What's Left In Speaker Design To Reduce Distortion/Increase Detail Retrieval?

[Purité Audio]

Ok, but why? If Genelec designs a speaker with essentially the same frequency range of the Salon 2, where would it be audibly better and why?

Really?
All the other stuff Matt, cardioid, constant directivity an FR +/-1.5dB between 20Hz-35kHz , phase coherent perfect step, can’t be done with passives, not quite true but much more effectively in a one box solution.
Keith

 

[ryanosaur]

I'm gonna be the low-brow here after looking at the review article...
But...

Did they really name a Speaker:

"Red Rocket?"

Spoiler: possibly NSFW

OK... but seriously.... Here is one:

Could the drivers plausibly be advanced to yield even lower distortion, with audible benefits?

I don't claim expert status at all, but why don't manufacturers really shoot for more finite and real methodology that allow for Xmax and Xmech to be absolutely true numbers where a design can be counted on to perform at full stroke without losing linearity? I see time and again how even though we know what Xmax should be, pushing drivers toward that limit increase compression and distortion. I know of one manufacturer who admits his drivers will sound like cr@p if pushed toward Xmax.
So when looking at the common piston design of most drivers, why can we not expect close to perfect linearity throughout the excursion of the piston? Why is it acceptable that we are given an Xmax which should be fully usable against the actual mechanical limits of the Driver but in reality is further limited due to other aspects of the Driver design such as motor strength and suspension which when designed properly should in my mind allow for fully usable Xmax in a known and repeatable manner?
In the end, after the design aspects of the Driver are accounted for, then implementation in a cabinet becomes the variable.

Just my 2¢. Corrections to my "understanding" are welcome!

 

[ocinn]

I mean the future of loudspeaker design is none at all. Audio sent directly to aural nerve with DSP to emulate soundstage in front of you + head tracking + 20-200hz woofers for tactile impact. BAACH-SP can already do this quite well for headphones. Now we just need to have someone with a cochlear implant that supports streaming to try it out, lol.

 

[goat76]

So for instance drums on a track on my system would be well placed in spatial terms, and I can hear if the drums were placed in a reverb. But the Estelon speakers just seem to effortlessly carve out precisely where the drums are in the soundstage and the precise acoustics or added reverb around the drums...and exactly where that reverb "ends" is more vivid and obvious. Basically there is this constant sense of more sonic information, presenting more precision about what is in the recording.

Just a thought or two...

If I remember this right, you have your speakers set up wider than an equilateral triangle. Nothing wrong with that if that's your preference, but it's possible you have sacrificed a little bit of that pinpoint accuracy that the stereo image will have when two speakers are working together, and creating a distinct-sounding phantom center.

As you know from your line of work, most sound objects in a stereo mix are precisely that, phantom sounds. If these phantom sounds will be “effortlessly carved out” precisely where they are in the stereo image, the two speakers must be positioned in a way so that they together can “project” a unified and perfectly non-blurred stereo image to your ears at the listening position.

Maybe it's just that, or...

My other theory is that a speaker pair that are better matched, frequency-wise, will have a better chance of a more precise and pinpoint stereo image.

 

[pablolie]

As expected, a range of replies, thanks.

So take for example something that measures great like the Revel Salon 2. What if anything might be done to improve it's performance? First, stick to material/mechanical
"improvements." For instance is it possible that some cost-no-object level of additional damping/bracing of the cabinet could yeild any audible benefits? (Anecdotal reports seem to suggest Acora is on to something with their nothing-special drivers placed in fully granite cabinets in that respect...maybe...).

Could the drivers plausibly be advanced to yield even lower distortion, with audible benefits?

Anything else mechanical?

Would making them active designs, in of itself, audibly improve on those speakers? Digital crossovers yield audible improvements?

I'm sure one could get right in to DSP which will can possibly improve the performance of many speakers in-room. Though I don't think I'm talking about that, so much as
the engineering associated just with the speakers themselves.

If your question is about potential, fundamental design breakthroughs that open a whole new world... my answer would be to not hold your breath. The improvements we have experienced over the last 30 years are great, but to a large degree they have revolved around trickle-down and "democratization" that resulted in much better performance at much lower price points. But the fundamentals are pretty much the same with all loudspeaker technologies - the improvements have been in materials and measurements - and have been incremental at the top end, at best.

We are not likely to see that ideal point source with a perfectly linear 0-24kHz frequency response anytime soon. But we clearly are at the very boundaries of what's hearable and what is not.

So what is next? Bio-engineering to improve our ears? Would that really make the music listening experience better?

Don't get me wrong - I love stuff like Dirac. I love educational websites just like this, which emphasize the fact that at some point in time your focus should shift from equipment to optimizing your room's response. Our gear is probably no longer the gate-holder to a fully-immersive music experience. Today I am in what is my "weekend cabin" with about $900 of audio gear investment in it, and was thoroughly enjoying listening to classical music, marveling at how awesome it sounds compared to what a similar investment would have bought me in the 80s or so.

 

[IPunchCholla]

Ok, but why? If Genelec designs a speaker with essentially the same frequency range of the Salon 2, where would it be audibly better and why?

I’m currently designing a speaker. It’ll be my first from scratch (I built new cabinets for drivers). I decided on a few arbitrary things to start: open baffle coaxial/full range and an enclosed sub. When I began modeling the drivers/enclosure it was pretty easy to get to +-3dB 30-20k even with cheap drivers with just 1st order crossovers. But with DSP, I don’t have to do a traditional crossover at all and even with those cheap drivers, I could get close to +-1 really easy. The trade off would be an equivalent sensitivity of 82dB/1 watt/1meter and any distortion the filters make.

Those are sim numbers so taken with a tablespoon of salt, but it did really point out to me some advantages of digital. To the point where I would never make passives. And that is without any of the other advantages of digital/active mentioned above.

 

[Curvature]

But we clearly are at the very boundaries of what's hearable and what is not.

I think this is mostly supposition. You are not the first one to say this in the thread. I'm not singling you out. I don't believe there is good evidence for it.

 

[Bjorn]

Traditional front firing speakers with a single dome/cone tweeter are IMO way behind horn loaded compression drivers or planar/ribbon/electrostatic drivers in this area. I believe this has to do with intermodulation distortion, but more than simply two tones.

Secondly, the interaction with the room is very crucial here as well. Less specular reflection (especially early ones but late ones also contribute) have a great impact on clarity, intelligibility and localication or "detail" and "resolution". That's partially why constant directivity horns and planars also do so well (since they normally have a narrow dispersion) but it's not the only one. I also find the CBT do very well in the area with the combination of multiple drivers to lower the distortion and the narrow vertical directivity.
Diffraction is another important element.

When you have become accustomed to listen to very detailed and high resolution speakers that dig out everything, it can be quite dull to go back to traditional speaker designs. Thankfully, we adapt over time though.

Obviously we can be fooled by a frequency response that isn't balanced and/or specular reflections that makes certain tones stick out.

 

[Ilkless]

The biggest step change in loudspeaker dynamic drivers in the last decade-ish are the Purifis. I'd imagine the next step (at least for drivers) would to bring that same level of thinking to a cost-no-object level. Imagine Purifi surround, cone and voice coil design, coupled to optimised, super-open cast frames and neodymium ala Scanspeak Illuminator/Ellipticor or SB Satori, with even more copper.

Then we have low-diffraction, computer-optimised waveguides that integrate seamlessly with a baffle, for the smoothest directivity at least horizontally, at whatever width is desired. Neumann, Genelec, GGNTKT, DD, JBL M2 are where things are going.

Or even point-source waveguides like the 8381. Danley Synergy is the pioneer in the space, but they are still erring towards PA high-SPL and narrow (and hard-bounded) directivity in design philosophy, rather than absolutely optimising for smooth directivity and FR with a gradual fall-off by the side.

We'll also see more coaxials by driver OEMs that try to close the gap with the absolute leaders in the space, which are the proprietary KEF/Genelec drivers.

 

[Kvalsvoll]

So for instance drums on a track on my system would be well placed in spatial terms, and I can hear if the drums were placed in a reverb. But the Estelon speakers just seem to effortlessly carve out precisely where the drums are in the soundstage and the precise acoustics or added reverb around the drums...and exactly where that reverb "ends" is more vivid and obvious. Basically there is this constant sense of more sonic information, presenting more precision about what is in the recording.

Always difficult to describe what we hear, so others can actually understand. Some will dismiss and say nonsense, while others will try to make sense, try to relate to own experiences with sound.

https://tidal.com/browse/track/9303700

Jordi Savall, la Moresca.

This is what I listened to today. Here, the tactile sense of the drum is also part of the realism. I listened in 2 different rooms, same speakers. The better room presents a better sense of presence of the instruments, and there is this solidity of the rendering of instruments, and huge contrast between how they appear. Some instruments just fill the room, other appear very localized and precise.

It is mostly the speaker, but also room acoustics. Both are required to achieve this sense of solid, real instruments playing.

 

[MattHooper]

Just a thought or two...

If I remember this right, you have your speakers set up wider than an equilateral triangle. Nothing wrong with that if that's your preference, but it's possible you have sacrificed a little bit of that pinpoint accuracy that the stereo image will have when two speakers are working together, and creating a distinct-sounding phantom center.

As you know from your line of work, most sound objects in a stereo mix are precisely that, phantom sounds. If these phantom sounds will be “effortlessly carved out” precisely where they are in the stereo image, the two speakers must be positioned in a way so that they together can “project” a unified and perfectly non-blurred stereo image to your ears at the listening position.

Maybe it's just that, or...

My other theory is that a speaker pair that are better matched, frequency-wise, will have a better chance of a more precise and pinpoint stereo image.

Good catch. I'd considered this. However I actually listen from different positions and distances. Sometimes I put the speakers closer together and also I put my sofa on big furniture sliders so I can move it approximately a foot or so back or forward (so sometimes I'm at the perfect equilateral triangle) . Likewise I listened to the Estelons both from further and closer to nearfield. So I think I have a general overall impression of the performance of each speaker. Of course I could be wrong, but that's an obvious caveat.

 

[MattHooper]

Really?

Of course. That's what we are here to discuss..

All the other stuff Matt, cardioid,

As in how Kii and others are implimenting this? How different, or how much audible improvement, would you expect against a well behaved speaker (e.g. pick your Revel) that is well set up in a good room?

constant directivity an FR +/-1.5dB between 20Hz-35kHz ,

Again...how audible do you think this will be? What would you expect to hear that you can't hear on a Revel Salon in a good set up?

phase coherent perfect step, can’t be done with passives, not quite true but much more effectively in a one box solution.

But that's pretty controversial. IIRC Toole et all downplay the audibility of time/phase coherence in real world set ups. (I have Thiel speakers which are passive Time/Phase coherent btw).

 

[Kvalsvoll]

As in how Kii and others are implimenting this? How different, or how much audible improvement, would you expect against a well behaved speaker (e.g. pick your Revel) that is well set up in a good room?

The differences you describe in post #1, if I understood correctly, that is, are attributed to midrange and low-mid frequencies, where a speaker with better directivity control makes a difference in reducing very early reflection level, which affect clarity and resolution in the mid - low mid.

This is not the only property that matters, but it is important for this realism and clarity sensation.

Now look at those Estelon speakers. See something that is different from a more typical speaker with front firing drivers? This speaker has a radiation pattern that is different, and this also makes the rendering of instruments different.

 

[cavedriver]

I'd like better/more methods of testing speakers. For example, and before everyone calls audiophiles satan worshippers, let's make recordings that intentionally have content 40~50 dB below complex musical content, maybe someone talking or something. Then test which speakers allow someone to understand the "quiet" speech. The recording should have the speech at several different volumes relative to the primary content.

I'll second the suggestions for more single- and double-blind testing in speaker development. I would love if audio events set this up for conference visitors. Maybe I'll pitch that to the organizers of the Capital Audio shows near me.

I'd like speaker designs with more effective point source presentation. KEF, Genelec, and others have been moving in this direction of course, but there's still good midbass integration missing from KEF's, and even though these units are using concentric drivers, does the sound really leave the driver in a way that perfectly matches what the original sounded like? The outer driver's cone's movement interferes with and modulates the diffraction of the tweeter's wavefront. Where can we get smooth ideal reproduction?

 

[MattHooper]

I had a listening experience today that put me in mind of this thread. "How much do you get with some of the newer approaches to speaker design?"

My Brother-In-Law (BIL) is in town. He's sort of an old-school Stereo Review/ASR oriented audiophile, not in to "woo-woo" stuff at all, more oriented towards studio level design, and interested more in advancements like active speakers. Though where he lives, he doesn't get to hear much of what he can read about. Today I brought him to a very accommodating local high end store (where I bought my Joseph speakers) who carries the Kii Audio 3 speakers which my BIL wanted to hear. It was a great time.
(BTW, he previously listened to my system and I brought him to my reviewer pal's place where he heard $65K Estelons).

I was able to listen to some Spendor classic (though new) 1/2s, which I'd really wanted to hear. I also heard some large floor standing speakers from Triangle (french). But we spent by far most of the time listening to a wide selection of music on the Kii 3 speakers (without the bass module). (A lot of prog rock too, since he loves that genre).

To compare first what I heard from the Spendors: The Spendors struck me with that wonderful rich but clear tone that I enjoy from the classic Spendor sound (though in this case a bit more neutral and forward than some older modes). The general sound of voices and timbre of instruments were very natural - there was a combination of density and body that gave voices a sense of in-room solidity, yet without edge, sounding appropriately human and soft. A very tricky balance IMO. Highs were open and airy, giving a sense of "happening live, not canned sounding." Eyes closed there was a good sense of hearing *something like* the gestalt of live voices and instruments. To my brain the timbre of the sound was generally "correct" to the impression I carry around in my head of live acoustic sources, but there was still some sense of homogenization to everything, a slightly softened presentation, a slight lack of purity and clarity. I found listening very compelling, but ultimately would be happier with some other speakers that
sound 'right' to my ear, but sound less homogenized.

The Kii Audio 3: I've heard them numerous times now and today just re-enforced previous impressions. They sound generally clean, mostly "invisible" as sound sources, yet unlike some speakers that "disappear" they don't sound wispy, the Kii Audio 3s have meat on the bones, with a very even balance from bottom to top. Where some speakers you might get more of the top end skin hits of a bongo (on some of my reference tracks) the Kii Audio 3s gave that detail too, but also the weight of the hits on the bongos as well. Midrange was satisfyingly complete, no sense of suck outs emphasizing one thing or another, so everything sounded balanced, and yet variations in recordings/vocals sounded obvious too. A monitor sound without being punishing. And of course they did that wonderful trick with such deep, controlled bass from a relatively small speaker.
I could hear more control, more tonal individuation between bass instruments, vs the same tracks on the two other speakers I listened to in the store.

And my Brother In Law? He was very impressed. He's not big on describing sound, but essentially he thought the Kii 3s were terrific and really just the type of gear and sound he would be looking for, if he's to purchase another speaker (he currently owns Sonus Faber speakers he bought ages ago). Of all the systems he has heard so far on his trip hear, I had the impression the Kii 3s would hit his bulls-eye the best.

Ok what did I find the Kii 3s, the more "advanced" active-based speakers brought to the table?

Well....I heard nothing at all paradigm changing. I've heard plenty of other audiophile passive speakers that impressed me a lot more, and that rendered a number of those tracks with more "you are there" scale, clarity, timbral realism etc (to my ears). For instance my old Thiel 3.7s dug out plenty of nuance and timbral precision, dynamics, clarity, imaging etc as well. I wasn't hearing anything "wow" at all or especially "advanced" from the Kii 3s. More subjectively, while the Kii clearly had some advantages in frequency range, clarity and precision, nothing sounded quite as natural (eyes closed test) to me as some of the stuff I heard on the old Spendor design. And frankly, after listening to the Kii speakers for a while I wanted to run home and hug my system. I find the sound soooo much more beautiful and beguiling at home. Not a fair apples to apples comparison, but the general character of the Kiis didn't want to make me keep listening.

So, on an ultra subjective level...purely based on "what I prefer" while I could hear the Kii speakers do some excellent things, they left me cold. On a subjective-trying-to-be-more-objective level, just in terms of "judging the overall sound quality" they sound "different" than many good passives but much "better?" Eh.

It sort of reminds me of the Single Ended amplifier blind test thread on this site, where a tube design with some of the most wretched measurements and poor SINAD is put up against a properly measuring design. The difference turned out to be subtle, some people even struggling to hear a difference. In the same vein, while clearly speaker design IS making advancements, in the end perhaps the latest designs won't always necessarily result in large sonic differences vs competing (or older) designs. A bit of us audiophiles making mountains out of sonic molehills...which is what we do, as enthusiasts.

I don't think that means of course nobody can find some new speaker design obviously much more to their liking, and be very enthusiastic about some of the problems they are solving. Even in the Big Picture, relatively small differences can seem really big to us when we care about them.

And of course the above is just a personal encounter with a few speakers. Nothing about it entails the question I posed is moot. It's interesting to ponder what aspects can still be improved, sonically, (even if taste is left to the listener) in modern speaker design.

That's my take today, FWIW.

 

[Timcognito]

Full range wall hung cabinet-less designs that do what the flat-screens did to CRTs

 

[b7676]

What passives are under 1%THD.

 

[RayDunzl]

What passives are under 1%THD.

My panels might be. My woofers struggle a bit. There's a hole around 48Hz that skews the measure. Panels cross from woofer at 180Hz.

At about 92dB at ten feet:

 

[Axo1989]

My panels might be. My woofers struggle a bit. There's a hole around 48Hz that skews the measure. Panels cross from woofer at 180Hz.

At about 92dB at ten feet:

View attachment 286673

I love seeing your measurements: your speakers are very good at what they do well.

Mine are 4-way with conventional cone drivers (but welterweights at 44 kg, compared to the 175 kg Genelec recently feted). Here is uncorrected 95 dB sweep (ref. 1 kHz) at ~2 metres on-axis. While distortion is under the -40 dB line, referenced to actual output they do break 1% (for second/third harmonic) at a few points on the graph.

Spoiler: THD relative and % H2-H3 and H4-H7

They don't match the electrostats (no surprise) but do ok. In their defence my room was a bit noisy >1 kHz and measurements are a bit rough (it's LP after all). But overall it's certainly possible to do passive and keep distortion down.

I'd listen ~10 dB lower (85 dB average is still pretty loud, really 75 dB average would be more relaxed listening) but we are interested in distortion at dynamic peaks so it's entirely reasonable to look at 95 dB (and even 105 dB) sweeps to get a full picture. Apply DSP correction (I usually do, that modal boom ~50 Hz is ugly, and the mid-bass needs some filling in) and we'll see increased distortion in any boosted areas (midrange goes down a bit with my target curve so that's lucky).

Apropos the thread topic, the new version of that speaker swaps the drivers for slightly less conventional variants without any centering spiders, the maker claims they produce less distortion. I've not seen measurements to verify that claim (maybe I'll try the upgrade).

The also change some of the cabinet materials, increasing the use of plastic honeycomb and ceramic foam in the structure, to manage resonances better (again, they claim). And they are doing an unconventional thin-film mid-range driver, with a meta-material backing, reckoning lower distortion again (on a high-end model, may trickle down if successful) but I've seen no review/measurements for that one at all.

Edit: I don't imagine KEF and others will stop development of their solutions any time soon.

 

[wiggum]

Detail retrieval isn't a term in ITU BS-2399. It is something used by recent audiophiles. It didn't exist even a decade ago I think and I have no idea what it means or its precise definition.

Anyways, as a non-designer, here is what I found out by reading several technical articles.

• Distortion is already below audible levels when listening to program material. It is still high in sub-bass & bass frequencies, but we tolerate it very well.
• The dynamic driver design is almost 100 years old and has survived when compared to other alternatives due to its durability.

We have extracted as much as possible from DD, only incremental changes are happening and it is something not always audible though it could be quantified.

All the research and investment seems to be on headphones and surround sound technology, not loudspeakers. With BT headphones all the rage and music being consumed through streaming, loudspeakers might become a niche that only grandpas listen to.