Do high-efficiency speakers really have better 'dynamics'?

[dlaloum]

To learn the slightest, You have to define the other end, namely the least effcient speaker. This is all so illogical. Why is my personal experience, that the claim of "more efficient == more dynamic" is not true, overtly dismissed?! Really, I've got specimen ranging from 80dB/Watt up to 100+dB/Watt in comparison. No difference between, nothing to backup the claim.

I state the missing difference, and I like You to acknowledge it. The other "audiophile" is plain wrong. What do You think about it?!

Electrostatics are relatively inefficient - and yet they are extremely dynamic... so the proposition that efficient = dynamic is easily debunked.

But in a purely dynamic cone speaker environment, that rule of thumb may well be about right.... you merely have to exclude everything that isn't a dynamic magnet/voice coil style speaker..... and once you do that, how meaningful is it really?

 

[fineMen]

... that rule of thumb may well be about right.... you merely have to exclude everything that...

Really, what merit has such a rule of thumb? Sounds good (pun intended) for the moment, but wouldn't hold another second. It's more into the science of human fallacy than into engineering. I personally argue, that-- referring to the original question, an "experienced audiophile" is the opposite of a reasonable person. Just by experience. You get what I mean?

So you agree with me that my test should produce a null result?

That makes 1 prediction for and 1 against, so far -- good evidence that the test is worth running!

Really? Please run the test on Your own. Do it, and let the community peer review it. Then we see. Or do You have any other idea? I mean => You do it! Agreed? Hereby You are fully entitled to affirm the null result, or otherwise. Welcome to science, dear fellow!
**
Just for clarification: I did already. No difference when the speakers are used within their particular specifications.

 

[Tom Danley]

Electrostatics are relatively inefficient - and yet they are extremely dynamic... so the proposition that efficient = dynamic is easily debunked.

But in a purely dynamic cone speaker environment, that rule of thumb may well be about right.... you merely have to exclude everything that isn't a dynamic magnet/voice coil style speaker..... and once you do that, how meaningful is it really?

Hi
So if one considers dynamics to be the audio equivalent to visual dynamic range, how bright is bright compared to dark, then one is making a judgement on the sum of every component's ability to pass the transient with out transmogrification.

At least up to the amplifier output terminals, an oscilloscope will reveal Voltage clipping (more common than many believe) but the loudspeaker is much harder because subjective loudness isn't the same as measured SPL.

Not only does our subjective loudness vs actual follow the equal loudness curves but we judge broader bandwidth signals at the same SPL as louder than a pure tone.
Thanks to our flight or flight instincts when one proportionally extends the low frequency corner music and natural events sound larger AND unless your listening outside, you also have a significant amount of room sound competing with the recording's image.

Here is where directivity comes into play, the larger the direct field is, the more you hear the recording and not the room.... and room sound is ALWAYS late.

A person with a good measurement mic might make a recording of a speaker and using audacity or other program, examine the signal you recorded vs the recording you played. Maybe even listen to the recording to see what a measurement microphone with no brain processing pics up and i bet that will be an interesting topic.

Tom Danley

 

[hardisj]

the larger the direct field is

Hey, Tom. When you say this, do you mean the larger the radiation pattern of the speaker is? Just making sure I follow.


Also...

transmogrification

I had to look that word up. I'm gonna try to use it on someone tomorrow and see what happens. Hopefully it doesn't get me punched. In the south, big words scare people, as you know. LOL

 

[kyle_neuron]

Hi
So if one considers dynamics to be the audio equivalent to visual dynamic range, how bright is bright compared to dark, then one is making a judgement on the sum of every component's ability to pass the transient with out transmogrification.

At least up to the amplifier output terminals, an oscilloscope will reveal Voltage clipping (more common than many believe) but the loudspeaker is much harder because subjective loudness isn't the same as measured SPL.

Not only does our subjective loudness vs actual follow the equal loudness curves but we judge broader bandwidth signals at the same SPL as louder than a pure tone.
Thanks to our flight or flight instincts when one proportionally extends the low frequency corner music and natural events sound larger AND unless your listening outside, you also have a significant amount of room sound competing with the recording's image.

Here is where directivity comes into play, the larger the direct field is, the more you hear the recording and not the room.... and room sound is ALWAYS late.

A person with a good measurement mic might make a recording of a speaker and using audacity or other program, examine the signal you recorded vs the recording you played. Maybe even listen to the recording to see what a measurement microphone with no brain processing pics up and i bet that will be an interesting topic.

Tom Danley

Hey Tom,

This might be of interest to yourself and the followers of this topic - I'm attending a conference held by the UK Institute of Acoustics next week, which has an interesting presentation on a new standard that I was unaware of until I read the brief of presented abstracts:

BINAURAL ASSESSMENT OF LISTENING EFFORT: INTRODUCTION, COMPARISON AND REALITY
Ossi Raivio, Jan Reimes, Head-Acoustics, Germany
ETSI recently introduced a new method for the prediction of perceived listening effort based on a mean opinion approach (TS 103 558). We will provide an overview about this method that was validated on an extensive amount of subjective listening tests. This method was designed to work with modern telecommunication systems whose signal processing is neither linear nor time independent. This is where existing, related methods, such as Speech Intelligibility Index (SII) and Speech Transmission Index (STI) show some weaknesses. To demonstrate the usage with typical and realistic use cases, we will simulate a PA system in a laboratory environment with standardised background noise and reverberation simulation systems, compliant with ETSI standards TS 103 224 and TS 103 557, respectively. The PA system performance is then evaluated by using various metrics, utilizing real speech signals and realistic background noises, whenever possible, without forgetting binaural processing.

The standard itself can be found here:

https://www.etsi.org/deliver/etsi_ts/103500_103599/103558/01.01.01_60/ts_103558v010101p.pdf

I'll report back once I've seen the presentation. If anyone has any questions they'd like asked by proxy, please read the standard document and fire away. I can't promise I'll ask, but if it's a good one I certainly am not afraid to put my hand up.

 

[Tim Link]

So you agree with me that my test should produce a null result?

That makes 1 prediction for and 1 against, so far -- good evidence that the test is worth running!

I have thought about testing a high efficiency speaker by hooking up a load resister in parallel with it. What we'd be testing then is really the amplifier, I think. If the amp is working harder to deliver the needed current to the speaker because of side leakage, what does that do to the sound, assuming we don't clip the amplifier? My guess is nothing if everything is level matched. I thought this experiment up after talking to a guy in the AudioNote room at the 2019 California Audio Show. He was telling me that high efficiency speakers sound better because of the reduced work by the amplifier.

 

[Tim Link]

Do high-efficiency speakers have better dynamics? About 33% of the time, yes. History shows high efficiency usually comes at the cost of a jagged FR. Thus if the destination note of an expressive figure hits at a FR peak, dynamics are literally expanded, by many dB in a typical case. If it hits at a FR trough, then the speaker is a compressor. In the middle, it's the same as any other speaker. Or is that thinking too simplistic?

This is something I have thought about also. How could we test this? Perhaps with a digitally manipulated song file to create a jagged frequency response, which is then played through a speaker with a very smooth frequency response to see if it makes that speaker sound more dynamic. This still may not tell the whole story because with horns the result of the jagged sound may be from diffractions coming from various directions. A simple test may be to create a strong comb filtering over some frequency ranges in a song to simulate mouth diffractions of horns and then play that through a very smooth direct radiator speaker and see how dynamic it sounds.

Edit: I created a couple sample files of a segment of a Basia song. One with comb filtering added. I tried to level match them as best I could but their dynamics are... different! What can you do?

I made these in Audacity. Listening there you can quickly toggle back and forth during the song. The comb filtering is made by combining the track with a 30 sample delayed version of itself reduced by 6db. Curious to know what you guys think of the difference. I had to upload MP3 because the WAV files were too big.

 

[Jim Matthews]

Hey, Tom. When you say this, do you mean the larger the radiation pattern of the speaker is? Just making sure I follow.


Also...



I had to look that word up. I'm gonna try to use it on someone tomorrow and see what happens. Hopefully it doesn't get me punched. In the south, big words scare people, as you know. LOL

I thought Transmorgrification was when a Dude dressed as a Morgan Fairchild for Halloween.

 

[richard12511]

I do wonder why I often end up listening to my more efficient speakers at louder volumes. With my 101dB sensitive speakers, I'll often settle in to a comfortable listening volume that's 5-8dB louder than I do with my less sensitive speakers. If I try to listen to the less efficient speakers at the same volume, it sounds a bit too loud, and I end up turning it down.

It's not a distortion thing(I don't think), as the less sensitive speakers have excellent distortion performance, and heavy handed limiters that warn you when even mild distortion is reached. Best guess is related to dispersion width. The smaller speakers have much wider dispersion, and I feel like my brain starts to get overwhelmed by reflections a bit at super loud volumes.

 

[mmi]

I’m just a consumer with little experience who understands next to none of the science, so please feel free to ignore.

I’ve recently been down a rabbit hole reading about Raal ribbons. There are a few forum threads around the web you can find with people arguing about their voodoo, if it exists and whether it can be explained by measurements. The designer Aleksandar has chimed in on some and said some really interesting stuff that I think relates to this topic. One thing he argues is that current (this was 2008 but should still stand) measurements don't give the complete picture, and a new approach is needed that quantifies information loss. Eg one aspect of his product he argues doesn't show up in measurements, within the context of systems with equally flat response, is that it deals better with low level signal behaviour.

"Any way you turn it, smooth response, low diffraction, good damping, short decay, reasonably low distortion, whatever the polars....none of these things tell you exactly what amount of information is masked or completely lost. Yes, we can compare them, but choosing the better one depends on personal views on the pros and cons in transducer design."

"If something presents more information to the listener, it will have a better chance of sounding better. It will definitely sound better if FR, decay and distortion are the same with a device that shows more information loss. And, it could be represented as a simple percentage, or bit-depth.

As we all know: bit dept, spectrum (bandwidth) of color and it's accuracy, contrast and frame decay speed, are invariably accepted as undoubtfuly accurate representations of which monitor will be better.

Analogous to that in psychoacoustics could be: bit depth (as it already is with digital media, but we need to measure it in speakers, too), bandwidth (which we already have) and it's accuracy (which we also have in form of FR flatness), dynamic tracking (analogous to contrast, which we do not simply measure yet) and impulse decay (that we have now, but we haven't standardized the bottom end as we did in room reverberation decay with RT60 simple number). Add to these a must-have combo like distortion x decay, and I assure you, things would be much clearer."

Edit: I guess my interest in this relates to detail (which to me translates to dynamics). I've never really seen how the measurements deal with or explains detail. As a layman what Aleksandar says makes intuitive sense, although my intuition means nothing really.

 

[tuga]

I'd be really ineterested in seing low-level performance measurements of compression and high-sensitivity drivers vs. low-sensitivity ones. (which measurements would best protray this I do not know, but definitely not FR)
Perhaps resolution or the accurate reproduction of very low level signals is the key to this anecdotal mystery?

 

[fineMen]

Hi
So if one considers dynamics to be the audio equivalent to visual dynamic range ...

I think this may irritate some, the analogy isn't perfect.

This post refers to some standard:

Do high-efficiency speakers really have better 'dynamics'?

To learn the slightest, You have to define the other end, namely the least effcient speaker. This is all so illogical. Why is my personal experience, that the claim of "more efficient == more dynamic" is not true, overtly dismissed?! Really, I've got specimen ranging from 80dB/Watt up to...

www.audiosciencereview.com

It makes sense to understand the concept fully, literally creates sense.

A caveat besides: in Your analogy a wave is shown with ridges and valleys. You say, the valleys are filled up by random noise, hence the difference between high and low points is decreased. Wouldn't the random noise add to both, the high point and the low point? So the contrast as a difference was maintained.

Of course that can be explained. It's only to clarify.

I add to Your argument. What is the "wave" made of? Think of an identifiable package of frequencies, an acoustic event, that stands out by its "gestalt". This is pushed out into some room. It travels on a direct path to my ear, but takes many other paths also, as the sound is reflected at the room's boundaries, and repeatedly many times.

At which point in time, or after how many reflections would the "gestalt" be lost? So that it wasn't identifiable any more, transformed into noise?

Is it possible for the hearing to once catch the "gestalt" and then subsume its reflected copies under that first appearance? I think of the precedence effect here.

Can a correlation be stated between precedence (de Haas) effect and perceived dynamics?

I feel this has a lot to do with just time, namely the size of the room. I argue there is a difference between a PA system for an icehockey stadion and a stereo in a living room. Me thinks, that MTF (modulation transfer function) is a concept mostly developed for railway stations (lots of noise) and larger venues like (European?) churches (reflections, time delay).

My motivation here is, that I couldn't yet measure an MTF in my room(s) that was anything below 'excellent'.
Another fellow here measured an MTF outside in free space--a misconception?

 

[Tim Link]

Electrostatics are relatively inefficient - and yet they are extremely dynamic... so the proposition that efficient = dynamic is easily debunked.

But in a purely dynamic cone speaker environment, that rule of thumb may well be about right.... you merely have to exclude everything that isn't a dynamic magnet/voice coil style speaker..... and once you do that, how meaningful is it really?

While I have always considered electrostatics to be detailed and airy, I never subjectively thought of them as particularly dynamic sounding. I thought standard cone direct radiator speakers had better dynamics even though they seemed a bit murky in direct comparison. What interested me about horns when I first heard them was that they had the clarity of electrostatics but all the slam and punch and even more of a dynamic direct radiator speaker. This is true for me of both hybrid horns with direct radiator bass but even more so for full range horn systems. So this make me think we are not all talking about the same thing when we refer to subjective dynamics. At the 2019 CAS I heard some big Soundlab electrostatics as well as some big Martin Logans. At another event I heard the Magnepan 30.7s. Dynamics was not a standout characteristic of any of these speakers in my mind. OTH the big Acapella horn speakers were intensely dynamic sounding, and other high efficiency speakers at the show also came across as highly dynamic, including Bernhagen-Porter and some that didn't use horns such as AudioNote and PureAudioProject. It was a revelation to me that all my favorite speakers at the show just happened to be high efficiency. I was especially surprised at my enjoyment of the Audio Note speakers because their design philosophy seems like rubbish - allowing the cabinet to resonate and sing along with the drivers in a complementary fashion. As much as I wanted to I couldn't deny that they were highly enjoyable to listen to.

 

[Kvalsvoll]

I do wonder why I often end up listening to my more efficient speakers at louder volumes. With my 101dB sensitive speakers, I'll often settle in to a comfortable listening volume that's 5-8dB louder than I do with my less sensitive speakers. If I try to listen to the less efficient speakers at the same volume, it sounds a bit too loud, and I end up turning it down.

It's not a distortion thing(I don't think), as the less sensitive speakers have excellent distortion performance, and heavy handed limiters that warn you when even mild distortion is reached. Best guess is related to dispersion width. The smaller speakers have much wider dispersion, and I feel like my brain starts to get overwhelmed by reflections a bit at super loud volumes.

My experience is similar - it gets better the louder you play, on other speakers it just gets louder, without adding much to the experience.

 

[Tom Danley]

I think this may irritate some, the analogy isn't perfect.

This post refers to some standard:

Do high-efficiency speakers really have better 'dynamics'?

To learn the slightest, You have to define the other end, namely the least effcient speaker. This is all so illogical. Why is my personal experience, that the claim of "more efficient == more dynamic" is not true, overtly dismissed?! Really, I've got specimen ranging from 80dB/Watt up to...

www.audiosciencereview.com

It makes sense to understand the concept fully, literally creates sense.

A caveat besides: in Your analogy a wave is shown with ridges and valleys. You say, the valleys are filled up by random noise, hence the difference between high and low points is decreased. Wouldn't the random noise add to both, the high point and the low point? So the contrast as a difference was maintained.

Of course that can be explained. It's only to clarify.

I add to Your argument. What is the "wave" made of? Think of an identifiable package of frequencies, an acoustic event, that stands out by its "gestalt". This is pushed out into some room. It travels on a direct path to my ear, but takes many other paths also, as the sound is reflected at the room's boundaries, and repeatedly many times.

At which point in time, or after how many reflections would the "gestalt" be lost? So that it wasn't identifiable any more, transformed into noise?

Is it possible for the hearing to once catch the "gestalt" and then subsume its reflected copies under that first appearance? I think of the precedence effect here.

Can a correlation be stated between precedence (de Haas) effect and perceived dynamics?

I feel this has a lot to do with just time, namely the size of the room. I argue there is a difference between a PA system for an icehockey stadion and a stereo in a living room. Me thinks, that MTF (modulation transfer function) is a concept mostly developed for railway stations (lots of noise) and larger venues like (European?) churches (reflections, time delay).

My motivation here is, that I couldn't yet measure an MTF in my room(s) that was anything below 'excellent'.
Another fellow here measured an MTF outside in free space--a misconception?

The wave part isn't quite right, preserving signal wave-shape is not the same as preserving the signal's amplitude envelope. Think about the Envelope Time Curve, In a gross sense and visually weighted to the HF side, the Energy Time Curve is a view of the amplitude envelope. It shows (again weighted to the high side) how fast the speaker can turn on and off. Maybe i am wrong but this to me seems like it might be related to what the topic is.

I don't think you understood the MTF reference it is from the optical domain (a hobby is taking pictures with a telescope) and is a way to measure resolution.


AS applied in audio, it is part of how the STIpa measurement system works (using 7 bands in the voice spectrum) but i am not suggesting STIpa is applicable to audio, only that the MTF method is what that is based on AND late room sound is part of what obscures the information as well as what the speaker does.


If you use ARTA which actually does measure MTF, you will not see perfection in anything real.

Consider the MTF to be more specific, this is a very rapid modulation of say 1KHz, on and off at a rate from say 1Hz up to 100Hz. The "filling in" is the sound filling in the quiet or off parts. In a room, the room sound has an effect of limit how "off" the sound can be and in commercial sound, when the late room sound reaches / exceeds the critical ratio, you can no longer understand random words.

My assertion is simply that hearing stereo is similar to understanding words in that our hearing system is filtering out noise in favor of information. Clearly with words (where there is an objective test ) that the interfering noise harms random word tests on people, clearly when that noise is particularly close reflections, that scrambles the stereo image.
I have seen systems that were dynamically limited because the power amp was clipping on peaks while the LED scale said peaks were below -20dB, i have seen crossovers and other parts of active systems clipping and the more parts in the chain, the more places for the unexpected to hide. Don't get me wrong this isn't clipping you hear but rather it's so short you can't hear it as a flaw.

In a room, loudspeaker directivity can extend the direct field and so, the subjective association with larger speakers and their directivty with "dynamics"

Hey Tom,

This might be of interest to yourself and the followers of this topic - I'm attending a conference held by the UK Institute of Acoustics next week, which has an interesting presentation on a new standard that I was unaware of until I read the brief of presented abstracts:


The standard itself can be found here:

https://www.etsi.org/deliver/etsi_ts/103500_103599/103558/01.01.01_60/ts_103558v010101p.pdf

I'll report back once I've seen the presentation. If anyone has any questions they'd like asked by proxy, please read the standard document and fire away. I can't promise I'll ask, but if it's a good one I certainly am not afraid to put my hand up.

Hi Kyle
Hey that sounds interesting soak it up!
In particular see what's the latest thought on that speech independent intelligibility prediction. In the large spaces, it is largely the room sound that limits intelligibility so then directivity is a key.
I went to a Synaudcon seminar on intelligibility some years back and asked Sander (the guy who was developing STPpa) about music and he was squishy and said well one would have to account for more frequencies to represent music bandwidth.
Do report back!
Tom

 

[BrokenEnglishGuy]

I do wonder why I often end up listening to my more efficient speakers at louder volumes. With my 101dB sensitive speakers, I'll often settle in to a comfortable listening volume that's 5-8dB louder than I do with my less sensitive speakers. If I try to listen to the less efficient speakers at the same volume, it sounds a bit too loud, and I end up turning it down.

It's not a distortion thing(I don't think), as the less sensitive speakers have excellent distortion performance, and heavy handed limiters that warn you when even mild distortion is reached. Best guess is related to dispersion width. The smaller speakers have much wider dispersion, and I feel like my brain starts to get overwhelmed by reflections a bit at super loud volumes.

What about their slope between 20hz-100hz?

 

[Tom Danley]

Hey, Tom. When you say this, do you mean the larger the radiation pattern of the speaker is? Just making sure I follow.


Also...



I had to look that word up. I'm gonna try to use it on someone tomorrow and see what happens. Hopefully it doesn't get me punched. In the south, big words scare people, as you know. LOL

Hi
I mean that the region where the direct sound is significantly greater than the reflected sound when you have more directivity vs less
Yeah it's a fun word, one of my favorite examples.

 

[dlaloum]

My experience is similar - it gets better the louder you play, on other speakers it just gets louder, without adding much to the experience.

Is that a good or a bad thing?

Is it something akin to stiction... you need to apply a certain amount of power to overcome inherent friction inefficiencies.... so once you get to a certain loudness, that has been overcome, and the distortions associated with overcoming the stiction end up negligible.

Whereas a "perfect" speaker, would simply sound the same, only louder.... ie: if it gets better as it gets louder, the question is - what was wrong that caused it to sound worse when softer? - and you have to ask yourself, what low level audio details are being lost along the way.

 

[Kvalsvoll]

Is that a good or a bad thing?

Is it something akin to stiction... you need to apply a certain amount of power to overcome inherent friction inefficiencies.... so once you get to a certain loudness, that has been overcome, and the distortions associated with overcoming the stiction end up negligible.

Whereas a "perfect" speaker, would simply sound the same, only louder.... ie: if it gets better as it gets louder, the question is - what was wrong that caused it to sound worse when softer? - and you have to ask yourself, what low level audio details are being lost along the way.

It is a good thing. The experience of the sound gets better because it is more exciting and involving, as tactile feel increases, and mechanisms of hearing deals with the louder peak transients in ways that affect how the sound is perceived. It is a consequency of how we perceive sound - though hearing and tactile bodily sensations - the sound as measured from the speaker is the same, only louder.

As I, and a few others have mentioned, it is not exactly the higher efficiency that creates this difference, it has to do with how the speaker radiates sound. Danley has made several excellent posts on this, in this thread (I believe I saw it in this thread).

This has nothing to do with stiction or other non-linear deficiencies in speakers. Once you have the capacity (combination of sensitivity, power handling and available amplifier power), the speaker can play loud enough, but some speakers simply do not create this excitement even if they have sufficient capacity.

It doesn't sound worse at lower volume - it sounds better even then. Because there is less "noise" from reflected sound.
it is so, that speakers that sound good when loud, also often sound better at low volume - they have a larger usable volume span.

 

[Tim Link]

I don't think you understood the MTF reference it is from the optical domain (a hobby is taking pictures with a telescope) and is a way to measure resolution.


AS applied in audio, it is part of how the STIpa measurement system works (using 7 bands in the voice spectrum) but i am not suggesting STIpa is applicable to audio, only that the MTF method is what that is based on AND late room sound is part of what obscures the information as well as what the speaker does.


If you use ARTA which actually does measure MTF, you will not see perfection in anything real.

Consider the MTF to be more specific, this is a very rapid modulation of say 1KHz, on and off at a rate from say 1Hz up to 100Hz. The "filling in" is the sound filling in the quiet or off parts. In a room, the room sound has an effect of limit how "off" the sound can be and in commercial sound, when the late room sound reaches / exceeds the critical ratio, you can no longer understand random words.

Art Noxon has spoken about MTF in reference to the Musical Articulation Test Tones he developed, which are a series of gated tones in the bass range that can be played through a speaker system and recorded at the listening position. At ASC we have software that can analyze the recording and compare how quiet the quiet spots get compared to the tone bursts, giving an articulation measurement in decibels for what amounts to 16th notes played with 16th note rests between over a range from 20Hz up to about 800Hz. On headphones the pulses can be heard clearly and distinctly, but in listening rooms there are almost always bad areas where articulation is lost. I say almost always because some highly treated studios and hifi rooms have measured incredibly well with no significant issues. We have found that anything below 5 dB tends to sound unclear. 10 dB is pretty good, and 15 dB or better is very tightly defined, dynamic sounding bass. Anything beyond that will not be noticeably better for most listeners, so it gives a pretty good idea of how well a room has "arrived" in terms of bass and lower midrange control. In REW, the C50 measurement happens to correspond pretty closely with our MATT analysis results.