I have a stand mounted speaker based on Scan-Speak 8545K midwoofer and a Satori TW29RN tweeter. Crossover is DSP. I really love my speakers, which I have been able to dial-in very well in my room through many hours of measurements, snacks and beer. 🙂
But you know how it is... you end up dreaming about a new project...
The main negative aspect of my current speaker is that it seems to loose it's "grip" of the music when I turn down the volume. I have to crank it up a bit before the bass and lower midrange really engages. Hard to explain, I was kind of hoping someone here shares my experience.
My question: What design choices should I be looking into if I want to design a speaker which performs really well at low volumes?
But you know how it is... you end up dreaming about a new project...
The main negative aspect of my current speaker is that it seems to loose it's "grip" of the music when I turn down the volume. I have to crank it up a bit before the bass and lower midrange really engages. Hard to explain, I was kind of hoping someone here shares my experience.
My question: What design choices should I be looking into if I want to design a speaker which performs really well at low volumes?
Yes, I am aware of this. I should add that I have the option of adding a separate crossover configuration for low volume listening. I have been entertaining this idea for a while.
I have read somewhere that small mildrange drivers around 4” provide better midrange performance at low levers, bit I might be mistaken.
- Dipole?
- Classic 3-way?
- MTM?
Hos do the different design principles behave at lower volume? Are there any general differences?
I have read somewhere that small mildrange drivers around 4” provide better midrange performance at low levers, bit I might be mistaken.
- Dipole?
- Classic 3-way?
- MTM?
Hos do the different design principles behave at lower volume? Are there any general differences?
with DSP. aren't multiple XOs or EQ contours readily achievable, or does it require more than a simple "toggle" ? Before re-engineering to any of the more complicated configurations pondered, I'd try that first
Last edited:
depending on the available dsp memory he could program an entire bank of presets to act like a "step" volume control with each step "eq" contoured.
Harmonic distortion is not a problem anymore at low levels.
Dispersion could be relevant. At lower sound pressure levels, reflections can drop below your hearing threshold. Therefore wide dispersion (actually: low directivity index) could be beneficial in order to still hear sufficient reflected sound.
Dispersion could be relevant. At lower sound pressure levels, reflections can drop below your hearing threshold. Therefore wide dispersion (actually: low directivity index) could be beneficial in order to still hear sufficient reflected sound.
Design one that performs well at higher volumes..
I have built many speakers with this failing. There is one good listening level, too loud or too quiet sounds different. Dynamic transparency is compromised and detail suffers. I put it down to small voice coils and/or low efficiency.
My two main criteria for a speaker which performs really "well" (i.e., which still sounds "alive" and not "dynamically muffled") at low volumes would be:
- Use drivers that require low excursion (i.e., use a large woofer for the low frequencies, and a solution for the mid-high frequencies that does not require the upper driver(s) to move much at all, i.e. either a waveguide/horn-loaded mid-tweeter or a dedicated mid-range + separate tweeter)
- Use a woofer with low mechanical losses (i.e., with low Rms relative to its Mms, which translates to a low (Fs/Qms) ratio)
You'll quickly realize that this short list automatically rules out all common 2-way standmounts equipped with a 4-8" mid-woofer that's designed/asked to be "flat" from 40-60Hz all the way up to 2-3kHz, whereupon it is directly crossed over to a 1" dome tweeter...
Marco
On mechanical losses (Rms and Qms)
First of all, the mathematical relationship:
(Rms/Mms) = 2*Pi*(Fs/Qms)
where:
Rms = mechanical resistance
Mms = moving mass
Fs = free air resonance frequency
Qms = mechanical 'quality' factor (= 1/bandwith of resonance peak)
And then a few quotes:
From an interview of a German DIY magazine editor (Google translated):
"Anyone who observes the Thiele-Small parameters of a woofer, turns his attention first on resonance frequency and Qts: These two parameters determine the achievable lower cutoff frequency. The second view is the equivalent volume - Vas: It reveals what cabinet size is necessary to achieve this goal.
These parameters say little about the sound character : whether the finished box will sound quick or slow, betrays - apart from the key priority of course crossovers and housing - the inconspicuous parameter Rms.
Nobody can imagine something concrete - mechanical Ohms, measured in kilograms per second, are something very abstract.
But that's not bad: the number alone before the unit counts. The lower it is, the lower the mechanical losses in the transformer, and the more vigorous result promises to sound. Why this is so, it is conclusively not (yet) explained - just the experience confirms this rule repeatedly. One promising approach is that the mechanical resistance in the speaker chassis, as opposed to an electrical resistance neither linear nor timeinvariant - it is changing its value as a function of signal level and frequency, and also shows a more or less distinct hysteresis behavior.
The cause of the mechanical losses are in the flow resistance, in all corners of a speaker chassis, for example, in the spider or too narrow Polepiece. Therefore, ventilation openings at all possible parts of a Speaker are so important. Also eddy currents, especially in voice coil carriers, increase the mechanical losses. This is the biggest advantage a Kapton-against an aluminium former.
Unfortunately, currently there are no usable investigations performed, so until further observation applies: Rms - The smaller the better."
and also:
"One of the crucial criteria for the quality of bass are the mechanical losses, expressed in parameter "Rms". Experiments repeatedly show a remarkably good correlation between low Rms and subjective as dry and accurately perceived bass reproduction. Rms is in relation to the membrane surface to set: As is a moving mass of 30 grams "a lot" for the 6 inch woofer ,and for the 12 inch woofer a "little", so is an Rms of 1.0 kg/s perfect for a 12 incher, while this would be only average for a 6 inch woofer.
Current experiments even indicates that the Rms factor has a qualitative impact to all frequency ranges - up to the tweeter. To keep this parameter low, two important design criteria must be met: The coil carrier must consist of an electrically non-conductive, or at least bad conductive material, so that no eddy currents may arise, which carry high mechanical losses for the biggest responsibility. Moreover, the mechanics of friction designed chassis, especially the vented side openings behind the spider are important in this context."
Also, from an Allan Isaksen (Wavecor) interview:
"So far I have not come across one driver with high mechanical losses that did not suffer from reduced dynamics and details and less natural sound. "
And, from a Tony Gee (TG-acoustics) interview:
Q: How about mechanical resistance in loudspeaker drivers. QMS and Rms. Important or overrated?
A: "Very important, one of the first things I look at when choosing an appropriate woofer. It increase dynamics and low level detail. All tg-acoustics designs use drivers with extremely low mechanical resistance. For example, the large 15-inch woofer in our up-coming reference speaker has mechanical losses normaly only found in small midwoofers!"
Finally, from a Joachim Gerhard (Audio Physic) interview:
"What many driver manufacturers have done the last years, is to increase the damping to make the frequency response more flat. But some old drivers, like the famous 6,5" paper woofer that Jan Paus at Seas made several years ago, (The Seas CA 17 RCY, ed. note) was optimized for low loss. So they made a compromise between frequency response and sensitivity. This driver was very good, and was used by Wilson Audio for many years. Later, in the 80's, manufacturers started to add more mass, they added more damping, and they made surrounds with high loss. That gave an extremely flat frequency response, but also a lot of energy storage. This compared, the old drivers were much quicker. They had some resonances, but you could get rid of that in the crossover. It was this run for flat response that gave a lot of modern drivers this dull, uninteresting sound. And you can also measure higher second and third harmonic distortion in some of them. If you compare the on-axis response between an old and new driver; you will see that the energy in the treble is far higher than in the new drivers. These so-called "modern" drivers often has a Qms of maybe 0.8 or 0.6. The old drivers had Qms values of maybe 5 to 7! We found that drivers with a very high mechanical Q sound more open, more clean and dynamic. And when you look at it, you find it is very simple, because they have less loss. The surround is easier to move, the spider is better constructed, they have better air flow, higher sensitivity. So a high mechanical Q is a very good indicator of energy storage behavior. This is one of our secrets. One of the many! "
First of all, the mathematical relationship:
(Rms/Mms) = 2*Pi*(Fs/Qms)
where:
Rms = mechanical resistance
Mms = moving mass
Fs = free air resonance frequency
Qms = mechanical 'quality' factor (= 1/bandwith of resonance peak)
And then a few quotes:
From an interview of a German DIY magazine editor (Google translated):
"Anyone who observes the Thiele-Small parameters of a woofer, turns his attention first on resonance frequency and Qts: These two parameters determine the achievable lower cutoff frequency. The second view is the equivalent volume - Vas: It reveals what cabinet size is necessary to achieve this goal.
These parameters say little about the sound character : whether the finished box will sound quick or slow, betrays - apart from the key priority of course crossovers and housing - the inconspicuous parameter Rms.
Nobody can imagine something concrete - mechanical Ohms, measured in kilograms per second, are something very abstract.
But that's not bad: the number alone before the unit counts. The lower it is, the lower the mechanical losses in the transformer, and the more vigorous result promises to sound. Why this is so, it is conclusively not (yet) explained - just the experience confirms this rule repeatedly. One promising approach is that the mechanical resistance in the speaker chassis, as opposed to an electrical resistance neither linear nor timeinvariant - it is changing its value as a function of signal level and frequency, and also shows a more or less distinct hysteresis behavior.
The cause of the mechanical losses are in the flow resistance, in all corners of a speaker chassis, for example, in the spider or too narrow Polepiece. Therefore, ventilation openings at all possible parts of a Speaker are so important. Also eddy currents, especially in voice coil carriers, increase the mechanical losses. This is the biggest advantage a Kapton-against an aluminium former.
Unfortunately, currently there are no usable investigations performed, so until further observation applies: Rms - The smaller the better."
and also:
"One of the crucial criteria for the quality of bass are the mechanical losses, expressed in parameter "Rms". Experiments repeatedly show a remarkably good correlation between low Rms and subjective as dry and accurately perceived bass reproduction. Rms is in relation to the membrane surface to set: As is a moving mass of 30 grams "a lot" for the 6 inch woofer ,and for the 12 inch woofer a "little", so is an Rms of 1.0 kg/s perfect for a 12 incher, while this would be only average for a 6 inch woofer.
Current experiments even indicates that the Rms factor has a qualitative impact to all frequency ranges - up to the tweeter. To keep this parameter low, two important design criteria must be met: The coil carrier must consist of an electrically non-conductive, or at least bad conductive material, so that no eddy currents may arise, which carry high mechanical losses for the biggest responsibility. Moreover, the mechanics of friction designed chassis, especially the vented side openings behind the spider are important in this context."
Also, from an Allan Isaksen (Wavecor) interview:
"So far I have not come across one driver with high mechanical losses that did not suffer from reduced dynamics and details and less natural sound. "
And, from a Tony Gee (TG-acoustics) interview:
Q: How about mechanical resistance in loudspeaker drivers. QMS and Rms. Important or overrated?
A: "Very important, one of the first things I look at when choosing an appropriate woofer. It increase dynamics and low level detail. All tg-acoustics designs use drivers with extremely low mechanical resistance. For example, the large 15-inch woofer in our up-coming reference speaker has mechanical losses normaly only found in small midwoofers!"
Finally, from a Joachim Gerhard (Audio Physic) interview:
"What many driver manufacturers have done the last years, is to increase the damping to make the frequency response more flat. But some old drivers, like the famous 6,5" paper woofer that Jan Paus at Seas made several years ago, (The Seas CA 17 RCY, ed. note) was optimized for low loss. So they made a compromise between frequency response and sensitivity. This driver was very good, and was used by Wilson Audio for many years. Later, in the 80's, manufacturers started to add more mass, they added more damping, and they made surrounds with high loss. That gave an extremely flat frequency response, but also a lot of energy storage. This compared, the old drivers were much quicker. They had some resonances, but you could get rid of that in the crossover. It was this run for flat response that gave a lot of modern drivers this dull, uninteresting sound. And you can also measure higher second and third harmonic distortion in some of them. If you compare the on-axis response between an old and new driver; you will see that the energy in the treble is far higher than in the new drivers. These so-called "modern" drivers often has a Qms of maybe 0.8 or 0.6. The old drivers had Qms values of maybe 5 to 7! We found that drivers with a very high mechanical Q sound more open, more clean and dynamic. And when you look at it, you find it is very simple, because they have less loss. The surround is easier to move, the spider is better constructed, they have better air flow, higher sensitivity. So a high mechanical Q is a very good indicator of energy storage behavior. This is one of our secrets. One of the many! "
Last edited:
This is why "loudness" tone controls were invented.
The human ear drops off high and low frequencies at low volumes.
The human ear drops off high and low frequencies at low volumes.
Could DSP be a good way to make an automatic loudness control? I've never used DSP, not sure what it's capable of, it is very difficult to do analogue
So, for a new design, a 3-way with a dome tweeter with good dispersion, a 4” midrange for good midrange dispersion and a woofer with low Rms, high Qms, low Mms and low Fs seems like a good starting point, right?
Let me thow up a hypothetical design for discussion:
Tweeter: ScanSpeak Illuminator D3004/6600 AirCirc Tweeter Textile Dome
midrange: Scanspeak Illuminator 12MU/4731T-00 4.5" Midrange - 4 ohm
Woofers: Dual ScanSpeak Revelator 18W/8531G-00 7" Mid Woofer 8 ohm
Any thoughts on this hypothetical setup for achieving a speaker to perform well at low volume?
In a sealed enclosure, I guess power compression is not a problem at low volumes? Would it make sense to try out an acoustic vent design or a transmission line? Any thoughts on low voulme performance?
Let me thow up a hypothetical design for discussion:
Tweeter: ScanSpeak Illuminator D3004/6600 AirCirc Tweeter Textile Dome
midrange: Scanspeak Illuminator 12MU/4731T-00 4.5" Midrange - 4 ohm
Woofers: Dual ScanSpeak Revelator 18W/8531G-00 7" Mid Woofer 8 ohm
Any thoughts on this hypothetical setup for achieving a speaker to perform well at low volume?
In a sealed enclosure, I guess power compression is not a problem at low volumes? Would it make sense to try out an acoustic vent design or a transmission line? Any thoughts on low voulme performance?
Last edited:
I’ve been thinking the same thing. Make bass and treble adjustments as a function of the volume.
Only problem - what SPL is selected as a ”reference” level for adjustment values from F-M curve? And won’t this be different for each recording?
BSS has a method to build a loudness curve with it’s DSP. I haven’t used it yet but plan to on the M2’s.
dbx has the best sounding LF augmentation I have ever heard in the 4800/4820’s. It is called Auto Warmth. It is very flexible and with a bit of tinkering it’s killer. This is NOT the subsonic synthesizer used in the DriveRack line. At a fraction of the price of a 4800 the dbx ZonePro line has Auto Warmth, but no delay functions.
The only other one I like is the old Audio Control Phase Coupled Activator. Not perfect but better than anything else I have heard except those mentioned above.
In all my systems this does more for the low level sound and listening enjoyment than anything else I have tried.
Barry.
P.S. I have no small systems.
dbx has the best sounding LF augmentation I have ever heard in the 4800/4820’s. It is called Auto Warmth. It is very flexible and with a bit of tinkering it’s killer. This is NOT the subsonic synthesizer used in the DriveRack line. At a fraction of the price of a 4800 the dbx ZonePro line has Auto Warmth, but no delay functions.
The only other one I like is the old Audio Control Phase Coupled Activator. Not perfect but better than anything else I have heard except those mentioned above.
In all my systems this does more for the low level sound and listening enjoyment than anything else I have tried.
Barry.
P.S. I have no small systems.
I think you may be on a good track there.
But I would probably change a couple of things, if possible:
(i) try and find a dedicated mid-range driver, rather than what is essentially a 4.5" mid-woofer with a lowish Fs (which you don't really need) and a large rubber surround (you don't need the 3.5mm Xmax either).
Something along these lines:
Audax HM100Z0 4" Aerogel Midrange
(ii) go for a larger woofer - still with a low (Fs/Qms) - if you can accommodate a larger box. Remember: at low frequencies, larger Sd = lower excursion = higher quality and more realistic bass.
If you wish to stay with ScanSpeak, this one could be a good candidate, for instance:
Scanspeak Classic 25W/8565-00 10" Woofer
EDIT: I just simmed this woofer and it would work well in a 50L closed box, giving a Qtc = 0.9 and an F-3 = 40Hz, which is very slightly peaked (+0.9dB) but I would argue excellent sounding when placed in a typical room, where room pressurization picks up below ~40Hz to yield an essentially flat response down to 20Hz.
Cheers,
Marco
Last edited:
Interesting to read about the driverack 4800/4820. I use the Driverack 360 - excellent performance, but no such feature (perhaps one can ask DBX to add it in a software update?).
But I would also like to mention the equalizer in the unassuming Sonos connect. I use the optical out to the driverack. I've tried many equalizers, and the simple one in Sonos - only "bass" and "treble" - is the one I like the most for adjusting the tonal balance on individual tracks and records, and for adapting to low volume listening late in the evening. With more complex equalizers, it's a hassle to get the loudness contour right. And it's easy to muck up the sound, in a way. Sonos just does it right. For me, I find that when I increase "bass" about 30 percent and "treble" about 10 percent in the Sonos, it becomes perfect for listening at a volume that doesn't wake up my neighbors. Recommended.
I support Marco's recommendation of more bass cone area. At least 2x8" and closed box with enough BSC! 12MU is not bad, but I support Audax HM100, I have a pair waiting... Only problem is that it's low end gets stressed very easily, xo around 4-500Hz? In that range usually baffle step effect might help mid but make things trickier to get woofers matching well
Last edited:
Caveat:
ALL-ELSE-EQUAL
1st:
"efficiency" (Sensitivity): The higher the better.
2nd:
Qe: The lower the better.
3rd:
Mms: The lower the better.
4th:
-"generically" Rms: the lower the better. (..I say "generically" because often at the smallest excursion it isn't the spider but rather the surround that dominates, and Rms is mostly about the spider.)
..of course there is a deep relationship between 1, 2, & 3.
Ex.:
Fostex FE208EZ 8" Full Range Sigma Series
ALL-ELSE-EQUAL
1st:
"efficiency" (Sensitivity): The higher the better.
2nd:
Qe: The lower the better.
3rd:
Mms: The lower the better.
4th:
-"generically" Rms: the lower the better. (..I say "generically" because often at the smallest excursion it isn't the spider but rather the surround that dominates, and Rms is mostly about the spider.)
..of course there is a deep relationship between 1, 2, & 3.
Ex.:
Fostex FE208EZ 8" Full Range Sigma Series
Last edited:
- Status
- Not open for further replies.