What's the thing where some systems sound better than other at low volumes?

[Megaken]

I've heard systems that sound excellent at normal volume but some of them sounded not as inspiring and present as other.

What affects that?

My own system sounded better at low volumes when I had a device with dynamic compensation or whatever it's called - the function that does it digitally. But aside from that?

 

[Doodski]

I've heard systems that sound excellent at normal volume but some of them sounded not as inspiring and present as other.

What affects that?

My own system sounded better at low volumes when I had a device with dynamic compensation or whatever it's called - the function that does it digitally. But aside from that?

The frequency response of the system at various volume levels makes the system sound different at varying levels. Sometimes the best levels is low or medium or high both not usually all of them.

 

[coonmanx]

I use a bit of EQ on all of my systems...

Downstairs the living room system runs a pair of Scott 166 speakers, which only have 6.5" woofers. So I just use the built in 5 band EQ on the MCS 3872 and that works pretty well. A bit of bass boost (not too much) and a slight downward slope. And that is set at the volume level that I listen at. I don't really vary that a whole lot. Not too soft, not too loud, just right...

Bedroom system upstairs has a pair of 8" two ways. So that system has a ten band EQ. Also not a very severe EQ in any way. Slight bass boost and then it stays fairly flat. Volume doesn't vary a whole lot. Don't listen too loud because my son is in the room next door so I have to keep it down. Sounds good at that level.

Finally the basement system, which used to be the living room system. Two pairs of 8" two ways. Also a mild EQ curve. Sound is powerful and I don't have to turn that up past a watt or two for it to get loud. And I also have an electric guitar down there so I can put on a CD and get to jamming. Did that today with the "Riding with the King" CD. Was fun.

So could I turn those up and still get good sound? Of course. But I might hit the EQ defeat switch for those louder levels. But I don't know since I don't crank it up these days.

 

[kemmler3D]

I've heard systems that sound excellent at normal volume but some of them sounded not as inspiring and present as other.

What affects that?

My own system sounded better at low volumes when I had a device with dynamic compensation or whatever it's called - the function that does it digitally. But aside from that?

This is caused by the loudness effect.

Basically, (due to human hearing) when you turn something up, the bass and treble seem relatively louder. When you turn it down, it seems to lack treble and bass.

Systems that have exaggerated treble and bass at normal volumes will seem to sound better at low volumes.

A way to deal with this is to have a loudness contour type control on your preamp / DAC or other processor. So it will automatically EQ the correct amount to adjust for the loudness effect.

 

[Blumlein 88]

I don't know. ESL's often had this quality of sounding good at low volumes. Yet they are iffy on bass anyway. You'd think at low volumes they would be worse due to that. Instead they seemed to usually provide a satisfying amount of detail so you didn't need to turn it up to hear it. At least with older speakers it often was a case of below some level the speaker just didn't seem to wake up, and sound good. ESL's stayed good until quite low sound levels.

 

[Keith_W]

This has already been briefly mentioned, but ... ISO 226 equal loudness contours (aka Fletcher-Munson curves).

As the volume gets softer, progressively more bass and treble are required for the tonality to be preserved. Without this type of compensation, music will seem to lose bass and treble, which will rob weight and "sparkle" from music.

Old fashioned preamps had a "Loudness" feature that was rather crude - it was either off or on. Modern playback software with volume control can implement this digitally, so the curve changes continuously as volume is changed. Examples - JRiver, Foobar, Acourate Convolver, Hang Loose Convolver. Notably, Roon does not have this feature.

In the old days, people used to talk about "lazy drivers" which needed some decent current to "wake up", as an explanation for why some speakers did not sound great at low volumes. They typically spoke of larger speakers with massive drivers that "did not like" being played at low volumes. There is apparently some kind of U-shaped curve where drivers perform in a linear fashion. I know nothing about the truth of this, I suppose there is a possibility that a small current can't get all that inertia from a giant cone moving. Maybe someone here can comment.

 

[Waxx]

This has already been briefly mentioned, but ... ISO 226 equal loudness contours (aka Fletcher-Munson curves).

In the old days, people used to talk about "lazy drivers" which needed some decent current to "wake up", as an explanation for why some speakers did not sound great at low volumes. They typically spoke of larger speakers with massive drivers that "did not like" being played at low volumes. There is apparently some kind of U-shaped curve where drivers perform in a linear fashion. I know nothing about the truth of this, I suppose there is a possibility that a small current can't get all that inertia from a giant cone moving. Maybe someone here can comment.

there is a truth in that, but that is mainly in very big subwoofers with a very heavy cone (high MMS) and little motor force (low BL). They may need a few watt before they start to act like it should. But those drivers are not used in hifi, and today even rare in pro audio systems as technology of today makes it easy to avoid that kind of problems. Older big woofers or subwoofers may still have that issue.

 

[KSTR]

In the old days, people used to talk about "lazy drivers" which needed some decent current to "wake up", as an explanation for why some speakers did not sound great at low volumes. They typically spoke of larger speakers with massive drivers that "did not like" being played at low volumes. There is apparently some kind of U-shaped curve where drivers perform in a linear fashion. I know nothing about the truth of this, I suppose there is a possibility that a small current can't get all that inertia from a giant cone moving. Maybe someone here can comment.

[...] mainly in very big subwoofers with a very heavy cone (high MMS) and little motor force (low BL). They may need a few watt before they start to act like it should.

This is a common misconception.

The real problem with "lazy" drivers is not a huge mass or a weak motor per se, it's the suspension which often has non-linear and generally unstable friction because it is mechanical friction in solids which has ugly stick-slip properties. The thick "fire hose" style rubber surround on many subwoofers is a typical specimen.

Nonlinear friction is the only mechanism which can cause "resolution" errors at very small signal/movement levels.

A loudspeaker is inherently a feedback system (when driven from a low output impedance) which tends to correct the suspension errors. Which means when the feedback factor is low (from a low (BL)²/Re) the correction is less efficient.

 

[Ze Frog]

Speakers tend to be the biggest factor there in my opinion. My D.I.Y bookshelf type sound so much better once at a specific sound level, maybe around 65dB and upward. Probably something to do with the motor structure and sensitivity, Tang Band W5 1685's, tried so many drivers before settling on those but even more expensive ones just couldn't match in a sealed cabinet. The Qts is around 0.63, not 0.43 as advertised and are perfect for sealed. Absolutely brilliant mid bass, my all time favourite but heavy at 3.4 kilos, mounting needs care.

Also DAC, I find a lot of DAC's really hit there dynamic sweet spot when up higher and will usually set amplifier input gain lower to allow for such.

 

[Waxx]

This is a common misconception.

The real problem with "lazy" drivers is not a huge mass or a weak motor per se, it's the suspension which often has non-linear and generally unstable friction because it is mechanical friction in solids which has ugly stick-slip properties. The thick "fire hose" style rubber surround on many subwoofers is a typical specimen.

Nonlinear friction is the only mechanism which can cause "resolution" errors at very small signal/movement levels.

A loudspeaker is inherently a feedback system (when driven from a low output impedance) which tends to correct the suspension errors. Which means when the feedback factor is low (from a low (BL)²/Re) the correction is less efficient.

This is the first time i hear this theory, so do you have scientific backings for that? Beccause what i say is considered proven by most in the diy world and as far as i know also in more professional world of speaker design. I would not mind to be corrected on this, but i'll only take it with scientific proof

 

[Keith_W]

This is a common misconception.

The real problem with "lazy" drivers is not a huge mass or a weak motor per se, it's the suspension which often has non-linear and generally unstable friction because it is mechanical friction in solids which has ugly stick-slip properties. The thick "fire hose" style rubber surround on many subwoofers is a typical specimen.

Nonlinear friction is the only mechanism which can cause "resolution" errors at very small signal/movement levels.

A loudspeaker is inherently a feedback system (when driven from a low output impedance) which tends to correct the suspension errors. Which means when the feedback factor is low (from a low (BL)²/Re) the correction is less efficient.

So there is some truth in that drivers with stiff suspension may not sound good at low volume?

 

[Blumlein 88]

So there is some truth in that drivers with stiff suspension may not sound good at low volume?

If so, makes sense an electrostatic speaker would excel at low volume in relative terms.

 

[DVDdoug]

My own system sounded better at low volumes when I had a device with dynamic compensation or whatever it's called - the function that does it digitally.

In the analog days most receivers had a loudness compensation button/switch, usually just labeled "loudness". When engaged, the bass would be boosted (relatively) as the volume control was turned-down. Most people didn't really understand what it was doing because when they were listening at low or "normal" volumes, pushing the button would boost the bass and make it "louder".

It rare now, but sometimes it's done digitally.

Here's a picture of an only receiver I found.

 

[ahofer]

I've heard systems that sound excellent at normal volume but some of them sounded not as inspiring and present as other.

What affects that?

My own system sounded better at low volumes when I had a device with dynamic compensation or whatever it's called - the function that does it digitally. But aside from that?

Sometimes the speakers have batwing FR as well, but that's obviously a sub-standard solution.

This is the reason I have the RME DAC. Dynamic loudness is the killer app.

 

[Ze Frog]

A lot of it likely comes down to the fact anything based upon a pistonic nature or indeed any kind of movement will have a specific point at which it is becomes more efficient. Highly likely a factor of lots of things in a speaker, voice coil, magnetic force, surround, driver membrane weight weight etc etc.

Bit like how extremely light membranes tend to be able to move with far more ease and usually results in increased sensitivity and less power required to achieve a specific volume. Also cabinet type, sealed tends to largely play better with a bit more power behind it than a identical ported enclosure because of the extra required force needed as the air is acting as a suspension as well.

 

[KSTR]

This is the first time i hear this theory, so do you have scientific backings for that? Beccause what i say is considered proven by most in the diy world and as far as i know also in more professional world of speaker design. I would not mind to be corrected on this, but i'll only take it with scientific proof

It immediately follows from simple physical first principles, and absolutely every professional driver designer knows this.

With true linear resistive friction (like from a resisitve air load in case of ESLs) the motional resolution is not limited. Mass and motor strength do not limit resolution.

 

[ahofer]

A lot of it likely comes down to the fact anything based upon a pistonic nature or indeed any kind of movement will have a specific point at which it is becomes more efficient.

Does this mean we would see highly variable speaker frequency response based on signal level? Looking at Erin's compression tests, we definitely see that, but only at the louder end.

 

[KSTR]

So there is some truth in that drivers with stiff suspension may not sound good at low volume?

Again, it's not the stiffness of the spring per se, it's the unwanted parasitic behaviors. It is true, though, that stiff rugged spiders and surrounds are typically more problematic than soft ones.

 

[Flaesh]

The real problem with "lazy" drivers is not a huge mass or a weak motor per se, it's the suspension which often has non-linear and generally unstable friction because it is mechanical friction in solids which has ugly stick-slip properties. The thick "fire hose" style rubber surround on many subwoofers is a typical specimen.

It's clear that the simple RMS parameter is a simplistic one. Are there any more detailed research\measurements? Including those demonstrating audible effects at low volumes.

 

[ChrisG]

It immediately follows from simple physical first principles, and absolutely every professional driver designer knows this.

No, it doesn't.
I would expect crossover distortion, or behaviour similar to a class B amplifier, and I would also expect Purifi would have covered this in their impressive white papers.


I'll try to dig out the measurements, or repeat them. The setup was this:

Beyerdynamic MC930 (chosen for low inherent noise and high voltage output)
KEF HTS3001SE, driven by a Cambridge CXA80 amplifier
MOTU M4, connected to laptop running REW

With the mic gain wide open and the mic close to the cone, I ran a sweep at a moderate SPL, probably 50-60dB in-room.

I then reduced the output level by 6dB and measured again.
Repeating that process, getting quieter and quieter. Eventually, I could only hear as it passed through the 5kHz area.
The next sweep, I couldn't hear it at all. The mic picked it up just fine, though.

For all sweeps that didn't clip the input on the M4's mic input, the frequency response was identical, albeit with a little more acoustic noise showing up on the last few sweeps.

Which rather suggests, IMO, that drivers will remain linear down to levels that we, as humans, cannot detect.

When I find time, I'll repeat the testing with my RME UCX-II, which has more input gain available.

Edit - here it is: https://www.diyaudio.com/community/...-output-capability.357209/page-3#post-6277664