• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Ascend Sierra Luna Mini-Monitor Review

echopraxia

Major Contributor
Forum Donor
Joined
Oct 25, 2019
Messages
1,109
Likes
2,696
Location
California
I wonder what the maximum SPL without compression the Sierra Towers with the larger 70-20xr ribbon tweeter could do when properly crossed over to subs? It's ~90dB sensitivity and the larger RAAL supposedly has much higher power handling than the 64-10 in the bookshelves.
You can see distortion measurements or the 70x20xr in the Ascend Horizon review here (the directivity plots of this horizontal center channel won’t likely correlate much with the towers, but it uses the exact same driver components so I expect the distortion measurements should be more or less the same):

1630600667519.png


Better, but still isn't going to compare with other high end tweeters at extremely high SPL.

Example of a very high performance tweeter from the Revel F328Be:

1630601155875.png


The dynamics difference between these is definitely audible when you listen loudly, or music with loud peaks (e.g. orchestral). However not everyone needs or even wants the ability to push 100-120db peaks effortlessly, as it at least risks annoying non-audiophiles during listening sessions :)

I didn’t think I wanted to ever push speakers to 100-120db until I heard speakers that can do it effortlessly. Of course you don’t listen at this level continuously, but experience dynamic peaks that occasionally reach high SPLs in movies or high dynamic range music.

Extreme SPL capability is IMO very important for “endgame” speakers. But not necessarily the most important thing when a compromise must be chosen. So it is IMO a totally fair compromise for Ascend to be making. I doubt the majority of people actually ever want their speakers in a domestic situation to exceed 96-100db often.

For Ascend to match the dynamics ability of other speakers like the F328Be would probably require them to move away from a RAAL tweeter, thus compromising the wide sound (to some extent) that they are very much loved for.

The Salon2’s show that it is possible to achieve without almost any compromise, but that it is very expensive and difficult to achieve. However in theory, I wouldn’t put it past Ascend to develop some kind of Revel-killer product now that they have an NFS :)
 

thewas

Master Contributor
Forum Donor
Joined
Jan 15, 2020
Messages
6,763
Likes
16,232
I don’t know where you are getting this idea that Ascend made bad speakers and suddenly now is making good ones with the NFS
Please read correctly what I wrote above:

I don't understand why a experienced loudspeaker designer would really need a Klippel NFS to suddenly make well measuring loudspeakers.

I didn't comment about the quality of the brand designs but that an NFS isn't necessary to make good ones and even without it design errors like of the Sierra Luna models could be easily avoided.
IMO it’s really unfair to take one bad review of an underperforming speaker and generalize that across the entire brand. JBL makes plenty of nasty sounding and measuring speakers, but for some reason that doesn’t prevent people from singing praises about the JBL M2 — nor should it, when the JBL M2 is objectively excellent!
I agree and didn't do that, on the other hand I personally haven't seen as poor measuring JBL loudspeakers as the Luna and I am not talking about some portable BT speakers but maybe you have seen and can link some?
 

thewas

Master Contributor
Forum Donor
Joined
Jan 15, 2020
Messages
6,763
Likes
16,232
sorry but this is a really ridiculous response.

the NFS makes it so people don't have to spend a day measuring the speaker, or spend a day measuring the speaker every time they changed something about the crossover.
People calling others responses as ridiculous often have limited thematic understanding themselves which proves directly in the same post. When you design/engineer a loudspeaker you measure the drivers responses without crossover in the desired loudspeaker enclosure and then can easily simulate/predict the output for any electric crossover configuration with dozens of simulation tools existing since decades. No loudspeaker designer is going to put a loudspeaker again on the NFS for many hours just for that but only for large acoustic changes and final confirmation.
 

echopraxia

Major Contributor
Forum Donor
Joined
Oct 25, 2019
Messages
1,109
Likes
2,696
Location
California
I agree and didn't do that, on the other hand I personally haven't seen as poor measuring JBL loudspeakers as the Luna and I am not talking about some portable BT speakers but maybe you have seen and can link some?
The “JBL One Series 104 Powered Studio Monitors” definitely rank up there among the all-time worst measuring speakers tested by ASR:
https://www.audiosciencereview.com/...-one-series-104-powered-monitor-review.11076/

I didn’t mean to misinterpret you when you say “I don't understand why a experienced loudspeaker designer would really need a Klippel NFS to suddenly make well measuring loudspeakers.” — but I suppose now I am confused as to what you’re actually trying to imply about Ascend here. When you say “to suddenly make well measuring loudspeakers”, it sounds like you are implying they did not make well measuring speakers prior to their purchase of the NFS. If not this, then what do you actually mean?

Ascend has been making well measuring loudspeakers for many years despite not having a Klippel NFS. Now that they do though, I do expect even better results from them (certainly in terms of NFS measurements). I don’t see anything hard to understand or surprising about any of this.
 
Last edited:

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,425
Likes
7,941
Location
Brussels, Belgium
No loudspeaker designer is going to put a loudspeaker again on the NFS for many hours just for that but only for large acoustic changes and final confirmation.

That's the point of the NFS, the fact that a robot does it means that you can actually afford to monitor the response for even the slightest changes.

what's the point of having an NFS if you only used it once in the development stage? it's a waste of a good robot.
 

richard12511

Major Contributor
Forum Donor
Joined
Jan 23, 2020
Messages
4,335
Likes
6,700
WOW, amazing results. Lots of respect to Dave from Ascend for making this big investment (buying a Klippel NFS) as a small indie speaker company, while many larger speaker companies still refuse to even publish any measurements, let alone NFS measurements.

TLDR for others who don’t want to dive into that thread; Ascend Acoustics has purchased an NFS and is using it to design much improved speakers coming soon. Here is the plot they posted of the upcoming Luna V2:

View attachment 151040

Horizontal plot:

View attachment 151041

Damn, that's some really wide dispersion. Good spin, too!
 

Shazb0t

Addicted to Fun and Learning
Joined
May 1, 2018
Messages
643
Likes
1,230
Location
NJ
You can see distortion measurements or the 70x20xr in the Ascend Horizon review here (the directivity plots of this horizontal center channel won’t likely correlate much with the towers, but it uses the exact same driver components so I expect the distortion measurements should be more or less the same):

View attachment 151101

Better, but still isn't going to compare with other high end tweeters at extremely high SPL.

Example of a very high performance tweeter from the Revel F328Be:

View attachment 151106

The dynamics difference between these is definitely audible when you listen loudly, or music with loud peaks (e.g. orchestral). However not everyone needs or even wants the ability to push 100-120db peaks effortlessly, as it at least risks annoying non-audiophiles during listening sessions :)

I didn’t think I wanted to ever push speakers to 100-120db until I heard speakers that can do it effortlessly. Of course you don’t listen at this level continuously, but experience dynamic peaks that occasionally reach high SPLs in movies or high dynamic range music.

Extreme SPL capability is IMO very important for “endgame” speakers. But not necessarily the most important thing when a compromise must be chosen. So it is IMO a totally fair compromise for Ascend to be making. I doubt the majority of people actually ever want their speakers in a domestic situation to exceed 96-100db often.

For Ascend to match the dynamics ability of other speakers like the F328Be would probably require them to move away from a RAAL tweeter, thus compromising the wide sound (to some extent) that they are very much loved for.

The Salon2’s show that it is possible to achieve without almost any compromise, but that it is very expensive and difficult to achieve. However in theory, I wouldn’t put it past Ascend to develop some kind of Revel-killer product now that they have an NFS :)
I'm not sure if HD is all that needs to be taken into account to determine maximum SPL though? There are speakers that can play at very high SPL with high distortion levels.
 

napilopez

Major Contributor
Forum Donor
Joined
Oct 17, 2018
Messages
2,111
Likes
8,446
Location
NYC
My working thesis is that the ideal dispersion width depends on the room's dimensions and the placement of the speakers. Wider room, further from walls = you want wider. Smaller room or closer to walls = you want narrower. The goal would be to maintain the same ratio of reflected sound in each case. Constant reflectivity? At least that makes sense to me. I would love to see more research done on this.

I'd like to see more research too, and I had the same initial hunches as you. After all the only reason directivity matters (aside from changing the sweet spot) is that it affects how loud reflections will be

But my actual experiences don't consistently line up. Both of my apartments since I started measuring speakers have been of the long and narrow kind, so I have tested having speakers both very close (2-3 feet) and quite far (~5-20 feet) from the side walls. I've also done some testing at my desk in a separate room, and I also now have access to a treated recording studio as well. In almost all cases, I prefer wider directivity speakers. Which is why I still ultimately end up preferring the Focal 806 and Polk Reserve R200 for my own listening to the likes of the Genelec 8341B lol. The D&D 8C remains the loan exception, which I think it because it is closer to constant directivity, which in a way makes it a bit more similar to wide directivity.

Of course, there could be some placebo or other psychological factors at play. I know I enjoy wide directivity enough now that it's hard for that not to be my predisposition, and no one is fully immune to bias. But that's been my impression.

Not to say that I don't think the room has any effect either. I do think there's something to adapting your speaker's directivity to a room. But I also think the other part of the equation is our own ability to adapt to rooms and hear a speaker through it. Would definitely love to see more research into this.

While I agree that the science of speaker design has been advanced by Ascend Audio's investigation into the port resonance, I'm not so sure that the customer community has directly benefited in the short-term. Yes, Ascend will come out with a 2.0 version of their Luna line but likely will expect a price increase due to the R/D expenditures. And to what end, only if I play a track that has a note ... ? Full disclosure - I had written off the Ascend Luna as a possible upgrade (based upon the ASR review) to my HT but will now reconsider the 2.0v but only if the expected price increase isn't objectionable.

Well, I was exaggerating a bit :):). Changes are that the issue would show up on more than just one track. I may not have quite Dave's experience but there are measurable issues that I don't hear all the time, and that's just within the realm of quasi-anechoic measurements. But just because I don't hear them in my own home doesn't mean that someone else, in a different room, with certain music, won't hear it as well.

Plenty of people obviously already liked the ascends, and they are some of the most popular internet direct speakers. That's great. But I've read plenty of complaints about them too. It could well be that some of the people that didn't enjoy the original speakers might instead enjoy the new ones.

I always return to something Dr Toole said; the best speakers are those with the 'least audible flaws.' It's not about making the most euphoric speakers, but rather making the speakers that are least like to create distractions from the music, to reveal themselves as wooden boxes. So any potential flaw is worth addressing. I do expect these to sound significantly better to those who have the chance to hear both.

Unless I missed it (I haven't read the full ascend thread) Dave didn't point to any significant price increases after acquiring the NFS. I guess we'll see but he always struck me as someone who prided himself on providing a great value.
 

ctrl

Major Contributor
Forum Donor
Joined
Jan 24, 2020
Messages
1,616
Likes
6,088
Location
.de, DE, DEU
That's the point of the NFS, the fact that a robot does it means that you can actually afford to monitor the response for even the slightest changes.
what's the point of having an NFS if you only used it once in the development stage? it's a waste of a good robot.

@thewas is right with his point of view, because once the individual drivers have been carefully and reliably measured, there is no need for further measurements during crossover development - apart from the finished product, of course.
The crossover simulation software is so good that the simulated frequency responses are almost identical to the measured ones.

Only if structural changes are made to the speaker (rounding of the side edges, BR port change,...), then a re-measurement of individual or all drivers may be necessary and then it is indeed as you say that the NFS makes life much easier.
 

napilopez

Major Contributor
Forum Donor
Joined
Oct 17, 2018
Messages
2,111
Likes
8,446
Location
NYC
I don't understand why a experienced loudspeaker designer would really need a Klippel NFS to suddenly make well measuring loudspeakers. The NFS just makes measurements more fast and efficient but as @napilopez has shown here and several decades of loudspeaker engineering, theory, books and designs have tought us, it is not a prerequisite. Of course it is a good investment for a big loudspeaker manufacturer but especially for smaller ones the $100k investment must be financed and in many cases could be possibly better invested in other directions.

I don't disagree with you, but I also think there's an important distinction between 'needing' an NFS to create a good speaker and greatly diminishing the potential of letting errors slip through the cracks. It returns to that point of making a speaker with the least audible flaws...

Especially reading through Dave's posts, it's not hard for me to imagine how if you've been creating speakers for a long time and you're used to doing things a certain way, and have gotten good results doing things that way, you could miss a flaw that shows up under specific conditions.

More specifically though, I find part 4 of Dave's post very interesting, because it closely reflects my own experience measuring speakers and comparing the results to the NFS. The quote below follows after Dave noticed the port resonance largely disappeared once he measured at a greater distance:

"The prominent high Q spikes at 600Hz and 1200Hz are mostly gone. I found this puzzling. As I mentioned previously, the Klippel NFS measures in the nearfield and the farfield response is calculated mathematically. My measurements clearly show that port resonance propagates differently than, for example, a woofer. Could there possibly be an issue such that NFS measurements are somewhat exaggerating front port resonance? There is no way for the NFS (or for any measurement system for that matter) to inherently know the difference between the output of a port or driver. Could the NFS be picking up all that nearfield turbulence and using it in the algorithms to determine the farfield response?

I didn’t know and the engineer I was working with at Klippel also could not confirm one way or the other. This would certainly explain some of the other front-ported speakers measured at ASR that showed prominent port resonances that didn’t show up in other published measurements. For example, the Salk WOW1 measurements published on the Salk website differ from ASR’s measurements in the area where the NFS picked up port resonance."

(Emphasis mine)

Later, he goes on to say:

"Based on much experimentation and various other measurement comparisons, I believe the Klippel NFS does slightly exaggerate front port resonance issues when calculating its far field measurements. The port resonance in the original Luna and Duo does indeed exist, as I mentioned and measured in Part 4, but I am not yet entirely sure if this is represented properly in far field measurements because this area of the far field measurements differs considerably between what the NFS calculates and what other measurement gear measures. Since there was no way for me to confirm this, one way or the other, the best approach was to simply eliminate / reduce the port resonance to the point where it does not show on both the NFS measurements and our other test systems. In this manner, I am assured it no longer exists."

Emphasis mine again.

So it seems Dave was aware of the peaks in the nearfield port response, but as it did not show up in his farfield data, he did not think it was a priority to fix these issues.

Although my super-nearfield measurements often show port issues too, and sometimes the issue does show up in the farfield too, other times no matter what I try, I cannot get the resonance to show up as dramatically as it does in the NFS measurement. Now, I'm not as confident as dave that this is an issue with the NFS, but I have long suspected that these port resonances are not quite as audible as they look. I am pretty sure they will be audible in certain situations as noted in an earlier post, but I find it very interesting that Dave found this issue too now that he has his own NFS.

It's something I've noticed since I first started comparing my measurements against the NFS, but I don't think I ever brought it up because I usually trust the NFS more than my own data. But in general, whenever I see a midrange resonance caused by the port, I assume it's prboably a bit less less audible than it looks. It should still be eliminated, but I don't usually consider it a dealbreaker.

It'll be interesting to see more research into this. It's just nice to see how the quality of measurements have shot up over the past couple of years =]
 

617

Major Contributor
Forum Donor
Joined
Mar 18, 2019
Messages
2,404
Likes
5,296
Location
Somerville, MA
I don't disagree with you, but I also think there's an important distinction between 'needing' an NFS to create a good speaker and greatly diminishing the potential of letting errors slip through the cracks. It returns to that point of making a speaker with the least audible flaws...

Especially reading through Dave's posts, it's not hard for me to imagine how if you've been creating speakers for a long time and you're used to doing things a certain way, and have gotten good results doing things that way, you could miss a flaw that shows up under specific conditions.

More specifically though, I find part 4 of Dave's post very interesting, because it closely reflects my own experience measuring speakers and comparing the results to the NFS. The quote below follows after Dave noticed the port resonance largely disappeared once he measured at a greater distance:

"The prominent high Q spikes at 600Hz and 1200Hz are mostly gone. I found this puzzling. As I mentioned previously, the Klippel NFS measures in the nearfield and the farfield response is calculated mathematically. My measurements clearly show that port resonance propagates differently than, for example, a woofer. Could there possibly be an issue such that NFS measurements are somewhat exaggerating front port resonance? There is no way for the NFS (or for any measurement system for that matter) to inherently know the difference between the output of a port or driver. Could the NFS be picking up all that nearfield turbulence and using it in the algorithms to determine the farfield response?

I didn’t know and the engineer I was working with at Klippel also could not confirm one way or the other. This would certainly explain some of the other front-ported speakers measured at ASR that showed prominent port resonances that didn’t show up in other published measurements. For example, the Salk WOW1 measurements published on the Salk website differ from ASR’s measurements in the area where the NFS picked up port resonance."

(Emphasis mine)

Later, he goes on to say:

"Based on much experimentation and various other measurement comparisons, I believe the Klippel NFS does slightly exaggerate front port resonance issues when calculating its far field measurements. The port resonance in the original Luna and Duo does indeed exist, as I mentioned and measured in Part 4, but I am not yet entirely sure if this is represented properly in far field measurements because this area of the far field measurements differs considerably between what the NFS calculates and what other measurement gear measures. Since there was no way for me to confirm this, one way or the other, the best approach was to simply eliminate / reduce the port resonance to the point where it does not show on both the NFS measurements and our other test systems. In this manner, I am assured it no longer exists."

Emphasis mine again.

So it seems Dave was aware of the peaks in the nearfield port response, but as it did not show up in his farfield data, he did not think it was a priority to fix these issues.

Although my super-nearfield measurements often show port issues too, and sometimes the issue does show up in the farfield too, other times no matter what I try, I cannot get the resonance to show up as dramatically as it does in the NFS measurement. Now, I'm not as confident as dave that this is an issue with the NFS, but I have long suspected that these port resonances are not quite as audible as they look. I am pretty sure they will be audible in certain situations as noted in an earlier post, but I find it very interesting that Dave found this issue too now that he has his own NFS.

It's something I've noticed since I first started comparing my measurements against the NFS, but I don't think I ever brought it up because I usually trust the NFS more than my own data. But in general, whenever I see a midrange resonance caused by the port, I assume it's prboably a bit less less audible than it looks. It should still be eliminated, but I don't usually consider it a dealbreaker.

It'll be interesting to see more research into this. It's just nice to see how the quality of measurements have shot up over the past couple of years =]

Deep high q cancellations are intrinsically sensitive to measurement location. The depth of the null could vary by 10db just by moving the mic around a few inches at these frequencies.

Just like when you invert a tweeter that is playing in phase with a woofer over a wide bandwidth. At the design axis you might see a crazy deep null, but if you move closer to the woofer or tweeter the depth of the null collapses.

This port stuff is a non issue. Even if it is 'real' and propagates into the far field like the Klippel insinuates, it is not an audible concern in practice.

How many reviews does Amir observe the port null and then admit he can't hear it? Remember the JBL 705 for example.
 

napilopez

Major Contributor
Forum Donor
Joined
Oct 17, 2018
Messages
2,111
Likes
8,446
Location
NYC
Deep high q cancellations are intrinsically sensitive to measurement location. The depth of the null could vary by 10db just by moving the mic around a few inches at these frequencies.

Just like when you invert a tweeter that is playing in phase with a woofer over a wide bandwidth. At the design axis you might see a crazy deep null, but if you move closer to the woofer or tweeter the depth of the null collapses.

This port stuff is a non issue. Even if it is 'real' and propagates into the far field like the Klippel insinuates, it is not an audible concern in practice.

How many reviews does Amir observe the port null and then admit he can't hear it? Remember the JBL 705 for example.

Well I agree that nulls this narrow are never audible to me. Peaks, however, are occasionally audible, and that's the main thing being addressed here. Just this week I was dealing with a very high Q peak that stuck out like a sore thumb when practicing on my digital piano. Not sure if it was from my room, the speaker, or even the piano sample itself (possibly a combination!) but it was always the same note, and only one of the various piano VSTs seemed to trigger it, but it was driving me nuts because I kept hitting that note in the specific piece I was practicing. Ended up reducing it by like 10 dB.
 
Last edited:
  • Like
Reactions: 617

ctrl

Major Contributor
Forum Donor
Joined
Jan 24, 2020
Messages
1,616
Likes
6,088
Location
.de, DE, DEU
So it seems Dave was aware of the peaks in the nearfield port response, but as it did not show up in his farfield data, he did not think it was a priority to fix these issues.

Well, I can't quite take the manufacturer's side in this case. If the sound pressure level of the resonances is already 5 dB louder than the BR port Helmholtz resonance in the near-field measurement and the baffle step correction leads to the unwanted resonances being around 11 dB louder and the BR port is located at the front, then one should question the design.

1630612594616.png

Especially since the resonances are also clearly visible in the frequency response of Amir's distortion measurements. These are, as far as I know, measured quite classically at a distance of 0.3 - 0.5m. This means that the manufacturer's distortion measurements may have looked similar.

1630613357401.png


So there have certainly been a lot of warnings that these resonances will be noticeable in the FR even without NFS.

Another thing, of course, is whether these resonances with high Q will be audible in the final product. But this can be easily checked with a sine wave generator - also by the owners of the loudspeaker if they really want to know ;)

Update: In any case, it is very commendable that the manufacturer has drawn consequences and is now optimizing its products.
 
Last edited:

napilopez

Major Contributor
Forum Donor
Joined
Oct 17, 2018
Messages
2,111
Likes
8,446
Location
NYC
Well, I can't quite take the manufacturer's side in this case. If the sound pressure level of the resonances is already 5 dB louder than the BR port Helmholtz resonance in the near-field measurement and the baffle step correction leads to the unwanted resonances being around 11 dB louder and the BR port is located at the front, then one should question the design.

View attachment 151126

Especially since the resonances are also clearly visible in the frequency response of Amir's distortion measurements. These are, as far as I know, measured quite classically at a distance of 0.3 - 0.5m. This means that the manufacturer's distortion measurements may have looked similar.

View attachment 151128

So there have certainly been a lot of warnings that these resonances will be noticeable in the FR even without NFS.

Another thing, of course, is whether these resonances with high Q will be audible in the final product. But this can be easily checked with a sine wave generator - also by the owners of the loudspeaker if they really want to know ;)

Update: In any case, it is very commendable that the manufacturer has drawn consequences and is now optimizing its products.

Fair enough, regardless the end result is better speakers =]

One thing that did surprise me was to read was him saying he believes we only hear in 1/6th, maybe 1/12. I thought the Toole/Olive research pretty firmly discredited that. I think perhaps 1/6th octave might be relevant to how we percieve overall tonality but I am fairly confident we can hear certain issues at a higher resolution. It just might be more difficult to hear them.
 

Sonny1

Active Member
Joined
Apr 5, 2020
Messages
252
Likes
350
Hats off to the guys at Ascend for their response to this review. Some companies deny negative analytical info and double down on the marketing speak and snake oil (you know these companies). I’m glad the folks at Ascend hit it head-on and fixed the speakers in question. This is what good companies do, no excuses, just action.

Side note, I heard their larger models with the RAAL tweeters and was very impressed. Was surprised to see the poor performance of these but it sounds like the 2.0 version is a major improvement. Impressed by their response.
 

aarons915

Addicted to Fun and Learning
Forum Donor
Joined
Oct 20, 2019
Messages
685
Likes
1,140
Location
Chicago, IL
WOW, amazing results. Lots of respect to Dave from Ascend for making this big investment (buying a Klippel NFS) as a small indie speaker company, while many larger speaker companies still refuse to even publish any measurements, let alone NFS measurements.

TLDR for others who don’t want to dive into that thread; Ascend Acoustics has purchased an NFS and is using it to design much improved speakers coming soon. Here is the plot they posted of the upcoming Luna V2:



Horizontal plot:

I could have swore a few years ago Dave was "schooling" me on how the CTA-2034 style measurements aren't the end all be all and don't necessarily mean a speaker will sound good. He even started off his current posts saying the same thing and basically refuting Toole's research about the audibility of resonances, claiming humans don't hear with greater precision than 1/6 octave, etc. If he really believed these things there would be no reason to buy a NFS, it's still good news that he bought one and deserves praise but his actions of buying a NFS and then reworking some of his speakers to measure better seem counter to his comments about how the measurements don't matter.

No, not always. Proof: Blind Listening Test - KEF R3 vs Ascend Sierra 2EX

Ascend’s preference for ultra wide beam speakers isn’t for everyone, but neither is KEF’s medium/narrow beam for everyone (as the results linked above show, where Ascend easily beats KEF in a blind test).

I do think Revel is a great default choice for most, however even Revel’s speakers below their flagship lines (PerformaBe, Ultima2) don’t necessarily beat Ascend (depending on which compromises are more or less important to you). Proof: Blind Listening Test - Revel F206 vs Ascend Sierra RAAL Towers

Ascend speakers aren’t perfect, but they competently achieve a very wide beam that you don’t get anywhere else at the price point. The Luna’s may not be Ascend’s best performing product, but now that Ascend owns a Klippel, I expect some amazing products coming up from them. And the Luna V2 (charts shown above) is one of the first examples of this.

If you take the entire beamwidth into account, the difference in dispersion is drastic but the early reflections have been shown to be most influential and they are only usually around 1 decibel different between the "ultra wide" RAAL and the narrow KEFs. Also remember this is just the sidewall reflection, the ceiling and floor bounce are going to be stronger in the KEF's and I've shown research in the past that shows that those reflections add spaciousness as well. Most of us have also seen the research of the mono vs stereo and how the spaciousness is basically equal once a 2nd speaker is added into the mix, to me that is evidence that we put way too much emphasis into "wide dispersion" since all of us use at least 2 speakers. I haven't personally noticed much difference in the RAAL 64-10 and the KEF LS50 in spaciousness (admittedly in a small room) and these are fairly extreme opposites of the dispersion spectrum. Just a side note also, the 64-10 has extremely wide dispersion but the 70-20 is just average but it does measure more neutral and is smoother off-axis which is more evidence to me that dispersion isn't that important but a neutral on and off-axis response is what mostly matters.
 

echopraxia

Major Contributor
Forum Donor
Joined
Oct 25, 2019
Messages
1,109
Likes
2,696
Location
California
If you take the entire beamwidth into account, the difference in dispersion is drastic but the early reflections have been shown to be most influential and they are only usually around 1 decibel different between the "ultra wide" RAAL and the narrow KEFs. Also remember this is just the sidewall reflection, the ceiling and floor bounce are going to be stronger in the KEF's and I've shown research in the past that shows that those reflections add spaciousness as well. Most of us have also seen the research of the mono vs stereo and how the spaciousness is basically equal once a 2nd speaker is added into the mix, to me that is evidence that we put way too much emphasis into "wide dispersion" since all of us use at least 2 speakers. I haven't personally noticed much difference in the RAAL 64-10 and the KEF LS50 in spaciousness (admittedly in a small room) and these are fairly extreme opposites of the dispersion spectrum. Just a side note also, the 64-10 has extremely wide dispersion but the 70-20 is just average but it does measure more neutral and is smoother off-axis which is more evidence to me that dispersion isn't that important but a neutral on and off-axis response is what mostly matters.
So you are saying the KEF R3 actually should have sounded more spacious than the Sierra 2EX? That’s not what happened though; the Sierra 2EX soundstage was more spacious and enveloping than the KEF R3 in the blind test. Just as my Genelec 8351B’s sound more focused but less spacious than the Revel Salon2’s. If you have an alternate theory explaining this, I’d be curious to hear it. I can’t say I know exactly what causes it, but the beam pattern of the mids and highs seems a prime candidate.

I could have swore a few years ago Dave was "schooling" me on how the CTA-2034 style measurements aren't the end all be all and don't necessarily mean a speaker will sound good. He even started off his current posts saying the same thing and basically refuting Toole's research about the audibility of resonances, claiming humans don't hear with greater precision than 1/6 octave, etc. If he really believed these things there would be no reason to buy a NFS, it's still good news that he bought one and deserves praise but his actions of buying a NFS and then reworking some of his speakers to measure better seem counter to his comments about how the measurements don't matter.
I remember some discussions like that on AVSForum which may be what you’re referring to, but I don’t recall it as Dave trying to “school” you. I think most of us were talking past each other in that conversation, because Dave certainly wasn’t saying measurements don’t matter, and also wasn’t saying CEA2034 doesn’t matter. I think he was pointing out some very valid weaknesses in CEA2034 that create an incomplete picture of how a speaker sounds, though I don’t remember what his exact argument was.

I recall arguing that CEA2034 clearly doesn’t tell the whole picture. I still believe this, and thanks in part to ASR, I don’t think this is at all controversial any more: CEA2034 tells a very big and important part of the picture, but it’s still missing a few other important variables that lead to large differences in speaker preference. These missing variables are probably just a matter of beam shape, distortion, linearity, dynamic compression, etc. But it’s pretty clear that CEA2034 alone does not tell us all we need to know about speaker, and so I find exploring that space of possible missing variables very interesting.

Unless you have other hypotheses to explain why I (and many others) prefer the spatial presentation of Revel and Ascend RAAL and Focal etc. for some kinds of content, versus KEF and Genelec and Neumann.

And of course I will reiterate that I love my Genelec 8351B’s. But they do have a distinctly different spatial presentation than wider pattern speakers like the Ascend RAAL speakers and like the Revel Salon2 that I’ve owned. Both are great in different ways.
 
Last edited:

aarons915

Addicted to Fun and Learning
Forum Donor
Joined
Oct 20, 2019
Messages
685
Likes
1,140
Location
Chicago, IL
So you are saying the KEF R3 actually should have sounded more spacious than the Sierra 2EX? That’s not what happened though; the Sierra 2EX soundstage was more spacious and enveloping than the KEF R3 in the blind test. Just as my Genelec 8351B’s sound more focused but less spacious than the Revel Salon2’s. If you have an alternate theory explaining this, I’d be curious to hear it. I can’t say I know exactly what causes it, but the beam pattern of the mids and highs seems a prime candidate.


I remember some discussions like that on AVSForum which may be what you’re referring to, but I don’t recall it as Dave trying to “school” you. I think most of us were talking past each other in that conversation, because Dave certainly wasn’t saying measurements don’t matter, and also wasn’t saying CEA2034 doesn’t matter. I think he was pointing out some very valid weaknesses in CEA2034 that create an incomplete picture of how a speaker sounds, though I don’t remember what his exact argument was.

I recall arguing that CEA2034 clearly doesn’t tell the whole picture. I still believe this, and thanks in part to ASR, I don’t think this is at all controversial any more: CEA2034 tells a very big and important part of the picture, but it’s still missing a few important variables that lead to large differences in speaker preference. These missing variables are probably just a matter of beam shape, distortion, dynamic compression, etc. But it’s pretty clear that CEA2034 alone does not tell us all we need to know about speaker, and so I find exploring that space of possible missing variables very interesting.

Unless you have other hypotheses to explain why I (and many others) prefer the spatial presentation of Revel and Ascend RAAL and Focal etc. for some kinds of content, versus KEF and Genelec and Neumann.

And of course I will reiterate that I love my Genelec 8351B’s. But they do have a distinctly different spatial presentation than wider pattern speakers like the Ascend RAAL speakers and like the Revel Salon2 that I’ve owned. Both are great in different ways.

I'm not saying the R3 would be wider but that I believe you have to take all reflections into account when talking about dispersion and not just the horizontal. I also said when more than 1 speaker was in the mix in Toole's mono vs stereo study the spaciousness differences were negated, when only 1 speaker was playing in mono they were much more evident. I haven't heard what you describe comparing the LS50 to RAAL tweeters but like I said it was a smaller room with no toe in and the sidewalls only about 3 feet from the speakers so the reflections wouldn't be that weak, I did notice a big difference in the vertical plane with the LS50 sounding bigger and more open which I have theorized that the vertical reflections are most likely why. I have heard the LS50 in a larger showroom away from walls and they did sound small in that room but I think that has a lot more to do with the lack of bass in that setup. I personally think a lot of this stuff may have something to do with room size and how many speakers are playing because I always listen to music in 3.1 so it's possible that this setup negates the need for "wide dispersion" speakers.

It seems to me something else could be at play because like I said the 70-20 in the Ascend towers isn't that wide of dispersion but since your listening tests show them to sound more spacious it's possibly because they are smoother off-axis and simply more neutral tweeters. Also, if wide dispersion was so important then you would think the 64-10 would be the best RAAL offering but I don't think I've ever heard someone claim to prefer it over the 70-20 which is more evidence to me that it's really neutrality, both on and off-axis, that is responsible for the preference.

Forgot to address Dave's comments. I could go back and find his comments but I don't really care but he has this condescending attitude that he knows better than anyone else and even talks about how he worked with Revel engineers in the past and seems to imply that he has surpassed them now, maybe it's just me but I find him pretty arrogant. He specifically mentioned designing a speaker that measured near perfect in the Spinorama yet sounded like crap as his evidence for why the standard is flawed. Funny that now here we are a few years later and he buys a Klippel and redesigns those speakers to measure better in the Spinorama...yet also claims that measurements don't tell you how a speaker sounds.
 

echopraxia

Major Contributor
Forum Donor
Joined
Oct 25, 2019
Messages
1,109
Likes
2,696
Location
California
I'm not saying the R3 would be wider but that I believe you have to take all reflections into account when talking about dispersion and not just the horizontal. I also said when more than 1 speaker was in the mix in Toole's mono vs stereo study the spaciousness differences were negated, when only 1 speaker was playing in mono they were much more evident. I haven't heard what you describe comparing the LS50 to RAAL tweeters but like I said it was a smaller room with no toe in and the sidewalls only about 3 feet from the speakers so the reflections wouldn't be that weak, I did notice a big difference in the vertical plane with the LS50 sounding bigger and more open which I have theorized that the vertical reflections are most likely why. I have heard the LS50 in a larger showroom away from walls and they did sound small in that room but I think that has a lot more to do with the lack of bass in that setup. I personally think a lot of this stuff may have something to do with room size and how many speakers are playing because I always listen to music in 3.1 so it's possible that this setup negates the need for "wide dispersion" speakers.

It seems to me something else could be at play because like I said the 70-20 in the Ascend towers isn't that wide of dispersion but since your listening tests show them to sound more spacious it's possibly because they are smoother off-axis and simply more neutral tweeters. Also, if wide dispersion was so important then you would think the 64-10 would be the best RAAL offering but I don't think I've ever heard someone claim to prefer it over the 70-20 which is more evidence to me that it's really neutrality, both on and off-axis, that is responsible for the preference.

Forgot to address Dave's comments. I could go back and find his comments but I don't really care but he has this condescending attitude that he knows better than anyone else and even talks about how he worked with Revel engineers in the past and seems to imply that he has surpassed them now, maybe it's just me but I find him pretty arrogant. He specifically mentioned designing a speaker that measured near perfect in the Spinorama yet sounded like crap as his evidence for why the standard is flawed. Funny that now here we are a few years later and he buys a Klippel and redesigns those speakers to measure better in the Spinorama...yet also claims that measurements don't tell you how a speaker sounds.
Well what makes beam width/height such a messy subject is that there truly may be no “right” answer. They’re just different flavors, with neither necessarily being more correct than the other — which dispersion pattern is best may depend more on room shape and listener soundstage style preferences.

For example, owning both the Genelec 8351B and Revel Salon2’s for a few years of focused listening across quite a few different rooms, I can honestly say now that I don’t think there’s an objectively “best” choice for everyone between medium vs wide beam, because some music tracks sound best on the medium beam presentation, while others best on the wider beam presentation.

There‘s a spectrum of different spatial presentation styles here, and I think this is an area where the one single style that’s best for you depends on personal preference. Or, if you’re fortunate enough to have multiple sets of speakers, you can have both :)
 
Last edited:
Top Bottom