One matter
But to me if you wanted a more complete picture, you should have information not based on a single power alone, but also on different loudness level. Some may claim that frequency response at 1nW, 1uW and 1W will be the same. But I believe it is very different. One aspect I noticed in cheap Japanese hifi sets that are ghetto blasters of the 80s, is they could not reproduce low level treble. They could reproduce high level treble though. So the sound lacks "air".
There should be other tests, that hasn't been thought of yet that could reveal more. Some people say the ability for a speaker to reproduce square wave is important. Some say how it respond in the time domain is important. I feel one test would be to feed the speaker an exotic waveform and see how it reproduces the waveform. The jury is out on this one.
As for good measurement speakers that sound poor. One category jumps out at me. Many multiway speakers with big diameter woofers when played soft at night, sounds quite lousy. However small full range drivers on the other hand sounds quite fine. This is an observation that has been observed by quite some friends. Why would that be?
Heavily damp heavy cone speakers will have flatter frequency response and looks as if it will perform better. Yet in practice, many (including some comments here) find a light cone, lowly damped, high efficiency speaker sound better.
Oon
I didn't say it can't be measured. What I said, the standard test that has been coined up that people usually use such as simple frequency response from a sine sweep and THD is not enough . Frequency response is certainly important but there are other important things too that is not considered too.
But to me if you wanted a more complete picture, you should have information not based on a single power alone, but also on different loudness level. Some may claim that frequency response at 1nW, 1uW and 1W will be the same. But I believe it is very different. One aspect I noticed in cheap Japanese hifi sets that are ghetto blasters of the 80s, is they could not reproduce low level treble. They could reproduce high level treble though. So the sound lacks "air".
There should be other tests, that hasn't been thought of yet that could reveal more. Some people say the ability for a speaker to reproduce square wave is important. Some say how it respond in the time domain is important. I feel one test would be to feed the speaker an exotic waveform and see how it reproduces the waveform. The jury is out on this one.
As for good measurement speakers that sound poor. One category jumps out at me. Many multiway speakers with big diameter woofers when played soft at night, sounds quite lousy. However small full range drivers on the other hand sounds quite fine. This is an observation that has been observed by quite some friends. Why would that be?
Heavily damp heavy cone speakers will have flatter frequency response and looks as if it will perform better. Yet in practice, many (including some comments here) find a light cone, lowly damped, high efficiency speaker sound better.
Oon
Last edited:
Purifi blog on hysteresis distortion
^^ This may be enlightening. They don't exactly spell out a solution, but I think it indicates that voltage and current drive are not exactly equivalent here. With voltage drive, the amplifier converts those voltage clicks into current and sound. With current drive, the voltage clicks float harmlessly without producing sound.
In fact, you could do a thought experiment, connecting a current amplifier to a voltage amplifier in place of the speaker, output to output, and something interesting will happen (assuming no technical issues with earth levels, etc.) :
Both amplifiers are configured to "fight" against each other. Say that 'Aj' is playing music, and 'Au' is trying to mute it. To maintain its 0V output, Au will have Aj's music playing through its output transistors in order to make that 0V output.
But if we reverse the situation so Au outputs a music voltage and Aj outputs 0A: Aj only makes tiny corrections for its terminal voltage to follow whatever voltage Au wants it to be, but almost none of Au's music signal is converted into current through Aj's transistors. This is potentially a significant difference.
Either way, it's the current that makes sound, so if a speaker is connected and it attempts to generate a spurious voltage, current drive should be a lot more benign than voltage drive, because it will let those voltages float around without getting converted into current and sound.
^^ This may be enlightening. They don't exactly spell out a solution, but I think it indicates that voltage and current drive are not exactly equivalent here. With voltage drive, the amplifier converts those voltage clicks into current and sound. With current drive, the voltage clicks float harmlessly without producing sound.
In fact, you could do a thought experiment, connecting a current amplifier to a voltage amplifier in place of the speaker, output to output, and something interesting will happen (assuming no technical issues with earth levels, etc.) :
Both amplifiers are configured to "fight" against each other. Say that 'Aj' is playing music, and 'Au' is trying to mute it. To maintain its 0V output, Au will have Aj's music playing through its output transistors in order to make that 0V output.
But if we reverse the situation so Au outputs a music voltage and Aj outputs 0A: Aj only makes tiny corrections for its terminal voltage to follow whatever voltage Au wants it to be, but almost none of Au's music signal is converted into current through Aj's transistors. This is potentially a significant difference.
Either way, it's the current that makes sound, so if a speaker is connected and it attempts to generate a spurious voltage, current drive should be a lot more benign than voltage drive, because it will let those voltages float around without getting converted into current and sound.
And current drive is as old as the road to Rome. I know quite sure KEF investigated it, likely for mass production, in the eighties. Don’t know if that was for Meridian. And to all who adore light cones: don’t let anyone stop you. Accept though your preference probably can‘t be backed up by science or logic. And therefore is merely like taste: nondebatable.
This might be, in essence, what I mean. You don’t like a flat response without resonances. I don’t know what they’re called now but once studios had Aphex aural exciters. They did (and do) a tremendous job in making music more enjoyable. Just by flicking a knob. Affordable too. Just an idea.
This might be, in essence, what I mean. You don’t like a flat response without resonances. I don’t know what they’re called now but once studios had Aphex aural exciters. They did (and do) a tremendous job in making music more enjoyable. Just by flicking a knob. Affordable too. Just an idea.
In theory it should not make a difference, in practice I am not so shure. I think the heavy one with mediocre motor can be made to present the same low level detail, by applying more current to the coil in order to have the same accelleration that enables the perception of low level detail. A heavier cone needs more power to do so, so more current. In that respect a woofer has no "speed". But in practice with higher currents, the weaker magnet field may deform and when this happens it may obscure the fine detail. That is what my gut feeling tell me, there must be scientific literature around that describes this, so if my gut feeling is proven wrong I will be happy to correct myself. I like a stiff magnet, so the movements of the coil will not deform the magnetic field in the gap that much. As in your example, both cars will bring you to Rome, but the Ferrari will pass on every bump to the driver and doing so creates a more intense happening, while the Mitsu will bring you there more comfortably albeit in a boring way.
https://link.springer.com/book/10.1007/978-1-4615-7361-6
That is a bargain!
Even normal price is worth more than just a bunch of "the best" drivers.
Last edited:
^ I'm not an engineer, nor am I very smart to be honest haha. I enjoy music a lot, I have a good set of ears and I like to learn things here and there. I'm not attempting to revolutionize anything, just stating what I like. To each their own. I come on this forum mostly to determine what not to do haha...
While there is nothing wrong with subjective impressions, I'm always baffled when people invest a lot of time and money into this kind of comparisons. You've invested hundreds of hours and tens of thousands USDs, yet in the end, let me put in bluntly, you are shooting blind. Driver and enclosure or horn interact in complex ways, and (with good enough drivers) resulting sound is defined almost completely not by intrinsic properties of driver, but by this very interaction. You can discard great driver as "bleh", but what if in 30% less volume and appropriate baffle step compensation it'll eclipse the others? Or, converseliy, what if that snappy, good for transients mid will be a great pain to make crossover for? You'll never know. Best horns always need precise EQ (just see for yourself here, raw FR is mostly unlistenable https://www.diyaudio.com/community/threads/my-first-crossover-2-way-w-vituixcad.400995/post-7394707 ), so just connecting horn and driver together will never show true sonic potential.
Also, it's next to impossible to make a 3-way crossover, especially passive, by ear.
Acoustic measurements now are easier and cheaper than ever before. Beginner setup can be $300 or lower for mic stand, electret measurement mic and audio interface, or about $1-2k for a really nice setup, and it takes about an hour or two to get familiarised with setup and begin taking measurements. Compared to funds and labor you are already invested, this is a trivial amount, and it will allow you to get your desired sound literally fifty times faster and easier. For a starting point, I prefer written guides like https://kimmosaunisto.net/Software/VituixCAD/VituixCAD_Measurement_REW.pdf , but if you are a video guy, YouTube is full of instrucions too: https://www.youtube.com/results?search_query=rew+acoustic+measurement
Also, it's next to impossible to make a 3-way crossover, especially passive, by ear.
Acoustic measurements now are easier and cheaper than ever before. Beginner setup can be $300 or lower for mic stand, electret measurement mic and audio interface, or about $1-2k for a really nice setup, and it takes about an hour or two to get familiarised with setup and begin taking measurements. Compared to funds and labor you are already invested, this is a trivial amount, and it will allow you to get your desired sound literally fifty times faster and easier. For a starting point, I prefer written guides like https://kimmosaunisto.net/Software/VituixCAD/VituixCAD_Measurement_REW.pdf , but if you are a video guy, YouTube is full of instrucions too: https://www.youtube.com/results?search_query=rew+acoustic+measurement
On a slightly different note, depending on how much a purist you are, adding a DSP and tuning your system will make a very big difference too. And it will be easier to achieve a high end sounding system rather than just swapping driver. Investing in a minidsp for use as a crossover and DSP will make a bigger difference, especially if it has DIRAC room correction. YMMV.
Oon
Oon
Still, I think @milezone gets a lot of it right. I suppose most of us haven't actually heard most of the drivers listed (I haven't), so these sort of anecdotes definitely have value if you don't want to go through a similar process of re-boxing and selling drivers at a significant cost.
Looking at the Satori MW19TX-8 vs. Visaton B200 pdfs
Wow. This gives clues as to the test set-up and preferences, and how well it might compare to my/your own "use cases".
The Satori obviously falls into the niche "we've given you jumbo Xmax at great cost so use it!"
While the B200 is the complete opposite, and looks like an underhung unicorn with its 4mm coil vs 8mm magnet gap.
That's not to doubt the Satori. We're just not dealing with compact 7" bookshelf speakers delivering 8mm of slamming bass. And FWIW, the analysis of the physics seems pretty accurate at a glance. The Satori has multiple opportunities for reflections to bounce around between the cone, the dome glued on top, and the soft-yet-low-damping rubber surround. Whereas the Visaton seems to do things in a more classical manner, but with other likely costs (large box for soft cushioning, low power, low max bass capability, etc.)
Looking at the Satori MW19TX-8 vs. Visaton B200 pdfs
Wow. This gives clues as to the test set-up and preferences, and how well it might compare to my/your own "use cases".
The Satori obviously falls into the niche "we've given you jumbo Xmax at great cost so use it!"
While the B200 is the complete opposite, and looks like an underhung unicorn with its 4mm coil vs 8mm magnet gap.
That's not to doubt the Satori. We're just not dealing with compact 7" bookshelf speakers delivering 8mm of slamming bass. And FWIW, the analysis of the physics seems pretty accurate at a glance. The Satori has multiple opportunities for reflections to bounce around between the cone, the dome glued on top, and the soft-yet-low-damping rubber surround. Whereas the Visaton seems to do things in a more classical manner, but with other likely costs (large box for soft cushioning, low power, low max bass capability, etc.)
Thanks for your elaborate write-up. It's one of the best I've come across in the western world.
In this case it's (also) best to ignore the comments about measurements.
There will always be people who seem to listen through a screen with measurement data.
Anybody with a sensible pair of ears, some processing intelligence and some frame of reference should appreciate your efforts.
That said, regarding compression drivers, the horns/waveguides do make a difference, as you'll know.
I share a lot of your findings/conclusions, a.o. regarding annular diaphragms and light cones.
You liked the 18Sound 15W500 and I can see why. I was tempted to buy a pair a few months ago (as new in box) that were offered locally.
It's one of the few modern pro drivers that share some characteristics with the best (sounding) drivers of the past > Altec, the old JBL instrument woofers, EV etc. I decided to forgo the 15W500s, because they lack a bit of motor power and damping for my taste (low-ish BL and relatively high-ish Qts) and I already have 2 pairs of 15" drivers with similar parameters (Fs: 50 Hz, Bl: 15.7, Mms: 51 g, Qts: 0.33). Xmax is extremely low, but you never put more than just a few watts in these.
The Purifi 6.5" PTT6.5M08 is a very good driver, as it should be considering the price.
Part of the sound quality is due to the cone/dust cap > made in China
Still, I don't think it sounds much better/more natural in the midrange than some oldskool 6.5" drivers from the 80s.
Last edited:
These are legitimate arguments to a certain extent.
It's pretty insane/useless to use inappropriate drivers in a given configuration, or to omit necessary corrections.
Unless you can circumvent the issue, something that applies to BSC.
Although I also use DSP, I still think that a well-designed passive crossover that's as minimalist as possible, combined with suitable drivers, has the edge.
"The best horns always need accurate EQ": it depends on your definition of the 'best horns'.
This raw response of a 70-90 years old midhorn + driver combo (in red) is hard to beat:
Last edited:
Many people in the DIY community seem to ignore/underestimate the impact of the electronics > DAC, amplifier, etc.
I've listened to several SOTA active speakers, such as the Grimm LS1be, as well as both Kii models (the Threes with and without BXTs).
While these systems are highly rated, I personally would only consider buying the Kii Seven... for 50% of the MSRP. Why?
Because these speakers all sound 'processed' compared to the best systems I know.
The key component to (both of the) Kii's relatively natural sound is the mid driver, which costs about $12.
From your website:
"In summation, a compression driver is most logically implemented as a tweeter. A compression driver horn combination is advantageous over direct radiating tweeters, amts, ribbons, etc. in that they project hf content into air / space with far greater velocity and controlled directivity, and thus further distances than a conventional direct radiating device. This is important as hf content rolls off faster than mids and lows. One can think of a hf waveform with its short fast wave length as more fragile and easily impeded by air, compared with more forceful mid and low frequency waves. Thus lf waves travel greater distances through air versus hf waves at equal volume. Air over distance from a sound source, as it effects sound waves, acts as a low pass filter. Increasing the velocity of hf output through the incorporation of horns and compression drivers for midfield and farfield sound reproduction reduces this effect and is thus worthwhile."
This is a good summary of how horns work, in short: the conversion of high pressure (force / area) to high volume velocity (particle velocity x area).
Although horns are the obvious choice if efficiency is important (always in my opinion), the 'listening experience' depends on many factors, including the space and listening distance as you rightfully mention. The size, shape, type of horn are equally important, as is the type of driver > oldskool drivers with long exit sections don't go well with modern waveguides.
VoxCelestial said: "Driver and enclosure or horn interact in complex ways, and (with good enough drivers) resulting sound is defined almost completely not by intrinsic properties of driver, but by this very interaction."
This is only partly true, because certain intrinsic characteristics that determine the 'listening experience' have little or nothing to do with the combination of driver + horn, such as breakup.
The subject of (diaphragm) materials alone is a science in itself.
Some people honestly believe that similar-sized compression drivers hardly sound different, once eq-ed...
Those people will probably skip this thread, which is wise/logical in their case.
"In summation, a compression driver is most logically implemented as a tweeter. A compression driver horn combination is advantageous over direct radiating tweeters, amts, ribbons, etc. in that they project hf content into air / space with far greater velocity and controlled directivity, and thus further distances than a conventional direct radiating device. This is important as hf content rolls off faster than mids and lows. One can think of a hf waveform with its short fast wave length as more fragile and easily impeded by air, compared with more forceful mid and low frequency waves. Thus lf waves travel greater distances through air versus hf waves at equal volume. Air over distance from a sound source, as it effects sound waves, acts as a low pass filter. Increasing the velocity of hf output through the incorporation of horns and compression drivers for midfield and farfield sound reproduction reduces this effect and is thus worthwhile."
This is a good summary of how horns work, in short: the conversion of high pressure (force / area) to high volume velocity (particle velocity x area).
Although horns are the obvious choice if efficiency is important (always in my opinion), the 'listening experience' depends on many factors, including the space and listening distance as you rightfully mention. The size, shape, type of horn are equally important, as is the type of driver > oldskool drivers with long exit sections don't go well with modern waveguides.
VoxCelestial said: "Driver and enclosure or horn interact in complex ways, and (with good enough drivers) resulting sound is defined almost completely not by intrinsic properties of driver, but by this very interaction."
This is only partly true, because certain intrinsic characteristics that determine the 'listening experience' have little or nothing to do with the combination of driver + horn, such as breakup.
The subject of (diaphragm) materials alone is a science in itself.
Some people honestly believe that similar-sized compression drivers hardly sound different, once eq-ed...
Those people will probably skip this thread, which is wise/logical in their case.
Just ordered a pair of 15W500's . Plan to cross around 650hz to a 16" horn with a monster 4"VC JBL CD. Fingers crossed. My main two considerations this driver seems to fit the bill for, 1: surprising smooth off axis and 2: good low listening perameters for sub watt listening nearfield. Milezone, did you end up going with these?? How's your experience so far. Rgds,
EliGuy
EliGuy
There are subjective observation/perception, measurements and technical clarifications.
Most technical phenomena can be measured, but there is a limit imposed by complexity, i.e. multiple interdependent factors at play at different output levels.
Large diameter woofers can sound fantastic at low listening levels, but this depends entirely on the technical composition of the woofer(s) used.
Large hi-fi woofers with rubber surrounds, or PA subwoofers with heavy cones and huge magnets will often sound poor at low levels.
Why? Losses!
Reasons: miniaturization (hi-fi) & power handling (PA).
Not to say I am doubting it, I a curious what are the drivers they are using.
I have heard the Kii.... I didn't quite like it. Sounds a bit unnatural. Quite agree with you, processed is the word. On the other hand I have heard the NAD C3050 playing on a PSB speaker with DIRAC correction, sounds incredible and life like...
So YMMV.
Oon
I cannot for the life of me imagine why a larger (and by extension, "stronger") magnet would sound worse at low power levels. If nothing else, the additional flux should help resist flux modulation.
EDIT: Assuming a flat/constant frequency response. A poorly-tuned system might have less bass with lower driver Qts, which would sound especially weak at low levels due to hearing sensitivity curve.
Yeah it is quite hard to think reasons why low output level would make worse sound than loud, other than if the system balance / boxes etc. were tuned for loud operation where heat changes things for example, which makes the low level listening just unoptimized.
Think this: a fullrange driver would quite likely compress quite much if one tries to listen any loud, which reduces low frequency output. Conversely, low frequency output is higher at low listening level when it is not compressing, which would compensate for Fletcher Munson curve like good old loudness button which might make impression the balance is fine for both low and loud listening. If this is true explanation for the phenomenon, this would not be so apparent with big system that doesn't compress and sound would be different but is is not fault of the system but due to ear having different frequency response at various levels, and the level which the system was optimized for. For bigger system one could use the good old loudness button, or perhaps modern systems can automatically compensate with position of volume knob.
My fullrange driver system sounded very good, except it didn't play party level which eventually killed the drivers. Compression was quite audible with it, it just didn't get any louder even if amp was cranked. On the other hand, my big multi-way system sounds very good both quiet and loud listening level, except frequency response balance is bit off either on low or loud, which is not a problem because I have buttons to optimize for any output level, and there is no compression even with good party level.
With this small experience, I'd say these low level listening sound are likely true in that sense that ear frequency response changes with level, while the reasons often used don't seem to make sense to me except through compression. Another point of view might be that with louder output there is more distortion, which makes the sound better to ear.
Think this: a fullrange driver would quite likely compress quite much if one tries to listen any loud, which reduces low frequency output. Conversely, low frequency output is higher at low listening level when it is not compressing, which would compensate for Fletcher Munson curve like good old loudness button which might make impression the balance is fine for both low and loud listening. If this is true explanation for the phenomenon, this would not be so apparent with big system that doesn't compress and sound would be different but is is not fault of the system but due to ear having different frequency response at various levels, and the level which the system was optimized for. For bigger system one could use the good old loudness button, or perhaps modern systems can automatically compensate with position of volume knob.
My fullrange driver system sounded very good, except it didn't play party level which eventually killed the drivers. Compression was quite audible with it, it just didn't get any louder even if amp was cranked. On the other hand, my big multi-way system sounds very good both quiet and loud listening level, except frequency response balance is bit off either on low or loud, which is not a problem because I have buttons to optimize for any output level, and there is no compression even with good party level.
With this small experience, I'd say these low level listening sound are likely true in that sense that ear frequency response changes with level, while the reasons often used don't seem to make sense to me except through compression. Another point of view might be that with louder output there is more distortion, which makes the sound better to ear.
Last edited: