Elac Uni-Fi 2.0 Review (bookshelf speaker)

[Diablogt]

I demoed the speakers when it first came out. The sound was good but not much to write home about. The biggest issue I had with it is that it sounds confusing with some tracks like multiple instruments fight against each other. However, no bad resonance detected even tho I played real loud. Is there a way to make sure this is a design issue or just a QC or shipping damage issue?

 

[MrPeabody]

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
View attachment 104888

Yes, as I said it is most likely a coincidence. I actually did read through the whole thread to see if there had been any discussion of this, but I apparently missed it. However I would not say that the woofer is "barely excited" at frequency 3x the crossover point. Also I would not exclude the possibility that acoustic energy from the midrange at that frequency travels inward through the port to excite the internal resonance. My inclination is that this is not likely, but it is only my inclination. I do not have sufficient reason to exclude this possibility, and given that the wavelength of the distortion peak is essentially a perfect match to this interior dimension, I would not exclude this possibility without having a pretty good reason to do so.

The installed tweeter has, if at all, only a small rear chamber and therefore very likely a resonant frequency above 1kHz.

Before I would conjecture what you are conjecturing, I would first need to have an alternative hypothesis for the cause of that little hiccup in the impedance graph. It had been my standing assumption that the resonance frequency of tweeters as generally reported in the spec sheets is measured for the assembled tweeter. It would not make sense for it to be otherwise, and to my recollection, typical resonance frequencies for tweeters is in the range of several hundred Hz to maybe 1 kHz, but not as high as 2 kHz, which would put the resonance within the tweeter's operating range. In any case, it is not the least bit unreasonable for the tweeter's resonance to be found at frequency this low, contrary to what you are saying (and saying without any particularly good reason to say it). If I were inclined to dismiss this hypothesis, I would be more inclined to do so on the basis that it should be suppressed by the high-pass filter, but I would not dismiss it on this basis either. For me to dismiss this hypothesis without definitive proof that it isn't correct, I would need to see a better hypothesis, or at least a really good reason to reject this hypothesis. Thus far I've seen neither.

I don't want to suggest any additional work for Amir, but the hypothesis that this is simply tweeter distortion could be easily tested, by simply putting a folded wool blanket over the tweeter and taking the distortion measurement again.

 

[Rock Rabbit]

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.



The crossover frequency to the tweeter is at 2kHz and noticeable harmonic distortion is only around 630Hz. There, HD2 is relatively high.
View attachment 104882
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?




Guitar is more complex and shows D4 too, when A4 is played - at least mine
View attachment 104879

From your capture

Peaks from left D2 D3 D4 A4...for resonance very little energy and two narrow resonances is a hint of membrane breakup
Modes http://hyperphysics.phy-astr.gsu.edu/hbase/Music/cirmem.html
https://www.acs.psu.edu/drussell/Demos/MembraneCircle/Circle.html

 

[w1000i]

Interesting how the distortion from the port clearly shows the problem:
View attachment 104762

Seems like the cabinet is amplifying it and then comes out of the port? It does not show up as a port resonance.

Can you test each driver alone ? without it crossover.

 

[AndreaT]

I just didn't see the point of showing other measurements when we had a core issue. Here it is anyway:

View attachment 104761

How unusual to see a crossover for a 2-way at 180 Hz or so. Ah, sorry, midrange and tweeter coaxial according to company website...

 

[amirm]

OK, guys, slept in this morning and seems like you all have been busy.

This morning I thought I should record the problem. When I put the speaker back on the stand and fed it with the youtube video, it didn't seem to happen. So I went back to Roon and at first it was fine there as well. I thought maybe this is level sensitive and indeed it is. I had to crank it up fair amount for it to occur again. Either this then makes the problem worse, or I get sensitized to it as I can then hear it at lower levels as well. It is also possible that using the speaker makes it do this more.

When I tried to record it, I realized it is not the breathing that is causing the problem but the letter "H" when she utters "he goes away." The H picks up a metallic overtone. I have enclosed a crappy recording on my Samsung phone. I can hear it there but it is far more subtle than what I hear in person.

I tried to get closer to the speaker for better recording and it became harder to hear and capture it as well!

FYI, I stuck a towel in the port and it made no difference. The issue is not the port. The port simply is allowing the distortion to be magnified. Audibility is elsewhere.

 

[amirm]

How unusual to see a crossover for a 2-way at 180 Hz or so.

It is a three-way as the top driver is coaxial. I was too lazy to change the notation there.

 

[amirm]

On the question distortion in the port, all of those measurements are at about 1 inch away from the radiating surface. I start with the tweeter and the lower the mic to center it with the next driver and then the port. So everything else is playing at the same time I am measuring the port. The frequency response measurements spit out distortion for free and that is where that data came from.

 

[napilopez]

Are you still refusing to evaluate speakers in pairs? I continue to maintain those sort of subjective evaluations are meaningless and not useful to the consumer.

On the point of mono vs stereo listening, , like Amir, I also measure and create Spinoramas (with less accuracy/resolution of course). I also listen in mono during evaluations, but I mostly listen in stereo because I use the speakers in my primary listening room setup. Usually it's at least a month, some it's more.

Though I am nowhere near Amir's pace, after a couple of dozen speakers, listening in stereo has almost never revealed something I could not glean from mono listening.

There are a few exceptions:

- Egregiously bad pair matching, a relative rarity these days. More common with ultra-budget speakers but even then rare to actually notice in stereo listening.

Ironically, it's easier to detect by listening to both speakers separately in mono. The time I noticed the worst pair matching, one of the speaker's dome tweeter had been pushed in, and I noticed the speakers sounded a bit 'dull' and perhaps slight asymmetrical, but it was mono listening and later measurements that clearly revealed the issue.

-Wireless speaker syncing: For wireless speakers, stereo is necessary to make sure the wireless connection is actually solid. Unfortunately, this is sometimes still not the case.

-Offset drivers: Speakers with offset, mirrored drivers will sound different depending on whether the tweeter is on the inside or outside edge of the speaker. It's harder to relate spatial presentation in mono to stereo for offset drivers in my experience.

-Stereo listening leads into a perceptual dip around 2kHz for the phantom center(Interaural crosstalk cancellation). Usually not a huge deal in a typically reflective environment, but it's a slight difference in tonality.

But far more important than these exceptions, in my opinion, is that listening in mono often reveals things that are much obscured in stereo without extended critical listening. An hour of mono listening feels like it tells me a few days worth of stereo listening. You get way more bang for your buck listening in mono when evaluating a speaker's flaws.

To make an exaggerated analogy, suggesting Amir should place the same or more weight on stereo listening is like suggesting a photographer should evaluate a camera's performance based on how the photos look on a smartphone after an instagram filter or two. That's how most consumers will see the photos, but it's not the best way to isolate performance issues.

Similarly, the net effect of stereo listening is making a speaker sound better, as per harman's research and my own experience. It does little to improve ones ability to critically evaluate a speaker.

Both is obviously better than just mono, but mono listening is very much more efficient than stereo. And at Amir's pace, it doesn't make sense to do otherwise imo.

 

[amirm]

On issue of something being loose, there are no fasteners in front. I don't want to try to pry up the ring around the drivers as the owner wants to return the speaker. Nothing is rattling or anything.

As to making intermodulation measurements, that is another option to pay for on Klippel. I could try to set it up and measure it on AP but really, I don't see my job as spending tons of effort troubleshooting the cause. I am here to report on what I test. There is enough data here for the company to investigate. Or contact the owner to buy his unit to test before he returns it.

In the past, I have reached out to Andrew and while he has answered my emails, he has not followed up on questions raised in my review. So I am not going to make a pest of myself by chasing him on this issue. I have ton of other stuff to measure. If he wants me to test another sample, I am here to do so.

 

[amirm]

Can you test each driver alone ? without it crossover.

No. I would need to fully take apart the speaker. It is not mine and the owner wants to return it so it is not something I can do.

 

[Rock Rabbit]

He goes away ....that explains everything
Is a vocal full swept in the resonance of mid range

 

[amirm]

I don't want to suggest any additional work for Amir, but the hypothesis that this is simply tweeter distortion could be easily tested, by simply putting a folded wool blanket over the tweeter and taking the distortion measurement again.

I put a blanket over the coaxial driver and I still could hear the problem. I then pulled it further down on top of the woofer and still could hear the problem.

Almost at the same time I hear the problem, the woofer goes way in and stays there for a second. The coaxial driver is also moving fair bit. Not sure if these are helpful hints but this is all I have.

 

[amirm]

BTW, this is the actual album:

The youtube version I linked to is from another album. While it showed the problem, it was not as severe there so maybe there is a difference in recording (or youtube compression changing it).

 

[Shazb0t]

As to making intermodulation measurements, that is another option to pay for on Klippel.

How much does the module cost?

 

[Juhazi]

Best way to hear this is to slowly sweep sine to find the exact freq that wakes up distortion. Then play that freq fixed and listen closely around the speaker and use fingers to feel it. Pretty much like Frank Dernie said.

 

[amirm]

How much does the module cost?

It is part of two other packages, one costing 3,600 and the other 4,700 euros. It just makes no sense to me to pay a fraction of that for the little use I get out of it. And free alternatives.

 

[amirm]

Best way to hear this is to slowly sweep sine to find te exact freq that wakes up distortion. Then play that freq fixed and listen closely and use fingers to feel it. Pretty much like Frank Dernie said.

If you had free time to burn, sure. Klippel has no interactive mode to do that. I would have to setup something with AP or REW, bring out measurement mic, setup the pre-amp for that, etc., etc.

The PEQ experiment I ran did that anyway. So not sure what all is gained out of stepping through a ton of tones to do it again.

 

[richard12511]

View attachment 104900He goes away ....that explains everything
Is a vocal full swept in the resonance of mid range

Can you explain more? I don't know what I'm looking at, or where the problem is.

 

[SmackDaddies]

Listening to one speaker will certainly allow one to hear many characteristics of a speaker, such as tonal balance. It seems to me that perspective, or presence of the speaker is more easily heard in a stereo performance .