SVS Ultra Bookshelf Speaker Review

[thewas]

The only conflict is between the listening impression and the score - not the objective data in itself. Would some of you prefer Amir lies about his impression to suit te score? How would that help anyone along?

I personally would rather have no panthers at all (especially not when they are related to the subjective opinion and not on the measurements with predefined limits) and the subjective part clearly separated at a forum called Audio Science Review and not Amirs Subjective Review but this discussion keeps repeating itself every time...

 

[whazzup]

If he would say something like "you look kinda pale, you probably have an X disease, take some of those pills" instead of sending me to do proper tests to determine what I actually have, then no, I definitely wouldn't go to that doctor again.
And this scenario is pretty much exactly what you pulled here.

Since we're having some fun with analogies :

Technically, Amir's saying 'Yo speaker, here are your test results. You look kinda pale, but it's not showing up on your tests. I've already given you some pills (eq) but you didn't respond to them.'

He shook his head, sent off the speaker, and promptly shared the test results on a publicly accessible server. Now people are reading the test results. Some thanked him for sharing, while others are screwing him over for accusing the speaker for looking pale. Amir gets angry and retorts that he's trained to observe for paleness in speakers.

Now, did anyone think to ask whether the speaker gave consent for everyone to see his test results?

*sorry, no more analogies, I promise...

 

[NDC]

If he would say something like "you look kinda pale, you probably have an X disease, take some of those pills" instead of sending me to do proper tests to determine what I actually have, then no, I definitely wouldn't go to that doctor again.
And this scenario is pretty much exactly what you pulled here.

Your argument hinges on Amir’s subjective listening not lining up with the preference score. But you’ve ignored much of the technically knowledgable replies to your consternation around Amir’s impression of the SVS as bright.

Of particular note were the observations regarding Olive scores, preference ratings and how these are imperfect and may not align with subjective impressions some 20% of the time. They are not perfect predictors of preference.

The directivity of the speaker has also been pointed out as problematic. Suggest you go back and read the information dispassionately. You seem wilfully obtuse at this point which is understandable given the emotive tone of your posts and the replies. But it’s just not kosher, so to speak.

 

[jazzendapus]

You're confusing me with someone else, I didn't mention preference score in any of my posts here. My initial point was that this speaker getting 3/5 compared to 5/5 that M106 got doesn't make sense given their overall measurements.

 

[lashto]

so much talk, so little content.

A) Another of my points. He's one guy. With an opinion. Throwing out a whole science based on one guy's opinion is...

... the right thing to do. (sometimes)
Many wave the science flag here but conveniently 'forget' that a single exception is enough to prove a theory wrong or incomplete. A single trained listener that can (controllably and repeatably) hear differently is enough to overthrow an entire acoustics theory.

...
The Olive score is but one small aspect of research about speaker preference. Somehow it has become *the thing* on our forum. It should not be.
...

This! The whole Toole/Olive research is basically the speaker equivalent of the DAC/Amp FR measurement. No way that is enough to fully assert the quality of a speaker, yet many use Toole's research as some sort of complete speaker-bible. Proper FR and directivity are just the first two things that must be done right. May be the most important but still nowhere near the only or the last todos.
My first reaction when reading Toole's reasearch some years ago was: excellent work! And the second was: why wasn't this obvious? Generally, I think Toole's research is overrated (at least around here lately). It is excellent work and it was obviously needed but it is just a (small) first step.

To talk about this speaker and the measures vs ears "controversy". On short: there is probably none.

Here's what the 2000 years old harmonics theory says about this speaker:

Matches @amirm's ear impressions almost 100% and does not contradict Toole's research (actually has nothing to do with it). Here's a bit more from the same Harmonics theory (additional to the very broad strokes of text on the pic above):

• the below 500Hz area will not only sound "warm" but will also be heard as a bit louder, even at the exact same FR level. That is another thing the H2 does, it's usually described by ear as "bloom". The H3 has the ~opposite effect, it kinda sucks all the "bloom" out of the fundamental and makes it sound somewhat more "precise" and less loud. Combining the two effects: this speaker will sound somewhat recessed in the ~500-2500Hz range, even if it had a 100% flat FR (this can be tested by EQ-ing a flat FR).
• the 1000-2000Hz range will sound particularly bad because it also has a pretty strong resonance at ~1500Hz (might somehow be related to the H3 dominance in that area). The H3 doesn't sound (that) bad by itself, I would bet on the combined H3+resonance effects as the reason why @amirm used "sharp" and "painful" to describe female vocals.
• edit/added. Male voices should sound better than female ones. Particularly deep ones like Cohen and Cash. Might still have a 'funny' timbre because of the issues in the 500-2500Hz area but should not be as sharp/painful.
• in spite of the "nice" H2 profile, the >2500Hz range should also sound quite bad/harsh/wrong because of the multiple resonances.
• it'll be impossible to fix this speaker with EQ. Something that might partially work: use @pkane's Distort tool and add extra H2 between 500-2500Hz.

Not convinced?
Here's another measurements vs ears "controversy" that happened not so long ago: Infinity IL10 vs Revel M106. The spinoramas are almost identical, yet the Infinity got a strong ear-dislike and the Revel got a lot of ear-praise. Same explanation, look at the HD graphs: the Revel has very orderly Hs (way above average from my experience) and the Infinity HD spectrum is quite 'mixed' (not as bad as this SVS but still below average). That kind of 'wrong' HD spectrum will give a speaker an inconsistent and (potentially) unpleasant timbre.
That's the main thing HD does, it changes the timbre between warm/pleasant/musical and cold/harsh/analytical. @amirm used "worlds apart" to describe the sound of IL10 and M106, that might only mean "HD-spectrum apart".

On short again: @amirm's trained ears were (most probably) right and there is no controversy anywhere.
(the whole Toole FR/directivity research was not contradicted anywhere as it does not even apply).

 

[lashto]

@amirm , from your measurements, this speaker is almost distortion-free (for humans at least). Was that evident on listening?

As for the harshness, I also suspect either a directivity error or a bad time alignment at crossover. If it is a directivity error ar 3 or 4 kHz, a high shelf around 10 kHz could make things even drier and fatiguing IME. I have spent litterally hundreds of hours trying to EQ my Epos K3 which have the same kind of error and gave up. Every EQ was making it worse.

Or it may be none of those and just another example of 'bad' HD. Do you have HD measurements of that K3 speaker?

 

[Thomas savage]

Prizes for continuing off topic nonsense include luxury thread bans , having your posts deleted and last but not least a holiday from ASR to a forum of your choice.

 

[Alexanderc]

Yeah I've seen that. I personally haven't heard the Prime Pinnacles yet, but I've been reading that SVS has managed to tone down their tweeter some with the Pinnacles.

In comparison to the Prime Towers here, as you can see - this would sear your ear drums out. This is old Klipsch RF series harsh.

SVS needs to go back to the drawing board instead of rehashing their tweeter and crossover network for their speakers. It was a surprise to me that they even released the Pinnacle Towers in the first place without addressing the elephant in the room first.

I’m asking because I genuinely don’t understand (the way this thread has gone I feel like I have to say this isn’t sarcasm): are you talking about the off-axis hump around 3-4khz? Isn’t that just wide dispersion in that area? It looks like the Prime Tower listening window is flat from 1200hz to 10khz at close to it’s rated sensitivity, which doesn’t look eardrum searing. Again, I’m not saying it isn’t, just looking for tips on how to read these graphs more intelligently.

 

[MZKM]

I’m asking because I genuinely don’t understand (the way this thread has gone I feel like I have to say this isn’t sarcasm): are you talking about the off-axis hump around 3-4khz? Isn’t that just wide dispersion in that area? It looks like the Prime Tower listening window is flat from 1200hz to 10khz at close to it’s rated sensitivity, which doesn’t look eardrum searing. Again, I’m not saying it isn’t, just looking for tips on how to read these graphs more intelligently.

I don’t know if it’ll indeed “sear your eardrums”, but that is in the presence region and it is ~5dB above the rest of the surrounding frequency range, so it will sound bright.

 

[AnalogSteph]

Sweeping the full audible band at five different levels we get:

I am not trusting the above graph much as it seems to show distortion his higher than lower levels??? I did have to re-calibrate my system so maybe something went wrong there.

As someone already wrote earlier, we seem to be noise limited; it is most obvious in the treble. If you start at 106 dB, there is an obvious steep drop above 10 kHz reflecting H2 leaving the ~20 kHz measurement BW. It's still there at 101 dB, but noise is starting to come up beyond ~12k, and by 96 dB the steep drop is barely visible if at all, with noise floor coming up around 5 dB for each 6 dB step below that (close enough for me).

Why there is an appreciable processing noise floor after the measurement bandwidth limiting is a good question for someone more familiar with the algorithm used.

 

[Sonny1]

I personally would rather have no panthers at all (especially not when they are related to the subjective opinion and not on the measurements with predefined limits) and the subjective part clearly separated at a forum called Audio Science Review and not Amirs Subjective Review but this discussion keeps repeating itself every time...

I appreciate your viewpoint but, and this is only my personal opinion, I really enjoy Amir’s subjective comments. This speaker is a good example. The measurements are pretty good but it appears to have some serious flaws. Other people have noted the flaws and Amir backed them up with measurements.

If they measured perfectly and he could not back up his conclusions, my opinion would be different. The measurements support Amir’s conclusion and he was unable to EQ the speakers to sound good.

Ultimately, people have the option of disregarding the subjective comments or questioning the measurements or methodology. Amir knows what he’s doing and always backs up his statements with facts. I don’t know any other reviewer in the industry who is willing to discuss and justify his conclusions as Amir has done. He’s one of the few reviewers I trust. The format and info provided by this site is welcome and very useful to me.

 

[lashto]

@ribosradagast I see that you have this SVS speaker and did read my HD post above. Would you care to share some impressions and confirm or disprove the sound predictions I posted?

P.S:
never heard or even seen this speaker, everything I posted is based solely on the ASR measurements and (my understanding of) the Harmonics theory.

 

[napilopez]

so much talk, so little content.

... the right thing to do. (sometimes)
Many wave the science flag here but conveniently 'forget' that a single exception is enough to prove a theory wrong or incomplete. A single trained listener that can (controllably and repeatably) hear differently is enough to overthrow an entire acoustics theory.


This! The whole Toole/Olive research is basically the speaker equivalent of the DAC/Amp FR measurement. No way that is enough to fully assert the quality of a speaker, yet many use Toole's research as some sort of complete speaker-bible. Proper FR and directivity are just the first two things that must be done right. May be the most important but still nowhere near the only or the last todos.
My first reaction when reading Toole's reasearch some years ago was: excellent work! And the second was: why wasn't this obvious? Generally, I think Toole's research is overrated (at least around here lately). It is excellent work and it was obviously needed but it is just a (small) first step.

To talk about this speaker and the measures vs ears "controversy". On short: there is probably none.

Here's what the 2000 years old harmonics theory says about this speaker:
View attachment 76587
Matches @amirm's ear impressions almost 100% and does not contradict Toole's research (actually has nothing to do with it). Here's a bit more from the same Harmonics theory (additional to the very broad strokes of text on the pic above):

• the below 500Hz area will not only sound "warm" but will also be heard as a bit louder, even at the exact same FR level. That is another thing the H2 does, it's usually described by ear as "bloom". The H3 has the ~opposite effect, it kinda sucks all the "bloom" out of the fundamental and makes it sound somewhat more "precise" and less loud. Combining the two effects: this speaker will sound somewhat recessed in the ~500-2500Hz range, even if it had a 100% flat FR (this can be tested by EQ-ing a flat FR).
• the 1000-2000Hz range will sound particularly bad because it also has a pretty strong resonance at ~1500Hz (might somehow be related to the H3 dominance in that area). The H3 doesn't sound (that) bad by itself, I would bet on the combined H3+resonance effects as the reason why @amirm used "sharp" and "painful" to describe female vocals.
• edit/added. Male voices should sound better than female ones. Particularly deep ones like Cohen and Cash. Might still have a 'funny' timbre because of the issues in the 500-2500Hz area but should not be as sharp/painful.
• in spite of the "nice" H2 profile, the >2500Hz range should also sound quite bad/harsh/wrong because of the multiple resonances.
• it'll be impossible to fix this speaker with EQ. Something that might partially work: use @pkane's Distort tool and add extra H2 between 500-2500Hz.

Not convinced?
Here's another measurements vs ears "controversy" that happened not so long ago: Infinity IL10 vs Revel M106. The spinoramas are almost identical, yet the Infinity got a strong ear-dislike and the Revel got a lot of ear-praise. Same explanation, look at the HD graphs: the Revel has very orderly Hs (way above average from my experience) and the Infinity HD spectrum is quite 'mixed' (not as bad as this SVS but still below average). That kind of 'wrong' HD spectrum will give a speaker an inconsistent and (potentially) unpleasant timbre.
That's the main thing HD does, it changes the timbre between warm/pleasant/musical and cold/harsh/analytical. @amirm used "worlds apart" to describe the sound of IL10 and M106, that might only mean "HD-spectrum apart".

On short again: @amirm's trained ears were (most probably) right and there is no controversy anywhere.
(the whole Toole FR/directivity research was not contradicted anywhere as it does not even apply).

Do you have hearing thresholds for all these factors? I think you are assuming too much audibility of these different types of distortion (though again, Amir is trained at this, so he may be an outlier).

Edit: As mentioned in my reply in the infinity thread I still think differences in frequency response and directivity are more likely culprits.

An interesting comparative experiment would be to compare a top-performing speaker in both FR and distortion metrics and compare it using three settings:

1) A control with the speaker performing as intended
2) A setting which slightly alters the frequency response
3) A setting which slightly alters the distortion profile

The exact amounts of distortion and deviation from neutral FR for a relevant comparison would be difficult to determine. Perhaps the test could be designed to randomize the differences in each factor. It would be interesting to plot not just which differences are more audible, but which ones more adversely affect the listening experience.

 

[lashto]

Do you have hearing thresholds for all these factors? I think you are assuming too much audibility of these different types of distortion (though again, Amir is trained at this).

Here's one DBT. Proves that approx -75dB THD is audible. That HD was added as mostly H2 and H3 which should be the least audible form of HD. This speaker (and all others) has HD components in the -50dB range, i.e. much more audible.
(might be best to continue the talk in the IL10 research thread)

 

[napilopez]

Here's one DBT. Proves that approx -75dB THD is audible. That HD was added as mostly H2 and H3 which should be the least audible form of HD. This speaker (and all others) has HD components in the -50dB range, i.e. much more audible.
(might be best to continue the talk in the IL10 research thread)

Yeah, we can move it there =] I'm sure distortion is audible. Pejorative is another matter though.

 

[daftcombo]

Or it may be none of those and just another example of 'bad' HD. Do you have HD measurements of that K3 speaker?

Not here but at home. Will post when I come back from holiday.

For now I can only show the THD profile:
https://www.audiosciencereview.com/forum/index.php?threads/epos-speakers.9282/post-299448

 

[daftcombo]

so much talk, so little content.

... the right thing to do. (sometimes)
Many wave the science flag here but conveniently 'forget' that a single exception is enough to prove a theory wrong or incomplete. A single trained listener that can (controllably and repeatably) hear differently is enough to overthrow an entire acoustics theory.


This! The whole Toole/Olive research is basically the speaker equivalent of the DAC/Amp FR measurement. No way that is enough to fully assert the quality of a speaker, yet many use Toole's research as some sort of complete speaker-bible. Proper FR and directivity are just the first two things that must be done right. May be the most important but still nowhere near the only or the last todos.
My first reaction when reading Toole's reasearch some years ago was: excellent work! And the second was: why wasn't this obvious? Generally, I think Toole's research is overrated (at least around here lately). It is excellent work and it was obviously needed but it is just a (small) first step.

To talk about this speaker and the measures vs ears "controversy". On short: there is probably none.

Here's what the 2000 years old harmonics theory says about this speaker:
View attachment 76587
Matches @amirm's ear impressions almost 100% and does not contradict Toole's research (actually has nothing to do with it). Here's a bit more from the same Harmonics theory (additional to the very broad strokes of text on the pic above):

• the below 500Hz area will not only sound "warm" but will also be heard as a bit louder, even at the exact same FR level. That is another thing the H2 does, it's usually described by ear as "bloom". The H3 has the ~opposite effect, it kinda sucks all the "bloom" out of the fundamental and makes it sound somewhat more "precise" and less loud. Combining the two effects: this speaker will sound somewhat recessed in the ~500-2500Hz range, even if it had a 100% flat FR (this can be tested by EQ-ing a flat FR).
• the 1000-2000Hz range will sound particularly bad because it also has a pretty strong resonance at ~1500Hz (might somehow be related to the H3 dominance in that area). The H3 doesn't sound (that) bad by itself, I would bet on the combined H3+resonance effects as the reason why @amirm used "sharp" and "painful" to describe female vocals.
• edit/added. Male voices should sound better than female ones. Particularly deep ones like Cohen and Cash. Might still have a 'funny' timbre because of the issues in the 500-2500Hz area but should not be as sharp/painful.
• in spite of the "nice" H2 profile, the >2500Hz range should also sound quite bad/harsh/wrong because of the multiple resonances.
• it'll be impossible to fix this speaker with EQ. Something that might partially work: use @pkane's Distort tool and add extra H2 between 500-2500Hz.

Not convinced?
Here's another measurements vs ears "controversy" that happened not so long ago: Infinity IL10 vs Revel M106. The spinoramas are almost identical, yet the Infinity got a strong ear-dislike and the Revel got a lot of ear-praise. Same explanation, look at the HD graphs: the Revel has very orderly Hs (way above average from my experience) and the Infinity HD spectrum is quite 'mixed' (not as bad as this SVS but still below average). That kind of 'wrong' HD spectrum will give a speaker an inconsistent and (potentially) unpleasant timbre.
That's the main thing HD does, it changes the timbre between warm/pleasant/musical and cold/harsh/analytical. @amirm used "worlds apart" to describe the sound of IL10 and M106, that might only mean "HD-spectrum apart".

On short again: @amirm's trained ears were (most probably) right and there is no controversy anywhere.
(the whole Toole FR/directivity research was not contradicted anywhere as it does not even apply).

That wouldn't explain for instance why @amirm didn't like the Kef R3 though:
https://www.audiosciencereview.com/forum/index.php?threads/kef-r3-speaker-review.12021/

 

[lashto]

Not here but at home. Will post when I come back from holiday.
For now I can only show the THD profile:
https://www.audiosciencereview.com/forum/index.php?threads/epos-speakers.9282/post-299448

Thanks. I assume that the blue/red/green lines on the bottom only show the TotalHD, not the individual H2/H3/H4. That would be useless for any timbre/HD 'predictions'. Hope you have something more detailed at home.
If you find a detailed H2/H3/H4/H5 measurement, I can provide some blind/theoretical 'predictions' and then you compare with your live impressions. And if we get a match, we won the prize

That wouldn't explain for instance why @amirm didn't like the Kef R3 though:
https://www.audiosciencereview.com/forum/index.php?threads/kef-r3-speaker-review.12021/

may be but it's very hard to say. The Kef HD graph is unclear, the H2/H3/H4 use almost the same color and are all cramped at the bottom. Can't 'read' anything in there.
Maybe @amirm can give us a better R3 graph to play with?!

 

[daftcombo]

Thanks. I assume that the blue/red/green lines on the bottom only show the TotalHD, not the individual H2/H3/H4. That would be useless for any timbre/HD 'predictions'. Hope you have something more detailed at home.
If you find a detailed H2/H3/H4/H5 measurement, I can provide some blind/theoretical 'predictions' and then you compare with your live impressions. And if we get a match, we won the prize

Yes, l already measured that and have it somewhere. But if distortion is low, no matter if it's H2 or H3 right?

 

[richard12511]

Your argument hinges on Amir’s subjective listening not lining up with the preference score.

I don't agraee that the "measurements don't agree with the subjective listen" argument hinges entirely on the Olive preference score. I think that even looking just at the spinorama, there is still a disagreement that exists between objective and subjective. Going forward in this thread, I think we should probably just ignore the Olive score altogether and focus on what's different in the actual measurements.

I think that carefully comparing the measurements of this speaker against the measurements of the M106, M105 and M55XC can help to show why this speaker doesn't sound as good as those other 3(at least post EQ). Given that those other 3 speakers sounded excellent, there must be some attribute that they share that the SVS lacks. Find the variable contained in A, B, and C, that's missing from D.

Ideas(many of these already mentioned by others):

1. Directivity Error(bottom half of spinorama)
Many have pointed to the directivity error as being the cause of the brightness, but I don't think it's that easy. Both the M106 and the M105 have the same directivity error, and both of those speakers still sound excellent. Some have suggested that the directivity error is due to trying to integrate a 6.5" woofer to a dome with no waveguide, but why does the M106 and M105(which do have a waveguide, and smaller woofer) still show that directivity error? Is there more going on there that I don't understand? I guess that's a question for those who know more than me.

Here is the SVS side by side with the M106, with the directivity error circled. Looking just at this spinorama, the magnitude of error seems about the same, or maybe 1dB worse on the SVS.

So, just looking at the spinorama, I don't think it's obvious that directivity error should be the sole cause of harshness here, otherwise it would have also doomed the M106 and M105. I do think directivity is a contributing factor, though.

Note, the spinorama groups vertical and horizontal together, so maybe there's an error in one that's masked by the other? That's one of my ideas I show later.

2. Overall tonality(top half of spinorama)
I would say that these two speakers have very, very similar spins. To my eyes, the M106 looks slightly more neutral on axis, while the SVS seems to have slightly smoother ER and SP measurements. I certainly don't see anything here in the comparison to suggest that one sounds excellent, while the other sounds bad, and the M105 looks very similar. I don't think anything in the top half of the spin is obviously to blame for the bad sound the SVS exhibited. The M55XC is another good comparison here, as (to me) it looks to have an overall much less neutral response. If the top half of the spinorama is to blame, then I think it would doom all 4 speakers, especially the M55XC. In fact, the M55XC does have what I would call bad tonality, but it makes up for it by having the most consistent dispersion of the speakers here.

*Notice the lack of directivity error in the M55XC!

Having looked at both the bottom half and the top half of the spinorama, I honestly don't think the error shows itself here, though I do think the directivity error is part of it. A, B, and D all have excellent top halves. C is the only one with an error free(mostly) bottom half, and yet A, B, and C are the ones that sound good, subjectively.

Going beyond the spinorama,

3. Distortion - at first glance, this speaker seems to have pretty good, low distortion. It's very comparable to the M106 and better than the M105. However, as @lashto pointed out, the SVS's distortion is mainly H3. I definitely think he could be on to something here, as both the M106 and M105 have lower H3. I definitely think this could be a contributing factor(good eye), but it can't be solely responsible, as it would have doomed the M55XC, which has comparable H3 with worse every other type.

5. Horizontal Dispersion - Here's where the M55XC really shines, and where (imo) it makes up most of the ground it lost with tonality and distortion. This is also where I think the SVS commits a sin that none of the other speakers do. The difference I see is, the SVS is firing off considerably more excess horizontal energy in the 2.5kHz - 8kHz region than it is in the 800Hz - 2.5kHz. Not sure how big this issue is, but it's something I noticed. Could the excess horizontal energy in the upper mids and treble - combined with the lack of horizontal energy in the lower mids - be a contributing factor towards that "harsh" sound that Amir(and many people who've heard this speaker) heard? Comparison between SVS and M106 below. I've drawn a blue line to help illustrate what I'm talking about.

As you can see, the M106 still has a similar horizontal directivity glitch, but it looks less severe(imo). The M106 looks much closer to a consistent width beam. Why doesn't this show up clearly in the spinorama's directivity index? The directivity glitches seem very comparable in the spinorama, but much less so here. The M55XC is actually the star here. It has by far the most consistent beam.

Anyways, those are just some ideas. I think there's probably multiple things going on here that cause the objective measurements to not be as near perfect as they seem at first glance. Likely it's several small magnitude errors that add up to a sizable change in sound. Even adding all the measurement errors together, though, I still can't honestly say I would expect one speaker to sound excellent and one to sound bad, but I have a lot to learn I suppose.

I've found that I generally prefer keeping the first reflections, but I think this could be a speaker that might sound better by absorbing them. Amir was unable to tame the harshness with EQ, but I wonder if he might have been able to tame it by absorbing the first reflections. This speaker is firing off extra treble energy to its left and right.