Revel Salon2 vs Genelec 8351B - Blind Test Preparations

[echopraxia]

It seems there is nothing remotely approaching consensus here on what is even an “acceptable” test procedure. At current rate, it looks like it may take a few decades or so for consensus to be reached here

Therefore, I will be canceling this test, until it’s understood what would be required in order for a victory of one speaker over the other to be accepted as valid (even if only as a weak data point with small sample size). In the current context, the significant amount of work required to make this happen is just not worth it as a weekend project.

Instead, in the near term, I am currently doing sighted tests for my own benefit. The guests that would have blind tested these will instead enjoy a more casual enjoyment-focused listening session from the winning speaker

But please by all means, let the debate continue! If this discussion leads to a conclusion and consensus on test procedure that would be accepted as a reasonable compromise vs the full fledged multi-million dollar Harman test facility, I will reboot the effort and we can start over. (However I’m skeptical this will happen.)

If there is no consensus, I would suggest some of you consider the irony that the only possible source of truth for your blind speaker tests is coincidentally restricted to only those produced by Harman (which is therefore non-reproducible, and un-falsifiable).

 

[echopraxia]

In order to remove the influence of differences in bass capability, did you consider just doing the abx with same subwoofer for both the genelec and the revel and crossing them over at, say, 80 hz? I'm assuming you don't want to penalize the genelec for not being a floorstander. Then you wouldn't need to worry about eq (at least eq in the bass region) since it would be the same sub in both comparators.

That was the plan except I see no reason not to EQ the sub to the room. As you can see, it’s quite lumpy and sounds very boomy prior to correction.

 

[DonH56]

It seems there is nothing remotely approaching consensus here on what is even an “acceptable” test procedure. At current rate, it looks like it may take a few decades or so for consensus to be reached here

Therefore, I will be canceling this test, until it’s understood what would be required in order for a victory of one speaker over the other to be accepted as valid (even if only as a weak data point with small sample size). In the current context, the significant amount of work required to make this happen is just not worth it as a weekend project.

Instead, in the near term, I am currently doing sighted tests for my own benefit. The guests that would have blind tested these will instead enjoy a more casual enjoyment-focused listening session from the winning speaker

But please by all means, let the debate continue! If this discussion leads to a conclusion and consensus on test procedure that would be accepted as a reasonable compromise vs the full fledged multi-million dollar Harman test facility, I will reboot the effort and we can start over. (However I’m skeptical this will happen.)

If there is no consensus, I would suggest some of you consider the irony that the only possible source of truth for your blind speaker tests is coincidentally restricted to only those produced by Harman (which is therefore non-reproducible, and un-falsifiable).

You don't need a consensus, just decide what you want to test. If you want to compare the midrange, roll off the bass to both and play whatever through the same signal path. Want to compare how the Genelecs with a sub hold up against Salon2's in the same configuration, then do that. Or both. It's a weekend project for fun, right? Don't waste the weekend here, go run your test, any test, and let us know what happened! I assure you the debate will be here, and just as "loud", after your test, no matter how you run it or how it turns out.

IME/IMO - Don

 

[preload]

This is my biggest criticism of Toole and Olive's research, who have developed a sound evaluation algorithm for free-standing loudspeakers with mono signals.
Many people who start with loudspeaker development make the mistake of tuning their first self-developed loudspeaker individually, standing freely in the room, with a mono signal - speak from experience

Can confirm.
A while back I constructed one of those DIY floorstanders on Troels site, complete with pricy scanspeak revelator drivers. I built the first one and it sounded great in mono listening. I compared it to a B&W CDM9NT speaker I had handy and I thought it sounded better. But when I built the second DIY speaker, listened in stereo, and compared it to the B&W's in stereo, I preferred the B&W's for most tracks. Quite a let down considering the DIY parts cost was more than a used pair of commercially built speakers. And these were supposedly "high end" drivers.

 

[preload]

That was the plan except I see no reason not to EQ the sub to the room. As you can see, it’s quite lumpy and sounds very boomy prior to correction.

That makes the most sense to me! Sounds like you've thought this through.

 

[Blumlein 88]

After re-reading that passage about mono testing, I'm still not convinced. The arguments are as follow:
* Various frequency response/timbre issues are partially hidden in stereo. Correct, but what's the difference between hidden and absent?
* Center channels in movie. I don't care, I say music is more important to most people here and quality is more often sought in that context.
* That segment quotes Klippel finding that most people associate "naturalness" and "pleasantness" with a "feeling of space", which explain mono's bias toward wide dispersion designs.

The entire passage is based on the unspoken axiom that more variance -> more correct assessment of the speaker. But I reiterate: what's the difference between a hidden shortcoming and no shortcoming?

This reminds me of this discussion with a Harbeth fanboy about a video showing that the Radial cone material was able to withstand a massive weight attached to it, unlike other materials: obvious "proof" that it was better. The same way that this carefully chosen criteria doesn't matter to sound quality, how does a mono difference matter in stereo when the entire thing is about more (and not less) differences between the two?

I may have more to say later. But I'm trying to measure the differences in something let us say how much it weighs. One method measures in pounds and another measures in ounces. Which method would I use? Even if the nearest pound is truly all I care about if by some strange twist it is easier to use the more accurate method measuring in ounces I'd certainly use that simpler method.

So I would not bother with 2 speakers if 1 is more discriminating even if the difference didn't always matter.

 

[echopraxia]

Subjective sighted shootout results:

Sighted impressions are no different from before. The only thing that’s changed since my same comparison in the same room is now I have better level matching and placement, and now the differences are even more clear.

Listened to this song (and others by them), as well as a variety of orchestral and rock songs.

In all ways but two very important ones, I find the Genelec 8351B to sound generally superior by a small margin. Instruments tonality/timbre sounds slightly more accurate/natural/correct. More realistic and enjoyable even, if we don’t count the points I’m going to mention below.

However the Salon2 pulls ahead in two ways. The first one is bass, though it’s not a huge factor in songs like this.

The second and more impressive one is soundstage. Put as simply as I can: The Genelecs sound like I am listening to speakers, but the Revel Salon2’s make me completely forget I am listening to music through speakers.

The location of instruments do not gravitate towards one speaker or the other as they tend to do more so on the Genelecs; with the Salon2, instruments sounds as if they are positioned wider and deeper, beyond the size of the room.

With the Genelec’s, the location of sounds seem to pull more towards the location of the speakers, inside the room.

With the Salon2’s, the borders of the room and the location of the speakers disappear.

If I ignore soundstage, the Genelecs win for what I perceive to be superior fidelity / accuracy / tonality / timbre.

But my final choice for this larger room goes to the Revel Salon2’s, because soundstage is SO important (to me, at least) when reclining enjoying music. Both have fantastic soundstage depth and width, to be clear. And the Genelec is superior in most other ways. But the Revel Salon2’s extremely “big” soundstage and ability for the speakers to disappear into the music adds a sort of “magic” to the experience in this room that pulls it ahead in terms of overall preference. Not that I believe it is magic — my best guess is this can be explained by the Salon2’s exceptionally wide horizontal response.

In my smaller home office, the Genelec 8351B plus a small sub sounds like perfection to my ears though It still doesn’t sound as “big” as the Salon2’s in a big room, but that’s not surprising. To me, the Genelec’s are perfection for small to medium rooms.

I love both very much for each scenario I normally use them for. These are absolutely, stunningly incredible speakers, and I enjoy them immensely.

(Sorry for the subjective descriptors, but this is the best I can do to explain the difference

 

[echopraxia]

You don't need a consensus, just decide what you want to test. If you want to compare the midrange, roll off the bass to both and play whatever through the same signal path. Want to compare how the Genelecs with a sub hold up against Salon2's in the same configuration, then do that. Or both. It's a weekend project for fun, right? Don't waste the weekend here, go run your test, any test, and let us know what happened! I assure you the debate will be here, and just as "loud", after your test, no matter how you run it or how it turns out.

IME/IMO - Don

Yep, I’m just enjoying them both now and comparing (sighted) in stereo with no subs at the moment (becuase it’s easy to configure this AB test via my MiniDSP SHD)

 

[Sancus]

It seems there is nothing remotely approaching consensus here on what is even an “acceptable” test procedure. At current rate, it looks like it may take a few decades or so for consensus to be reached here

If Amir waited for a consensus on the best way to test and write reviews ASR would never have gotten started

 

[ctrl]

Can confirm.
...I built the first one and it sounded great in mono listening. I compared it to a B&W CDM9NT speaker I had handy and I thought it sounded better. But when I built the second DIY speaker, listened in stereo, and compared it to the B&W's in stereo, I preferred the B&W's for most tracks.

Welcome to the club! In this case you have to retune the crossover or listen to it permanently with an equalizer or DSP.

A loudspeaker that sounds very pleasant to a bit boring in mono will probably sound optimal in a stereo setting.

If two loudspeakers are at almost the same level in terms of sound, you will end up listening in mono, most likely in a different order than you would in stereo.

 

[echopraxia]

If Amir waited for a consensus on the best way to test and write reviews ASR would never have gotten started

Yep, and I now understand and can relate a lot more tangibly why Amir doesn’t do his listening comparisons blind: It’s a lot of work — and as proven here — for little to nothing (since no matter what, no majority consensus will believe the results anyway)

Only real advantage is if it benefits you personally, but I’ve listened to these speakers for so many hours now (in different rooms in the case of the Genelec) I think I have a pretty reliable sense of their differences without any bias creeping into the picture. Of course, I can’t prove that for sure without a blind test, but I’m satisfied that my sighted evaluation is accurate enough.

 

[BenB]

...

In all ways but two very important ones, I find the Genelec 8351B to sound generally superior by a small margin. Instruments tonality/timbre sounds slightly more accurate/natural/correct. More realistic and enjoyable even, if we don’t count the points I’m going to mention below.

However the Salon2 pulls ahead in two ways. The first one is bass, though it’s not a huge factor in songs like this.

The second and more impressive one is soundstage. Put as simply as I can: The Genelecs sound like I am listening to speakers, but the Revel Salon2’s make me completely forget I am listening to music through speakers.

The location of instruments do not gravitate towards one speaker or the other as they tend to do more so on the Genelecs; with the Salon2, instruments sounds as if they are positioned wider and deeper, beyond the size of the room.

With the Genelec’s, the location of sounds seem to pull more towards the location of the speakers, inside the room.

With the Salon2’s, the borders of the room and the location of the speakers disappear.

If I ignore soundstage, the Genelecs win for what I perceive to be superior fidelity / accuracy / tonality / timbre.

But my final choice for this larger room goes to the Revel Salon2’s, because soundstage is SO important (to me, at least) when reclining enjoying music. Both have fantastic soundstage depth and width, to be clear. And the Genelec is superior in most other ways. But the Revel Salon2’s extremely “big” soundstage and ability for the speakers to disappear into the music adds a sort of “magic” to the experience in this room that pulls it ahead in terms of overall preference. Not that I believe it is magic — my best guess is this can be explained by the Salon2’s exceptionally wide horizontal response.

I think it would be informative if you could do a substantial amount of single-speaker listening (mono). From what you've said here, I suspect you might prefer the Genelec for single speaker listening where soundstage advantages can't be realized to the same degree. I posted about the perception of wide horizontal dispersion (with limited vertical dispersion) vs wide vertical dispersion (with limited horizontal dispersion), and found that I liked wide horizontal dispersion for the soundstage improvements. However, in single-speaker listening, I could barely tell a difference.

https://www.audiosciencereview.com/forum/index.php?threads/mono-vs-stereo-speaker-evaluation.15503/

 

[echopraxia]

I think it would be informative if you could do a substantial amount of single-speaker listening (mono). From what you've said here, I suspect you might prefer the Genelec for single speaker listening where soundstage advantages can't be realized to the same degree. I posted about the perception of wide horizontal dispersion (with limited vertical dispersion) vs wide vertical dispersion (with limited horizontal dispersion), and found that I liked wide horizontal dispersion for the soundstage improvements. However, in single-speaker listening, I could barely tell a difference.

https://www.audiosciencereview.com/forum/index.php?threads/mono-vs-stereo-speaker-evaluation.15503/

Hmm. Even if you are correct, what good would that be though when the result that actually matters (stereo listening preference) favors the wide dispersion?

In other words, let’s say the Genelec 8351B and Salon2 preference becomes less different (or reversed even, due to the Salon2’s preference being due to soundstage overpowering the other superiorities of the Genelec) when listening in mono. That doesn’t actually make the Genelec any more preferred because I and every audiophile I know does not listen to music mono on a single speaker.

 

[q3cpma]

I may have more to say later. But I'm trying to measure the differences in something let us say how much it weighs. One method measures in pounds and another measures in ounces. Which method would I use? Even if the nearest pound is truly all I care about if by some strange twist it is easier to use the more accurate method measuring in ounces I'd certainly use that simpler method.

So I would not bother with 2 speakers if 1 is more discriminating even if the difference didn't always matter.

Well, it would be true in the ideal case where it's the only difference between two speakers, but it might mislead you when weighing the pros and the cons.

Subjective sighted shootout results:

Sighted impressions are no different from before. The only thing that’s changed since my same comparison in the same room is now I have better level matching and placement, and now the differences are even more clear.

Listened to this song (and others by them), as well as a variety of orchestral and rock songs.

In all ways but two very important ones, I find the Genelec 8351B to sound generally superior by a small margin. Instruments tonality/timbre sounds slightly more accurate/natural/correct. More realistic and enjoyable even, if we don’t count the points I’m going to mention below.

However the Salon2 pulls ahead in two ways. The first one is bass, though it’s not a huge factor in songs like this.

The second and more impressive one is soundstage. Put as simply as I can: The Genelecs sound like I am listening to speakers, but the Revel Salon2’s make me completely forget I am listening to music through speakers.

The location of instruments do not gravitate towards one speaker or the other as they tend to do more so on the Genelecs; with the Salon2, instruments sounds as if they are positioned wider and deeper, beyond the size of the room.

With the Genelec’s, the location of sounds seem to pull more towards the location of the speakers, inside the room.

With the Salon2’s, the borders of the room and the location of the speakers disappear.

If I ignore soundstage, the Genelecs win for what I perceive to be superior fidelity / accuracy / tonality / timbre.

But my final choice for this larger room goes to the Revel Salon2’s, because soundstage is SO important (to me, at least) when reclining enjoying music. Both have fantastic soundstage depth and width, to be clear. And the Genelec is superior in most other ways. But the Revel Salon2’s extremely “big” soundstage and ability for the speakers to disappear into the music adds a sort of “magic” to the experience in this room that pulls it ahead in terms of overall preference. Not that I believe it is magic — my best guess is this can be explained by the Salon2’s exceptionally wide horizontal response.

In my smaller home office, the Genelec 8351B plus a small sub sounds like perfection to my ears though It still doesn’t sound as “big” as the Salon2’s in a big room, but that’s not surprising. To me, the Genelec’s are perfection for small to medium rooms.

I love both very much for each scenario I normally use them for. These are absolutely, stunningly incredible speakers, and I enjoy them immensely.

(Sorry for the subjective descriptors, but this is the best I can do to explain the difference

Might have preferred the quite wider 8260A, then. It should be quite competitive with the 8351B, as the real "problem" of the 8351A was the woofer design.

 

[BenB]

Hmm. Even if you are correct, what good would that be though when the result that actually matters (stereo listening preference) favors the wide dispersion?

In other words, let’s say the Genelec 8351B and Salon2 preference becomes less different (or reversed even, due to the Salon2’s preference being due to soundstage overpowering the other superiorities of the Genelec) when listening in mono. That doesn’t actually make the Genelec any more preferred because I and every audiophile I know does not listen to music mono on a single speaker.

It matters because single-speaker evaluations are quite prevalent, and Harman research on speaker preferences indicates that a single speaker is sufficient. Without evidence to the contrary, there's no real logical reason to be skeptical of the practice of evaluating speakers individually in mono (other than a gut feeling). Personally, I defer to evidence and research over my gut feelings, and you may even find places where I support the practice of single-speaker evaluations. However, I'm starting to think that this practice is flawed. It's possible that good speakers today are similar enough that secondary aspects of performance are starting to matter more than in the past. It's possible that stereo evaluation is imperative now, and wasn't when the research was done. Regardless of how we got here, if speakers today need to be evaluated in pairs with stereo listening, that's important information for the industry right now.

 

[HooStat]

Ok, here are measurements with no sub, 1/24 smoothing, 10hz - 22khz:

Revel Salon2:
View attachment 79308

Genelec 8351B:
View attachment 79309

What useful info do you think a listening test between these (as such) would conclude?

Just curious as to what is going on at 60 and 100 Hz. There is clearly a room mode, but I am wondering why the Genelec is so much more susceptible to it. You can see that the mode at 40Hz is identical. Perhaps it has to do with the placement on top of the Revel somehow?

 

[echopraxia]

Just curious as to what is going on at 60 and 100 Hz. There is clearly a room mode, but I am wondering why the Genelec is so much more susceptible to it. You can see that the mode at 40Hz is identical. Perhaps it has to do with the placement on top of the Revel somehow?

Yeah I wondered about that too. It could be, but I wonder if the smoother Salon2 bass response might be due to three 8” woofer spread out though space rather than just a single one that might be more likely to hit any single mode.

It matters because single-speaker evaluations are quite prevalent, and Harman research on speaker preferences indicates that a single speaker is sufficient. Without evidence to the contrary, there's no real logical reason to be skeptical of the practice of evaluating speakers individually in mono (other than a gut feeling). Personally, I defer to evidence and research over my gut feelings, and you may even find places where I support the practice of single-speaker evaluations. However, I'm starting to think that this practice is flawed. It's possible that good speakers today are similar enough that secondary aspects of performance are starting to matter more than in the past. It's possible that stereo evaluation is imperative now, and wasn't when the research was done. Regardless of how we got here, if speakers today need to be evaluated in pairs with stereo listening, that's important information for the industry right now.

Yeah but let’s say I tested this and confirm your hypothesis. Who here will really put any weight in those results, even if it’s a blind test?

It seems many people here are unwilling to accept even incremental evidence that contradicts Harman research, unless it is all in one big leap a greater body of work than the entire sum total of all Harman research. Unless of course the incremental result confirms Harman research, in which case the same people accept the incremental evidence. Not sure they realize the irony and why that’s a problem, but I don’t think I’m going to be able to change those attitudes regardless.

 

[preload]

It matters because single-speaker evaluations are quite prevalent, and Harman research on speaker preferences indicates that a single speaker is sufficient. Without evidence to the contrary, there's no real logical reason to be skeptical of the practice of evaluating speakers individually in mono (other than a gut feeling). Personally, I defer to evidence and research over my gut feelings, and you may even find places where I support the practice of single-speaker evaluations. However, I'm starting to think that this practice is flawed. It's possible that good speakers today are similar enough that secondary aspects of performance are starting to matter more than in the past. It's possible that stereo evaluation is imperative now, and wasn't when the research was done. Regardless of how we got here, if speakers today need to be evaluated in pairs with stereo listening, that's important information for the industry right now.

Right. Stereo performance is the gold standard. Performing blinded listening tests with a stereo pair is far less practical. Harman therefore did experiments to demonstrate that mono listening tests are "close enough" to stereo listening tests to be valid. But the intent is for mono testing to act as a surrogate for the less practical and more expensive stereo testing. By no means is mono testing superior to stereo testing in predicting stereo performance/preferences.

 

[aarons915]

Just curious as to what is going on at 60 and 100 Hz. There is clearly a room mode, but I am wondering why the Genelec is so much more susceptible to it. You can see that the mode at 40Hz is identical. Perhaps it has to do with the placement on top of the Revel somehow?

100Hz is probably the classic floor bounce which 3 woofers will help to smooth out, it is definitely 1 real benefit of multiple woofers even if you don't need the output.

 

[preload]

Welcome to the club! In this case you have to retune the crossover or listen to it permanently with an equalizer or DSP.

I think you're right. I decided to save some money and not use the ultra expensive crossover parts spec'd by Troels and probably the DCR was off on a couple of the coils. Either that or Troels didn't take into account directivity in his designs back then. I think it was around that time he started introducing diy plans that incorporated a wave guide. This was a looooong time ago.