Revel Salon2 vs Genelec 8351B - Blind Test Preparations

[watchnerd]

Are you SURE about that?

Finally, let's go back to Rega and KEF. If the overall ranking is based on the sum of Sound Quality and Spatial Quality (as is suggested by Klippel, whose paper is cited in Toole's book), again we see a change in order of speaker preference. In Mono the Rega's combined score clearly exceeds that of second-place KEF, but in Stereo the combined scores for Rega and KEF are TIED.

In the category of Sound Quality, there was no change of ranking within this sample size of three. I speculate that this data set is too small to reliably conclude that the order of speaker preference does not change from Mono to Stereo in this category, and it is too small to conclude that it sometimes does.

To my eyes, the data is more tightly clustered in stereo, too.

 

[Duke]

To my eyes, the data is more tightly clustered in stereo, too.

I hadn't noticed that, but I think you are right.

 

[March Audio]

Are you SURE about that?

I think I can show at least one change in the order of speaker preference, within the extremely limited data we have access to:

View attachment 79504

On the Rega, Sound Quality stays the same and Spatial Quality goes up in Stereo.

On the KEF, Sound Quality goes down very slightly in Stereo, look very closely at the relative height of its black block compared to the Rega's, though I doubt this is statistically significant. But now look at Spatial Quality: In Stereo, the KEF overtakes the Rega in Spatial Quality. So there we have an example of a change in the order of speaker preference.

On the Quad, the Sound Quality score improves greatly and the Spatial Quality score improves enormously in Stereo, though not enough to overtake either the Rega or the KEF. Would the Quad have passed up speakers whose Mono scores were lower than those of the Rega and KEF? Without more data, we don't know. I think it is possible, and I think you disagree, but I could be wrong on both counts.

Finally, let's go back to Rega and KEF. If the overall ranking is based on the sum of Sound Quality and Spatial Quality (as is suggested by Klippel, whose paper is cited in Toole's book), again we see a change in the order of speaker preference: In Mono the Rega's combined score clearly exceeds that of second-place KEF, but in Stereo the combined scores for Rega and KEF are TIED.

In the category of Sound Quality, there is no change in ranking from Mono to Stereo within this sample size of three. I speculate that this sample size is too small to reliably conclude that the order of speaker preference does not change from Mono to Stereo in this category, and too small to conclude that it sometimes does.

YES.

Overall rating, not just one aspect. As you mention you do need to be statistically significant.

Please watch the video.

Listen to the man himself tell you 09:30 "Every time, no exceptions, nothing special about stereo"

and dont forget that people are more discerning in mono, will be more critical and come out with more consistent/accurate judgements than in stereo.

 

[Blumlein 88]

To my eyes, the data is more tightly clustered in stereo, too.

Yes, because stereo makes us more forgiving of issues apparently. The judgement of quality is over a wider scale in mono, and less in stereo. Stereo is less discriminating.

 

[DJBonoBobo]

Like me, and Dr. Toole, and many others? Here is an old, old post of mine, take for what it's worth. This seems to be another "religious" topic where most are already convinced of their beliefs and no amount of argument will change that.

I use subs, and have for decades, for all the usual reasons:

- Very (perhaps extremely) few "large" speakers actually play well below 40 Hz let alone 20 Hz. They distort heavily when presented with large bass signals (which most are -- see Fletcher-Munson) and driving them down low robs headroom for higher frequencies and causes distortion well above the fundamental signal frequency (harmonic and nasty intermodulation). Subs typically enable the mains to operate with much lower distortion.

- Very rare is the room setup such that the best place for stereo imaging and soundfield is the best place for the subs (or deep bass drivers) to counter room modes and such. Having independent subs provides placement options to smooth the in-room response. It is almost impossible to counter a null without subs (typically must move the MLP or change the room's dimensions though there are purpose-built panels that can also work). This is one of the things that led me to subs despite having quite capable mains.

- Powered subs offload the main amplifiers of the need to provide deep bass energy, providing more headroom and cleaner sound from the amplifiers.

- Music (let alone action movies) often contains deep bass content even if it is not real obvious. Kick drums, tympani, organ, sure, but also piano hammer strikes, plucked strings, beat patterns from instruments playing together, etc. May not really notice when they are there but usually obvious when they are taken away. Having subs fill in the bottom octave or three can make a difference.

- Purpose-built subs can provide high output cleanly at relatively low cost. The amplifiers and drivers need only cover a fairly limited frequency range so have fewer constraints upon them than woofers in a full-range system.

I do prefer main speakers with fairly deep bass and always have. Crossovers are not brick walls so a fair amount of energy still comes from the mains an octave below the crossover frequency. Higher-order crossovers allow you to reduce the overlap, but I still like having the capability. I have never really understood the idea of running "passive" bi-amping as implemented by an AVR (sending full-range signals to multiple channels and letting the speaker's crossovers separate frequency bands -- wastes amplifier headroom and seems to me of little benefit). Nor do I agree with the "plus" setting putting subs and mains in parallel; again, my idea has always been to isolate the two for the reasons above.

My first sub was a DIY design using an Infinity IRS woofer with my own control box to provide the crossover and a servo circuit using the second voice coil of the woofer. I had a Hafler DH-220 around so also incorporated a circuit to bridge it for use as a subwoofer amp. It worked well and the -3 dB point was ~16 Hz. I now run four small (F12) Rythmik subs using a similar (but updated) servo design with my Revel Salon2's and am happy with the result.

FWIWFM/IME/IMO/my 0.000001 cent (microcent) - Don


Stereo subs:

I have gone back and forth on stereo subs over many years (since ~1979/1980 when I built my first sub) and ultimately decided it is not worth it. It limits placement and correction options, almost no stereo content exits at sub frequencies (remember a wavelength is >11 feet at 100 Hz, >22 feet at 50 Hz, just how much stereo separation can there be in a normal listening situation?), and the end result was always much better when I ran the subs mono and placed them optimally for best in-room bass response. If your crossover is so high and/or filter roll-off so low that your subs intrude into the lower midrange you might appreciate stereo but I have always rolled off well below the point at which I could localize the subs. For years I ran stereo subs but many tests blind and otherwise convinced me stereo subs are just an unnecessary hassle that actually reduced my system's performance and sound.

Thanks for your answer, but i didn´t mean to question why someone would use subs. I absolutely get the advatages. What i meant with my probably poorly translated (from german to english) question was: If you are already sure that you will use subs, would you buy a speaker like the Salon2? They seem to have much more than "fairly" deep bass. And i thought, maybe they are meant for people who don´t use subs?
The context was my interest in a comparison between "Genelec + sub" vs. "Salon2 without sub". Maybe this could be an interesting test with regard to the "quasi-religious" controversity you mentioned?

 

[ctrl]

Anyone can test it out. By your logic a pair of speakers that sound good in stereo will sound dull if used in mono. Anyone can test this out for themselves at home.

Several effects come together. Because the loudspeaker stands freely in the room, the influence of lateral reflections is greatly reduced - with corresponding consequences if the loudspeaker (and the lateral walls) does not show an "optimal" radiation pattern.

My experience is that if a free-standing loudspeaker, listened to in mono, is adjusted to the best possible transparency of the sound, then it sounds too bright when listened to in stereo.
I'm not a sound engineer, but what happens to a hi-hat pinned extremely left in the stereo mix, listened to in mono, compared to the entire signal?

But as already mentioned, everyone can easily test this for themselves and then draw their own conclusions.

They have analysed people's preferences and found that people like a certain sound. This happens to be a flat anechoic on axis response with smooth off axis response. The algorithm correlates the objective measurements to the observed preferences to allow prediction of what speakers people will like.

In order to understand my criticism of the Harman/Olive results and the score, we have to go into more detail.

What did Toole find when comparing speaker ratings for mono vs. stereo. This was already linked in this thread, I would like to show it here again for the sake of completeness:

Just by the way (I don't want to go into this in more detail here), for me this evaluation shows that the ranking can easily change (for loudspeakers rated similarly in mono) when switching from mono to stereo.
But don't get me wrong, I agree with the general statements made by Toole.

UPDATE: Sorry, I completely overlooked in the long preparation for this post that @Duke has already gone into this aspect in detail.

What is usually not mentioned when these results are cited? How did they come about.
Let us take a look at the test setup used by Toole:

This certainly corresponds quite well to a "typical" listening room and the placement of the speakers should also be realistic.
I have nothing to complain about and the results should be very well transferable to typical listening situations.

How was the monophonic hearing test performed? Toole writes about this in his book
Source: Sound Reproduction

The experiment was done in stereo and also in monophonic form, using only the left loudspeaker.

So the left speaker was heard as shown in the pictures above.


Now let's look at how Harman/Olive conducted their hearing tests - what was the setup there?

Do any of you notice anything different?
For the evaluation, the loudspeakers were listened to free-standing in an acoustically treated room, far away from lateral boundaries.

Thus, the results and conclusions of Toole's experiment are not transferable at all, since the influence from the side walls is almost completely missing and thus also all additional spatial information associated with it.

Therefore I say that Harman/Olives results and developed score are not wrong, but are only valid for free standing loudspeakers in mono.

Mono vs stereo. I think this is going to be a very hot topic. There's most likely gonna be lots of debate back in forth, and no matter what you choose, half of the people are gonna say your test is useless . Do what you think is most logical.

I don't think anyone in their right mind would call ABX testing in stereo "useless" - perhaps "unnecessary", since ABX testing in mono, "done right" (as Toole described it - see above), should produce almost identical results.
If two loudspeakers are listened to in mono, almost equally rated, each will probably compare the loudspeakers in a stereo setting as "ultima ratio" and then decide.

 

[aac]

Was there any kind of proof that preference on selected tracks guarantees preference on a wastly larger number of tracks (ideally the majority of tracks existing, but it is not possible, so at least maybe some selection of the most popular ones, like top of the charts)?
to me it looks like the most important thing for these experiments to be really meaniningful.

 

[farcurse]

My 2 cents:

Sorry for the late reply, but I strongly recommend that you do go forward with this listening test. I think that your proposed methodology (arrange speakers so that they sound similar in sighted listending, EQ to remove low frequency room modes, etc.) makes complete sense. I feel that the debate about whether this is "scientific" or not misses the point. In a Bayesian sense, I'm pretty sure that we'll have more data coming out of the experiment than going in. This to me is what we're after, and we get value even if the sample size is too small to provide 95% null hypothesis rejection, or something like that. As for EQ, you can't win in either case. Either you provide no EQ, are at the mercy of your particular room, and can't generalize to any other listening spaces, or you do, and are tied to a particular EQ methodology. I think that the latter is better because there are fewer EQ methodologies than different room shapes - if you just EQ so that bass is flat that would correspond to what many listeners here would do anyway.

Please do the test!

 

[March Audio]

Several effects come together. Because the loudspeaker stands freely in the room, the influence of lateral reflections is greatly reduced - with corresponding consequences if the loudspeaker (and the lateral walls) does not show an "optimal" radiation pattern.

My experience is that if a free-standing loudspeaker, listened to in mono, is adjusted to the best possible transparency of the sound, then it sounds too bright when listened to in stereo.
I'm not a sound engineer, but what happens to a hi-hat pinned extremely left in the stereo mix, listened to in mono, compared to the entire signal?

But as already mentioned, everyone can easily test this for themselves and then draw their own conclusions.



In order to understand my criticism of the Harman/Olive results and the score, we have to go into more detail.

What did Toole find when comparing speaker ratings for mono vs. stereo. This was already linked in this thread, I would like to show it here again for the sake of completeness:
View attachment 79520
Just by the way (I don't want to go into this in more detail here), for me this evaluation shows that the ranking can easily change (for loudspeakers rated similarly in mono) when switching from mono to stereo.
But don't get me wrong, I agree with the general statements made by Toole.

UPDATE: Sorry, I completely overlooked in the long preparation for this post that @Duke has already gone into this aspect in detail.

What is usually not mentioned when these results are cited? How did they come about.
Let us take a look at the test setup used by Toole:

View attachment 79526
View attachment 79525

This certainly corresponds quite well to a "typical" listening room and the placement of the speakers should also be realistic.
I have nothing to complain about and the results should be very well transferable to typical listening situations.

How was the monophonic hearing test performed? Toole writes about this in his book
Source: Sound Reproduction

So the left speaker was heard as shown in the pictures above.


Now let's look at how Harman/Olive conducted their hearing tests - what was the setup there?

View attachment 79527
View attachment 79528

Do any of you notice anything different?
For the evaluation, the loudspeakers were listened to free-standing in an acoustically treated room, far away from lateral boundaries.

Thus, the results and conclusions of Toole's experiment are not transferable at all, since the influence from the side walls is almost completely missing and thus also all additional spatial information associated with it.

Therefore I say that Harman/Olives results and developed score are not wrong, but are only valid for free standing loudspeakers in mono.




I don't think anyone in their right mind would call ABX testing in stereo "useless" - perhaps "unnecessary", since ABX testing in mono, "done right" (as Toole described it - see above), should produce almost identical results.
If two loudspeakers are listened to in mono, almost equally rated, each will probably compare the loudspeakers in a stereo setting as "ultima ratio" and then decide.

No.

Simple fact is that they have tested the same speakers in mono and stereo configurations on multiple occasions. To quote Toole "without exception" the same speakers were preferred in both scenerios. This is the case regardless of what you are saying the speaker/room position configurations were.

I'm not really interested in arguing with forum members about this. It's a bit tedious TBH. Your opinion simply doesnt carry weight against that of Toole et Al.

And no I haven't experienced your findings when "tuning" speakers because I wouldn't do it that way. I would (and have) developed speakers to have a flat anechoic on axis response (smooth off axis) which sound fine in both mono and stereo.

Your comments actually don't make much sense to me. A free standing mono speaker is likely to have less bass, ie a brighter overall balance than a stereo pair positioned traditionally nearer walls where boundary enforcement takes effect. This is contrary to what you suggests happens.

Whilst the room boundary absorptive / reflective properties have influence at higher frequencies, its the on axis response (anechoic) that is of primary importance in terms of the basic tonal balance

 

[preload]

I'm still digesting the discussion, but wanted to point out in the meantime that if we're using the Toole paper from the 1980's to draw conclusions about the validity of mono testing, there are two big time problems I see (at least at first glance)
- only 3 speakers were tested
- the speakers were early 1980's technology
I'm having some hesitation about drawing conclusions from that, considering how much loudspeaker technology has improved over 35 years, particularly in how stereo speakers can create a realistic soundstage.
I want to also mention, as others have, that we don't listen to mono speakers ordinarily. So what is of ultimate interest to us is how speakers sound as a stereo pair, in an optimal room position. The imaging and soundstage matter. And I think @echopraxia confirmed this when he concluded that even though the genelec sounded noticeably better technically, when he factored in soundstage in stereo, he overall preferred the revel.

 

[richard12511]

I'm still digesting the discussion, but wanted to point out in the meantime that if we're using the Toole paper from the 1980's to draw conclusions about the validity of mono testing, there are two big time problems I see (at least at first glance)
- only 3 speakers were tested
- the speakers were early 1980's technology
I'm having some hesitation about drawing conclusions from that, considering how much loudspeaker technology has improved over 35 years, particularly in how stereo speakers can create a realistic soundstage.
I want to also mention, as others have, that we don't listen to mono speakers ordinarily. So what is of ultimate interest to us is how speakers sound as a stereo pair, in an optimal room position. The imaging and soundstage matter. And I think @echopraxia confirmed this when he concluded that even though the genelec sounded noticeably better technically, when he factored in soundstage in stereo, he overall preferred the revel.

I think most here are assuming that they've continued to test the theory since then. Unfortunately(afaik), that data is still private.

 

[watchnerd]

I want to also mention, as others have, that we don't listen to mono speakers ordinarily.

But I listen to mono content on 2 speakers * a lot*.

Like probably 20% of my listening is mono.

How does "double mono" change things versus stereo?

 

[echopraxia]

But I listen to mono content on 2 speakers * a lot*.

Like probably 20% of my listening is mono.

How does "double mono" change things versus stereo?

Oh BTW this reminds me of another interesting subjective impression from comparing these speakers.

If I listen to mono content or configure the miniDSP to play mono through both speakers, on the Revel Salon2 it sounds exactly as if the sound is coming from a phantom center source in front of me. There is almost no sense that the phantom center image is not real, or localized to the left/right speakers.

In the Genelecs, the phantom center experience is not nearly as realistic. I’m not sure why, but I think it is related to what my ears hear as I turn my head left and right slightly. If I turn my head to the left or right, it becomes even more obvious that the Genelecs sound more localized to the speakers, while the Revel Salon2 phantom center image remains very impressively anchored in place.

 

[watchnerd]

Oh BTW this reminds me of another interesting subjective impression from comparing these speakers.

If I listen to mono content or configure the miniDSP to play mono through both speakers, on the Revel Salon2 it sounds exactly as if the sound is coming from a phantom center source in front of me. There is almost no sense that the phantom center image is not real, or localized to the left/right speakers.

In the Genelecs, the phantom center experience is not nearly as realistic. I’m not sure why, but I think it is related to what my ears hear as I turn my head left and right slightly. If I turn my head to the left or right, it becomes even more obvious that the Genelecs sound more localized to the speakers, while the Revel Salon2 phantom center image remains very impressively anchored in place.

Only if I keep my head in a virtual vice and facing perfectly directly ahead does the phantom center image of the Genelec match the Salon2’s.

That's weird.

My Contour 20 behaves just like how you describe the Salon 2. It's spooky, even. I have to put my ears right to the speaker to localize the sound.

Subwoofer integration issue, maybe?

 

[echopraxia]

Only if I keep my head in a virtual vice and facing perfectly directly ahead does the phantom center image of the Genelec match the Salon2’s. The moment my head turns, the illusion of the Genelec crumbles far sooner than the illusion from the Salon2. I have to have my head turned almost 90 degrees to the side before the phantom center image falls apart from the Salon2, while for the Genelec it’s more like 45-60 degrees. But it’s not abrupt, it’s a continuous degradation of the illusion.

These numbers also are suspicuously in alignment with the dispersion patterns of these speakers. Not sure if that’s just coincidence or not.

 

[echopraxia]

That's weird.

My Contour 20 behaves just like how you describe the Salon 2.

Subwoofer integration issue, maybe?

Same exact result with no subwoofer. It’s definitely not related to bass, and works even for acoustic music with very little bass content.

This is very interesting. I wonder what would explain this.

 

[stunta]

I agree with the folks who suggested you go ahead with it as a best effort. Worst case, there could be some useful learnings for everyone with this exercise. Best case, it is a good experiment that works well enough for domestic environments. Wish you all the best.

 

[echopraxia]

I keep coming back to this phantom center image test because it is such a stark and audible difference, as well. I wonder if this might make a candidate for part of a blind test! Also, it’s so quick and easy to do we could very reasonably physically blindfold people for the short amount of time it takes.

 

[echopraxia]

I agree with the folks who suggested you go ahead with it as a best effort. Worst case, there could be some useful learnings for everyone with this exercise. Best case, it is a good experiment that works well enough for domestic environments. Wish you all the best.

I will, but I am going to give it some time so I don’t feel rushed. I was planning to do this soon but instead I will wait and consider various options, including even maybe building a rotating turn table and visually opaque screen/curtain etc.

My reasoning for this is simple: I’m already pretty satisfied with what I’ve personally learned from my sighted comparisons this weekend. Therefore, a blind test at this point would be for the benefit of others, not me. I do want to do this, but as a result I also want to make sure I do this right. I know I will never please everyone, but for a test of speakers at this level of near perfection, and given the amount of work it will already require to host a blind test, I will want to go as far as I possibly can in terms of rigorous controls to ensure the test is truly blind. Maybe even double blind.

 

[watchnerd]

I keep coming back to this phantom center image test because it is such a stark and audible difference, as well. I wonder if this might make a candidate for part of a blind test! Also, it’s so quick and easy to do we could very reasonably physically blindfold people for the short amount of time it takes.

I would definitely include that as a test.