Request for help understanding Nubert Nuvero Nova measurements

[Leopardy]

Hello everyone! I followed the site for quite some time but I am very new to actually trying to understand the "details" of speaker measurements. I just watched amir's video yesterday read a few things and wanted to test my new knowledge with some slightly different looking graphs. I tried to find some on Nubert's new top line (Nuvero Nova) of passive speakers. The book shelf speaker amir reviewed a while ago, is a tier below that now. I found graphs of the stand mount Nuvero Nova 14 (134cm tower speaker for roughly 6000€ a pair), in a magazine which brings me to my questions.

The first thing I noticed is that the first graph covers an absolutely huge frequency range while being quite small in size. I assume that means less details are visible. Does this already make it unreliable or is this still detailed enough to get a good idea of a speakers performance? Is what we see at 500 Hz a "directivity error"? The directivity still looks pretty good to me. I suspect that the format of this graph helps it look flat a lot.

The next thing I noticed is the little "Clio" in the edge. I assume that Clio also produces some near field measurement system or some kind of different trickery to get anechoic measurements without the actual thing. Not sure though... EDIT: these are actually measurements in a "certified room with low reflections".

I think the third (EDIT second not third) graph shows the effects of different settings of the speaker. The standard setting seems to be very bright. The grey response looks better to me.

The distortion graph is a little confusing to me again. The German text says that the distortion measurement at 95 db shows how confident the speaker handles distortion since it is basically identical to the distortion measurement at 95 db. What is that supposed to tell me?? I assume there is some kind of typo involved here. The way this is framed makes me believe the lower graph is 95 db and the blue one something higher. That would make these distortion measurements absolutely excellent, right?

I hope this is the correct space for a post like this.

Greetings
Jakob

 

[Leopardy]

I found a graph for the even bigger (four way!) nuvero nova 18. Aren't these insanely good? I feel like I am somehow lead astray here

Red is on axis
Green 30° horizontal
Blue 45 ° horizontal
Grey 60 ° horizontal

 

[Doenerkunde]

A few general points about measurements by Hifi magazines:

Anechoic rooms are not neccessarily accurate across the entire audio spectrum. For anechoic rooms to be accurate down to 20hz they would have to be gigantic and isolated extremely well. For example the hifi magazine Soundstage has a cooperation with Canadas National Research Council and can utilize their anechoic room, that is accurate down to 100hz.

One of the best "non Klippel NFS" sources for measurements: https://www.soundstagenetwork.com/index.php?option=com_content&view=article&id=16&Itemid=18

But most Hifi magazines don´t have access to an anechoic room like the one of the NRC. So the magazines have to build „somewhat anechoic“ measurements themselves. Most often they do this by „splicing“ the measurements: For mids and highs they typically measure like withing 1m distance. Sometimes they „gate“ their measurements, meaning the mic only picks up a few miliseconds and everything that comes after that gets cut off. This gets rid of room reflections, but also lowers the resolution of the measurements. For the low frequencies they measure the bass drivers and port within a few cms distance. Sometimes they also use „ground plane measurements“ for mid bass, meaning they place the speaker and mic close to the floor. In the end they take their seperate measurements for the different frequency bands, slice them up, put them together and build a „full range frequency response“. Every magazine has their own way of doing this. Another aspect is, that they employ variing degrees of smoothing to their measurements. Notice the "1/3 octave smoothing" on the Nuvero Nova 18 measurements? This is pretty heavy smoothing and will potentially mask "small ripples" (eg resonances) in the frequency responses.

So the problem with hifi magazine measurements is, that they all use different means to achieve their frequency responses and not a standartised protocoll across all publications. This is where the "CEA2034" you see above Amirs NFS measurements comes into play, because it defines and sets a standard:
https://www.audiosciencereview.com/forum/index.php?attachments/ansi-cta-2034-a-pdf.45978/

Some examples:

Nubert Nuvero 60 measured by Stereoplay:

Nubert Nuvero 60 measured by Amir:

They are a pretty good fit, but notice how the Stereoplay measurement looks much smoother, because they employed heavier smoothing. This masks the midrange ripples visible in Amirs NFS measurements.

Philharmonic Audio BMR Monitor measured by Stereophile:

Source: https://www.stereophile.com/content/philharmonic-audio-bmr-monitor-loudspeaker-measurements

Philharmonic BMR Monitor measured by Erin:

Source: https://www.erinsaudiocorner.com/loudspeakers/philharmonic_bmr_v2/


Notice the difference in the bass region. Stereophile is known for publishing measurements with somewhat exaggerated bass, because of the way they handle, splice and stitch together their different measurements.

 

[AnalogSteph]

I think the third graph shows the effects of different settings of the speaker. The standard setting seems to be very bright. The grey response looks better to me.

You mean the second? Agreed on all points then. Although I wouldn't call a treble boost of +3dB "very bright" yet. Dispersion seems to look good, it's about time Nubert caught on.

The German text says that the distortion measurement at 95 db shows how confident the speaker handles distortion since it is basically identical to the distortion measurement at 95 db. What is that supposed to tell me?? I assume there is some kind of typo involved here.

For sure. I presume they meant to say 85 dB. And I'd guess blue = H2, red = H3 (at least dome tweeters generally are dominant H2 affairs).

I found a graph for the even bigger (four way!) nuvero nova 18. Aren't these insanely good? I feel like I am somehow lead astray here

Red is on axis
Green 30° horizontal
Blue 45 ° horizontal
Grey 60 ° horizontal

I think those labels are incorrect, going by treble dropoff I'd say green is on axis and red is 30°.

As mentioned, these graphs are quite heavily smoothed, but aside from a slight irregularity around 6 kHz they look promising enough.

 

[Arindal]

Is what we see at 500 Hz a "directivity error"? The directivity still looks pretty good to me.

Directivity cannot be judged by solely looking at a limited number of graphs from a very limited measuring window, in this case 3 graphs maximum 30deg horizontally. To identify a directivity error or estimate the speaker´s behavior in a room, you at least need a bunch of response graphs covering angles like 75, 90 or 120 deg both horizontally and vertically. Or even better, measurements from the complete sphere to 180deg, a directivity index calculation (like the ´Sound power DI´ graph mentioned above), spinorama, polar plot or alike.

The 30deg horizontal graphs indicates a broad radiation hence wide listening window, that's all. In case of a speaker concept like this, with several midrange/midwoofers, particularly vertical angles of 40...100deg are of interest, as typical outcome of vertical lobing, cancellation between the drivers, can be seen in this window.

The standard setting seems to be very bright. The grey response looks better to me.

I have heard some comments on YT that this increase in SPL towards higher treble represents the idea of compensating typical diffuse soundfield behavior in a room. Due to both increasing absorption and directivity index towards higher frequencies, overall in-room response typical decreases. Personally I find it implausible to compensate this by increasing level on axis, particularly if we are talking about the frequency band 5-8K, in which the human brain can distinguish direct from indirect sound, and on the other hand a single tweeter on a convex baffle is expected to show pretty broad radiation pattern hence low directivity index.

 

[Leopardy]

You mean the second?

Yes. I put in an edit.

Thank you all for the long and helpful responses! I will take time to chew through it in detail tomorrow!

 

[Keith_W]

You have to shift gears a bit when you try to interpret a graph like this, which is a measurement of a speaker at different off-axis angles: 0, 30, 45, 60deg respectively.

If the measurements are bunched tightly together, it means that the 60deg measurement is the same as on-axis. The speaker has wider directivity at these frequencies. If they included a 180deg measurement, and if that was the same as the 0deg (on-axis), then you would say that the speaker is omnidirectional at this frequency.

If the measurements start to diverge, as in this case above 200Hz and especially above 5kHz, it means that the directivity is increasing / the radiation arc is getting more narrow. If you were to toe in your speakers so that you listen to it on-axis, you would hear the green curve. If you toed them out so that you are listening to it at 30deg, you would hear the red curve. Or if you sit off-centre, you would hear the green curve from one speaker, and the red curve from the other.

It is much better to do a standard CEA 2034 Spinorama measurement that Amir and Erin do, but this requires a very expensive measurement setup (>$100k) and there are no audio magazines on this planet who own one. It's so expensive that most speaker manufacturers (nearly all!) don't own one. There are workarounds to obtain these measurements, but they are difficult to do and require time and skill. Magazine publishers usually don't bother - the general public don't know how to read these measurements anyway.

 

[kharan]

This is mostly likely a very good speaker.

Looks like seriously built. Big surrounds on the woofers suggest long excursion, driver size, spacing and crossover regions make sense, big round overs to control diffraction. Now the question is whether they yolo'ed it or if they properly simulated everything. Likely the second.

 

[Leopardy]

I think those labels are incorrect, going by treble dropoff I'd say green is on axis and red is 30°.

I assume you are right. I double checked the video I pulled this from and it says red is on axis. I asked for clarification in the comments. The guy on the right is the designer. It's a pretty interesting interview if you take everything with a pile of salt; german though.

The video also says that the graph is from Thomas Bien who is the designer. I assume that the three graphs out of the magazine are company measurements too then.

 

[chips666]

This is mostly likely a very good speaker.

Looks very promising indeed
I just need to find 6000 euro...
Enjoy...

 

[Leopardy]

Enjoy...

I wish No space for huge speakers like this atm. Even if, I am not sure I would get this over the KEF R11 Meta.

 

[That_guy]

Measure well but god they are awful looking.

 

[Arindal]

Measure well

It is really surprising to me that these speaker meet critical acclaim here.

Without having heard them, I would assume that a speaker showing decreasing directivity index in the treble region compared to midrange while having increasing on-axis response in this band as well, would result in a very different tonality than what is usually accepted as ideal here. Take recent KEF models as an opposite example which combine narrowing directivity and significantly (some models like Meta Q11) decreasing on-axis level.

I would expect these two extremes to sound very very different in a room, to a degree I would assume that not both at the same time can be neutral, natural and ´right´ in scientific terms. Maybe someone can explain this, or has heard both.

 

[Doenerkunde]

It is really surprising to me that these speaker meet critical acclaim here.

Without having heard them, I would assume that a speaker showing decreasing directivity index in the treble region compared to midrange while having increasing on-axis response in this band as well, would result in a very different tonality than what is usually accepted as ideal here. Take recent KEF models as an opposite example which combine narrowing directivity and significantly (some models like Meta Q11) decreasing on-axis level.

I would expect these two extremes to sound very very different in a room, to a degree I would assume that not both at the same time can be neutral, natural and ´right´ in scientific terms. Maybe someone can explain this, or has heard both.

The on axis rise in treble can be mitigated via switches on the terminal.

See the grey curve in treble with the switch set to "soft" for the Nova 14:

Whats really impressive about them is their low frequency extension. The -3dB point of the 6k/pr Nova 14 extends into the high 20s. The 12k/pr Nova 18 is basically linear down to the very limit of our hearing.

So while they don't strictly meet the Harman criteria of steadily increasing directivity in the top two octaves, I think it's still appropriate to call them "well" measuring.

 

[waofner]

View attachment 502282

You have to shift gears a bit when you try to interpret a graph like this, which is a measurement of a speaker at different off-axis angles: 0, 30, 45, 60deg respectively.

If the measurements are bunched tightly together, it means that the 60deg measurement is the same as on-axis. The speaker has wider directivity at these frequencies. If they included a 180deg measurement, and if that was the same as the 0deg (on-axis), then you would say that the speaker is omnidirectional at this frequency.

If the measurements start to diverge, as in this case above 200Hz and especially above 5kHz, it means that the directivity is increasing / the radiation arc is getting more narrow. If you were to toe in your speakers so that you listen to it on-axis, you would hear the green curve. If you toed them out so that you are listening to it at 30deg, you would hear the red curve. Or if you sit off-centre, you would hear the green curve from one speaker, and the red curve from the other.

It is much better to do a standard CEA 2034 Spinorama measurement that Amir and Erin do, but this requires a very expensive measurement setup (>$100k) and there are no audio magazines on this planet who own one. It's so expensive that most speaker manufacturers (nearly all!) don't own one. There are workarounds to obtain these measurements, but they are difficult to do and require time and skill. Magazine publishers usually don't bother - the general public don't know how to read these measurements anyway.

only if you are listening to near field! especially above 10 kHz the frequency responce decreases with listening distance to flat or even lack of high frequency.

 

[waofner]

I do not understand how it is possible that, at least some of the most praised klippel measured speakers, get this good evaluations when these speakers are not able to repoduce the frequencies above 15 kHz correctly at listening place.
simple test: 20€ Truthear GATE IEM - for direct comparison.

 

[Keith_W]

I do not understand how it is possible that, at least some of the most praised klippel measured speakers, get this good evaluations when these speakers are not able to repoduce the frequencies above 15 kHz correctly at listening place.

I have a different take on it. It's not ideal, but most of us won't notice. Only a microphone will hear a 2dB lift above 15kHz. A teenager or young adult MAY be able to detect it, but since there is so little musical information up there, I doubt if even they would notice. The rest of us on ASR, who are mostly in our 50's or older, have no chance of hearing it. And don't forget, the ERB is pretty wide up there. It's 1644Hz wide at 15kHz according to the formula ERB=0.108f+24.7.

The bottom octave is FAR more important. Almost everybody can hear it, even people with hearing loss (which affects high frequencies in most people). And look at that - this speaker is flat down to 20Hz.

My take: it's not perfect, but it's probably going to be fine. There are much worse speakers out there for sure.

 

[waofner]

I have a different take on it. It's not ideal, but most of us won't notice. Only a microphone will hear a 2dB lift above 15kHz. A teenager or young adult MAY be able to detect it, but since there is so little musical information up there, I doubt if even they would notice. The rest of us on ASR, who are mostly in our 50's or older, have no chance of hearing it. And don't forget, the ERB is pretty wide up there. It's 1644Hz wide at 15kHz according to the formula ERB=0.108f+24.7.

The bottom octave is FAR more important. Almost everybody can hear it, even people with hearing loss (which affects high frequencies in most people). And look at that - this speaker is flat down to 20Hz.

My take: it's not perfect, but it's probably going to be fine. There are much worse speakers out there for sure.

I am in my 50's , the thing for me is, there are recording's which really need the 15kHz to top end to get the "image" of the room.
one good example: Claudia Koreck "Gute Nacht Freunde"(Qobuz) sorry , i am not aware of a good english spoken example right now.

listen to it on decent IEM's or at least good headphones and compare to played on speaker's- that's why i find my Linton's in best case "looks fantastic- but only good enough for background music" and my ADAM S3V's (although i really hate their look's in my living room) the best i ever heard- and measured.

 

[waofner]

measurement ADAM s3v listening place

 

[Arindal]

So while they don't strictly meet the Harman criteria of steadily increasing directivity in the top two octaves, I think it's still appropriate to call them "well" measuring.

My assumption would rather be we have a rare case of decreasing directivity between 2K and 6K as it is typical for a d´ Apollito style concept with very compact midrange drivers (even more so equipped with a convex baffle, I assume). So even with reduced treble EQ jumper setting, I would expect treble in the room to rise in amplitude. Would not call this ´not strictly meeting´ but rather ´being a completely opposed concept´ to Harman philosophy in general and certain lauded products in particular.

Don´t get me wrong, I would not support either of these extreme concepts but rather something in between, but my guess would be these two are very very far from each other and in a real-world listening room, at least one would be perceived as tonally imbalanced (or rather both in opposite directions).

get this good evaluations when these speakers are not able to repoduce the frequencies above 15 kHz correctly at listening place.

Would not overly worry about frequencies above 15K, but rather focus on the 3-8K band. A small tweeter in a convex baffle tends to exhibit a very broad directivity pattern, opposed to a 2-driver d´Apollito-style midrange array for the octaves below.

Whats really impressive about them is their low frequency extension. The -3dB point of the 6k/pr Nova 14 extends into the high 20s.

Certainly true, but I would not consider this to be that problematic given it has a vented system, the number of designated subwoofers and sheer enclosure volume.