Hi Chris
I agree with what you are saying. I was more arguing against a purely instrumental approach, which I think might only be possible when instruments are able to measure the whole phenomenon, which is not possible at the moment. Certainly it was the scientific approach that allowed the improvement of sound diffusion systems and should be the reference method for design. Sometimes, however, we are faced with two loudspeakers with the same characteristics but different sounds: there listening (and taste) come into play.
Those who believe that the purely listening-based approach leads to the creation of good-quality speakers usually do not achieve worthy results, partly because as you said people's taste does not always go in the direction of greater realism. One day while talking to a Powersoft executive about the M-Force I heard him tell me that as soon as the system came out, a PA manufacturer had requested to add some distortion to the system as people, in his opinion, were not used to such a clean and undistorted sound in the extreme bass area.
Similarly, in a youtube video a B&C executive tells how many manufacturers continue to demand an HF driver that has a 6-7db breakup peak at about 15000Hz. The new model has corrected this problem but they like it less: some people like the "crisp" sound of the old model!
I have no trouble believing that.
Franco
Last edited:
Here is another source for Axi2050 measurements:
https://data-bass.ipbhost.com/topic/1381-measurements-from-big-cds-and-horns/
it looks good to me but at the same price or cheaper there are a lot of options and perhaps a 2" throat is not the best for home use.
https://data-bass.ipbhost.com/topic/1381-measurements-from-big-cds-and-horns/
it looks good to me but at the same price or cheaper there are a lot of options and perhaps a 2" throat is not the best for home use.
The problem Joseph pointed out about AXI2050 is the multi-tone distortion behavior at higher frequency range.
Can the measurement from other site correlate to the multi-tone measurement?
Can the measurement from other site correlate to the multi-tone measurement?
My problem with the 2" is that it won't fill out a horn in the VHF as well as a 1". I seriously considered this driver for awhile, but I want to build something that's going to have even coverage for as much as the spectrum as possible.
Well, there is nothing to suggest that you can not get perfect dispersion with a 2 inch entry either as long as one can keep the wavefront under control. I described this here:
https://www.diyaudio.com/community/...rivers-today-2022.382609/page-37#post-7321453
https://www.diyaudio.com/community/...rivers-today-2022.382609/page-38#post-7321983
Here is an 1,4 entry horn, and I know for a fact that what we see here is a smooth dispersion rolloff only slightly offset by the Wave front distortion at high frequencies, caused by the phase plug.
https://www.diyaudio.com/community/...rivers-today-2022.382609/page-37#post-7321453
https://www.diyaudio.com/community/...rivers-today-2022.382609/page-38#post-7321983
Here is an 1,4 entry horn, and I know for a fact that what we see here is a smooth dispersion rolloff only slightly offset by the Wave front distortion at high frequencies, caused by the phase plug.
I agree with this. My comment is that we need both listening and measurements--for the reasons given.
This actually isn't correct. The driver is used from 340 Hz - 20 kHz unchanged in the flagship Klipsch Heritage Jubilee (on a K-402). The best I can say is that it is your opinion, not fact. I listened to one covering 275 Hz--20 kHz in a K-402 for a month with the other stereo loudspeaker running a TAD TD-4002/K-402 covering 475 Hz-20 kHz (as detailed in the link just below). The subjective differences were very small, and in a blind test, it would be very difficult to show any differences. However, lobing that is present when using a separate tweeter, especially one crossed above 1.5 kHz due to having more than 1/4 wavelength vertical or horizontal separation between the midrange and tweeter, will be extremely audible as you move around the room.
There's more on this driver being used full-range here, in case you missed it.
Any driver (compression or not) will experience FM distortion (FMD) of the type that's dominant for wide-band drivers. The Axi2050 covers more than 6.5 octaves on a K-402. High frequency FM distortion will be there regardless of what driver is used over the same bandwidth. Drivers with higher efficiency will have diaphragms that move less across its passband, thus lowering their FMD. Horn loading reduces driver FM distortion by at least the ratio of horn loading vs. direct radiating (for the same driver). That's at least 15 dB, i.e., anyone using direct radiating drivers will experience at least 15 dB higher levels of modulation distortion if using a cone or dome driver rather than horn-loading them.
See https://www.klippel.de/fileadmin/kl...es/AN_10_Loudspeaker_FM_and_AM_Distortion.pdf, the figure on page 6 for explanation).
The 1.4 moves the onset of HF narrowing phenomenon from ~6.9 kHz (2" throat) to 9.8 kHz (1.4" throat), but the phenomenon is fairly gradual. The phase plug extension by Klipsch extends that up to 14 kHz (but it's patent pending, as detailed in the first link in this post.
Chris
Last edited:
I’m agree with you with the problematic of lobing effect with separates tweeter that’s why it ´s more easy to get a two way speaker sounds right , but a good coacial like the BMS well filtered and time aligned could do the job . In my experience mixing a 15inch cone driver with a 2inch is a mistake you fast a 6,5 or 10inch driver between to make the transition more transparent but it is only my experience and i have made and will make mistakes 😉
I agree with this. I use a bi-amped BMS 4592ND in the prototype K-402-MEH with dual Crites CW1526C woofers. Getting all the drivers on the same horn aperture and within 1/4 wavelength at crossover is key.
Chris
I bought my Celestions used for a good price but could have gotten used Radian 950 cheaper still.
I had heard the Celestions in a mates DIY horns and was impressed, I havent heard the BMS or Radians. So I bought the Celestion axis. Havent regretted yet 😉
Last edited:
Thanks for the information. FM and AM distortion reminds me the article from purifi
https://purifi-audio.com/2019/12/07/amfm/
The AM distortion is much pronounced to me.
Is the multitone test from Joseph mainly relayed to FM distortion? Seems that's hard to judge from spectrum.
The reason why I linked the Klippel article was to answer that question. Almost all of the high frequency modulation distortion will be FMD, while AMD is mainly found at bass and midrange frequencies.
The way that you test for modulation distortion determines whether or not you can separate AMD from FMD. The Klippel NFS with suitable distortion measurement add-on kit can apparently separate these two distortion types, but the multitone RTA approach used by other measurement apps apparently cannot separate the two distortion forms, AMD and FMD.
So until we get a measurement app like the NFS, we will have to abide by the rule of thumb that Klippel mentioned on page 6 of his AN10 Application Note, linked above. What this means is that, for now, when using multitone RTA, we really have to confine our analytic efforts to below 1 kHz in terms of the contribution of the design of the drivers, since FMD is basically sensitive to only diaphragm radiating area and horn loading, and pass band width. If you change out compression drivers, using like designs of diaphragms, the FMD measurements at higher frequencies should be approximately the same.
What Troy Crowe was apparently originally trying to do in his IMD article was to measure threshold AMD and FMD levels at low drive levels, and compare them to some stressing form of music for masking lower harmonics of an alto flute (which is a very soft instrument to begin with, in real life).
It turns out that his original horn used apparently caused the Axi2050 driver to produce higher levels of FMD at higher frequencies, and fail his arbitrary -65 dB level pass/fail criterion (a threshold insensitive to frequency band or output SPL/sound power). When he changed horns, he revealed to me that he found the Axi2050 passed that standard, but he didn't actually say that in the article.
In my experience using compression drivers, threshold levels of IMD aren't the objectionable form of this type of distortion, but rather when found at higher relative sideband levels--at higher SPL. This is one of two major differences between direct radiating drivers and horn loaded (the other being compression distortion).
Chris
Last edited:
I am slightly worried that a couple of important things may possibly slip by here.
Firstly, what causes FM IMD will typically be a relative fraction of the speed of sound as speed of diaphragm. I did some work on this a few years back, and concluded that in most cases, a trained listener will not be able to hear cone speeds that does not exceed 0,3m/s even under optimum conditions. This equals a 1001Hz tone being modulated to 1000 and 1002Hz.
Try it here:
https://onlinetonegenerator.com/
Try to plot in 1001Hz, push play, and see if you can hear when you change it to 1000 or 1002Hz
If we translate this to a compression driver, and the fundamental is 686Hz. We push it to 0,1mm, which is seriously loud, then the fundamental is only at 0,39m/s. It a driver has a peak excursion of 0,5mm and can handle 100 watts at say 1kHz (which fits well with drivers I have done extensive measurements on). This driver typically delivers around 110dB at 1W, and at 100W, it is still not all the way up to 0,5mm and delivers 130dB at a single frequency. If we imagine then that 0,1mm is just 12dB down from 130dB, we are at 118dB at 0,1mm excursion.
My point is, in most setups, you will be playing extremely loud before you can even begin to talk about doppler effect.
Secondly, we can not really discuss a certain frequency range as being where this or that type of IMD occurs. IMD is and will allways be a type of distortion that involves at least two different frequencies. One will typically be the fundamental and the other will typically be the one distorted by the first. Off course this goes both ways, but it is uncommon that high frequency tones affect low frequency tones significantly.
As two tones must be involved, it is off course also important that they are being reproduced by the same driver.
What we typically see is that for practical use, the treshold of FM IMD varies depending on frequency. We are less sensitive to variations in tonality outside the range where the ear has the highest sensitivity. The practical limit to FM modulation audibility seems to be way over 1m/s, while for a typical bass/midrange that is crossed over at less than 800Hz, up to 1m/s will typically be ok. But drivers being crossed over at that point are normally quite large, meaning that 1m/s is seriously loud, and typically only occurs when the driver has its peak excursion.
The message is really that for mosst practical applications, FM AMD as audible distortion in loudspeakers is almost non existent, while AM AMD is the single largest source of audible distortion after frequency response and possibly dispersion.
Firstly, what causes FM IMD will typically be a relative fraction of the speed of sound as speed of diaphragm. I did some work on this a few years back, and concluded that in most cases, a trained listener will not be able to hear cone speeds that does not exceed 0,3m/s even under optimum conditions. This equals a 1001Hz tone being modulated to 1000 and 1002Hz.
Try it here:
https://onlinetonegenerator.com/
Try to plot in 1001Hz, push play, and see if you can hear when you change it to 1000 or 1002Hz
If we translate this to a compression driver, and the fundamental is 686Hz. We push it to 0,1mm, which is seriously loud, then the fundamental is only at 0,39m/s. It a driver has a peak excursion of 0,5mm and can handle 100 watts at say 1kHz (which fits well with drivers I have done extensive measurements on). This driver typically delivers around 110dB at 1W, and at 100W, it is still not all the way up to 0,5mm and delivers 130dB at a single frequency. If we imagine then that 0,1mm is just 12dB down from 130dB, we are at 118dB at 0,1mm excursion.
My point is, in most setups, you will be playing extremely loud before you can even begin to talk about doppler effect.
Secondly, we can not really discuss a certain frequency range as being where this or that type of IMD occurs. IMD is and will allways be a type of distortion that involves at least two different frequencies. One will typically be the fundamental and the other will typically be the one distorted by the first. Off course this goes both ways, but it is uncommon that high frequency tones affect low frequency tones significantly.
As two tones must be involved, it is off course also important that they are being reproduced by the same driver.
What we typically see is that for practical use, the treshold of FM IMD varies depending on frequency. We are less sensitive to variations in tonality outside the range where the ear has the highest sensitivity. The practical limit to FM modulation audibility seems to be way over 1m/s, while for a typical bass/midrange that is crossed over at less than 800Hz, up to 1m/s will typically be ok. But drivers being crossed over at that point are normally quite large, meaning that 1m/s is seriously loud, and typically only occurs when the driver has its peak excursion.
The message is really that for mosst practical applications, FM AMD as audible distortion in loudspeakers is almost non existent, while AM AMD is the single largest source of audible distortion after frequency response and possibly dispersion.
https://www.toutlehautparleur.com/moteur-de-compression-faitalpro-hf1460-8-ohm-1-4-po
nice discount here for the Faital 1460 , very tempting 🙂
nice discount here for the Faital 1460 , very tempting 🙂
It doesn't need a tweeter, Just discussing this in another thread, heres the Axi on a horn it doesn't perform the best on....Clearly makes it to 20khz
Actually there might not be a better 2" exit driver...
The elephant in the room.
I knew it! Pretty shady to not point that out I think....but at the end of the day, he is just another salesman, selling his own products... Doesn't look good that the SOA driver doesn't play well with his products? I really can't figure out the motivation to skew the results....
He didn't to point it out really... The driver he said, passed his test, actually tested worse
The Driver on the right "passed" his test, the Axi on the left actually has lower IMD
I honestly sense something divisive here.... For the longest time I could not figure out why the top end response of the Axi looked as such, until I finally noticed the Axi measurement goes well beyond 20khz
Last edited:
This thread is actually drifting towards the "modulation distortion" thread that was opined above as a necessary addition to any discussion involving subjective performance of drivers and how to make measurements that correlate to those subjective assessments.
Presently, I need to do my homework using multi-tone RTA in REW before I feel comfortable with the subject, but I can say that "better is the enemy of 'good enough'". I don't believe we should ignore IMD measures because it's hard to be both concise and complete. I think it's better to just start the task of exploring IMD through measurements and correlating to subjective means, and see where it takes us (in the absence of external standards and procedures that may or may not help us). I think this is possibly the one last largely unexplored territories of loudspeaker design, and that it has the potential to really upend current loudspeaker design practice (for those that value realism or even just accuracy in reproduction).
So the bottom line for me is: Let's continue and see if we can collectively find audibility thresholds, and be careful that we're not confounding IMD effects with other loudspeaker or driver distortions at the same time (as I see a lot of with "independent reviews").
The reason why I think we should continue these discussions is because, as I listen to different compression drivers, and even lower frequency drivers, I can hear differences. With the absence of measurements and correlations to IMD, I think that a major piece of the puzzle has been omitted from the conversation. Why do we hear well-designed and well-implemented horn loaded drivers in a qualitatively different way than direct radiating drivers? I think the subjects of IMD and compression distortion (or should I say, its audible absence) makes the sound quality different, and even in small amounts, it may turn out to be.
I do think that FMD is not really a big player in horn-loaded drivers (as you pointed out), but I do think that AMD is the elephant in the room that everyone has been ignoring--for a long, long time.
Chris
Last edited: