Play 4,5kHz and 9kHz together and 3kHz and 9kHz together. You will notice the difference. Celef is right. Been there, done that. Even playing a keyboard helps, play C4 and C5 and compare to C4 and E5.
I meant what I wrote: H4 is lower than H2 and H5 lower than H3. Since loudspeakers are not linear, I do not compare the fundamental products but the frequency of the distortion signal and it’s cause. I think we do not have to discuss about what the cause is of any harmonic distortion product as we all agree on the theory that it is the excitation by a (not the) fundamental.
I meant what I wrote: H4 is lower than H2 and H5 lower than H3. Since loudspeakers are not linear, I do not compare the fundamental products but the frequency of the distortion signal and it’s cause. I think we do not have to discuss about what the cause is of any harmonic distortion product as we all agree on the theory that it is the excitation by a (not the) fundamental.
"Alex Voishvillo — in his extensive range of papers [3],[4] and other presentations on nonlinearity in speakers, their audibility, and the latest methods of measurements and assessment—clearly shows distortion that manifests itself at low and moderate signal levels is easily noticeable and the most objectionable. Similar conclusions are reported by Earl Geddes and Lidia Lee.".. Igor Levitsky
The above is very true, low signal level nonlinearities are highly audible, but also rare. In mechanical systems like loudspeaker low level nonlinearities are exceedingly rare, but can and have occurred. In electronics they are not as rare (crossover distortion,) but have been basically under control for a log time.
So audibility of nonlinear distortion while listening to music is a non issue? Are there any exceptions to this? Let's say a well recorded acoustic piano or violin recording?
Of course there are exceptions. A bad design can have audible distortion, most small speakers will distort audibly if an excessive SPL is set, so it can happen. But the point is that it can also be designed out to the extent that it is not audible for the vast majority of situations. Not that this is the case however as I am always surprised by just how bad many systems are.
So a metal diaphragm producing relatively(compared to softer diaphragm materials) high amplitude, higher order distortion in the 1 to 5khz range is okay as long as it's what, 30db to 40db below the fundamental? Does your research also account for long term aural fatigue that may not be perceptible in the short term?
Last edited:
Your premiss is incorrect. Why would a metal diaphragm produce "higher amplitude, higher order distortion"?
And there is no simple answer to your question because there are too many other variables, like what level are the other orders at, ...?
Listener fatigue has never been studied to my knowledge, but I believe that distortion is probably a significant effect, but poor frequency response is also a distortion and that is more likely the root cause of fatigue.
I would imagine that the long term aural fatigue(Due to high order nonlinear distortion) would be impossible to control for due to the variables involved, and will always be an open question. Does that sound fair?
In our study I was able to generate any amount of any order and control them individually, so what you say "can't be done" and has been done and in my mind it is not an "open question".
Got it. I always wanted to ask you that and thank you so much for making yourself available for questions. It's so cool of you to tolerate and educate anyone who will listen. You have infinite patience🙂
Last edited:
In terms of fatigue - I am of the opinion it isn't from the higher odd order peaks from the L15 you can see in my graphs above. The distortion does "taint" the sound. I could hear this on acoustic guitar.
However - I've found the fatigue is a flat frequency response in the 1 - 5KHz range. I found a BBC dip of sorts (effected by pushing the LR slopes apart for mid/tweeter) created a broad 3dB onaxis dip which made this much better for long term listening.
I need to get in room and gated power responses to see what was going on - but regardless of where I crossed the L15 - fatigue was there with the tweeter playing "too hot" (i.e. flat).
I also intend to disconnect the tweeter and just play the L15 to see if anything offensive is there for fatigue.
Edit: The fatigue is also quite dependent on the recording I am playing. "harsh" recordings on my lesser speakers are more listenable, but they don't extract the detail from the great recordings. Vice-versa for my more detailed speakers.
However - I've found the fatigue is a flat frequency response in the 1 - 5KHz range. I found a BBC dip of sorts (effected by pushing the LR slopes apart for mid/tweeter) created a broad 3dB onaxis dip which made this much better for long term listening.
I need to get in room and gated power responses to see what was going on - but regardless of where I crossed the L15 - fatigue was there with the tweeter playing "too hot" (i.e. flat).
I also intend to disconnect the tweeter and just play the L15 to see if anything offensive is there for fatigue.
Edit: The fatigue is also quite dependent on the recording I am playing. "harsh" recordings on my lesser speakers are more listenable, but they don't extract the detail from the great recordings. Vice-versa for my more detailed speakers.
Last edited:
exactly the same feeling with that range... flat here is too much hot for my taste whatever the room. Think people are slighty different here.
However I wish I could test with different phase integration and filter slopes to be sure about that on diferent loudspeakers...
However I wish I could test with different phase integration and filter slopes to be sure about that on diferent loudspeakers...
Although off topic - phase alignment is an interesting one. It seems the overall power and in room response is what dominates our listening experience. This might seem obvious... but it tells me <1% distortion or 10 degrees of phase difference... don't actually mean anything. The phase alignment results in the frequency response.. Put it this way - I wonder if I could tell in a blind test a BW vs. LR aligned system having identical frequency responses at the same listening position?
... you could listen to the cone break-ups faster (sooner) with the LR12 ? I just wanted to say than harmonics being a part of how the brain are checking the instruments accuracy, one maybe want those harmonics are coming in time to avoid listening fatigue. But I just don't know, it' just I was thinking about what Earl Geddes said about the 1K hz range and above : don't miss it in your design. Imo it stays on topic with the object of this thread.
If most of the (acoustical) music and its spl live below 1 K hz, you must need the above range for the accurate reproduction (as original perception)... harmonics (so also disonances of some odd harmonics at the right spl and timing to make instruments acurate ! Hence the question if that can be wasted by the harmonics mask of the speaker/driver/filter, harmonics break-ups of hard cones, sub harmonics break-ups of soft dome, etc ! So just an opinion, I'm not a scientist.
But perhaps it slight towards the theory of human hearing...too much. Sorry for that, let's back to the topic.
I would like to ask aboy the cone break-ups of a 8" hard-dome, dunno if I dare !
If most of the (acoustical) music and its spl live below 1 K hz, you must need the above range for the accurate reproduction (as original perception)... harmonics (so also disonances of some odd harmonics at the right spl and timing to make instruments acurate ! Hence the question if that can be wasted by the harmonics mask of the speaker/driver/filter, harmonics break-ups of hard cones, sub harmonics break-ups of soft dome, etc ! So just an opinion, I'm not a scientist.
But perhaps it slight towards the theory of human hearing...too much. Sorry for that, let's back to the topic.
I would like to ask aboy the cone break-ups of a 8" hard-dome, dunno if I dare !
Easy to test with dsp-crossover eg. minidsp. been there, done that and yes, we are extremely sensitive for spl changes in 2-3kHz area. Heard of Flwtcher-Munson curve...? One more reason is that many rooms have high RT in that range too.
Not to forget that between 2-3kHz a lot of speakers cross over to a tweeter that radiates 180 degrees (2π), whereas a lot of midranges and bass-midranges do not. So you get a sound power discontinuity in that region, certainly with acoustical untreated room boundaries. No news though.
often refered in Loodspeaker Cook-books as SD too big mismatch between drivers in multiways designs...
I studied filters in tech school many years ago, sadly without benefit of advanced math skills or software analysis, so it's often difficult to follow along (presently on page 4). But despite my ignorance I hope to make sense from your knowledge or at least speculations of all here on how Pierre perfected the filter needed to eq the Radian 745Be driver to this TH-4001 horn. And also the eq for extending this beryllium driver's HF response past its breakup range and out to ~ 17kHz.
Please also share any other comments, subjective or otherwise, about what must be superbly performing speakers. See posts # 15266, 15276.
https://www.diyaudio.com/community/threads/beyond-the-ariel.100392/page-764
I'll be chatting with Pierre on this and more today, but I'm also interested in your ideas, and also any experience you might have had with beryllium compression drivers, especially in two-way systems like this.
