PS Audio got a Klippel NFS

[Chris Brunhaver]

I can predict the following changes in PS Audio speakers based on this:

1) Change from planar transducers to point-source, moving coil drivers, since a point source provides more symmetrical radiation and a moving coil platform suffers less nonlinear distortion.
2) Introduction of a HF and possibly MF waveguide, for controlled and matched dispersion.
3) Driver arrays will group LF drivers closer to the HF and MF components instead of clustering woofers of a tower speaker near the floor.
4) A coincident midrange-tweeter array will provide symmetrical vertical and horizontal radiation. Possibly with cardioid radiation of the rear wave to control horizontal dispersion beyond the ability of the waveguide at lower midrange frequencies.
5) Powered cardioid bass in more complex assemblies.
6) Hard metallic-oxide diaphragms instead of fiber constructions

CEA2034 methodology favors speakers exhibiting a specific set of behaviors, so the only speakers that will appear to measure 'well' are those that fully embrace its dictates and assumptions. If your weather vane is properly calibrated, you will see which direction the wind is blowing. Many formerly perfectly acceptable speakers are now landfill.

Well, you're right about a few things, certainly, even in the last few months of R&D.

1. I am am looking at other tweeter solutions and want to use the best tweeter for the job, regardless of nostalgia or uniqueness in the marketplace. I'm doing some rigorous evaluation of AMTs and, yes, domes.
2. The reason for #1 is partly number 2. Our tweeter doesn't work perfectly on a waveguide due to a non-planar wavefront occurring at about 10 kHz and the front magnet structure forming a cavity resonance around 15 kHz. I've been testing some custom AMTs and domes that perform essentially textbook perfect.
a. Our large planar mid performs very well (when the crossover point to it is sufficiently low to keep directivity uniform) so I'm planning on using it for some upcoming designs , with others less expensive models using quality cone mids.
3. The CTA-2034 measurements don't factor in the effect of floor bounce. I was seeking to minimize floor bounce (more effectively in some towers than others) with the baffle placement I chose previously but I think you're right that tightly flanking the mid with woofers (in a three way) is typically a better approach with the higher than typical crossover points we use on our planar mids.
4/5 I would eventually like to work on this but there are major "packaging" challenges with this when using acoustic waveguides. Yes, a cardioid or essentially a damped dipole could be an interesting options though, to extend that low enough would require an active solution. I haven't don't research on this and this isn't in the next product cycle, though I am working on an active speaker. Cardioid isn't necessarily a panacea. There are some great implementation of it from Gradient (for many years), Dutch & Dutch, Kii but, looking at Amir's measurements of the Palmer speaker, getting the implementation right can be challenging and sometimes integrating the lower monopole bass section a little more challenging. You can still make a speaker sound great with a traditional radiation pattern.
6. I like aluminum woofers but don't love aluminum mids most of the time because the driver's own harmonic distortion can excite the cone modes. Purifi has a good paper on how to minimize current based distortion with passive notch filters for this though (https://purifi-audio.com/blog/app-notes-2/low-distortion-filter-for-ptt6-5x04-naa-11) It's really hard to beat a good paper cone, as there is so much adjustment that can be made to the stiffness and damping properties.
For our eventual high end line refresh, I have been working with some spread tow carbon fiber (similar to Textreme but not from that vendor) with good results and I'll likely go that direction. I have been working with a driver vendor who works with Sonus Faber, Mofi and others who I've been developing some less expensive paper cone drivers that I'm very happy with the cone performance on (they're incredibly smooth and extended).
Don't get me wrong, there are a lot of good metal cone drivers out there. KEF does something quite interesting where they took the technology from their mechanical uni-Q (there the center dust cap mas a mechanical crossover where the surround decouples form the cone at HF and used that simulation tool to make a compliant attachment from the former to cone on some of their aluminum mids where it decouples the first breakup mode of the cone. Anyway, aluminum can be made to be very high performance but it's not my first choice, generally. You can still get great performance, distortion and directivity with other materials.

At some point, I'll do a posting to share some of the klippel data on the new drive units and show the in-house FEA tool and sim work that went into them. I promised to do some videos and some of the stuff mentioned here would be good topics to dig into.

 

[ppierre]

3) Driver arrays will group LF drivers closer to the HF and MF components instead of clustering woofers of a tower speaker near the floor.
....
CEA2034 methodology favors speakers exhibiting a specific set of behaviors

3. The CTA-2034 measurements don't factor in the effect of floor bounce

Kippel and CTA-2034 methodology, could be amended to take in account rear wall or floor for speakers designed to exploit them ?

Like Sigberg Audio Saranna or d&d 8c for rear wall bounce or floor bounce like the PS Audio Aspen.

In wall speakers are measured in a baffle. A wall or a floor cloud be used or simulated ?

 

[TimVG]

Kippel and CTA-2034 methodology, could be amended to take in account rear wall or floor for speakers designed to exploit them ?

Like Sigberg Audio Saranna or d&d 8c for rear wall bounce or floor bounce like the PS Audio Aspen.

In wall speakers are measured in a baffle. A wall or a floor cloud be used or simulated ?

This can all be done in Vituixcad.

Would also recommend keeping overall sound power DI regression slope in mind. Having recently done a test in private between two well behaved loudspeakers. There is a discernable difference in enjoyability outside of the sweet spot between a DI slope of 1,12db/octave and 1,35db/octave. Apart from stereo image, good sound shouldn't be limited to the sweet spot.

 

[AcousticTheory]

Well, you're right about a few things, certainly, even in the last few months of R&D.

Thanks for responding.

3. Planar midrange drivers seem to suffer the worst from nonlinearity due to having more diaphragm excursion, which is the bane of planar drivers. The workaround is increased diaphragm area, but this results in even worse dispersion. A pistonic cone midrange with low moving mass is superior.
4. This is the most 'optional' area of the overall CTA-2034 formula, but it is the area where manufacturers can differentiate themselves from the existing offerings.
5. The only reason to pursue this is if it improves Spins, or improves in-room response without harming Spins. Kii and Ascilab are going down this road.
6. Metallic or not, the key is to be pistonic, and push breakup out of the passband, and metal cone or ceramic cone drivers do this. If the cone is breaking up, the edge is resonating, the surround is whipping, etc., that harms consistency of dispersion. A pistonic radiator behaves predictably in this regard.

The fastest manufacturers to embrace the new methodology will suffer the least in these pages. Some are several steps ahead of PS in fielding 'The One' speaker. After 'The One' speaker has been created, the only thing left to do is put a different paint color on the common reference design, so the first to get there will be rewarded by the market.

 

[Sigberg Audio]

Thanks for responding.

3. Planar midrange drivers seem to suffer the worst from nonlinearity due to having more diaphragm excursion, which is the bane of planar drivers. The workaround is increased diaphragm area, but this results in even worse dispersion. A pistonic cone midrange with low moving mass is superior.
4. This is the most 'optional' area of the overall CTA-2034 formula, but it is the area where manufacturers can differentiate themselves from the existing offerings.
5. The only reason to pursue this is if it improves Spins, or improves in-room response without harming Spins. Kii and Ascilab are going down this road.
6. Metallic or not, the key is to be pistonic, and push breakup out of the passband, and metal cone or ceramic cone drivers do this. If the cone is breaking up, the edge is resonating, the surround is whipping, etc., that harms consistency of dispersion. A pistonic radiator behaves predictably in this regard.

The fastest manufacturers to embrace the new methodology will suffer the least in these pages. Some are several steps ahead of PS in fielding 'The One' speaker. After 'The One' speaker has been created, the only thing left to do is put a different paint color on the common reference design, so the first to get there will be rewarded by the market.

I think you are simplifying how people choose which speakers to buy quite a bit.

 

[AcousticTheory]

I think you are simplifying how people choose which speakers to buy quite a bit.

Not at all. The goal is to arrive at the optimal design, as confirmed by having the best CTA-2034 Spin results, across a product line incorporating a range of sizes of the device. Then you have a set of speakers in which each model is the best-designed among peer companies. "The One" is maybe a bit grandiose; it is more of a unifying set of design practices and selections of common components that might be reflected across a product line with similar 'DNA' than it is a single speaker. But either you do the very best that you can to satisfy the measurement criteria, or you don't, and that part is very simple. The competition is between minds that best incorporate the considerations of the measurement suite, because the interpretation of that set of measurements defines the criteria for excellence, and the company that does the best, fields the best product. After that, everyone else is merely catching up to avoid embarrassment.

 

[LTig]

Not at all. The goal is to arrive at the optimal design, as confirmed by having the best CTA-2034 Spin results, across a product line incorporating a range of sizes of the device.

This however is not the criterion (yet) by what the majority of buyers chose the speaker to buy (sadly).

 

[TimVG]

The goal is to arrive at the optimal design, as confirmed by having the best CTA-2034 Spin results

While I absolutely appreciate the spinorama, a good spinorama should be, in my humble opinion, the result of a well designed loudspeaker and not a design goal by itself.

 

[kimmosto]

While I absolutely appreciate the spinorama, a good spinorama should be, in my humble opinion, the result of a well designed loudspeaker and not a design goal by itself.

Key questions are what is good spinorama and what it misses or could miss though result is good in readers' opinion. Not impossible to find publications using NFS giving wrong impression or irrelevant/illogical subjective notes about DUT, in my opinion.

 

[dfuller]

Not at all. The goal is to arrive at the optimal design, as confirmed by having the best CTA-2034 Spin results, across a product line incorporating a range of sizes of the device.

Let's keep in mind Goodhart's Law: When a measure becomes a target, it ceases to be a good measure. Let's not assume that conforming to visually good CEA2034 data is the only requirement for a good speaker, and even within that there's a great deal of wiggle room on what makes for a "good" speaker.

 

[TimVG]

Key questions are what is good spinorama and what it misses or could miss though result is good in readers' opinion. Not impossible to find publications using NFS giving wrong impression or irrelevant/illogical subjective notes about DUT, in my opinion.

Well a spinorama for me is a diagnostic tool with regard to resonances, directivity errors, tonal regression.. and I believe that's what it was developed for as well.
There is, as far as I'm aware, nothing in the literature that stipulates what a good or bad spinorama is, which is exactly the reason why it should not be used as end goal (again, in my opinion)

Now of course you know better than I do what it doesn't show, and then on top of that there is the interpretation of the same data by different people as well The "score" essentially has nothing to do with the CEA 2034-A standard as well, and by now has a life of its own.

 

[kimmosto]

which is exactly the reason why it should not be used as end goal (again, in my opinion)

Success in ASR's bench may be important or even the most important goal to some. I have also experimented a bit by designing using ASR's test criteria for CTA-2034, i.e. linearization of ON and PIR using SM and NBD, but that goal has not yielded better result. At the very least, concept and quality/properties of the drivers should be taken into account and, if necessary, courage to deviate from exactly linear PIR and horizontal ON. Basically I knew that before, but had to test in practice when suitable prototype was available.

 

[Chris Brunhaver]

Thanks for responding.

3. Planar midrange drivers seem to suffer the worst from nonlinearity due to having more diaphragm excursion, which is the bane of planar drivers. The workaround is increased diaphragm area, but this results in even worse dispersion. A pistonic cone midrange with low moving mass is superior.
4. This is the most 'optional' area of the overall CTA-2034 formula, but it is the area where manufacturers can differentiate themselves from the existing offerings.
5. The only reason to pursue this is if it improves Spins, or improves in-room response without harming Spins. Kii and Ascilab are going down this road.
6. Metallic or not, the key is to be pistonic, and push breakup out of the passband, and metal cone or ceramic cone drivers do this. If the cone is breaking up, the edge is resonating, the surround is whipping, etc., that harms consistency of dispersion. A pistonic radiator behaves predictably in this regard.

The fastest manufacturers to embrace the new methodology will suffer the least in these pages. Some are several steps ahead of PS in fielding 'The One' speaker. After 'The One' speaker has been created, the only thing left to do is put a different paint color on the common reference design, so the first to get there will be rewarded by the market.

Of course. I appreciate the feedback

3. I can share some distortion data on the midranges but, things are a bit more complex than that. Yes, at low frequency where they are displacement limited, planers are poor but I don't use them there. Where they shine is that the planar drivers have essentially no inductance (resistive impedance), as our 250 mm mid only has 4 turns of wire and no steel near the coil. This eliminates inductance modulation and hysteresis distortion etc. There can be edge resonances and cavity resonances to address but there is no cone breakup and midrange distortion and decay can be truly excellent. For instance, at 90 dB SPL/1m, THD is ~0.05% at 1khz, which is very good. My old colleague Igor Levitsky wrote some interesting stuff on the subject as a bit of an advertisement for his work for Dai Ichi/Radian Audio here: https://audioxpress.com/article/thin-film-high-frequency-transducers-part-1-the-advantages https://audioxpress.com/article/thi...sducers-part-1-the-lt2-lt6-ribbon-transducers
4/5 Well, the CTA-2034 doesn't cover a lot of aspects of speaker performance. It does cover the most important aspects of sound quality/tonality but It is only looking at linear distortion and nothing to do with nonlinearity. IMO, optimizing output, non-linear distortion has less psychoacoustic study around it (though there is extensive work from Earl Geddes, Alex Voishvillo and others) but that is a big area of differentiation that isn't being addressed completely.

Purifi and high end driver design and pushing forward the state of the art for small driver. I've been working on custom high excursion, low distortion drivers (currently using XBL^2 and other techniques though evaluating other too for future use. While Bruno at Kii has his purifi drivers at his disposal, they are using rather inexpensive Tymphany drivers for the Kii products. They seem to be prioritizing european manufacturing (and all of their other design know-how) above the raw driver quality. Yes, they are still getting great performance but I'd love to see what even more serious drivers could do in there.

6. Yes, though both a high break-up and what happens as this breakup are important. That why I am interested some of these other composites etc. is that the breakup is high and tunable.

I do agree that there are lots of opportunities to do better for PS speakers. Cosmetic design/implementation is a huge one too. However, I would disagree that there's is "the one" anything. It's like telling every car owner that they should buy the new corvette because it is offers the best price/performance. I do have an engineering crush on some of the companies who have been doing this for years and are at a further stage of refinement with their design concepts but I think there's room for everyone as long as you are offering some distinct value or even novelty. Of course, I want much of the value to be on how well things are engineered but some buyers have different sensibilities and value things beyond pure performance.

 

[Paul7777x]

Don’t know if this is allowed… but if so has anyone any info on these?

Alleged new three way actives from PS Audio buy no details.

 

[Chris Brunhaver]

Don’t know if this is allowed… but if so has anyone any info on these?

Alleged new three way actives from PS Audio buy no details.

Well, I’m happy to answer questions.

This was supposed to be a proof of concept (not representative of actual cosmetics and performance) of an active speaker to bring to axpona and survey people about their interests in products like this and the idea of products like this from us. I can post the flier we made with some details if you’re interested in giving feedback.

 

[Paul7777x]

Well, I’m happy to answer questions.

This was supposed to be a proof of concept (not representative of actual cosmetics and performance) of an active speaker to bring to axpona and survey people about their interests in products like this and the idea of products like this from us. I can post the flier we made with some details if you’re interested in giving feedback.

Thanks for the reply. I heard what o think was your name on the video? Is that right, as the designer?

My interest lies in the fact that it’s time for what will be my last speakers.

And, despite my interest in all thing Lyngdorf and Purifi at the moment, my true holy grail is a pair of wonderfully good three way active speakers.

Are there any plans to produce them? Or have someone produce them perhaps?

 

[iscorpio71]

This speaker looks amazing! must sound too, and simplifies a setup. But, as with all-in-one solutions, it doesn't permit for flexibility.

I'm looking to build an audio-video room, and am almost locked on the Omega's Dynaten J25 Expert.

 

[RexrothPigeon]

it doesn't permit for flexibility.

I haven't looked into these, but I'm curious what flexibility it precludes?

 

[iscorpio71]

I haven't looked into these, but I'm curious what flexibility it precludes?

The speakers already have a DAC, preamp, and amp sections selected for you. So, you can't mix and match, sell different parts for upgrades, or fine-tune your preferred sound, for example, using an R2R DAC with a tube amp.

 

[RexrothPigeon]

The speakers already have a DAC, preamp, and amp sections selected for you. So, you can't mix and match, sell different parts for upgrades, or fine-tune your preferred sound, for example, using an R2R DAC with a tube amp.

Ah, yes, they don't make it easy to degrade the sound, that's true. But with DSP you can find ways to make it sound worse if you try!