither can I successfully configure a waveguide that would work well in an enclosure (OSSE). I only have a very basic understanding of what you guys are doing when you design anything that isn't freestanding axisymmetric. I was also surprised to learn how crossovers and EQ affect the waveguide frequency bandwidth after playing with the VituixCAD simulations (viewing the non-normalized polars). Now that I've learned that I see there is huge value in using ATH/ABEC to simulate a waveguide plus enclosure then exporting the polar response data into VituixCAD. Your software is an excellent educational tool for people like me. Thank you for writing it.
I'll need to order more 3D filament. I burned through all of mine making variants of just this one waveguide for the two compression drivers (DE250, DE111) and how they fit to match internal rings. As you mentioned in a previous post, part of 3D printing is learning how to construct a model that will successfully print on your 3D printer without warping, fitting together nicely and easily, etc.
(If anyone wants the 3D models I've made I'm happy to post or send them via direct message. Just don't rely on them 100% because I think they could be made more precise in the conical throat section of the drivers. But if you want a basic starting point they are good enough for that.)
I'll need to order more 3D filament. I burned through all of mine making variants of just this one waveguide for the two compression drivers (DE250, DE111) and how they fit to match internal rings. As you mentioned in a previous post, part of 3D printing is learning how to construct a model that will successfully print on your 3D printer without warping, fitting together nicely and easily, etc.
(If anyone wants the 3D models I've made I'm happy to post or send them via direct message. Just don't rely on them 100% because I think they could be made more precise in the conical throat section of the drivers. But if you want a basic starting point they are good enough for that.)
<3I can't speak for others, but don't believe that I really know what I'm doing...
I could not believe the R7 wouldn't sound dull. I would be surprised. Maybe the KVAR, free-standing asymmetric is worth a printing experiment for @CinnamonRolls? It modifies the power response much more gentle than this EQ.
There's no reason to switch to asymmetric, IMO. You can set just about any DI curve with axisymmetric as well. His waveguides are a perfect example that axisymmetric can be top-notch.
Here is the config file for the DE111 R-OSSE. I started modifying it but then decided it would be better to let you guys do it because you have a better understanding of how it will interact with the 8 inch driver enclosure/crossover. I'll print the version you think would make a good test and do my best to give you feedback.
It does sound dull at first. I created a playlist of songs Sean Olive (Harman) uses to evaluate speakers in Amazon Music and listened through it with the crossover variants. Over time I drifted toward R7. Maybe I'll drift back toward a flatter on-axis (less dull) variant as I continue listening. After listening and adapting to R7 I wouldn't call it dull anymore.
Also, here is Sean Olive's piano evaluation playlist.
My memory is terrible but I think the way I got to those numbers is I saw you say something about a Rot=3 in conjunction with a different config setting widened the directivity in the higher frequencies. That made me curious about Rot so I over-exaggerated Rot to see what would happen. I distinctly did not care about optimizing for higher frequencies because I can't hear them. My hearing is non-existent above 12kHz and that's a stretch to even claim 12kHz. However, the simulations looked good and it was a smaller + easy to print waveguide. Which was great because I wanted to see what freestanding R-OSSE waveguides sound like in real life. They sound as good as the simulations look. Now I'm very interested to read feedback from people who build larger waveguides like the ones you're working on. I know these little waveguides sound beautiful, imagine how good the large ones designed by people with knowledge and experience are going to sound.
First thing I tried was to decrease 'Rot' and it seems to do what we need. This is for your settings (left) and for Rot=0 (right):
The on-axis DI is higher roughly by 3 dB at 10 kHz, which is what is takes to make the on-axis FR more or less flat in your R7 version, while retaining a similar power response.
It would be really helpful if it was finally tested what are the subjective differences, if any. It can be even source material dependent of course, I wouldn't expect a generally valid conclusion.
Last edited:
So as I indicated in a previous post I didn't understand the window setting meny correctly so my posted picture where not correct. Now that I'm better informed I think your used window looks like this and you did set it so that the wiggle in the red circle was outside the window, so as too take it out of the measured result. My pitfall was that the windows where sizes (ms) and not time stamps (ms), like the Ref.
Sorry for the confusion!
//
I realized afterwards that by removing the rotation, the overall diameter will be of cource smaller (hence the lower DI below 1.5 kHz). It could be actually a better match for your woofer but to assure the rest equal for the experiment, perhaps you might want to set also R=111 mm, which compensates for this. Now the DI makes a gradual rise against the current version.
So the proposed change is:
R = 111
Rot = 0
Last edited:
@mabat Attached a csv file for the Le'Cleach horn profile of my horn, feel free to simulate
I am curious how my design looks from a sim perspective. Little bit scary for me of course, maybe it is not as good as I perceive... but hey that is the way to progress. The rudimentary measurements I took looked good, though.
I dofeel free to simulate
Obviously, these are impedance-optimized devices. I'm wondering what they can be good for... Probably not a high-fidelity reproduction. But maybe a garden - martens won't be pickey of what they get scared of
Last edited:
Haha funny how a device to scare martens doubles as a hifi device.. but I can confirm that the "sound" matches the sim, I like it even more now