Elac Adante AS-61 Speaker Review

[andreasmaaan]

YOu mean this region?

View attachment 50492

If so, the four curves above track each other so if you subtract them, you don't see a peak.

I'm just relying on my eyes here, but I see a dip of around double the intensity (4dB vs 2dB) and half the Q in the sound power vs on-axis (red curve vs black curve in your screenshot I believe).

I would have expected the Sound Power DI to show a corresponding peak.

 

[MZKM]

Thanks again. Do you have a link to the literature that specifies how the score is calculated, please?

Also, would you mind fleshing out your statement that floor and ceiling reflections are not overly audible? Is this based on the fact that HRTFs show lower sensitivity to up/down sources in the high frequencies, or is it based on specific studies in which listeners were asked to give preference ratings of specific speakers? If the latter, do you have any specific links, perhaps?

And finally, one (I think) last question about the ELAC: The Spinorama shows a rather large dip in the Sound Power, Early Reflections, and Predicted In-Room Response between 1kHz and 2kHz. These dips are significantly wider and deeper than the small dip in the same frequency range in the on-axis response, yet there is no peak in this range in the Sound Power DI or Early Reflections DI. This seems odd to me - are you aware of how this could be?

Sorry also to be dominating this thread with my attempts to understand the preference ratings. Above all, I'd just like to read some detailed literature on how the preference rating is computed - then I'll happily shut up

In the Notes portion of my preference score database it has a link to an explanation,
direct link: https://www.audiosciencereview.com/...quivalent-sinad-discussion.10818/#post-302047

No links, but Toole has done research on the effect of the different reflections, it’s gonna be in his book somewhere.

I see a small rise in DI from 1kHz-2kHz.

 

[andreasmaaan]

For all passive speakers, they are fed 2.83 volts. The analyzer is smart and adjusts amplification gain until it gets this precise voltage level (amp output gets routed through the analyzer on the way out to the speaker). CEA-2034 stipulates 2.83 volts and not a fixed SPL for passive speakers and hence this strategy.

So depending on efficiency of the speaker, the actual SPL level varies some. Most have been similar though so until we test ultra efficient speakers or something, the variation in SPL is small.

For active speakers, the situation gets much more complicated. CEA 2034 wants 79 dB at 3 meters (or is it 2?). I premeasure this at 0.33 meters since my analyzer robotic arm can't go a lot past that point. I computed something like 94 dB required at this distance. For some reason though, with active speakers, the CEA-2034 graph shows completely wrong SPL information (unlike passive speakers). I think it is confusing the voltage to the active speaker with voltage on a passive one. I have to contact them to see what is going on with this.

Note that the concept of "at what distance measurements were taken" do not apply to Klippel NFS measurements. The system measures in near-field and computes the distance on its own based on optimized number of points (1004 points in this case). Once it computers the sound field mathematically, then it projects the sound field to far field and reports it at 1 meter. No actual measurement is occuring at 1 meter. It is like having the formula Y = 2 * X +5 and plugging whatever you want in "X" and solving for "Y."

Hope this is clear.

Thanks for this. It is a bit odd that it's the voltage input and not the sound pressure output that CEA stipulates, as it's the latter that matters to the listener. A penalty for efficient speakers!

And a slight advantage for active speakers, as the average passive loudspeaker produces more than 94dB SPL at 33cm from a 2.83V input.

Would you consider @amirm standardising the distortion measurements to a specific, slightly higher SPL? Perhaps something like 96dB/1m (105.63dB/0.33m), for example? I think this would be more revealing of relvant differences between speakers, and also ofc fairer to efficient passive speakers.

 

[napilopez]

Thanks again. Do you have a link to the literature that specifies how the score is calculated, please?

Also, would you mind fleshing out your statement that floor and ceiling reflections are not overly audible? Is this based on the fact that HRTFs show lower sensitivity to up/down sources in the high frequencies, or is it based on specific studies in which listeners were asked to give preference ratings of specific speakers? If the latter, do you have any specific links, perhaps?

And finally, one (I think) last question about the ELAC: The Spinorama shows a rather large dip in the Sound Power, Early Reflections, and Predicted In-Room Response between 1kHz and 2kHz. These dips are significantly wider and deeper than the small dip in the same frequency range in the on-axis response, yet there is no peak in this range in the Sound Power DI or Early Reflections DI. This seems odd to me - are you aware of how this could be?

Sorry also to be dominating this thread with my attempts to understand the preference ratings. Above all, I'd just like to read some detailed literature on how the preference rating is computed - then I'll happily shut up

Interpreting the measurements is definitely and score a bit of an art @andreasmaaan and it's clear the score isn't perfect, but to answer your questions, the score is derived from this Sean Olive paper: A Multiple Regression Model for Predicting Loudspeaker Preference Using Objective Measurements: Part II - Development of the Model. Part 1 is also worth reading too, as it actually calculates a more accurate preference rating for bookshelf speakers alone.

@MZKM breaks down the individual components of the score in this thread.

Vertical reflections contribute less to spatial qualities of the sound and maybe tonality, so they seem to carry less weight based on what I've read, though there's some contention here.

Studies on audibility aside, with regards to the score, it's ultimately a matter of how much the vertical measurements influence the graphs used in calculating the preference score. The two operations on the predicted in-room response constitute the bulk of the score, and the PIR itself is made up of 12 percent listening window, 44 percent early reflections, and 44 percent sound power. Horizontal measurements are weighed more heavily in the early reflections and listening window curve and equally in sound power, so ultimately horizontal measurements end up being weighed more strongly by the preference formula anyway.

 

[andreasmaaan]

In the Notes portion of my preference score database it has a link to an explanation,
direct link: https://www.audiosciencereview.com/...quivalent-sinad-discussion.10818/#post-302047

No links, but Toole has done research on the effect of the different reflections, it’s gonna be in his book somewhere.

I see a small rise in DI from 1kHz-2kHz.
View attachment 50498

Thanks @MZKM, my recollection is that Toole's finding was that up/down reflections do contribute significantly to perceived tonal balance (it was only in respect of spatiousness that these reflections were found to be relatively unimportant). This is why I was a little puzzled. My memory may be faulty, of course.

Also, I see from that post that you've already done a huge amount of thinking about some of these matters. Well done with all this

 

[bobbooo]

Thanks again. Do you have a link to the literature that specifies how the score is calculated, please?

If you're not an AES member and so can't view the preference formula paper in the above link, you can find it here (scroll down for the correct paper). Well worth a read.

 

[andreasmaaan]

Interpreting the measurements is definitely and score a bit of an art @andreasmaaan and it's clear the score isn't perfect, but to answer your questions, the score is derived from this Sean Olive paper: A Multiple Regression Model for Predicting Loudspeaker Preference Using Objective Measurements: Part II - Development of the Model. Part 1 is also worth reading too, as it actually calculates a more accurate preference rating for bookshelf speakers alone.

@MZKM breaks down the individual components of the score in this thread.

Vertical reflections contribute less to spatial qualities of the sound and maybe less to timbre, so they seem to carry less weight. Basically though, it's ultimate a matter of how much the influence the graphs used in calculating the preference score. Horizontal measurements are weighted more heavily in the early reflections portion of the score (which is a portion of the predicted in room response).

Thanks @napilopez, those links are exactly what I was after.

Concerning the weighting of floor/ceiling reflections, I agree 100% re: their contribution to perceived spaciousness, but it's the following (in bold) I can't recall ever coming across any evidence of:

"Vertical reflections contribute less to spatial qualities of the sound and maybe less to timbre..."

I don't recall ever reading a study that suggested that the contribution to timbre was significantly less than for lateral reflections. It seems that both you and @MZKM recall reading it though. If anyone around here knows where it comes from, I'd love them to post a link to the relevant studies.

 

[andreasmaaan]

And PS to clarify, not contributing to spatiousness would of course not be any reason to reduce the weighting of these reflections on listener preference. Hence my questions on this point.

EDIT: I went too far with this statement here, I quickly realise now I should have said "...would not be any reason to greatly reduce the weighting of these reflections".

 

[napilopez]

Thanks @napilopez, those links are exactly what I was after.

Concerning the weighting of floor/ceiling reflections, I agree 100% re: their contribution to perceived spaciousness, but it's the following (in bold) I can't recall ever coming across any evidence of:

"Vertical reflections contribute less to spatial qualities of the sound and maybe less to timbre..."

I don't recall ever reading a study that suggested that the contribution to timbre was significantly less than for lateral reflections. It seems that both you and @MZKM recall reading it though. If anyone around here knows where it comes from, I'd love them to post a link to the relevant studies.

Yeah, I qualified that statement because the research seems to be contentious .

I specifically was thinking about ceiling vs floor reflections. Ceiling reflection everyone seems to agree can be bad. But Toole seems to believe the floor reflection isn't a huge contributor to our perception because we've adapted to it. Chapter 7.4.7 from the latest edition of his book:

"Very early in my explorations of loudspeaker/ room interactions I took note of the measurable effect of the floor reflection in steady-state room curves. It seemed like a problem that needed attention, so I devoted some time to modifying loudspeakers to minimize it, and followed through with subjective evaluations. I cannot claim to have been exhaustive, but I soon became frustrated when the curves looked better but the sound seemed not to have changed very much."

Furthermore:

"The Fraunhofer Institute in Germany constructed an elaborate listening room in which different room surfaces could be changed (Silzle et al., 2009). “Regarding the floor reflection, the audible influence by removing this with absorbers around the listener is negative— unnatural sounding. No normal room has an absorbent floor. The human brain seems to be used to this.”

He then goes on to surmise that we've adapted to floor reflections so it doesn't make sense to try to treat it. This doesn't necessarily mean that we don't prefer good frequency response for the floor reflection, but it suggests we're less sensitive to it than others.

However, other research has said that it's the worst of reflections though, so who knows =]

 

[amirm]

Would you consider @amirm standardising the distortion measurements to a specific, slightly higher SPL? Perhaps something like 96dB/1m (105.63dB/0.33m), for example? I think this would be more revealing of relvant differences between speakers, and also ofc fairer to efficient passive speakers.

I have to sort out the issue with SPL levels for active speakers. Once there, yes, I will need to standardize on something.

 

[andreasmaaan]

I have to sort out the issue with SPL levels for active speakers. Once there, yes, I will need to standardize on something.

Many thanks @amirm

 

[lszomb]

The design is quite elegant with a lot of thoughts. Now, I think we should put ATC SCM20ASL Pro (I own a pair) for testing to see if these stand for the price.

 

[goldark]

Any chances of reviewing the Sierra 2-EX? It would be a competitor to the Adante/Carina so might be interesting. Ascend already has posted its own measurements but always good to get another data point to confirm: http://www.ascendacoustics.com/pages/products/speakers/SRM2EX/srm2exmeas.html

 

[andreasmaaan]

Yeah, I qualified that statement because the research seems to be contentious .

I specifically was thinking about ceiling vs floor reflections. Ceiling reflection everyone seems to agree can be bad. But Toole seems to believe the floor reflection isn't a huge contributor to our perception because we've adapted to it. Chapter 7.4.7 from the latest edition of his book:

"Very early in my explorations of loudspeaker/ room interactions I took note of the measurable effect of the floor reflection in steady-state room curves. It seemed like a problem that needed attention, so I devoted some time to modifying loudspeakers to minimize it, and followed through with subjective evaluations. I cannot claim to have been exhaustive, but I soon became frustrated when the curves looked better but the sound seemed not to have changed very much."

Furthermore:

"The Fraunhofer Institute in Germany constructed an elaborate listening room in which different room surfaces could be changed (Silzle et al., 2009). “Regarding the floor reflection, the audible influence by removing this with absorbers around the listener is negative— unnatural sounding. No normal room has an absorbent floor. The human brain seems to be used to this.”

He then goes on to surmise that we've adapted to floor reflections so it doesn't make sense to try to treat it. This doesn't necessarily mean that we don't prefer good frequency response for the floor reflection, but it suggests we're less sensitive to it than others.

However, other research has said that it's the worst of reflections though, so who knows =]

Yes, excellent points re: Toole and I recall this now also.

I think it's worth noting though that in the context of this anecdote he gives, he seems to be referring primarily to the first cancellation null, typically in the bass or low-midrange, and present not only in the context of speakers, but also in the context of any real acoustical source (i.e. that would be the null he would have been trying to EQ out from the steady state response; we know this because we know he doesn't believe in EQing the steady state response above the room's modal frequency range).

But it is in the mid and high frequencies that non-coincident-driver multiway loudspeakers (and not - unless by chance - real acoustical sources) lobe.

Not an argument, just some food for thought...

Interesting re: the Bech paper, thanks

EDIT: I now see in the Bech paper that, "The threshold of detection is suggested to be determined by the spectral changes in a dominant frequency range of 500 Hz - 2 kHz."

So perhaps Toole's anecdote and Bech's findings are in fact quite compatible.

 

[amirm]

Any chances of reviewing the Sierra 2-EX? It would be a competitor to the Adante/Carina so might be interesting. Ascend already has posted its own measurements but always good to get another data point to confirm: http://www.ascendacoustics.com/pages/products/speakers/SRM2EX/srm2exmeas.html

Send them a note and see if they will send in a sample for review. It is only good marketing for them if it produces the same results they are publishing.

 

[napilopez]

Yes, excellent points re: Toole and I recall this now also.

I think it's worth noting though that in the context of this anecdote he gives, he seems to be referring primarily to the first cancellation null, typically in the bass or low-midrange, and present not only in the context of speakers, but also in the context of any real acoustical source (that would be the one he would have been trying to EQ out from the steady state response).

But it is in the mid and high frequencies that non-coincident-driver multiway loudspeakers (and not - unless by chance - real acoustical sources) lobe.

Not an argument, just some food for thought...

Interesting re: the Bech paper, thanks

A very good point! One is an inherent lobing to basically all things that make sound, another is built into the speaker. Seems like something that merits more research

 

[andreasmaaan]

A very good point! One is an inherent lobing to basically all things that make sound, another is built into the speaker. Seems like something that merits more research

It actually seems like the Bech study partially addresses this indirectly (sorry for the repeat, this was a late edit in my previous post).

Bech says:

"The threshold of detection is suggested to be determined by the spectral changes in a dominant frequency range of 500 Hz - 2 kHz."

This is indeed outside of the room's modal range, so it may well be compatible with Toole's anecdotal finding that EQing the steady-state modal response is not perceptually useful.

And of course, it is roughly in the range of a typical non-coincident-driver loudspeaker's lobing.

 

[Erik]

Please Review Triangle I am about to buy Titus ez or comet series..

Triangle loudspeakers are bad.



The graphs are from AUDIO magazine.

 

[Daverz]

Send them a note and see if they will send in a sample for review. It is only good marketing for them if it produces the same results they are publishing.

Worth requesting samples of the S300 and S400 from Buchardt as well. ASR is worth at least as much buzz for them as all those Youtubers that did Buchardt videos.

 

[andreasmaaan]

@napilopez having now read the Bech study, re-read Toole, and pondered a little further, I thought I'd throw some further ideas in the mix.

Firstly, regarding the Fraunholfer research you quoted:

“Regarding the floor reflection, the audible influence by removing this with absorbers around the listener is negative— unnatural sounding. No normal room has an absorbent floor. The human brain seems to be used to this.”

This teaches that removing the floor reflection entirely is bad, but is agnostic as to whether or the extent to which the spectral balance of the first reflection contributes to perceived timbre.

In other words, it may be important to have the reflection present, but it may be equally important that it exhibit smooth response (at least above the room's transition frequency range) .

Secondly, some more detail regarding Bech. He writes:

"The results have confirmed...that the floor reflection will contribute on an individual basis to the timbre of a noise signal....The introduced filtering of the individual transfer functions has a significant effect for reflections where spectral changes occur in the mid and high frequency ranges."

He later specifies that he infers from the test results that the "mid and high frequency range" is approximately 500-2000Hz.

This would suggest that (according to Bech) a speaker's negative vertical polar response in the mid and high frequency ranges is of primary importance in terms of perceived timbre.

Thirdly, re-re-reading Toole, I'm now less sure about my assumption that he was interested primarily in modal frequencies. He doesn't really specify either way, but I think it could equally be presumed that he was referring to the super-modal range also. Perhaps @Floyd Toole you're around and could elaborate a little on this anecdote?