revel ultima salon2 vs focal sopra no. 2

[Art Vandelay]

However, the best off axis performance I have seen is the Philharmonitor BMR. The red trace is on-axis and the green trace is for 80 degrees off axis and never deviates more than 5dB!

http://philharmonicaudio.com/BMR Philharmonitor.html

Excellent, yes, but only valid in the horizontal plane. And noting that long ribbon tweeters are very directional in the vertical plane.

T-M-W designs in general can't avoid a messy vertical axis, although high order crossover filters do help to some extent.

My B&W 800Di's present some problems with their 2nd order acoustic slopes - and very asymmetric driver roll-off off axis. I needed to install acoustic diffusers on the ceiling at the first reflection points for perfect mid and treble balance, and they needed to be physically large enough to diffuse > 4kHz. That diamond tweeter is almost isotropic at the crossover frequency.

 

[Art Vandelay]

FTR, Here's the results of my own MLS measrements.

Speaker under test is the b&w 800 diamond.

On axis @ 1m (b&w reference height & axis)

30 degrees off-axis

Due to the large diameter of the B&W FST and the high crossover frequency, the power response dips between 2 and 4 kHz.

My preference is to listen on-axis.

 

[Irvrobinson]

I might have read Absolute Sound about once in my entire lifetime.

FTR, I've heard the Salon 2's paired with different amplification, ranging from ML Class D to Gryphon pure class A, and ditto preamps and sources. Even in the same listening room and playing the same CD source I've heard them sound bright and edgy as well as smooth and seductive, all due to changes in amplification.

WRT 800D3's, my local dealer has them setup in his prime demo room, driven by Krell amps and front end. I've heard many speakers in that room, and many had triple figure x 1000 price tags, but none have sounded as good as the 800D3's.

Art, we are on different pages altogether. You are claiming to hear substantial differences in a given speaker's performance from electronics alone, and I think that's just your imagination. While I have owned various high-end amplifiers and pre-amps (Krell Class A and Mark Levinson while under Madrigal), the differences have never been as overt as you describe, and with good reason. The distortion and timing artifacts you talk about with differences in output stage classes or feedback is 70db below the fundamentals in most cases, and often 80db down. While I do believe artifacts like that can result in longer-term listening issues, and I believe I experience them, the whole idea that they can cause obvious issues that would certainly be measurable in a simple frequency response measurement (like "edginess") is beyond belief.

As for speakers sounding "good", that's a concept over the years I've moved away from. Accuracy is all that matters to me in the long run. Good is irrelevant. Speaker colorations can make certain recordings sound "better" to some listeners, though the same speakers tend to similarly color all recordings, and since hearing is a trainable skill, once you know what to listen for even pleasant colorations can drive you nuts. I find various Magicos are like that, for example. I figure they're voiced to be impressive in an audition. I'll have to try to hear the 800D3s. I spent considerable time with the 800D1s and the D2s, though at $34K/pair I fear B&W has lost their value proposition recipe.

 

[Art Vandelay]

Art, we are on different pages altogether. You are claiming to hear substantial differences in a given speaker's performance from electronics alone, and I think that's just your imagination. While I have owned various high-end amplifiers and pre-amps (Krell Class A and Mark Levinson while under Madrigal), the differences have never been as overt as you describe, and with good reason. The distortion and timing artifacts you talk about with differences in output stage classes or feedback is 70db below the fundamentals in most cases, and often 80db down. While I do believe artifacts like that can results in longer-term listening issues, and I believe I I experience them, the whole idea that they can cause obvious issues that would certainly be measurable in a simple frequency response measurement (like "edginess") is beyond belief.

As for speakers sounding "good", that's a concept over the years I've moved away from. Accuracy is all that matters to me in the long run. Good is irrelevant. Speaker colorations can make certain recordings sound "better" to some listeners, though the same speakers tend to similarly color all recordings, and since hearing is a trainable skill, once you know what to listen for even pleasant colorations can drive you nuts. I find various Magicos like that, for example. I figure they're voiced to be impressive in an audition. I'll have to try to hear the 800D3s. I spent considerable time with the 800D1s and the D2s, though at $34K/pair I fear B&W has lost their value proposition recipe.

That I imagined the difference is possible, and I'm sure we've all been guilty of that at times too, but in the instance referred to above there were 2 other people present who heard exactly the same as I did, and expressed it before I voiced any opinion. This increases my confidence that what I heard was real and not imagined.

FTR, I'm definitely in agreement with you though when you say that most differences between amplifiers require a longer term listener - appliance relationship in order to be fully appreciated.
This has definitely been my experience too.

Also, when I say that the 800D3 sounds better or best I do mean that it sounds most transparent / least colored. For myself, accuracy is the end goal.

Of course all speakers do add color to some extent, and over time any critical listener will readily identify the signature or signatures that are common from one recording to the next. The same applies to amplifiers, sources, and of course the listening environment itself.

Accuracy is something that almost any connoisseur of audio demands and strives for, but ultimately will never be quite able to achieve. Ultimately, one needs to settle for something close to accurate and with a flavor of color that is deemed to be least objectionable. This is one area where subjectivity rules.

So perhaps we're not on completely different pages after all.

Fwiw, I've only ever heard Magico speaker at audio shows, so I haven't formed a definitive personal view, other than a suspicion that the treble region was a bit too bright for my liking. I think it was the Q5 on both occasions.

 

[Swerd]

where does one find that off axis response curve? I'd like to see that data for other speakers too!

Those graphs of audio response are done by the Canadian National Research Council and are found at SoundStage. The NRC's speaker measurements are widely considered the best available set of speaker performance measurements available online.
http://www.soundstagenetwork.com/index.php?option=com_content&view=article&id=16&Itemid=18
Here is the page for the Revel Ultima Salon2.
http://www.soundstagenetwork.com/measurements/speakers/revel_ultima_salon2/

Those measurements published by Stereophile are also good, but I think you should avoid reading the review text and go straight to the measurements page. It is widely understood, that Stereophile avoids making overtly negative comments about speakers even if the measurement indicates they are deserved. What isn't said can become important. So sometimes you must read between the lines, or have a lawyer read it with you .

 

[skypickle]

thank you. I am still deciphering the graphs.

 

[Irvrobinson]

Those measurements published by Stereophile are also good, but I think you should avoid reading the review text and go straight to the measurements page.

I completely agree, and that's what I do.

 

[Dennis Murphy]

Excellent, yes, but only valid in the horizontal plane. And noting that long ribbon tweeters are very directional in the vertical plane.

T-M-W designs in general can't avoid a messy vertical axis, although high order crossover filters do help to some extent.

My B&W 800Di's present some problems with their 2nd order acoustic slopes - and very asymmetric driver roll-off off axis. I needed to install acoustic diffusers on the ceiling at the first reflection points for perfect mid and treble balance, and they needed to be physically large enough to diffuse > 4kHz. That diamond tweeter is almost isotropic at the crossover frequency.

The RAAL is pretty good vertically--as good as I would want from a normal listening distance. There's not much point in maximizing ceiling reflections. And, as you reference, any two or three-way speaker with vertically arrayed drivers will have big dips in the crossover region(s) as you move above the tweeter. So a somewhat erratic vertical response is pretty much a fact of life for most speakers. What you want to avoid is vertical dispersion that is so narrow that you lose the highs when you stand up.

 

[Dennis Murphy]

thank you. I am still deciphering the graphs.

One thing to keep in mind when studying the CNRC graphs is that the vertical axis demarcation is 10 dB. So any deviation would appear twice as large if the plot used the standard 5 dB grid. Still, these measurements are about the best you'll find, particularly since they don't have to engage in any questionable splicing of near-field port and woofer response with 1-meter anechoic measurements higher up. The distortion plots are also very useful due to the use of a genuine anechoic chamber.

After making the rounds of an audio show with expensive speakers that don't sound right to me, I can always see why when I check the CNRC plots. And they show why I've never used AMT high frequency drivers. You can spot the ragged response in the plots--particularly when you visualize what they would look like with a 5 dB grid.

 

[Art Vandelay]

Those measurements published by Stereophile are also good, but I think you should avoid reading the review text and go straight to the measurements page.

Measurement come with limitations too, and it's usually necessary to look beyond the basic on-axis or waterfall responses.

Most of the Stereophile on-axis response measurements are centred around tweeter height and at a distance of 50", which is good from the pov of consistency but not necessarily the reference height or appropriate measurement distance for every speaker.

There's no shortage of forum threads dedicated to stereophile measurements of speakers that exhibit an uneven on-axis response where participants are quick to conclude that the design is flawed and as a result the speaker is going to sound mediocre.

OTOH, if Stereophile had measured at a different height or axis it's quite possible that they would see a very flat response, and that's also what the listener would hear after fine tuning the amount of toe-in and tilt.

Actually, there are many speakers that are designed with a specific reference axis in mind, which often compensates for driver beaming and a rising on-axis response, while at the same time preserving a linear power response.

 

[Irvrobinson]

OTOH, if Stereophile had measured at a different height or axis it's quite possible that they would see a very flat response, and that's also what the listener would hear after fine tuning the amount of toe-in and tilt.

Some years ago I met JA at an audio show in LA. It was only a short conversation, but I came away thinking that this was a guy who takes measuring full range speakers very seriously. (I've never seen him measure a sub.) I doubt anyone here is going to significantly out-think him in this regard. (Okay, Floyd Toole posts now and then, maybe him.) I also came away thinking JA is one very smart guy. The subjective stuff in Stereophile is the usual commercial malarky, but when JA gets serious about measuring or recording, the guy has my respect.

 

[Swerd]

Measurement come with limitations too, and it's usually necessary to look beyond the basic on-axis or waterfall responses.

True. But there is real value in the total body of measurements done with a standardized method, if done over a wide range of different speakers. These 'speaker performance databases' from the Canadian NRC and JA of Stereophile constitute the only reliable way we consumers can compare speakers without stepping into the minefield of reviewers' comments. (In Stereophile, JA does the measurements, but the reviews are always written by others.)

Some years ago I met JA at an audio show in LA. It was only a short conversation, but I came away thinking that this was a guy who takes measuring full range speakers very seriously. (I've never seen him measure a sub.) I doubt anyone here is going to significantly out-think him in this regard. (Okay, Floyd Toole posts now and then, maybe him.) I also came away thinking JA is one very smart guy. The subjective stuff in Stereophile is the usual commercial malarky, but when JA gets serious about measuring or recording, the guy has my respect.

Agreed. JA seems to understand the value of standardized measurement methods done over a wide range of different speakers. I always first look at his measurements, and then scan the reviewer's comments to see where they conflict with each other .

I would argue that DM also knows and understands the value and the limitations of speaker measurements. Because I know him, I am aware he also understands what measured acoustic performance features actually translate into audibly better sounding speakers to human listeners.

Floyd Toole takes it to another level of sophistication – he not only understands what measured acoustic performance features actually translate into audibly better sounding speakers to human listeners, he knows how to measure the listeners preferences in scientifically valid listening tests.

 

[Swerd]

thank you. I am still deciphering the graphs.

Be patient. And listen to a lot of speakers. It takes a while before you begin to understand what the graphs can and cannot tell you.

Edit:
I tend to agree with Irvrobinson when he says "Accuracy is all that matters to me in the long run."

As illustrations, compare these 3 extremely different looking speaker measurements. Can you guess which is more accurate?

http://www.soundstage.com/measurements/ethera_vitae/

http://www.soundstage.com/index.php?option=com_content&view=article&id=835:nrc-measurements-fluance-xl7f-loudspeakers&catid=77:loudspeaker-measurements

http://www.soundstage.com/measurements/zucable_druid/

 

[Irvrobinson]

One more comment on accuracy, which I've made before, but I want to reiterate anyway, is that you can't use commercial recordings to judge accuracy. The vast majority of commercial recordings are mixed to sound a certain way, augmented (reverb, EQ, whatever), and you weren't there at the performance anyway. Even when you are there, unless it was a purist recording made with a pair of mics and no augmentation, you still can't use it. Even for non-commercial small-ensemble recordings, I've witnessed several cases where vocals are augmented and magnified, percussion is faded back, and the sax is mic'ed right on the horn flare. Ugh.

To judge accuracy your best tool is a digital recorder in your own room. Music quality doesn't matter, so get a high school musician to play, record speaking voices, record non-musical sounds. If they sound real on playback, your speaker is probably accurate. If varying kinds of sounds across the spectrum sound real, you probably spent a lot of money and use a lot of space to make that happen. You also need to prepare yourself mentally for this, because unless you have the right room and the right speakers, and you don't run your subs 10db hot, you're apt to be sorely disappointed at what you hear.

 

[skypickle]

@Swerd i would say the ethera have the most accuracy given the consistent linearity of their frequency response.

 

[Art Vandelay]

True. But there is real value in the total body of measurements done with a standardized method, if done over a wide range of different speakers. These 'speaker performance databases' from the Canadian NRC and JA of Stereophile constitute the only reliable way we consumers can compare speakers without stepping into the minefield of reviewers' comments. (In Stereophile, JA does the measurements, but the reviews are always written by others.)

Actually, JA does still review occasionally, and fwiw he does often say that measurement anomalies should be treated with caution. speakers with 1st order crossover as well as many panel speakers can present measurement anomalies that look evil but don't really exist for real world listening.

For basic FR plots it's now relatively easy to do your own these days anyway, and there's no substitute for being able to measure on-axis or off-axis yourself at various heights, which makes it possible to know the optimum V and H axis for listening. Of course it helps to have a good measurement mic such as an Earthworks 31, which costs a bit, but there plenty of MLS software available at low cost.

From personal experience, the measurement that best aligns with a subjective interpretation is the averaged response at the listening position, and if you really want a ruler flat response it's now possible to achieve it with room correction software. I've tried a few including Dirac Live, but I don't use it because the response is close to flat anyway, and ultimately I preferred the sound without the correction in line, but I think it would probably be beneficial in most installations, - though you still need to know what you're doing to some extent in order to get the best out of it.

That said, it should be possible to make a Revel Salon sound almost the same as a B&W 800 Diamond or Focal Sopra etc with good room / speaker correction software.

 

[Art Vandelay]

One more comment on accuracy, which I've made before, but I want to reiterate anyway, is that you can't use commercial recordings to judge accuracy. The vast majority of commercial recordings are mixed to sound a certain way, augmented (reverb, EQ, whatever), and you weren't there at the performance anyway. Even when you are there, unless it was a purist recording made with a pair of mics and no augmentation, you still can't use it. Even for non-commercial small-ensemble recordings, I've witnessed several cases where vocals are augmented and magnified, percussion is faded back, and the sax is mic'ed right on the horn flare. Ugh.

You're implying here that accuracy no longer matters if you only listen to commercial recordings.

I agree to an extent but would add the qualification that commercial quality varies wildly. However, we do now live in a world where dynamic compression rules.

The main problem IMO is that most commercial recordings are designed for playback on small ear buds and multimedia systems, so they're EQ'ed with a smiley face response as well as dynamically compressed / limited to the max.

 

[Irvrobinson]

You're implying here that accuracy no longer matters if you only listen to commercial recordings.

That wasn't my intent at all. That thought didn't even crossed my mind. All I'm saying is that to discover what is accurate in your room, the best way to do it is with familiar reference recordings. The process still isn't perfect, because with the recording you're hearing the room acoustics twice, but it's as close as I can think of getting, so far.

 

[Art Vandelay]

That wasn't my intent at all. That thought didn't even crossed my mind. All I'm saying is that to discover what is accurate in your room, the best way to do it is with familiar reference recordings. The process still isn't perfect, because with the recording you're hearing the room acoustics twice, but it's as close as I can think of getting, so far.

I don't believe it's necessary, aside from the fact that it would be difficult to do without introducing additional variables and room interactions.

If you can verify that the response is reasonably flat at the listening position - with tools and measurements, you can pretty much sit back and listen with confidence that everything is fine.

The greater objective is to hear the recording as you would have heard it live in the venue itself.

It's never achievable of course, even when the recording process itself is also trying to achieve that, but we can always dream.

 

[Dennis Murphy]

I don't believe it's necessary, aside from the fact that it would be difficult to do without introducing additional variables and room interactions.

If you can verify that the response is reasonably flat at the listening position - with tools and measurements, you can pretty much sit back and listen with confidence that everything is fine.

The greater objective is to hear the recording as you would have heard it live in the venue itself.

It's never achievable of course, even when the recording process itself is also trying to achieve that, but we can always dream.