Epos ES 14 N - best passive Speaker in SpiNorama.org so far? (7.4/10 with equalisation without subwoofer)

[thewas]

Genelec are generally very good designs there is no doubt, but you have to take some compromises when you only have a small volume and want to use bass reflex, so the walls of the Genelec box can't be that thick and the placement of the damping and port can't be optimize compared with a bigger box. Even if you build the box as clever as you can you hit at some point the physical wall. A bigger cabinet doesn't necessarily have to be built that cleverly to be in total better.

These measurements you are citing with attaching a single accelerometer are heavily dependant on the area where it is attached to the box some cm left or right and you can get a totally different results. To find every weak spots like the bass reflex port, the tweeter plate, terminals or a spot where the inner cable are slightly rattling is very time consuming and you can't be sure if you missed one weak spot.
Since you don't have any none curved walls with a Genelec box it isn't easy to accurately measure the real acceleration of the surface anyways. So I am very sceptical if these measurements with attaching an accelerometer can be used to compare very different speaker cabinet in a meaningful way.

Locally an acceleration sensor doesn't care if the surface is flat or not so the measurements can be very well compared independent from that factor. A factor for example which can make such single point comparisons problematic is having a local mode knot although this doesn't happen to all modes at the same point. Rattling terminals or ports are not dependend of the wall construction so I don't see a reason why a Genelec should be considered inferior there compared to classic MDF constructions.

 

[totti1965]

BTW, it’s great when engineers engage here, but it is weird when there is a third person acting as an oddly strong advocate (near salesperson) and the engineer avoids direct questions and occasionally speaks in riddles or challenges. That is setting off my BS detector.

I’m ready to believe it’s the language barrier, but that’s what I’m seeing here.

Sorry, but if your pointing at me: I have open the thread only because I was / am Interested in this speaker specifically as a potential customer and since I wanted to get as much information as possible, I tried to talk Karl-Heinz into sending a Pair to @amirm or Erin.
The opinion (and detailed measurements!) of Amir are of high value for me, but I can’t afford the USD 600,— to send a Pair of speakers (or 4.600,— dollars, because I don’t own the Speakers now - they are just on my shortlist!) to Seatle!
Karl-Heinz is the Developer. He sales the speaker worldwide. But there is no USA Distributor. So I wanted to connect Amir and KHF, hoping that perhaps Madrona digital could be a Distributor. But unfortunately KHF is not interested to build a USA Distribution for the Epos Brand. At least not this year.
He told me that he was a bit surprised that Pierre Aubert took the Spinorama plots from his private Blog without asking him.
That is the whole story.
If that is all bullshit for you - o.k.
But at least I learned many things the last 6 days.
And perhaps the most important thing was that KHF seems to be an authentic honorable Person:
He answers if you send him a p.m., which seems to be in Germany from the „Big Dudes in the Industrie“ much less common than in the US.

 

[sigbergaudio]

I think it is possible if the difference between two Spinorama.org rankings is more than let‘s say 2.0 out of 10.

If the measurements are obviously problematic, then it would likely be possible yes.

 

[YSC]

Sorry, but if your pointing at me: I have open the thread only because I was / am Interested in this speaker specifically as a potential customer and since I wanted to get as much information as possible, I tried to talk Karl-Heinz into sending a Pair to @amirm or Erin.
The opinion (and detailed measurements!) of Amir are of high value for me, but I can’t afford the USD 600,— to send a Pair of speakers (or 4.600,— dollars, because I don’t own the Speakers now - they are just on my shortlist!) to Seatle!
Karl-Heinz is the Developer. He sales the speaker worldwide. But there is no USA Distributor. So I wanted to connect Amir and KHF, hoping that perhaps Madrona digital could be a Distributor. But unfortunately KHF is not interested to build a USA Distribution for the Epos Brand. At least not this year.
He told me that he was a bit surprised that Pierre Aubert took the Spinorama plots from his private Blog without asking him.
That is the whole story.
If that is all bullshit for you - o.k.
But at least I learned many things the last 6 days.
And perhaps the most important thing was that KHF seems to be an authentic honorable Person:
He answers if you send him a p.m., which seems to be in Germany from the „Big Dudes in the Industrie“ much less common than in the US.

I always think, past a certain goodness threshold, like default preference score of 5, then the rest is just how much time and effort you can put into your room, and you can choose with personal taste with other non audio attributes like look and brand name. Tbh if not geek picking and blind testing you likely won’t regret the sound anyway.

But geekly comparing which one is better or worse should be based on the objective performance data

This es14n is no doubt one very good speaker, but when other things considered I would not believe it’s the best

 

[test1223]

Locally an acceleration sensor doesn't care if the surface is flat or not so the measurements can be very well compared independent from that factor. A factor for example which can make such single point comparisons problematic is having a local mode knot although this doesn't happen to all modes at the same point. Rattling terminals or ports are not dependend of the wall construction so I don't see a reason why a Genelec should be considered inferior there compared to classic MDF constructions.

The coupling of a typical acceleration sensor with the box is critical. There is plenty of scientific literature which shows multiple aspects of it like the shift of the resonance frequency of the sensor since the coupling acts as a mass loaded spring or the loss of transfered energy due to the mounting technique. Since the sensor had a none curved surface you need more wax or other coupling material which isn't that good.
With every proper sensor you get these hints in its manual with an estimate impact on your measurements.
The measurements from the review is better than nothing but is is certainly not that good to compare two speaker this way without all the context which most readers don't have.

The port can often act like a sound transmitter from the inside due to the relatively lage surface area and typically thinner walls. So the problem is not only the lack of damping of the air inside the port. If you have more volume you can build a thicker port or more bracings inside to support the port walls.

Since you have to have little space with a typical near field monitor a lot of companies including Genelec built the electronic with jackets and knops directly into the volume of the loudspeaker box. This introduces plenty of potential issues of unwanted resonances and sound transmissions. Even if you are aware of this problem, which Genelec certainly is, it is very hard to avoid all potential issues with it. If you have more space you will most likely built a separated cabinet part for the electronic.

2) over-engineering a non-audible phenomenon by making speakers ridiculously huge and heavy is more about marketing than engineering

So you also include the heavy and optimized boxes from Revel as over-engineered? The biggest model weight about 80kg or 180lbs. If you put the weight in relation to the volume it is clear that it is a very very heavy box.
I don't think so. You simply need a lot of volume and weight to hinder sound from transmission and building up resonances.

 

[Karl-Heinz Fink]

Well, let’s clarify, are you saying speaker resonances don’t show up in the standard suite of measurements here - FR, distortion, and decay, but matter to listeners?

My points are:
1) They are audible or not, and that’s what matters
2) over-engineering a non-audible phenomenon by making speakers ridiculously huge and heavy is more about marketing than engineering

Do you disagree with one or both of these?

Yes, of course I do.

Now my question: Have you ever worked on speakers with identical drivers and crossovers and identical measurements, only with different cabinet constructions?
If your answer is "No" - what are we discussing here? Now I could ask about a blind A/B comparison result to prove your theory. Did you ever do that? Well, I did and only a deaf person would not have heard the difference.

My main business always was consulting for other companies....you make a quotation for a design and of course, the faster we can get it perfect, the better for us.
For designing drivers and making crossovers, we have already done simulations for a long time, so we know what we get when the speaker is standing in the listening room for the first time. Whenever the result was not good, we started to investigate and in 90% of all cases, there was a cabinet colouration problem. That's, of course, a nightmare for a project.
The problematic frequencies are above 400Hz up to 1-3kHz. Those radiating panels are in the most sensitive area of our human hearing, radiating with different Q to the side and to the top. The BBC made a lot of experiments with cabinet walls .......but one has to understand that this was made mainly for speech and that's why the solution with thin Plywood and thick Bitumen worked best for them, shifting the problem to low frequencies. I would not call the BBC engineering department a marketing team, and I don't think they were trying to improve non-relevant effects.
So for me, this was the starting point and it took a while before we worked out what was essential and what was not so important. These cabinet constructions are now made in simulations with COMSOL and validated later with a laser scanner.....just measurements, no Voodoo.
This is not about sledgehammer engineering like some High-End companies like to do....yes, it works, but it is expensive. This is about reducing the output from cabinets in a sensible way.

Just my 2c...

 

[dogmamann]

Depends what you want to fix, for example linear phenomena can be usually corrected, non linear usually not.

What are linear and non linear phenomena?

 

[totti1965]

Doesn't the Ascend Sierra LX have a preference score of 6.8 without EQ? I don't see anyone getting excited about them, and some comments are they sound bright.

Great speakers, indeed! Ascend uses the Klippel NFS since 2020 or so. Erin seems to have an Ascend Audio Speaker in the pipeline as quoted in a post in the middle of this thread (The Sierra LX?). The Sierra LX is also on my shortlist. The Epos ES 14 N seems to have a tiny bit more beef / it’s the larger speaker and also seems to have the more massive Cabinet……. But it is 2.5 times as costly as the Sierra LX.

 

[totti1965]

What are linear and non linear phenomena?

Linear = Frequency responses
Non linear = Distortions
But: the Difference between direct sound from short distances and
the „Spinorama Sound“ may be varying / changing from frequency to frequency.
Bang! Here we have a non linear phenomenon also.
You can‘t do anything against it with Equalisation.
This is the reason, why some speakers have higher rates in the Spinorama ranking list
(„With equalisation“) and other speakers have lower ratings.
What is ranked is the „ability“ to equalize. The KI three, the Dutch & Dutch 8c or the Epos ES 14 N
or the Ascend Audio Sierra LX all have high abilities in this regard.

 

[thewas]

The coupling of a typical acceleration sensor with the box is critical. There is plenty of scientific literature which shows multiple aspects of it like the shift of the resonance frequency of the sensor since the coupling acts as a mass loaded spring or the loss of transfered energy due to the mounting technique. Since the sensor had a none curved surface you need more wax or other coupling material which isn't that good.

Shifting the modal frequencies of the structure (not the sensor) in the case of a loudspeaker enclosure more a theoretical problem as there exist very small mass sensors and also the tiny shift doesn't undermine the qualitative analysis which is important, meaning nobody cares about the absolutely exact modal frequency but which modes exist and are significant in amplitude compared to other. Also such sensors can be also glued with special glue, wax just gives less effort but its not the only way.

The measurements from the review is better than nothing but is is certainly not that good to compare two speaker this way without all the context which most readers don't have.

Still better though than no measurements at all and just claims about some supposed superiority.

The port can often act like a sound transmitter from the inside due to the relatively lage surface area and typically thinner walls. So the problem is not only the lack of damping of the air inside the port. If you have more volume you can build a thicker port or more bracings inside to support the port walls.

Since you have to have little space with a typical near field monitor a lot of companies including Genelec built the electronic with jackets and knops directly into the volume of the loudspeaker box. This introduces plenty of potential issues of unwanted resonances and sound transmissions. Even if you are aware of this problem, which Genelec certainly is, it is very hard to avoid all potential issues with it. If you have more space you will most likely built a separated cabinet part for the electronic.

I know from Neumann what effort they spend to avoid such issues (and am sure for Genelec it isn't less) which usually is not performed by typical Hifi manufacturers, in the end such theoretical discussions don't bring us further, only measurements count and from what he have till now those brands show flawless perfection which is almost never seen in comparable Hifi products.

 

[dogmamann]

But: the Difference between direct sound from short distances and
the „Spinorama Sound“ may be varying / changing from frequency to frequency.

Can you get me an article which mentions this ?

 

[dogmamann]

Shifting the modal frequencies of the structure (not the sensor) in the case of a loudspeaker enclosure more a theoretical problem as there exist very small mass sensors and also the tiny shift doesn't undermine the qualitative analysis which is important, meaning nobody cares about the absolutely exact modal frequency but which modes exist and are significant in amplitude compared to other. Also such sensors can be also glued with special glue, wax just gives less effort but its not the only way.


Still better though than no measurements at all and just claims about some supposed superiority.


I know from Neumann what effort they spend to avoid such issues (and am sure for Genelec it isn't less) which usually is not performed by typical Hifi manufacturers, in the end such theoretical discussions don't bring us further, only measurements count and from what he have till now those brands show flawless perfection which is almost never seen in comparable Hifi products.

After using all these machines many hifi manufactures fail to meet the level of perfection that Neumann and genelec were able to acheive.

What I see as the difference in these cases always is: the level of knowledge and scientific apporach those engineers generally use. Most hifi speaker developers work on their “idea” of ideal sound, without proper scientific methods. Many of them don’t even have idea about things which many forum members here posses. Yes, like any businesses some people are just lucky to cling on and some are clever at marketing. As they go rich, they have the money to buy “state of the art” equipments for mentioning in their websites as part of marketing yet, they won’t be using it for everything because they simply lack the know how about how to use it.

 

[dogmamann]

For example, see what that engineer(?!) has to say about cone material and so. His is passionate but half of the things he mentions is to make people believe that his inaccurate speaker is the right way to reproduce vocals. LS3/5a is a good way to reduce the influence of his mediocre recording mics and mixer. But claiming that eq of upper mid dip is the solution to always accurate reproduction of vocals is total BS and yet people fall for his lovely accent and passion in saying the wrong things with authority.

No matter what they do, people keep buying those shoe boxes for insane prices.

 

[test1223]

Shifting the modal frequencies of the structure (not the sensor) in the case of a loudspeaker enclosure more a theoretical problem as there exist very small mass sensors and also the tiny shift doesn't undermine the qualitative analysis which is important, meaning nobody cares about the absolutely exact modal frequency but which modes exist and are significant in amplitude compared to other. Also such sensors can be also glued with special glue, wax just gives less effort but its not the only way.

You shift the resonance frequency of the sensor plus mounting as a system which is very significant. Have a look at manuals and documentation of pcb or bruel & kjaer e.g. here

Model 355B02 | PCB Piezotronics

www.pcb.com

If you use a sensor which is intended zu use below the resonance frequency and it shifts in your frequency range of interest you can get huge differences (figure 1 of the manual).

 

[YSC]

Yes, of course I do.

Now my question: Have you ever worked on speakers with identical drivers and crossovers and identical measurements, only with different cabinet constructions?
If your answer is "No" - what are we discussing here? Now I could ask about a blind A/B comparison result to prove your theory. Did you ever do that? Well, I did and only a deaf person would not have heard the difference.

My main business always was consulting for other companies....you make a quotation for a design and of course, the faster we can get it perfect, the better for us.
For designing drivers and making crossovers, we have already done simulations for a long time, so we know what we get when the speaker is standing in the listening room for the first time. Whenever the result was not good, we started to investigate and in 90% of all cases, there was a cabinet colouration problem. That's, of course, a nightmare for a project.
The problematic frequencies are above 400Hz up to 1-3kHz. Those radiating panels are in the most sensitive area of our human hearing, radiating with different Q to the side and to the top. The BBC made a lot of experiments with cabinet walls .......but one has to understand that this was made mainly for speech and that's why the solution with thin Plywood and thick Bitumen worked best for them, shifting the problem to low frequencies. I would not call the BBC engineering department a marketing team, and I don't think they were trying to improve non-relevant effects.
So for me, this was the starting point and it took a while before we worked out what was essential and what was not so important. These cabinet constructions are now made in simulations with COMSOL and validated later with a laser scanner.....just measurements, no Voodoo.
This is not about sledgehammer engineering like some High-End companies like to do....yes, it works, but it is expensive. This is about reducing the output from cabinets in a sensible way.

Just my 2c...

Sorry about the first question, imo it’s impossible to have two speakers with identical driver and crossover, only different cabinet to have the same measurement, on axis yea, that could be done but when it goes to directivity and distortion, it should always be different simply with different cabinet resonant and volume, be it distortion or say front baffle related effect due to directivity or even bass extension, if all you try to say is about on axis measurements I would say we didn’t disagree on that, but if things like overall spin I don’t feel like there could be unmeasurable attributes only able to be heard by ear

 

[thewas]

You shift the resonance frequency of the sensor plus mounting as a system which is very significant. Have a look at manuals and documentation of pcb or bruel & kjaer e.g. here

Model 355B02 | PCB Piezotronics

www.pcb.com

If you use a sensor which is intended zu use below the resonance frequency and it shifts in your frequency range of interest you can get huge differences (figure 1 of the manual).

That's why I wrote above that a sensor can be glued, in that figure than the resonance of the mounting is far above the typically investigated loudspeaker enclosure range.

 

[test1223]

That's why I wrote above that a sensor can be glued, in that figure than the resonance of the mounting is far above the typically investigated loudspeaker enclosure range.

I was explaining the whole time why such measurements of the magazine you linked have to be taken with conscious. They obviously can't glue the sensor to the cabinet.

Edit: The whole point was that I want to empathize that especially none flat surfaces are more difficult to measure without getting none comparable measurements. Even if you use glue with such surfaces you change the resonance frequency significantly more compared with a flat surface, since you have much more glue between the sensor and surface compared with a flat surface.

Edit edit: if you use wax or some removable adhesive which is typically worse than wax you get a shifts with this specific sensor in significant frequencies where the resonance is about 6 to 8kHz on a flat surface with a curved surface you get in the region of hand probe measurements which is a resonances frequency of about 2kHz

 

[DSJR]

For example, see what that engineer(?!) has to say about cone material and so. His is passionate but half of the things he mentions is to make people believe that his inaccurate speaker is the right way to reproduce vocals. LS3/5a is a good way to reduce the influence of his mediocre recording mics and mixer. But claiming that eq of upper mid dip is the solution to always accurate reproduction of vocals is total BS and yet people fall for his lovely accent and passion in saying the wrong things with authority.

No matter what they do, people keep buying those shoe boxes for insane prices.

I'll bite -

Decent reproduction of speech IS important for our listening as we've evolved our hearing around human voices primarily. Get that wrong and although I daresay some of us can adapt (most men are tone deaf apparently), it makes for more concentrated and sometimes more fatiguing listening for pleasure at home.

Forget LS3/5A's! They were designed to magnify distortion and hiss in a BBC OB van initially (told me by one of the design team back then) and pleasurable music reproduction was NOT part of their initial portfolio. the fact that an entire sub-sect of 'audiophiles' worship the bloody things beats me frankly, but they do and my local dealer praises the current Falcon model's 'detail' when I winced and had to leave the room....

My Harbeths from 2007 also have a slight dip in the lower kHz region engineered in and also, a slightly more 'organic' subjective bass tone compared to the current Plus and XD variants. These things were done deliberately to enable long term listening without fatigue, starting with speech reproduction (which I do a hell of a lot these days playing audio-books) and also I reckon to aid lower level listening (something else I do more now as my hearing can't take very high levels now any more than my next door neighbour can!). Sure, the remaining audiophool in me would like a more 'tactile' reproduction as the best active speakers can give, but looking at the ragged upper mid responses of so many models here (to hell with 'directivity' for a moment!), I'm not sure I want to trade pretty accurate timbres of midrange instruments for a hyper 'detailed' ragged noise, no matter how these speakers are crossed over or amplified!

As for the Harbeth P3, it's a delightful small box to listen to music through. It can go much louder than a LS3/5A can before audible complaint and it 'does' brushed cymbals just as one hears them live (not the one-note 'shhhh' sound some other boxes can give). yes, I agree they're very expensive, but believe me, their owners adore them and tend to hang on to them as well. Yes, it's expensive now, but many people even here can afford such expense without stressing their bank balances.

Seriously fellas, a flat Klippel style response and directivity isn't the entire picture and subjective fine tuning of crossover points and driver levels as well as phase tracking is important as well. LISTENING to the sodding things is vital really for the final 'tuning' of the blending of disparate drivers and obviously the box they're fitted in as well.

 

[totti1965]

Can you get me an article which mentions this ?

Just take a look at the detailed frequency plots at spiNorama.org. / Erins audiocorner / or audiosciencereview.com. Meanwhile there are 500 of them.

 

[totti1965]

For example, see what that engineer(?!) has to say about cone material and so. His is passionate but half of the things he mentions is to make people believe that his inaccurate speaker is the right way to reproduce vocals. LS3/5a is a good way to reduce the influence of his mediocre recording mics and mixer. But claiming that eq of upper mid dip is the solution to always accurate reproduction of vocals is total BS and yet people fall for his lovely accent and passion in saying the wrong things with authority.

No matter what they do, people keep buying those shoe boxes for insane prices.

Partly you may be right. But Bernd Timmermanns from German Magazine Hobby HiFi who measured thouthands of speakers mentioned that the OLD KEF B110 speaker Chassis, which is build in in the 3/5a speakers (it is from 1974 or so!) is for its 110 mm size still the world record holder in mechanical losses Rm.
So you can simply measure und you find something out about the Faszination this speaker has for generations of listeners!