Elac Uni-Fi 2.0 Review (bookshelf speaker)

[Jmudrick]

Hello Jack,

Nice to get your professional opinion on this.

Yes quite possibly, I've had narrow band distortion- turn out to be drivers not fully screwed into the baffle...

Remembering all the fuss with the broken Kali IN-8, I am likewise doubtful this is a design flaw. These people (Kali, Elac) measure.

 

[Frank Dernie]

@amirm I'm not sure how much time you want to devote to this issue but you could stick a mic in front of the speaker when it is making this egregious noise, that might clarify the severity of the issue. Not something I'd normally suggest but it might be an interesting point for discussion.

When I was doing noise and vibration research we always listened to the sensor signal as well as analysing it, usually an accelerometer or velocity transducer. It was amazing how often its sound allowed us to recognise quickly where problems came from.
It was something my original boss suggested.
He had done it ever since struggling to analyse a vibration signature on a warship transmission system. After spending a long time inconclusively analysing signals, and this would be 55 or 60 years ago, so with much slower, more expensive and difficult to use kit than nowadays.
He decided to listen to the signal and heard unmistakable interference from the ship intercom system. It would have saved days of analysis and head scratching if he had done it as soon as the analysis looked strange, so he told us young engineers to listen if the signal analysis was difficult to interpret.
I never had intercom but recognisable sounds like gear whine and so forth saved time focusing our work.
Fun days.

 

[Steve Dallas]

That is interesting. This is a different measurement than you originally published, which showed much lower system distortion at 86 dB and higher distortion (over 4%) at 96 dB. How was this conducted? Is it a nearfield of the port? I'm just interested in how you isolated the distortion product to the port. In any event, this seems to be another strike against front-mounted ports.

I certainly know very little about the subject matter here, but my Revel F206s do not exhibit any of the oft repeated ills of front ports. Nor do my KEF Q100s. I purchased those specifically for their front ports, as they sit in an actual bookshelf. It seems it can be done well.

[shrugs with bewilderment]

 

[Xyrium]

Another mediocre one for the record books. Thanks for reviewing this one Amir...and thanks to the owner for sending it in!

It is interesting that the speaker could impart noise at that SPL. I wonder if this is just a typical mass production / less than stellar QC problem.

 

[ROOSKIE]

if that's the case then it's kind of sad for concentric designs.. we all know how fedex and UPS handle packages..

I for one am grateful that the people who work for the delivery services are willing to lift and carry the endless stream of packages that come my way. Not to mention seeing how many other packages arrive up and down just my one street.
We are as gods to receive such packages delivered to our doorsteps. (My LSIm707's weighed 120lbs each and I recieved 6 of them here)
My Tai Chi instructor worked @ups loading trucks to make ends meet. He was possitive about it but there was no doubt it was extremely hard work. He is man who teaches a dozen martial arts and is very muscular and in shape, as in break you in half with ease shape.
I applied to work part time there back in my 20's. I took one look at the conveyor belts and the ear bleedingly loud machinery surrounding my work station and I left.
Beyond that.
Every single manufacturer knows what shipping entails and decides how to address this.

 

[Helicopter]

I saw an interview video at a show where Andrew Jones is demonstrating a new Unifi speaker he designed and he says he is pleased to hear it for the first time. They used new drivers, and had designed them and the speaker on a computer then got a pair for the show from China. They never did a prototype due to time constraints. I think that was an earlier one. I can't find the video now.

I did find a video in which he said he wants the performance of his speakers to 'surprise' people. Looks like he at least did that for many with this one.

Having seen that show video, I am not surprised there is a major resonance problem. This speaker reinforces my impression that some Elacs are great, like the new DBRs; but not all of them are great, like Revels, or good to great, like Focals.

Thanks for the review Amir.

 

[ROOSKIE]

...



My issue with the Uni-Fi 2.0 is that I didn't find the treble very pleasing, nor did I find it overly pleasing on the Debut Reference. I said this in my video review. I've come to the conclusion that they're using pretty much the same tweeter on both. I haven't spent too much time looking at the Klippel measurements, but is there anything there that might explain why the treble response might sound somewhat harsh?...

Yah, subjectively I have noticed something about the ELAC highs. (5.2, 6.2, UB5)

Nothing bad per say, I just found them to have an "etched" sound. It is my belief that some really like this quality and some to do not. (and of course some prolly don't consider it at all)

This is something both myself an my GF heard consistently, including in direct comparison with some other speakers.

None of these speakers were "bright", just etched & from memory and direct comparison the UB5 had less of it.



As far as what measurements describe this I am still unsure myself. Possibly HD and or IMD. Possibly something to do with the material of the dome.



With some tracks I actually like the etched quality and I did like ELAC speakers at low volumes more than normal. In the end none of the ELAC speakers stayed and I do not miss them. I still think they (all three sets i had here) are very well engineered for the prices and while not my top choices are certainly good speakers overall.

A lot of people are jumping to conclusions without knowing for sure if the issue is widespread or specific to this one speaker. To me that's the danger of taking these at face value. It's like the thing with the Denon review. It's important to analyze the data to figure out what it's telling us. Science is about continually asking questions surrounding the data given to us via measurements. It's important that we interpret the data correctly (using the scientific approach) if we want to come to the right conclusions.

I deff do not get the negative anti-hype, I mean even this pair measures quite well if it didn't have that resonance. We have a sample size of one right now here along with several others who did not notice this issue on other sites. Hopefully ELAC send another one in/looks into it.

 

[MZKM]

Remembering all the fuss with the broken Kali IN-8, I am likewise doubtful this is a design flaw. These people (Kali, Elac) measure.

I believe for Kali they said each unit is measured. I highly, highly doubt that ELAC is measuring each unit, their volume of sale is decently larger.

 

[Jmudrick]

I believe for Kali they said each unit is measured. I highly, highly doubt that ELAC is measuring each unit, their volume of sale is decently larger.

I wasn't suggesting each unit was measured, rather that it seems surprising such resonance issues would escape their design lab where they certainly would be measuring for such flaws.

 

[thewas]

I believe for Kali they said each unit is measured. I highly, highly doubt that ELAC is measuring each unit, their volume of sale is decently larger.

AFAIK the till few years ago only existing made in Germany Elac models were all measured in the factory, but they are also more expensive than the recent parallel "Elac USA" models which are made in China.

 

[SMc]

Also, how in the world did Andrew Jones become the only speaker designer people know by name? Not to take anything away from him, but it's very strange to me.

Paul S. Barton comes to mind for living designers.

 

[Helicopter]

Yah, subjectively I have noticed something about the ELAC highs. (5.2, 6.2, UB5)

Nothing bad per say, I just found them to have an "etched" sound. It is my belief that some really like this quality and some to do not. (and of course some prolly don't consider it at all)

This is something both myself an my GF heard consistently, including in direct comparison with some other speakers.

None of these speakers were "bright", just etched & from memory and direct comparison the UB5 had less of it.



As far as what measurements describe this I am still unsure myself. Possibly HD and or IMD. Possibly something to do with the material of the dome.



With some tracks I actually like the etched quality and I did like ELAC speakers at low volumes more than normal. In the end none of the ELAC speakers stayed and I do not miss them. I still think they (all three sets i had here) are very well engineered for the prices and while not my top choices are certainly good speakers overall.



I deff do not get the negative anti-hype, I mean even this pair measures quite well if it didn't have that resonance. We have a sample size of one right now here along with several others who did not notice this issue on other sites. Hopefully ELAC send another one in/looks into it.

The tweeter cover looks similar to the one on the Elac Uni Fi UB5. On the UB5, the cover is supposed to function as an ultrasonic waveguide. I have serious doubts about the importance of ultrasonic performance, but that thing might be doing something more to the tweeter in general. If the mid isn't sensitive, I can still sorf of understand not wanting to pad the woofer too. Then again, maybe they only used simulations to design this again and overestimated the simulations.

 

[Steve Dallas]

I saw an interview video at a show where Andrew Jones is demonstrating a new Unifi speaker he designed and he says he is pleased to hear it for the first time. They used new drivers, and had designed them and the speaker on a computer then got a pair for the show from China. They never did a prototype due to time constraints. I think that was an earlier one. I can't find the video now.

I did find a video in which he said he wants the performance of his speakers to 'surprise' people. Looks like he at least did that for many with this one.

Having seen that show video, I am not surprised there is a major resonance problem. This speaker reinforces my impression that some Elacs are great, like the new DBRs; but not all of them are great, like Revels, or good to great, like Focals.

Thanks for the review Amir.

I remember that interview as well. It was a head-scratcher to be sure. From computer straight to production. Did they do any durability testing?

 

[MrPeabody]

It is not surprising to see the high distortion peak at 585 Hz, given that the impedance graph clearly indicates a resonance of some sort at this frequency. Surely the two are related.

The wavelength corresponding to this frequency will be 1.9 feet, which is essentially twice the interior height of the enclosure. (And to a high degree of exactness when considering excitation of resonances.) This suggests a standing wave between the bottom and top interior surfaces of the enclosure may be playing a prominent role in the full understanding of the phenomenon. Or this may just be a coincidence.

The frequency is harmonically related to the frequency of the upper impedance peak seen at about 2.4 kHz. I'm inclined to think that this doesn't have anything at all to do with the phenomenon, and mention it only for the sake of completeness. (Presumably this impedance peak corresponds roughly to the crossover point, since at the crossover point, each of the parallel circuit paths through the crossover will exhibit high impedance; there will be no path with low impedance that would mask the high impedance of the other paths.)

In the simplest hypothesis that I'm able to come up with, 585 Hz is the tweeter resonance. It isn't pronounced in the impedance plot because it is masked by the low impedance of the midrange circuit which is in parallel with the tweeter circuit. It is plausible that when the speaker is driven to high SPL, the tweeter's desire to resonate overwhelms the ability of the high-pass filter to damp the resonance, and the result is the conspicuously non-linear behavior evidenced in the distortion plot. Potentially, the tendency for the tweeter to do this may be in some way exacerbated by the concentric arrangement. Given the concentric arrangement it is natural to wonder about this, but I don't know of any substantive reason to think that something along these lines is occurring. A more compelling question is whether the tweeter resonance and the interior cabinet resonance are mutually sympathetic, each stimulating the other. I would not entirely exclude this possibility, but I think it is not particularly likely.

Bottom line, my hypothesis is that this is simply a resonance issue with a tweeter that isn't adequately damped intrinsically and that demands a steeper high-pass filter that it wasn't given. This is the simplest hypothesis that comes to mind and it is consistent with all the evidence that I see.

 

[DSJR]

I'm hoping/suspecting it's a fault and not a design issue (the slightly offset tweeter reported earlier may well be a factor in this review).

A similar tale possibly... Decades back, there was a popular Tangent 2cu size speaker (RS4) with a deliberate 'batman shape' response if memory serves (they'd be annihilated here, long before any listening was done). A pal had a pair for exclusively orchestral and choral music reproduction and they threw a wonderful soundstage with this music, the odd rersponse not getting in the way of venue acoustic I noticed. Twenty five years or so went by with little issue but suddenly on one or both, soprano voices became very distorted at certain notes. The Sony V-FET amp was checked and found to be ok (it destroyed itself and a different pair of speakers some years later ) and a friendly speaker maker took them in and tested the drivers. Turns out the 8" Audax bass units had suffered adhesive fatigue and there was a massive issue at 900Hz or so. In these more enlightened times, I'm sure the bass drivers (now obsolete) could have been rebuilt by a couple of trusted businesses in the UK, but back then we had little idea of how to proceed. Said pal changed the speakers and no idea what happened to the Tangents sadly.

So, hopes that the issue here is a transit based? fault rather than a design issue...

P.S. UK based, I don't know Andrew Jones so much (blame 'noaudiophile' for my education on the matter ). Peter Comeau (started as co-owner of Heybrook, noiw a wizz at IAG), Spencer (r.i.p.) and son Derek Hughes, Robin Marshall (with whom I shared employer back in the 70's) Billy Woodman and the 'low-cost-speaker-king' Andy Whittle I definitely know and know of

 

[F1308]

Maybe, sometimes, Amir makes a joke I think 52 in de name stands for 52000 dollar for stereo.

Yes, It is 25000$ per unit, 50000$ stereo with those specifications !!!!
A DAC placed well in that red area will suffice.
Obviously a present from a novice.

 

[MrPeabody]

I believe for Kali they said each unit is measured. I highly, highly doubt that ELAC is measuring each unit, their volume of sale is decently larger.

Good point. In standard QC, you only have to test a specific percentage of units in order to have confidence that the percentage of faulty units will be below a threshold of your choosing. For the given percentage of faulty units that you are willing to tolerate, the percentage that you have to test decreases as the quantity of units increases, up to a point where, in accordance with the law of diminishing returns, the percentage you have to test does not decrease further as the quantity of units increases.

 

[ctrl]

Inhaling notes (?), please add the guitar notes. There's D4 with harmonics D5 + inhaling (all at same time),

You're right, the midrange is primarily where the guitar plays (D4 and A4 is both from the guitar).

Tried again to identify the sound described by @amirm when Eva is inhaling and I think I have found the relevant frequency range.
Very roughly, the breathing noise is audible in the range 1.2-5kHz (mainly 1.8-4kHz). If this frequency range is lowered, then this noise disappears almost completely. But that would be a completely different frequency range than @amirm suspects.

As an attachment one second from the track to demonstrate the said music passage.
Perhaps @amirm can confirm that this is indeed the strange sounding music passage.

And effectively is the midrange metallic cone with a double resonance mode, typical membrane modes (modes 2.1 and 0.2 probably, 2.135f and 2.295f)

The crossover frequency to the tweeter is at 2kHz and noticeable harmonic distortion is only around 630Hz. There, HD2 is relatively high.

Can you describe how the membrane resonances are excited when the harmonic distortion is insignificant and the midrange is already 6dB damped at 2kHz?
What do you mean by (2,1) mode? Third order modulation distortion with 2*f1 - f2?

Obs.: your FFT shows d4 and a4...but instruments play harmonics ;-)

Guitar is more complex and shows D4 too, when A4 is played - at least mine

 

[ctrl]

It is not surprising to see the high distortion peak at 585 Hz, given that the impedance graph clearly indicates a resonance of some sort at this frequency. Surely the two are related.

The wavelength corresponding to this frequency will be 1.9 feet, which is essentially twice the interior height of the enclosure. (And to a high degree of exactness when considering excitation of resonances.) This suggests a standing wave between the bottom and top interior surfaces of the enclosure may be playing a prominent role in the full understanding of the phenomenon. Or this may just be a coincidence.

As previously stated, the crossover frequency of the woofer is 200Hz and at 585Hz the woofer is barely excited and thus no standing wave in the cabinet - so it is most likely coincidence

In the simplest hypothesis that I'm able to come up with, 585 Hz is the tweeter resonance.

I would see that as very unlikely. The installed tweeter has, if at all, only a small rear chamber and therefore very likely a resonant frequency above 1kHz.

 

[uwotm8]

Excuse me but what if it's just a broken/damaged unit? The issue is quite big to remain undetected on the way from prototype to series and pass QC.