FYI: the left channel's measurements show a burst in odd-order distortions at 900Hz, going all the way up to 10 kHz, with 3rd harmonic at approx -40dB:
For these relatively high distortions, my subjective assessment of audio quality and the sine sweep measurements agree.
BTW, the right channel is smooth.
AFAIK, there are 2 drop-in replacements for 4K211: HM100C0 and HM100Z0. Both have been reviewed on diyaudio:
https://www.diyaudio.com/community/threads/audax-hm100z0.6367/post-4528750
https://www.diyaudio.com/community/threads/audax-hm130c0-opinions.153064/post-1944482
Alas, Z0 can not be crossed at 250Hz.
The measurements are nice in both cases (nothing like what is above) but the subjective opinions of people who used them are firmly on the negative side:
"I have used the HM100CO's, which I had crossed over to the tweeters at around 3200 Hz. They had a noticeable sibilance with certain female vocalists. The detail was incredible, but I couldn't get past the sibilance which I attribute to cone breakup."
"This is my experience of the Aerogels as well. They do sound good and have a very appealing type of sound to them. It's only when you get quite critical of the sound that you realise you can do better."
Enough for this particular case.
For these relatively high distortions, my subjective assessment of audio quality and the sine sweep measurements agree.
BTW, the right channel is smooth.
AFAIK, there are 2 drop-in replacements for 4K211: HM100C0 and HM100Z0. Both have been reviewed on diyaudio:
https://www.diyaudio.com/community/threads/audax-hm100z0.6367/post-4528750
https://www.diyaudio.com/community/threads/audax-hm130c0-opinions.153064/post-1944482
Alas, Z0 can not be crossed at 250Hz.
The measurements are nice in both cases (nothing like what is above) but the subjective opinions of people who used them are firmly on the negative side:
"I have used the HM100CO's, which I had crossed over to the tweeters at around 3200 Hz. They had a noticeable sibilance with certain female vocalists. The detail was incredible, but I couldn't get past the sibilance which I attribute to cone breakup."
"This is my experience of the Aerogels as well. They do sound good and have a very appealing type of sound to them. It's only when you get quite critical of the sound that you realise you can do better."
Enough for this particular case.
kapelli, I'll try to give you some useful advice on the choice of a midrange unit using measurements. This would be different, though related, to measurements to determine a good sounding speaker system.
Harmonic & Intermod Distortions are actually quite low on this list so don't worry too much if your unit has 1% THD at 94dB spl. I've designed speakers with this yucky level of THD which have been described as sounding "very low distortion" compared to famous speakers with very low measured distortion in DBLTs
even by some of the best ears in da business.
So all da stuff about more linear motor systems bla bla bla, while I've spent a LOT of time trying to improve them, I'm under no illusion about where the emphasis should be.
- The response (@1m on axis bla bla) should be smooth and not have big yucky peaks outside its operating frequency range. This is a particular sin of metal diaphragms / cones / domes. You need this so your crossover doesn't have extra complication. eg dealing with the 'bass resonance' of the mid or the aforementioned peaks. Supersonic peaks (ALWAYS found in metal dome treble units) might be OK
- It should have a nice 'waterfall' in its operating range and bit beyond. If the 'cliff' in the waterfall drops 20dB, it is probably blameless in this respect. Look up da false prophets Toole & Olive on the audibility of various 'delayed resonances' at these levels or better still ... the papers we wrote on the subject which they quote. There's loadsa stuff on interpretation of 'waterfalls' which I won't go into now except to say most 'waterfalls' are useless below 1kHz.
Harmonic & Intermod Distortions are actually quite low on this list so don't worry too much if your unit has 1% THD at 94dB spl. I've designed speakers with this yucky level of THD which have been described as sounding "very low distortion" compared to famous speakers with very low measured distortion in DBLTs
even by some of the best ears in da business.So all da stuff about more linear motor systems bla bla bla, while I've spent a LOT of time trying to improve them, I'm under no illusion about where the emphasis should be.
Can you link to the measurements of these units you are considering?
I think you can find most of them here, including waterfalls
https://hificompass.com/en/speakers/measurements
Hi mikets42,
On my speakers, I thought the tweeters were terrible. This in comparison to others, these speakers have aluminum domes. So my opinion of them was pretty low. Then I updated my amplifier and all that was gone. The issue was the amplifier and very accurate speakers. The fault was the amplifier. Yes, the amplifier measurements were also better once I was done.
To give specifics, the speakers are PSB Stratus Gold (original) and the amp in question is a Marantz 300DC. Now people remark on how clean and easy to listen to the system is.
What I'm trying to illustrate is that the speakers may be really good, but showcase shortcomings in the electronics. So I would be careful of subjective opinions without knowing what the rest of their system is. Therefore I would not worry too much about opinions if the measurements are otherwise good.
Another case, I bought a set of Klipsch THX 6000 speakers for my bench without auditioning them (no one had them). You'll see they aren't cheap. Anyway, they perform better than expected and even measured very smooth for frequency response, two people, different calibrated mics and computer programs. Anyway, I'm more than happy with them. I have the older equivalent to the Klipsch RP8000f in the bedroom, really nice.
So you can actually trust some measurements and get pretty much what you want. Mind you, I am very familiar with Klipsch products. If I wasn't, I would definitely listen to them first. The pools would be narrowed by examining the measurements.
On my speakers, I thought the tweeters were terrible. This in comparison to others, these speakers have aluminum domes. So my opinion of them was pretty low. Then I updated my amplifier and all that was gone. The issue was the amplifier and very accurate speakers. The fault was the amplifier. Yes, the amplifier measurements were also better once I was done.
To give specifics, the speakers are PSB Stratus Gold (original) and the amp in question is a Marantz 300DC. Now people remark on how clean and easy to listen to the system is.
What I'm trying to illustrate is that the speakers may be really good, but showcase shortcomings in the electronics. So I would be careful of subjective opinions without knowing what the rest of their system is. Therefore I would not worry too much about opinions if the measurements are otherwise good.
Another case, I bought a set of Klipsch THX 6000 speakers for my bench without auditioning them (no one had them). You'll see they aren't cheap. Anyway, they perform better than expected and even measured very smooth for frequency response, two people, different calibrated mics and computer programs. Anyway, I'm more than happy with them. I have the older equivalent to the Klipsch RP8000f in the bedroom, really nice.
So you can actually trust some measurements and get pretty much what you want. Mind you, I am very familiar with Klipsch products. If I wasn't, I would definitely listen to them first. The pools would be narrowed by examining the measurements.
I feel this is too dismissive of magnetic non-linearities and fails to make the discrimination between different distortion mechanisms.kapelli, I'll try to give you some useful advice on the choice of a midrange unit using measurements. This would be different, though related, to measurements to determine a good sounding speaker system.
These are my 'first order faults' and are the most important audible ones. You'll note the papers I quote are a lot about audibility of distortions. After all, if most (all?) people can't detect them, why strive for supa dupa performance here when you can spend $$$ on reducing the audible distortions?
- The response (@1m on axis bla bla) should be smooth and not have big yucky peaks outside its operating frequency range. This is a particular sin of metal diaphragms / cones / domes. You need this so your crossover doesn't have extra complication. eg dealing with the 'bass resonance' of the mid or the aforementioned peaks. Supersonic peaks (ALWAYS found in metal dome treble units) might be OK
- It should have a nice 'waterfall' in its operating range and bit beyond. If the 'cliff' in the waterfall drops 20dB, it is probably blameless in this respect. Look up da false prophets Toole & Olive on the audibility of various 'delayed resonances' at these levels or better still ... the papers we wrote on the subject which they quote. There's loadsa stuff on interpretation of 'waterfalls' which I won't go into now except to say most 'waterfalls' are useless below 1kHz.
Harmonic & Intermod Distortions are actually quite low on this list so don't worry too much if your unit has 1% THD at 94dB spl. I've designed speakers with this yucky level of THD which have been described as sounding "very low distortion" compared to famous speakers with very low measured distortion in DBLTs
So all da stuff about more linear motor systems bla bla bla, while I've spent a LOT of time trying to improve them, I'm under no illusion about where the emphasis should be.
Can you link to the measurements of these units you are considering?
I can make a good case that the last 25 years or so of improvements in moving coil drivers have not arisen from the control of resonances (although as your point 1 remarks, several approaches have made matters worse!). I argue instead that the last 25 years of improvements have been dominated by better magnetic/motor non-linearity. Manufacturers would simply not have bothered with the extra work and expense if there were no subjective benefits to be gained.
Whether a matter is immediately more obvious than another is not the question. As far as I am aware, chasing a mythical single measure of "sound quality" had been eschewed in this thread? It is unlikely to be successful for the reasons I have cited. More importantly here, however, is that prioritizing what is known is never likely to push the envelope forward. Progress in moving-coil driver design has simply made lower level distortions more discernible, hence why I believe magnetic non-linearities are now an example of that warranting more attention.
Hi markbakk,
Where is your money going?
Well, of course! So, to you, what is the weakest link? This thread is pretty much all about progress in electromagnetic transducers. We have seen the biggest improvements in this field over the last few decades since the 1970's. Dramatic improvements.
Where is your money going?
The important measurement is DBLT. I was going to go into the 'distortions' that are really audible in Jim's new thread compared to those which DBLTs show are quite minor. Alas that's fallen into a nit-picking pit of how to measure TSPs when practically ALL methods today to dream these up are good enough to do good box design.
And I HAVE contributed to the theory of box design. As makers, we used different numbers which highlight what's important to us for consistent production. Somewhere I have 7 pages of algebra which showed the equivalence of my model to Thiele - Small etc. a mind numbing experience I'm unlikely (and now probably unable) to repeat
That's not to say I haven't spent a LOT of effort trying to improve magnetic non-linearities bla bla. But the HUGE AUDIBLE DISTORTIONS that we (and to some extant, da false prophets Toole & Olive) have identified, remain even in the 'best' speakers today.
If you have links to proper DBLTs which show the audible importance of these magnetic non-linearities, I am certainly interested.
Anyone come up with the units that they claim have audible Barkhausen noise on a sweep?
Better directivity control and better low end reproduction in normal listening (aka living) rooms. I find non-linear distortion hardly an issue these days. And while electromagnetic transducers certainly have improved, I think there's a lot of marketing blabla going on in DIY and in high end.
Could not agree more with markbakk.
Currently there seems be an obsession with:
a) the desire for ridiculous TSP parameter accuracy and vented design precision.
b) the -claimed- "sonic characteristics" of individual drivers, yet completely ignoring system design, equalization and system directivity characteristics. The more expensive, the better seems to be the fashion of the day.
Three extraordinary drivers do not make a good 3-way.
Currently there seems be an obsession with:
a) the desire for ridiculous TSP parameter accuracy and vented design precision.
b) the -claimed- "sonic characteristics" of individual drivers, yet completely ignoring system design, equalization and system directivity characteristics. The more expensive, the better seems to be the fashion of the day.
Three extraordinary drivers do not make a good 3-way.
Last edited:
Thats the first thing everyone do, calculate a box with some calculator, then choose driver that makes 1db more output at some hz While it feels important thing to do, it really isn't very important but actually is the simplest thing one can do, the lowest energy thing one can do, an entry point. Or reading others posts and adverticements, watch influenser on youtube say what you need to buy for instasuccess(tm), and then go and transwer the wealth. Driver arrives, model the box and slap it together. Sound, nice!
All the other stuff beyond this requires considerable amount of energy to pull through, like building the damn thing successfully so that it is allowed to reside in family living room. Let alone making any measurements of a speaker, succesfully, and then having any understanding of what the measurements show just to realize the system is poo and must be rebuilt from scratch. At prototype number four one notices the box calculator is not even needed anymore and all energy goes into some other things, like trying to figure out what kind of a system I need for my room and practical positioning, in order to sound be at ears such that my auditory system is fooled nicely that my perception is as good as it gets. And, one could do 100 speakers, but if never developing listening skill and trying to figure what is important and what is not, because if not knowing what to strive for then it's just random. Well, everyone have fun in their own way As long as it's fun it's the right thing to do in my opinion, gotta start somewhere. After enough energy consumed, system would develop to better and better, basically system develops with listening skill and understanding of concepts in first place, and then successfully implementing them, and evaluating the system really went forward, and not just randomly different.
While it feels important thing to do, it really isn't very important but actually is the simplest thing one can do, the lowest energy thing one can do, an entry point. Or reading others posts and adverticements, watch influenser on youtube say what you need to buy for instasuccess(tm), and then go and transwer the wealth. Driver arrives, model the box and slap it together. Sound, nice!All the other stuff beyond this requires considerable amount of energy to pull through, like building the damn thing successfully so that it is allowed to reside in family living room. Let alone making any measurements of a speaker, succesfully, and then having any understanding of what the measurements show just to realize the system is poo and must be rebuilt from scratch. At prototype number four one notices the box calculator is not even needed anymore and all energy goes into some other things, like trying to figure out what kind of a system I need for my room and practical positioning, in order to sound be at ears such that my auditory system is fooled nicely that my perception is as good as it gets. And, one could do 100 speakers, but if never developing listening skill and trying to figure what is important and what is not, because if not knowing what to strive for then it's just random. Well, everyone have fun in their own way
As long as it's fun it's the right thing to do in my opinion, gotta start somewhere. After enough energy consumed, system would develop to better and better, basically system develops with listening skill and understanding of concepts in first place, and then successfully implementing them, and evaluating the system really went forward, and not just randomly different.
Last edited:
Hi kgrlee,
I've designed enough speaker systems over the years, before computers and with computers. The basic physics have not changed but our understanding has improved considerably.
Hi markbakk,
In high end audio, there is always a lot of talk done by folks who haven't a clue, and certainly marketing types. So in other words, blah, blah, blah. However that hasn't anything to do with the folks who actually do know what they are talking about. You'll find these people don't have much time to post and yak.
Directivity control has been a concern since the very early days for one and better low end reproduction has always been a top concern. So I'm not sure what your point is. Everything matters in a balance, including individual driver quality. Even back in the late 1970's, I agonized over drive quality to design good sounding systems. It is always important.
Hi Boden,
To your point "b". A balanced approach is always taken by any serious designer. That is a cornerstone of engineering.
What is talked about in forums open to the public is often a reflection of various marketing campaigns and "white papers". This also speaks to your first point, anyone who has designed a speaker system and measured it under various conditions is well aware that the other variables make "perfect" unattainable. VC temperature, air pressure and density along with other things will shift tuning. So the only people who talk about "perfect tuning" are those who have no experience.
I think you've made your feelings known. But Toole's work was solid and extensive. I don't know Olive, but if they are associated with Floyd Toole, I'll bet their work is also founded in careful science and experimental procedure as well. What I hear is someone who disagrees with work done by people maybe they can't compete with. Maybe you can drop the descriptors and simply refer to them as "Toole & Olive". We all know who you mean. You can attack the idea, but not the person. Very unprofessional.
I've designed enough speaker systems over the years, before computers and with computers. The basic physics have not changed but our understanding has improved considerably.
Hi markbakk,
In high end audio, there is always a lot of talk done by folks who haven't a clue, and certainly marketing types. So in other words, blah, blah, blah. However that hasn't anything to do with the folks who actually do know what they are talking about. You'll find these people don't have much time to post and yak.
Directivity control has been a concern since the very early days for one and better low end reproduction has always been a top concern. So I'm not sure what your point is. Everything matters in a balance, including individual driver quality. Even back in the late 1970's, I agonized over drive quality to design good sounding systems. It is always important.
Hi Boden,
To your point "b". A balanced approach is always taken by any serious designer. That is a cornerstone of engineering.
What is talked about in forums open to the public is often a reflection of various marketing campaigns and "white papers". This also speaks to your first point, anyone who has designed a speaker system and measured it under various conditions is well aware that the other variables make "perfect" unattainable. VC temperature, air pressure and density along with other things will shift tuning. So the only people who talk about "perfect tuning" are those who have no experience.
Agreed on that, with the exception of the directivity control and low end reproduction being top on the list. That may be true to some extent in studio monitors, but not in the consumer market, not even hi end. You asked where my money went. Not in ridiculously overpriced drivers or crossover components.
I know you asked Mark, but I will give my answer:
I believe that 90% of loudspeaker goodness comes from the frequency response, both on and off axis. Properly controlling and managing the on-axis, sound power, early reflection, and predicted in-room responses is 90% of the battle... maybe 95%.
Only after we do a competent job on the frequency response do the other aspects matter... things like lowering distortion, eliminating small resonances, minimizing diffraction secondary sources, lowering power compression, etc.
If we design a speaker that has a flat on-axis response that is +/- 1 dB from 40 - 20k, has a DI that smoothly rises from 0 dB at 100 Hz to 10 dB at 10k, has a power response that smoothly falls at 0.8 dB/octave from 100 Hz to 10k... that speaker is going to sound very good... even if it has average "nothing special" performance from a distortion and power compression standpoint.
So let's get the important things right, and make that the first job. Then we can worry about esoteric concerns like very low motor distortions, extreme cabinet resonance suppression, advanced diaphragm materials. Those aspects ARE important in high performance designs... the kind of designs that do an excellent job of getting the FR/DI/Sound-Power aspect right.
j.
The basic question of the topic... My answer is yes of course.
But if one is designing a speaker the most important and first questions should be what spl and low end limit is needed. Then one must determine directivity aimed. Then one should know how many drivers/ways will be needed and approximate size and xmax of drivers and size of the box needed to achieve the goal.
Then one can start choosing drivers and considering eg. xo points and slopes, possible reflex system etc. within the limits of the drivers chosen. Impedance, Fs, Sd and Xmax are basic T/S numbers to watch, and in most cases factory specs are good enough. Then look at cone resonances etc. distortions also at least one octave past the intended xo point. There are many sites that offer measured data. Still it is very important to measure the chosen drivers in intended boxes to find flawed units and to see box resonances etc.
Magnet/coil system, cone/dome material and surround type and material might have some sonic characters, but effects are minimal when a driver is used in it's comfort zone.
DIY pages and forums seem to be focused on small BR 2-ways or subwoofers (or fullrangers). Both are actually quite challenging to design and to make very good. A moderately sized woofer for a 2-way is almost mission impossible, Purifi seems to be the best performer there. 2-way tweeters must use waveguides and they are also very challenging to make and match with.
But if one is designing a speaker the most important and first questions should be what spl and low end limit is needed. Then one must determine directivity aimed. Then one should know how many drivers/ways will be needed and approximate size and xmax of drivers and size of the box needed to achieve the goal.
Then one can start choosing drivers and considering eg. xo points and slopes, possible reflex system etc. within the limits of the drivers chosen. Impedance, Fs, Sd and Xmax are basic T/S numbers to watch, and in most cases factory specs are good enough. Then look at cone resonances etc. distortions also at least one octave past the intended xo point. There are many sites that offer measured data. Still it is very important to measure the chosen drivers in intended boxes to find flawed units and to see box resonances etc.
Magnet/coil system, cone/dome material and surround type and material might have some sonic characters, but effects are minimal when a driver is used in it's comfort zone.
DIY pages and forums seem to be focused on small BR 2-ways or subwoofers (or fullrangers). Both are actually quite challenging to design and to make very good. A moderately sized woofer for a 2-way is almost mission impossible, Purifi seems to be the best performer there. 2-way tweeters must use waveguides and they are also very challenging to make and match with.
Last edited: