Take that chart, and tilt it with the anchor point being the upper most frequencies. When you put in the appropriate passive network the response almost tilts the whole response and you're left with a flat response.
You can kinda get a visual example of that with heissmanns waveguide tests.
https://heissmann-acoustics.de/en/test-vifa-xt-300-xt25tg-waveguide-wg-300/
Getting higher sensitivity on a driver, almost always correlates with less distortion at a given SPL, more so if the SPL is over 100dB@1m.
Now, there are other kinds of distortions, of course. Fortunately Earl Geddes (GedLee speakers) is our savior.
Not only that, but especially a horn will make sure that higher harmonics are basically absent.
As a small penalty that 2nd order is usually a bit higher.
Which aren't really that important luckily and in general already very low.
Not sure what mechanisms @stoneeh had in mind, but I'm thinking that phase plugs may be so closely coupled to the driver membrane (within 1 mm?) that any significant displacement will modulate the shape of the horn. So, although a drive unit could boast high sensitivity down to a few hundred Hz, max power might only be a few mW before the cone displacement brings amplitude modulation to noticeable levels.
Last edited:
I asked the author of that post in question what kind of distortion that he was referring to (post #19). I haven't gotten a reply thus far.
I think when you go look at the relative sideband levels at higher SPLs (and certainly anything above 100 dB/1m), you will see that the higher the sensitivity of the drivers/horns, the lower the higher order sidebands will be due to modulation distortion. In general, these sidebands are about 20-25 dB lower in amplitude than the same drivers being used in direct radiator mode. So, as the sensitivity goes up, the most objectionable form of distortion (modulation distortion sidebands) actually disappear as concerns for the DIY loudspeaker designer.
A lot of people seem to be fixed on harmonic distortion, but the real issue is the lower side-band modulation distortion products--the higher order products--that are most audible and certainly most objectionable, since they are non-harmonic in nature.
Chris
I think it's important to understand that it's not really the higher harmonics themselves that are audible, but rather the higher harmonic side bands (i.e., modulation sidebands) that are most objectionable--making the resulting sound quality opaque and harsh sounding at higher SPLs.
I think when you go look at the relative sideband levels at higher SPLs (and certainly anything above 100 dB/1m), you will see that the higher the sensitivity of the drivers/horns, the lower the higher order sidebands will be due to modulation distortion. In general, these sidebands are about 20-25 dB lower in amplitude than the same drivers being used in direct radiator mode. So, as the sensitivity goes up, the most objectionable form of distortion (modulation distortion sidebands) actually disappear as concerns for the DIY loudspeaker designer.
A lot of people seem to be fixed on harmonic distortion, but the real issue is the lower side-band modulation distortion products--the higher order products--that are most audible and certainly most objectionable, since they are non-harmonic in nature.
Chris
Last edited:
Bingo.
A source for the described phenomenon: JBL Tech Note - Characteristics of High-Frequency Compression Drivers
While not explicitly focusing on it, this document also describes the basic correlation between increased sensitivity and increased nonlinear distortion in high frequency horn systems.
Klippel's work on nonlinearities is also helpful: https://www.klippel.de/fileadmin/_m...linearities–Causes_Parameters_Symptoms_01.pdf
A source for the described phenomenon: JBL Tech Note - Characteristics of High-Frequency Compression Drivers
While not explicitly focusing on it, this document also describes the basic correlation between increased sensitivity and increased nonlinear distortion in high frequency horn systems.
Klippel's work on nonlinearities is also helpful: https://www.klippel.de/fileadmin/_m...linearities–Causes_Parameters_Symptoms_01.pdf
I don't believe I agree with the above sentence. Perhaps I missed the exact words you are referring to.
In terms of power vs. distortion in compression drivers/horns, home hi-fi levels (typically less than 105 dB/1 m from all the loudspeakers in your array), you're putting less than 1 watt into each driver. That's really small in terms of power input, even at 10% acoustic efficiency (i.e., 90%--0.9 watt-- is going into heat).
The thermal issues occur at PA and (perhaps) commercial cinema levels, but in my experience--not home hi-fi levels (with possibly one exception that I've run across in my long DIY experience...if that individual can still hear at all...which I doubt).
We may be talking past each other, in that you may be substituting the word "sensitivity" for "intensity". The kind of SPL that is needed to start seeing thermodynamic second-order harmonic issues at a horn throat is well above anything that I care to consider for home hi-fi.
Chris
Last edited:
It's not so bad. It happens more often than not that an entire community completely misunderstands a certain issue.
That JBL paper explicitly says this phenomenon is a property of compression drivers. The higher non-linear distortion arises from the very high acoustic pressure between the diaphragm and phase plug. They also say this is made worse by narrow dispersion (and presumably deeper) horns. Neither of these conditions is true for HiFi tweeters on shallow waveguides. I don't know what the OP had in mind, but we should be clear what format we're talking about.
Per the horn-design engineer that I described in post #16, a customer preference for 90 x ~60 degree coverage in-room (horizontal x vertical) are by far the most desired polar coverage angles for both home hi-fi and commercial cinema. As such, any horns that cover less than this revealed requirement, in my experience, ends up producing insufficiently sized in-room acoustic images that do not result in a convincing continuous stereo or multichannel image (evaluated from listening anywhere from side wall-to-side wall at a typical 10-15 feet [3-4.5 m] listening distance. This produces subjective "holes" in the projected soundstage image between loudspeaker channels that are quite easy to hear by leaning slightly left and right at the listening position(s).
I have also found that those claiming that narrower coverage horns sound better in their listening room have room acoustics issues. Most notably, this is trying to use a listening room that's likely too small in its basic dimensions to set up a robust stereo or multichannel acoustic image for home hi-fi listening.
By taking the two constraints of: 1) home hi-fi listening SPLs, together with 2) a minimum listening room size of ~17.5 width x 22 deep x 9 feet high (mean dimensions as self-reported by a representative population of 65 diyAudio users, with link above), I believe any apparent perceived "nonlinear distortion crisis" of horns or compression drivers is suitably avoided. YMMV.
Chris
I have also found that those claiming that narrower coverage horns sound better in their listening room have room acoustics issues. Most notably, this is trying to use a listening room that's likely too small in its basic dimensions to set up a robust stereo or multichannel acoustic image for home hi-fi listening.
By taking the two constraints of: 1) home hi-fi listening SPLs, together with 2) a minimum listening room size of ~17.5 width x 22 deep x 9 feet high (mean dimensions as self-reported by a representative population of 65 diyAudio users, with link above), I believe any apparent perceived "nonlinear distortion crisis" of horns or compression drivers is suitably avoided. YMMV.
Chris
Last edited:
The design criteria is favourable for designing 90 x 90 and the presentation is as good as the effort you put in to the speaker and room combination.
That's not to say that you can't do 90 x 60 properly, however while a room may suit those angles that doesn't consider that it's harder to get right.
That's not to say that you can't do 90 x 60 properly, however while a room may suit those angles that doesn't consider that it's harder to get right.
For one, a wider angle gives economy of baffle real estate vs lowest frequency and driver spacing. For another, an axisymmetrical waveguide is conducive to reducing HOM due to distortion when changing the wavefront shape. It also makes the job of designing in the converging frequency less critical, to match the woofer.
This is the part that I'm not agreeing with.
What baffle?
By the time the direct arrivals get to the listening position, there is far more "wavefront shape distortion" than any horn itself can impart, unless you're listening inside an anechoic chamber.. Even earphones suffer from wavefront shape distortion just getting through the ear's penna, and eardrum bounce.
Take an in-room measurement at 1m in front of the horn mouth, then one at 2-4 m, and look at the difference between them.
The above is the actual sentence you were trying to quote, not the one you quoted in part. When you took a fragment of that sentence, unfortunately you changed that particular sentence's meaning.
Look at what you left out to answer your questions about what was said.
Chris
A goal is to minimise early reflections around or behind the speaker.
The first potential source of reflection, a fact that contributes to it's importance, is HOM produced due to mismatch at the wavefront throat.
(To your other points, I disagree. Ask if you'd like clarification.)
The first potential source of reflection, a fact that contributes to it's importance, is HOM produced due to mismatch at the wavefront throat.
(To your other points, I disagree. Ask if you'd like clarification.)
Experimentally and from my own experiences, this is true, although there seems to be a tradeoff in use of nearfield absorption (within 1-2 m of the front of the loudspeakers) with subjective liveness/ambience of the listening room. I've found that the human hearing system is very easily distracted by early reflections just around the front of loudspeakers in most home hi-fi sized listening rooms.
Additionally, if your loudspeakers lose directivity control above the room's Schroeder frequency (typically loudspeakers with direct radiating bass have issues from 100-550 Hz) then you've got the added problem of having early reflections from around the rear of the loudspeakers. Effective near field absorption at these frequencies is both large and bulky (i.e., much larger than horn-loaded bass bins). Diffusers in this frequency range are typically unacceptably large, and compete in room real estate with higher frequency diffusers.
Fortunately, not all loudspeakers lose directivity in these four to five critical bands between 100-550 Hz. For instance, my 5.2 array avoids these issues such that low frequency absorption is not required in the nearfield space around the rear of the loudspeakers.
What is the threshold of audibility of these "HOMs" (higher order modes) of the horns used in loudspeakers, and what exactly does it sound like? Do you have a particular technical study in hand?
There seems to be a lot of talk here to this phenomenon but no study of threshold of audibility. A lot of angst. This seems to be wildly over-compensated by some here, but without consideration for "how much is actually audible?". Simple blind A-B tests would be a good starting point. My own experiences say that this is wildly overthought and actually inaudible.
Chris
Last edited:
The Yamaha NS 777 is an example of a very affordable but very well engineered /implemented 3 way speaker using wave guides for the mid and tweeter.They sound way better than they should for their price and perhaps the wave guides help give them a surprisingly seamless and coherent sound.
Maybe because it can be difficult to build. KEF are doing great, though, in my opinion - by adding more stationary space(non moving) for the waveguide in the center of their newer Coax's.