The 20khz is as wide as a 200hz round tractrix and similar performance is had through the first 30 degrees, lower in frequency...I anticipated this much, still the throat shape is throwing me for a loop plus there's a bunch of smoothing on the builders measurements...
Also, My opinion of the so called "head in vice" just isn't as dramatic as people claim.... I really don't get it as I am the one who usually over exemplifies things lol...If you are outside of the sweet spot, thats user error....
I spent yesterday evening listening back n forth to my AXi on the150hz "E-tractrix" vs the Acoustic Elegance 15m 16ohm......High passed at 300hz and low passed at 700hz using steep slopes.... Response was..."similar"...similar enough for me to figure out what I believe the main difference to be...Direct energy..... I preferred what I hear coming out of the horn, during that experiment....
300hz to 700hz
More experimentation is needed but at least at this point I believe that there is more direct energy, lower, with the horn...and that the horn is cleaner sounding....The burst decay showed a little more resonance on the top side of that small passband...and the horn had more of its decay leaning towards the lower side, in a gradual manner. I think that this is the reason for my some of my perception ideas....
I had one experience where the compression on a vocal show'd its self much more obvious on the horn vs the woofer....The woofer obviously has less HF content regardless of the very similar HF roll off I was able to create with the low pass filter.
Attachments
Last edited:
Some Western Electric horns seem to have similar transitional sections, markedly less diamond-shaped and more elongated/flattened oval, because WE never made elliptical horns.
A genuine Tractrix horn can never have full CD characteristics without (band) aids, such as diffraction slots.
This is one of the main reasons why (traditional) horns are so popular among afficionados of - mostly classical, vocal music.
Last edited:
Both statements are correct but are describing different things. I do not see a traditional diffraction slot in your picture, the horn walls do not constrict in to cover part of the driver in an attempt to control directivity to a higher frequency than the size of the driver would normally allow.
The longer dimension can control directivity to a lower frequency but that says nothing of the pattern of the directivity which depends on wall angle/curvature.
At 1m the measurements look fine. Measure it at 3m and it will not look the same. The fact that there is such a high degree of rolloff in all curves suggests that the profile is not typical of horns that return the on axis back closer to flat without EQ through significant beaming.
Now you have it measure it both ways, equalize and listen to and see which you prefer.
This may help you take a look at the first two pages where they describe the 2344 Bi-radial horn
Rob
http://www.audioheritage.org/vbulle...2-Improvements-in-Monitor-Loudspeaker-Systems
The cross section per se doesn't mean much - the profile, which is far more important, can still be arbitrary. What you show is a mere consequence of the horn having an elliptical mouth.
Here's shown the cross section of the ST260-KVAR, so yes, I've seen that before (and it really doesn't mean anything) -
Last edited:
There are currently serious doubts about your measurement, the setup and your interpretation on my side.
First, there is a nearly flat section from 6k to 9k in your measurement, imo this is simply impossible for such a large diaphragm to turn from a constant decay starting at about 2k5 into a linear section and then decay again. I own these drivers myself and when you measure them the decay is nearly perfectly linear from 2k5 down to about 15k (ripple smoothed).
As usual, we know less about what you are doing. For such a quasi nearfield measurement with a very long horn you have to work very precisely. Please show us a picture of your measurement setup, how you accomplished that all measurements for different curves had the same distance to sound origin and how you exactly determined the angle (for which it is unclear if it should be half or full angle).
Then we know nothing about your hardware (mic, mic calibration, amplifier, audio interface) and software (smoothing, gating, etc.) you used and of any passive components or dsp you had in the middle.
Concerning the throat section shape I assume that it is basically related to the production process how to transform the round shape into an elliptical. I doubt that any BEM optimization took place. For directivity control overview you basically could draw the round profile of such a long tractrix horn as 2D on paper and simply draw some straight lines from each throat boundary to the side wall up to where the line touches the opposite side wall and beyond. With this you get a basic idea of the beaming effects. There is not much diffraction present in these horns so you are basically limited with this opening angle for higher frequencies. For the elliptical shape you will have a little bit more directivity control for the hrz plane compared to the round shape but also less control for the vrt plane.
Last edited:
Very glad to see that the horns have arrived and there's some actual progress on this project. Only took 500+ pages!
I know I've said this before in another thread and provided links to support it, but you can’t accurately measure the directivity of an acoustic device in the near-field.
At the very least, you can't use the data taken in the acoustic near field to predict the output at other distances. The data is valid only for that specific distance, no farther and no nearer.
A good description and diagrams can be found here: https://www.prosoundtraining.com/2010/06/28/far-field-criteria-for-loudspeaker-balloon-data/
It's interesting that there's actually an HF limit based on measurement distance, which affects larger horns or acoustic devices more.
With such a large mouth dimension, even 1 metre is too close to be in the far-field. Let alone 1".
Near-field measurements of horns can however be useful to find reflections from within the horn itself. If the axial length is 85cm, and your mic is placed 20cm from the mouth, then mouth reflections should appear at 8ms after the start of your measured IR data.
Calculated from (0.85 * 3 + 0.2)/343 because the reflected sound has to travel from the driver to the mouth, bounce back to the throat, then back along the horn length and finally the distance to the mic capsule). Your initial impulse peak 'should' arrive at 3ms. This is over-simplified as the horn isn't a straight tube, but you get the idea.
Edit: adding some images from a few papers by Di Cola et al. to clarify:
The red portion shows the mouth reflections from a straight pipe of 30cm in length.
I'd suggest that you get some experience looking at 'normal' quasi-anechoic or in-room IR data before this becomes intuitive or even useful.
If you can't take the horn outside yet @camplo then for now, place the mic at your intended listening distance from the centre of the mouth. You can use a piece of taut string or fishing wire, cut to length or marked, to ensure that you're maintaining this distance when moving the mic for the off-axis measurements.
You can add another couple of pieces of string in an x across the mouth to give you a reference position. It'll be so small that the error in the data will be negligible compared to reflections from your mic stand or even the microphone body.
The distances from the horn mouth and the mic position(s) will dictate the time period for your right-hand window or gate. It should be less than twice the distance to the nearest boundary - this also dictates the low-frequency cutoff for valid 'anechoic’ frequency data:
Using the maths from before, your window period needs to be shorter than 8ms if your nearest large surface is less than 1.325m away from either the mic or horn mouth.
If this seems too short, then you might want to try a Hann window shape for a more gradual roll-off of the reflected sound. It'll let some reflected sound into the measurement data, but at a reduced amplitude. I think REW defaults to a Tukey 0.25 window shape for both sides, which is quite steep.
Also, ask the horn manufacturer if they will send you either the ARTA .pir file or a .WAV export. Then you can compare your measured data to theirs. That will give you the best chance to figure out what is horn and what is room in your own measurements - although they'll also have a reflection at twice the distance from the horn mouth to the ground in that impulse response, same as you.
I know I've said this before in another thread and provided links to support it, but you can’t accurately measure the directivity of an acoustic device in the near-field.
At the very least, you can't use the data taken in the acoustic near field to predict the output at other distances. The data is valid only for that specific distance, no farther and no nearer.
A good description and diagrams can be found here: https://www.prosoundtraining.com/2010/06/28/far-field-criteria-for-loudspeaker-balloon-data/
It's interesting that there's actually an HF limit based on measurement distance, which affects larger horns or acoustic devices more.
With such a large mouth dimension, even 1 metre is too close to be in the far-field. Let alone 1".
Near-field measurements of horns can however be useful to find reflections from within the horn itself. If the axial length is 85cm, and your mic is placed 20cm from the mouth, then mouth reflections should appear at 8ms after the start of your measured IR data.
Calculated from (0.85 * 3 + 0.2)/343 because the reflected sound has to travel from the driver to the mouth, bounce back to the throat, then back along the horn length and finally the distance to the mic capsule). Your initial impulse peak 'should' arrive at 3ms. This is over-simplified as the horn isn't a straight tube, but you get the idea.
Edit: adding some images from a few papers by Di Cola et al. to clarify:
The red portion shows the mouth reflections from a straight pipe of 30cm in length.
I'd suggest that you get some experience looking at 'normal' quasi-anechoic or in-room IR data before this becomes intuitive or even useful.
If you can't take the horn outside yet @camplo then for now, place the mic at your intended listening distance from the centre of the mouth. You can use a piece of taut string or fishing wire, cut to length or marked, to ensure that you're maintaining this distance when moving the mic for the off-axis measurements.
You can add another couple of pieces of string in an x across the mouth to give you a reference position. It'll be so small that the error in the data will be negligible compared to reflections from your mic stand or even the microphone body.
The distances from the horn mouth and the mic position(s) will dictate the time period for your right-hand window or gate. It should be less than twice the distance to the nearest boundary - this also dictates the low-frequency cutoff for valid 'anechoic’ frequency data:
Using the maths from before, your window period needs to be shorter than 8ms if your nearest large surface is less than 1.325m away from either the mic or horn mouth.
If this seems too short, then you might want to try a Hann window shape for a more gradual roll-off of the reflected sound. It'll let some reflected sound into the measurement data, but at a reduced amplitude. I think REW defaults to a Tukey 0.25 window shape for both sides, which is quite steep.
Also, ask the horn manufacturer if they will send you either the ARTA .pir file or a .WAV export. Then you can compare your measured data to theirs. That will give you the best chance to figure out what is horn and what is room in your own measurements - although they'll also have a reflection at twice the distance from the horn mouth to the ground in that impulse response, same as you.
Last edited:
Thanks. Yes, I intended to cross a CD+horn at 500-ish hz, so I guess this horn would not do then. The other one I looked at is Goldwood GM-450PB, which appears to reach 500 hz (provided that the CD can handle it).
That Goldwood horn lists a cutoff frequency of 500Hz, that is not the frequency it should be used down to. Where a horn has a defined cutoff the rolloff is quite steep with associated phase change at that point. Half an octave to an octave higher than the cutoff is generally a reasonable guess for where you can cross without difficulty.
I saw it somewhere (though can't find it currently) that based on measurement, the Goldwood horn can solidly reach to 300hz.
Have a read of some horn theory to see why a horn with a 500Hz cutoff cannot solidly reach to 300Hz. If in reality it does then the spec of 500Hz cutoff is wrong.
https://www.grc.com/acoustics/an-introduction-to-horn-theory.pdf
https://www.grc.com/acoustics/an-introduction-to-horn-theory.pdf
This JBL 2386 clone can be used from about 500-600Hz, depending on the driver.
It's used by DonVK in this project.
Last edited:
The horn was apparently good enough for this (outermost) G.I.P. system that costs €32.000... each
Last edited:
Why not? Up to and above 200 Hz (1.72 meters), typical MTM artifacts hardly play a role.
Besides, 120 dB will be a piece of cake.
Last edited:
