Here's an example of how I used the method I previously described here to evaluate equipment. Yesterday I visited a very familiar dealer's listening room--one where I have quite frequently auditioned equipment in the past--to audition the new version of circa-$15k/pair loudspeakers. A particular well-known reviewer uses the prior version of these as "reference" speakers in his surround sound system. They are moderately large three-way floor standing box speakers with an elegant/high-tech design. This pair was represented as fully broken in, having been played at another store for several weeks before being transferred to this store. Electronics were all from a single high-end brand and are the electronics I usually use to audition speakers at this dealer. I listened for less than 45 minutes because that was all it took for me to reach what I considered a fairly firm conclusion using my method.
First up was 88 Basin Street with the Count Basie band. Bass purred with warmth and was well defined and deep. Very dynamic presentation, with open, clear, large presentation. Those were as they should be. Soundstage was not as well developed or defined as it should have been, however, and cymbal crashes seemed forward.
I tried another big band recording from my current travel pack, the early Reference Recordings "Your Friendly Neighborhood Big Band." Again very dynamic presentation and this time there was great depth of stage, as this recording should sound. But cymbals again crashed brashly and lacked shimmer. The female vocal also sounded too sibilant and almost lispy.
I tried Clark Terry's "Live From the Village Gate" on Chesky, moving my chair closer to get close to 90-degree separation on this coincidently miked stereo recording. Cymbals again were overly prominent, and I further zeroed in on the problem as too much ring (mid-treble) relative to shimmer (top octave) and tap (midrange and lower). Terry's trumpet was also edgier than it should be. Peaky octave from 5 kHz to 10 kHz was my thought.
I moved to female vocal/piano recital on a "pedestrian" DGG recording, Salzburg Recital. Piano sound was decent, but not as clear and real as it can sound, and not as clearly focused on the stage in back of the vocalist as it should sound. Kathleen Battle's voice was surprisingly unrealistic, almost unrecognizable as her: very sibilant with the delicate tender qualities almost totally covered up. When she hit the high note at the end of "O Had I Jubal's Lyre," yes, it was dynamically contrast-y enough, but rather than raise the hackles on the back of my neck and take my breath away, it made me cringe a bit. Also, the movements of her voice on the stage were not as well defined in their moment-to-moment unsettledness as I know they can be, jerking suddenly far left at the climax. Very odd.
I move to a pop recording, Eva Cassidy, Live at Blues Alley. Again, Eva's voice is quite sibilant and the litheness of her technique, the way she sings around the center tone of a note from moment to moment, is almost absent. Cymbals again ring too much and lack shimmer, and thus seem overly prominent and forward in the mix as in "hey, listen to me!" Ugh.
I could have stopped the audition there, maybe 15 or 20 minutes after it began, and crossed these speakers off my list, since I really cannot abide speakers with mid-treble peaks. But I proceeded to more fully evaluate other aspects of their performance. Any other problems would prevent me from later questioning my conclusions and wanting to re-audition them at a later date.
I quickly moved on to small vocal ensemble, The Waverly Consort. Nice deep acoustic, lots of hall venue, as it should be. Vocals fairly clean, but not as natural sounding and clean as the Harbeth sound I know well. Ensemble work of the voices was not as cleanly blended with yet distinct voices as it should/can be. A bit of grunge on the ensemble vocal sound, actually. Good, but not breath-taking, as I know it can be. I miss the Radial cones of my Harbeths.
Unusually, I decide not to test the speakers with large chorus (RR Rutter Requiem) since if it can't handle small vocal ensembles well, large choruses won't be up to snuff either.
On to RR's early orchestral recording, Fiesta. Good solid bass drum whacks and nice depth of stage, as appropriate on this one. But the bells are a bit fuzzy, not as bell-like as they should be. Again, the lack of Radial midrange and the signature of this model's midrange cone is raising its not-so-pretty head.
More bass: the Gnomus cut of the Dorian organ version of Pictures at an Exhibition. Excellent bass depth and definition for this size speaker unaided by subwoofer. When I turned it up toward what I regard as realistic levels for this cut, something in the room or speakers "took off" a bit with bass resonance, but I will charitably blame the room. Good down to the 30s I'd say, but not like the acoustic suspension bass room-lock effect of my JL Audio f113 subs which are flat to way below that and can play this cut much louder without any distress whatsoever.
On the Mercury Living Presence "The Composer and His Orchestra," I verify that the speakers are rolling the top octave compared to the next octave down since they don't well reveal the peak centered at about 11 kHz caused by the old microphones. And, oddly, this recording sounds more "civilized," than it should, as if the speakers were tuned around this recording to take out the recording's top-octave emphasis and not reveal the peakiness in the speakers' mid-treble I've observed on other recordings. That's not impossible, since this recording has an audiophile reputation of being of reference quality. This recording also sounds oddly non-open, compressed. and undynamic. Tape hiss is relatively innocuous, a bit too much so, being only a quiet "ssss" with no lower "shussh" overlay. Soundstage is good, but not as good as I know it can be on this recording.
Just for kicks I try to evaluate the speakers' ultimate soundstaging capabilities with a coincidently miked Sheffield Lab recording of Wagner and others. I move forward again to get the best possible focus. Good, but not world-class solidity of staging, but the speakers could probably do better if I'd angled them in a bit more. Low cellos are nicely differentiated. But . . . the violins are steely, which definitely should not be on this recording.
Enough.
I'm sure the midrange on these new speakers will sound just dandy to many. But we Harbeth owners are used to a much higher standard of naturalness and truly clean clarity without "edge." The Kevlar "crack" or "cringe" sound to the midrange is still very much there on these newest models using that material for the midrange cone. Voices don't sound nearly as natural as with Harbeths and small vocal ensembles are just not very well defined as individual voices and tend to sound a bit grungy. And Harbeth's natural-sounding high-frequency balance and the near-perfect integration of tweeter and midrange was missing. The mid highs were just too strong, making the ringing sound of cymbals overly prominent at the expense of shimmer and strongly coloring female voices, so strongly that much of the singer's identity, not just the beauty and subtlety of the particular singer's art, is erased.
Frankly, these sounded about the way the old version measured in figure 4 in this review except that in these new ones the 4 kHz peak might even be a bit more prominent. To my ears, to sound maximally natural on a decent range of commercial recordings in this type of on-axis measurement, the speaker response should not exceed the 1 kHz reference level in the range from 2 kHz on up. And given the way most commercial recordings are made, I actually tend to prefer a bit of a downsloping response from bass to treble, with even a bit of a dip-and-return from 2 to 6 kHz thrown in for good measure.
To me, peaks in the 2 kHz to 10 kHz region are anathema to natural sound, whether the speakers cost $150, $15,000, or $150,000. Sure, with more money, you buy less distortion, higher SPLs, more bass extension, etc. But the basic on-axis frequency balance still has to be correct in order for the speakers to sound realistic, at least to me. Such response obviously can and has been built into speakers before, such as here for very cheap and here for a fraction of what these new speakers cost.
Sometimes--more and more lately--I think that consumer speaker design is devolving rather than evolving. Perhaps that shouldn't be surprising given what I see as a general dumbing-down trend in the wider culture. A "hi-fi" sound that is sui generis--with little in common with the sound of live unamplified instruments and voices--has become the "house sound" of many manufacturers, even those with histories in pro-audio monitoring.
Yes, consumer-oriented speakers are getting better at playing louder cleanly with reasonable amount of power, presenting a decently large, open soundstage, portraying instruments in depth and superficially sounding clean and very detailed. The bass region, particularly, seems better handled in terms of power handling, punch, and differentiation.
But can't they at least get solo female vocals to sound beautifully intimate without excess sibilance and edge? Female vocals are a staple of audiophile demo material and a speaker which can't even handle such actually easy-to-reproduce-decently material with beauty and grace certainly is not worth $15k a pair. And most modern recordings--female vocals included--are already too hot in the treble for their own good. What are they thinking?
And cymbal shimmer is, or ought to be, a sound common enough to many listeners' experience to be conspicuous in its absence. Everyone has heard a drum kit. But then I remember that today most drum kits are in isolation booths even when heard live and really do have very little shimmer compared to ring because the PA system is inadequate or the booth blocks the shimmer. And then there is the ubiquitous sampled noise which passes for a drum kit on many current recordings.
And miked trumpets ARE very likely to sound too trebly/edgy . . . .
So it goes. Art more and more mimics technology, rather than the other way around. Quite depressing. It's enough to make me want to dive further into vintage audio. Back in the day, Acoustic Research knew how to get a natural-sounding frequency balance from a home system quite well. No, it wasn't perfect, but when it erred, it erred in the direction of just moving you backward in the concert hall rather than throwing unnatural tizz at your ears. Sophisticated high art translated well through relatively "primitive" technology by tastefully intelligent designers who knew well the sound of the real thing and used every trick then available to them to emulate it. Now, we have speaker technology and development tools undreamed of 40 years ago, but apparently many speaker designers and consumers could care less or know less about the art.
On the way back from yesterday's audio-store audition, I tuned my car's HD FM radio to the local college jazz music station. I instantly heard a more life-like portrayal of cymbal sound--tap, ring, and shimmer all in fairly decent balance--from my factory-installed auto system which, believe me, in no way represents state-of-the-art sound. At least it cheered me up to know that I was NOT imagining the lack of high-frequency realism I'd just heard in a pair of $15k speakers.
First up was 88 Basin Street with the Count Basie band. Bass purred with warmth and was well defined and deep. Very dynamic presentation, with open, clear, large presentation. Those were as they should be. Soundstage was not as well developed or defined as it should have been, however, and cymbal crashes seemed forward.
I tried another big band recording from my current travel pack, the early Reference Recordings "Your Friendly Neighborhood Big Band." Again very dynamic presentation and this time there was great depth of stage, as this recording should sound. But cymbals again crashed brashly and lacked shimmer. The female vocal also sounded too sibilant and almost lispy.
I tried Clark Terry's "Live From the Village Gate" on Chesky, moving my chair closer to get close to 90-degree separation on this coincidently miked stereo recording. Cymbals again were overly prominent, and I further zeroed in on the problem as too much ring (mid-treble) relative to shimmer (top octave) and tap (midrange and lower). Terry's trumpet was also edgier than it should be. Peaky octave from 5 kHz to 10 kHz was my thought.
I moved to female vocal/piano recital on a "pedestrian" DGG recording, Salzburg Recital. Piano sound was decent, but not as clear and real as it can sound, and not as clearly focused on the stage in back of the vocalist as it should sound. Kathleen Battle's voice was surprisingly unrealistic, almost unrecognizable as her: very sibilant with the delicate tender qualities almost totally covered up. When she hit the high note at the end of "O Had I Jubal's Lyre," yes, it was dynamically contrast-y enough, but rather than raise the hackles on the back of my neck and take my breath away, it made me cringe a bit. Also, the movements of her voice on the stage were not as well defined in their moment-to-moment unsettledness as I know they can be, jerking suddenly far left at the climax. Very odd.
I move to a pop recording, Eva Cassidy, Live at Blues Alley. Again, Eva's voice is quite sibilant and the litheness of her technique, the way she sings around the center tone of a note from moment to moment, is almost absent. Cymbals again ring too much and lack shimmer, and thus seem overly prominent and forward in the mix as in "hey, listen to me!" Ugh.
I could have stopped the audition there, maybe 15 or 20 minutes after it began, and crossed these speakers off my list, since I really cannot abide speakers with mid-treble peaks. But I proceeded to more fully evaluate other aspects of their performance. Any other problems would prevent me from later questioning my conclusions and wanting to re-audition them at a later date.
I quickly moved on to small vocal ensemble, The Waverly Consort. Nice deep acoustic, lots of hall venue, as it should be. Vocals fairly clean, but not as natural sounding and clean as the Harbeth sound I know well. Ensemble work of the voices was not as cleanly blended with yet distinct voices as it should/can be. A bit of grunge on the ensemble vocal sound, actually. Good, but not breath-taking, as I know it can be. I miss the Radial cones of my Harbeths.
Unusually, I decide not to test the speakers with large chorus (RR Rutter Requiem) since if it can't handle small vocal ensembles well, large choruses won't be up to snuff either.
On to RR's early orchestral recording, Fiesta. Good solid bass drum whacks and nice depth of stage, as appropriate on this one. But the bells are a bit fuzzy, not as bell-like as they should be. Again, the lack of Radial midrange and the signature of this model's midrange cone is raising its not-so-pretty head.
More bass: the Gnomus cut of the Dorian organ version of Pictures at an Exhibition. Excellent bass depth and definition for this size speaker unaided by subwoofer. When I turned it up toward what I regard as realistic levels for this cut, something in the room or speakers "took off" a bit with bass resonance, but I will charitably blame the room. Good down to the 30s I'd say, but not like the acoustic suspension bass room-lock effect of my JL Audio f113 subs which are flat to way below that and can play this cut much louder without any distress whatsoever.
On the Mercury Living Presence "The Composer and His Orchestra," I verify that the speakers are rolling the top octave compared to the next octave down since they don't well reveal the peak centered at about 11 kHz caused by the old microphones. And, oddly, this recording sounds more "civilized," than it should, as if the speakers were tuned around this recording to take out the recording's top-octave emphasis and not reveal the peakiness in the speakers' mid-treble I've observed on other recordings. That's not impossible, since this recording has an audiophile reputation of being of reference quality. This recording also sounds oddly non-open, compressed. and undynamic. Tape hiss is relatively innocuous, a bit too much so, being only a quiet "ssss" with no lower "shussh" overlay. Soundstage is good, but not as good as I know it can be on this recording.
Just for kicks I try to evaluate the speakers' ultimate soundstaging capabilities with a coincidently miked Sheffield Lab recording of Wagner and others. I move forward again to get the best possible focus. Good, but not world-class solidity of staging, but the speakers could probably do better if I'd angled them in a bit more. Low cellos are nicely differentiated. But . . . the violins are steely, which definitely should not be on this recording.
Enough.
I'm sure the midrange on these new speakers will sound just dandy to many. But we Harbeth owners are used to a much higher standard of naturalness and truly clean clarity without "edge." The Kevlar "crack" or "cringe" sound to the midrange is still very much there on these newest models using that material for the midrange cone. Voices don't sound nearly as natural as with Harbeths and small vocal ensembles are just not very well defined as individual voices and tend to sound a bit grungy. And Harbeth's natural-sounding high-frequency balance and the near-perfect integration of tweeter and midrange was missing. The mid highs were just too strong, making the ringing sound of cymbals overly prominent at the expense of shimmer and strongly coloring female voices, so strongly that much of the singer's identity, not just the beauty and subtlety of the particular singer's art, is erased.
Frankly, these sounded about the way the old version measured in figure 4 in this review except that in these new ones the 4 kHz peak might even be a bit more prominent. To my ears, to sound maximally natural on a decent range of commercial recordings in this type of on-axis measurement, the speaker response should not exceed the 1 kHz reference level in the range from 2 kHz on up. And given the way most commercial recordings are made, I actually tend to prefer a bit of a downsloping response from bass to treble, with even a bit of a dip-and-return from 2 to 6 kHz thrown in for good measure.
To me, peaks in the 2 kHz to 10 kHz region are anathema to natural sound, whether the speakers cost $150, $15,000, or $150,000. Sure, with more money, you buy less distortion, higher SPLs, more bass extension, etc. But the basic on-axis frequency balance still has to be correct in order for the speakers to sound realistic, at least to me. Such response obviously can and has been built into speakers before, such as here for very cheap and here for a fraction of what these new speakers cost.
Sometimes--more and more lately--I think that consumer speaker design is devolving rather than evolving. Perhaps that shouldn't be surprising given what I see as a general dumbing-down trend in the wider culture. A "hi-fi" sound that is sui generis--with little in common with the sound of live unamplified instruments and voices--has become the "house sound" of many manufacturers, even those with histories in pro-audio monitoring.
Yes, consumer-oriented speakers are getting better at playing louder cleanly with reasonable amount of power, presenting a decently large, open soundstage, portraying instruments in depth and superficially sounding clean and very detailed. The bass region, particularly, seems better handled in terms of power handling, punch, and differentiation.
But can't they at least get solo female vocals to sound beautifully intimate without excess sibilance and edge? Female vocals are a staple of audiophile demo material and a speaker which can't even handle such actually easy-to-reproduce-decently material with beauty and grace certainly is not worth $15k a pair. And most modern recordings--female vocals included--are already too hot in the treble for their own good. What are they thinking?
And cymbal shimmer is, or ought to be, a sound common enough to many listeners' experience to be conspicuous in its absence. Everyone has heard a drum kit. But then I remember that today most drum kits are in isolation booths even when heard live and really do have very little shimmer compared to ring because the PA system is inadequate or the booth blocks the shimmer. And then there is the ubiquitous sampled noise which passes for a drum kit on many current recordings.
And miked trumpets ARE very likely to sound too trebly/edgy . . . .
So it goes. Art more and more mimics technology, rather than the other way around. Quite depressing. It's enough to make me want to dive further into vintage audio. Back in the day, Acoustic Research knew how to get a natural-sounding frequency balance from a home system quite well. No, it wasn't perfect, but when it erred, it erred in the direction of just moving you backward in the concert hall rather than throwing unnatural tizz at your ears. Sophisticated high art translated well through relatively "primitive" technology by tastefully intelligent designers who knew well the sound of the real thing and used every trick then available to them to emulate it. Now, we have speaker technology and development tools undreamed of 40 years ago, but apparently many speaker designers and consumers could care less or know less about the art.
On the way back from yesterday's audio-store audition, I tuned my car's HD FM radio to the local college jazz music station. I instantly heard a more life-like portrayal of cymbal sound--tap, ring, and shimmer all in fairly decent balance--from my factory-installed auto system which, believe me, in no way represents state-of-the-art sound. At least it cheered me up to know that I was NOT imagining the lack of high-frequency realism I'd just heard in a pair of $15k speakers.
. I was guessing. You're right about the reference. A lot of people will deem the kind of speakers I like best as boring or laid-back. I consider both of those very desirable qualities.
