The Harbeth Monitor 40.2 speakers are, for me, at least, end-game speakers. With these I can sail happily through my Golden Years. They have the sound of music as I've always dreamed it would sound at home.
Back to the Future
Last December I began listening again seriously to my Harbeth Monitor 40.1s in my downstairs music room. The size of the presentation, the richness, the AUTHORITY and NATURALNESS of the big Harbeths! No, even dialed in to face the listening position they didn't image and stage as precisely as the Janszen Valentina Actives in my upstairs audio room. But they just sounded REAL, even spatially. It struck me that at some basic level, for all their clarity, precision, and seemingly natural tonality, the Janszens, by comparison, sounded artificial and small.
So, as has happened again and again since my acquisition of the original Harbeth Monitor 40s back so many years ago, speakers come, speakers go, but I keep coming back to the big Harbeths because they, and only they, seem to have the sound of real, live, unamplified music and have it in such an abundance, no matter the material. I bought my Monitor 40s around the year 2001, within a year or two after Robert E. Greene's (REG) seminal review of them in The Absolute Sound #116 (February/March 1999). I waited to buy mine until there was a North American distributor in the United States, and thus waited until Fidelis AV in New Hampshire (which remains the North American Harbeth distributor to this day) acquired distribution rights and established a dealer network so I could hear the Monitor 40 before making a purchase decision. Before that, Winter Tree Audio in Canada seemed to be the sole North American distributor/dealer for Harbeth, despite REG's walk-on-water review. Since I've owned the Harbeths, I've acquired and then sold:
AR-3a (vintage)
AR-5 (vintage)
AR-303a (vintage)
Rectilinear III (vintage)
KLH 12 (vintage)
Spendor SP1/2 (vintage)
EPI 100 (vintage)
Ohm Walsh 5 Mk III
Linkwitz Orion
Sanders 10C
Gradient 1.3
Gradient 1.5 Helsinki
Gradient Revolution Active + Gradient SW-T triple tower subwoofers
Stirling Broadcast LS3/6 + AudioKinesis Swarm
Now I've sold the Janszen Valentina Actives as well.
System Changes and Context
In my prior home, after my first wife passed away in 2010, I had up to nine complete serious audio systems at once in seven different rooms. Since I remarried in 2015, in our new home, I've had two rooms dedicated to audio for the past couple of years.
But now I've decided it's time to slim down further audio-wise. I have now consolidated to a single dedicated audio room. The lifestyle changes involved in now having a large family for the first time in my life (my wife is one of seven siblings, for example) means that even I need to limit the space dedicated to purely audio pursuits. And the rewards of having a large family are very much worth it. Thus, the downstairs Music Room where my Harbeth Monitor 40.1s resided is now a living room without any serious music system.
When the Monitor 40.1s were developed, I had the original pair which was making the rounds of reviewers (the same pair reviewed in Stereophile and TAS) in my home for a couple of weeks. As a result of that audition, I traded in my M40s through Fidelis AV for the Monitor 40.1s.
Some (REG, for instance) say that the original M40s are better than the subsequent 40.1s. REG objects to the Monitor 40.1's emphasis in response around 1 kHz which he thinks needs to be equalized out in order for the later speaker to be competitive in accurate reproduction with the earlier version. I disagree, finding this emphasis benefits the subjective results on most music even if the objective response measurement is a bit elevated in this region. One thing REG and I agree about as to the original M40 is that the tweeter guard on that model needs to be removed for best sound, sound competitive with the openness of the high frequencies of the later models. See my discussion "A Delicate Operation: Removing Tweeter Guards From Harbeth Monitor 40s."
But I certainly think that both the M40 and M40.1 are GREAT speakers, so who am I to argue with REG's preference? Even Paul Seydor, who reviewed the new Monitor 40.2 for The Absolute Sound says that REG’s superbly dialed in M40s yield a reproduction of symphonic music second to none in Seydor’s experience.
I'd heard the new Monitor 40.2 two years running at AXPONA in Chicago and was impressed at both hearings. Designer Alan Shaw definitely seems to be onto something in his new designs, something which allows his newest creations, the SLH5+ and now the M40.2, to sound simultaneously very accurate and yet forgiving enough to allow the 90% of recordings which are not sonic exemplars to still be deeply satisfying listening experiences. Now, with the help of Mike Kay at Harbeth dealer Audio Archon in Illinois, I've traded in my Monitor 40.1s for a new pair of Monitor 40.2s.
I have also now paired the M40.2s with the bespoke 14.25"-high Ton Trager Reference stands for the M40.2s. For the first couple of weeks with the M40.2s I was using my similarly sized tried-and-true Something Solid XF MkII stands I was using with the M40.1s. Power is from a brand new pair of Benchmark AHB2's run in bridged mono mode. I'm high on the Benchmark electronics right now, if you hadn't noticed.
Here's a rundown of the equipment I'm now using in my dedicated upstairs audio room (a converted bedroom):
Sources: Oppo UDP-205 for disc playback; Auralic Aries G2 for streaming Tidal, internet radio, and Airplay from my iPhone. Both are new to my room.
The update from my former EVS-modified Oppo BDP-105D means that I can no longer decode my Reference Recordings HDCD discs, but all of these are available decoded through Tidal. And since I exclusively use the HDMI output of the disc player so as to be able to input high-resolution disc sources (SACD, Blu-ray Audio) to my Benchmark DAC, I wanted the HDMI de-jittering feature which is new to the latest Oppo UDP-205.
Since I rely on streaming from the internet for a lot of my listening these days, I've chosen one of the top new internet streamers to replace my classic Logitech Squeezebox Touch and Apple Airport Express. See the linked Aries G2 thread for my comments on the improvements the Aries G2 brings about in streaming sound quality.
HDMI De-Embedder: Kanex Pro HDMI Audio De-Embedder (converts the HDMI output of the Oppo to coaxial digital for input the my Benchmark DAC).
DAC: Benchmark DAC-3 HGC (driving speaker amps) and Benchmark DAC-3 DX (driving headphone amp). One of the coax digital inputs of the HGC is configured by moving an internal jumper to be a digital pass-through output to the DX.
Speaker Amplifiers: A pair of Benchmark AHB2 amps, run in mono. I'll have more to say down the road about these amps in another thread.
Headphone Amp: SimAudio Moon Neo 430 HA, fed a balanced analog signal from the Benchmark DAC-3 DX. The DAC in the headphone amp is not used.
Headphones: Audeze LCD-4 with aftermarket balanced cabling in the form of the Moon Audio Silver Dragon premium cabling for the Audeze LCD series.
Speakers: Harbeth Monitor 40.2
Stands: The sources all sit centered atop an Ikea Lack table enhanced and leveled with felt pads under its legs plus Bright Star Little Rock damping weights (themselves with felt padding on the underside contacting the tabletop). The Oppo sits atop a large Little Rock which damps the top of the Ikea Lack table. A smaller Little Rock sits on the cover of the Oppo; it is small enough that it does not block all the ventilation holes of the Oppo Chassis. The Auralic Aries G2 sits atop that smaller Little Rock; that Little Rock is large enough for the footprint of the Aries G2. This position for the Auralic and Oppo provides the strongest wireless signal reception from the wireless Netgear Nighthawk X8 router (with Comcast Extreme 105 service) on the downstairs main floor almost directly under these units.
The headphone amp sits under the Ikea Lack table and atop a large Bright Star Little Rock which has three Bright Star Isonode sorbothane feet between it and the wood floor. The top of the headphone amp is damped with another Little Rock which also serves as the base for the two stacked Benchmark DACs. The top DAC is damped with an issue of The Absolute Sound magazine. The Kanex Pro HDMI De-Embedder sits atop this magazine on three small nylon button feet.
The Benchmark amplifiers each sit on separate Mapleshade 4"-thick maple platforms, themselves sitting on three Mapleshade Isoblock 1 rubber feet, two under the front two corners and one in the center rear. These small amps sit at the front of the maple platforms so that the cables attached to the rear are supported on the platforms and have enough room so as not to hit the acoustic panels behind the amps. The speaker cables and all other cables are also routed so as to avoid contact with the room's carpet to minimize electrostatic charges on the cabling.
The speakers now sit on the bespoke Ton Trager Reference stands for the Harbeth M40.2 speakers.
Cabling: With one exception, the cabling is either from Blue Jeans Cable or Benchmark. The exception is the Oyaide NEO d+ Class A Rev 2 USB cable connecting the Aries G2 to the Benchmark DAC-3 HGC.
The Oppo is connected to the DAC-3 HGC via the Oppo's audio-only HDMI 2 output by Blue Jeans Belden Series FE HDMI cable into the Kanex Pro HDMI Audio De-Embedder. The output of the Kanex to the DAC-3 HGC uses Blue Jeans Cable coaxial digital cable.
The two Benchmark DACs are connected by a Blue Jeans Cable coaxial digital cable. The balanced analog output cables from both DACs are by Benchmark.
The speakers cables are Benchmark, with Speak-On connectors at the amp end and locking banana plugs at the speaker ends, the configuration Benchmark recommends for its AHB2 amp when used with non-Benchmark speakers.
Gain Structure: To maximize signal-to-noise ratio and minimize distortion through optimized system gain structure, the Benchmark DAC-3 HCG is used without output padding so it is putting out a high pro-audio-level signal and the Benchmark amps are used at their low-gain setting.
Electrical: All electronics are driven from two dedicated 20-amp circuits (one for the amps, another for lower-draw equipment), each of which feeds a single quad of wall outlets behind the audio equipment. Both of these circuits are fed from the same phase of the home's 220-volt service.
Acoustic Room Treatment: A combination of 4"-thick female-pattern Sonex and Pi Audio AQD Diffusers are used. The diffusers are at the first reflection points of the speakers on the side walls and wall behind the speakers as seen from the listening seat. The Sonex is used for reflection control in the corners behind the speakers, on the wall behind the listening seat to cover the first reflection points, and on the ceiling first reflection points. Wooden blinds cover the room's window and the wooden furniture and CD racks in front of the speakers is arranged to be at least five feet from the speakers. The listening position is also at least five feet from any of the walls or furniture in front of the speakers; this minimizes "early" reflections audible from the listening position.
Other Tweaks: All non-soldered electrical contacts I can reach are treated with Caig Audio DeOxit Gold GL100, the brush-on liquid stuff, not any of the spray varieties. Most equipment (Squeezebox Touch and speaker terminals excepted) have EVS Ground Enhancers added.
Further details about the equipment and set-up are discussed in other threads in Tom's Corner, including:
Kanex Pro HDMI Audio De-Embedder
Sennheiser HD 800 S Headphones + SimAudio Moon Neo 430 HA Headphone Amplifier
Apple Airport Express as Internet Audio Streaming Receiver
Benchmark DAC3 DX
The Lowly Toslink
EVS Oppo BDP-105 Mods, Ground Enhancers, Black Discus & Mounting Tweaks
Electronic Visionary Systems (EVS) Ground Enhancers: Can You Spare $30?
Contact Cleaning: The Right Stuff
If It's Spring, It Must Be Time for New Speakers (and More): Janszen Valentina Active
Electronic Equalization
Sure, in my small (11' x 13') audio room I may need to use electronic equalization to best tame the big Harbeth low end. We'll see; so far on most program material there is no obnoxious flatulence. Designer Alan Shaw has further domesticated the low end of his big guys. But, hey, I've equalized the Harbeth low end many times before and I've long since become an expert at flattening the bass of the 40.x with such EQ, if needed. I've also owned a lot of equalizers in my time, most of them since I've had the Harbeths:
Cello Palette Preamp
Z-Systems rdp-1
Legacy Steradian (for Legacy Whisper speakers)
Rives PARC (the first equalizer I used with the M40s)
Rane DEQ-60L
TacT RCS 2.2XP AAA, both stock and fully Maui-modded versions
Audient ASP231
DSPeaker Anti-Mode 2.0 Dual Core (both 2012 and 2013 models)
Behringer DCX2496 + DEQ2496 (also provided stereo shuffling processing)
ART EQ355
RoomPerfect (in Lyngdorf TDAI-2170)
Z-Systems rdq-1
DSPeaker X4 (coming soon, I’ve again been recently promised)
If I decide that I do need or would just like to try electronic equalization, my trusty Z-Systems rdq-1 is waiting in the wings if the DSPeaker X4 does not soon materialize at my doorstep.
The Ton Trager Stands
Harbeth designer Alan Shaw is not in the habit of recommending speaker stands. To my knowledge, the Ton Tragers are the first speaker stands to get a solid recommendation from him, a video recommendation, no less, which can be viewed here.
My M40.2s are set atop the Ton Trager stands per Ton Trager's recommendations here. Following those directions as closely as I can seemed to make a positive difference.
Without meticulous adherence to those directions, the sound I was getting with my Something Solid XF MkII was at least the overall equal of that produced with the Ton Tragers. Those preferring a bit more high frequency zip and air and the largest possible stage presentation and greatest feeling of envelopment from near-field listening will probably still prefer the Something Solids even if the directions are meticulously followed.
But get the speakers positioned just so atop the Ton Tragers and give the stand feet a few days to really settle into the carpet and what you get is better depth, increased mid and high frequency smoothness and realism, as well as increased natural detail with all traces of excess tizz eliminated. Bass seems deeper and better defined. Most of all, at high playback levels, the speakers seem cleaner. It sounds like room and/or speaker box resonances or overload are reduced. The stage shape and size changes more from one recording to another.
At just under $1,400 a pair, the Ton Trager stands are expensive. The Something Solid XF MkII, the next best stand by far which I have used with any of the M40 series, is still a close second and is much less expensive, less than 1/3 the cost, even counting shipping charges from British dealer Deco Audio.
Appearance-wise, yes, I suppose the Ton Tragers are more handsome. Certainly my wife thinks so. But there is something to be said for the vestigial appearance of the Something Solids with their much more open framework. That more open framework also makes it very simple to approximately adjust the toe-in of the speakers to get each toed in equally by just looking at the frame from the listening position.
I should note that these comments are in the context of not using spikes with either of these speakers stands. The stands "float" atop the carpet and carpet pad atop the wood floor.
I generally do not like the sound of spiking the big Harbeths or any other speakers, for that matter, to the floor. Not using spikes allows the carpet and weight of the speakers to considerably damp a metal-frame stand like the Something Solids compared to the awful metallic ringing you get when the metal stand is spiked to the floor. With the Something Solids the entire lower rectangle of the framework of the stand contacts the carpet. With the Ton Trager, only the extended tenon "tone bed" makes contact with either the floor or the speaker. When plucked or tapped with a finger with the M40.2s on board, both the Ton Trager and Something Solid stands resonate a bit. The Ton Tragers sound like wood with an apparently lower-frequency resonance, the Something Solids like metal with a higher-frequency resonance, but neither has a long resonant "tail" to the excitation produced when tapped or plucked.
Even with heavy speakers on board, the Something Solid stands are fairly easy to move on the carpet by small amounts to get the speakers adjusted just so with respect to the walls and listener. The much smaller contact points of the Ton Trager stands makes such movements considerably more difficult, but not impossible. You just have to lift up a bit and concentrate on moving a particular corner at a time, rather than swiveling or sliding the entire stand. Obviously the whole stand does in fact move; it's just the technique that is different.
I also should mention that the Ton Trager stands apparently were designed to be mounted directly on hardwood floors or, even better yet, on slate rock atop the hardwood floor. That is not possible in my room given the room size and desired carpet damping of the listening room floor. Anyway, in my experience, you definitely want a thick carpet and pad at the first floor reflection between you and the speaker drivers as viewed from the listening seat. Thus, even if you mount the speaker stands on a wood floor, you'd better have carpet begin very close to the stand and the carpet should extend from there to and through the entire listening area.
Actually, Sonex damping of the floor sounds more wonderful yet, but is a dangerous tripping hazard, especially in a darkened room. But in my younger and even more idealistic years, I sometimes had not only 4" Sonex covering the first floor reflection areas, but as much as 12" of Sonex damping the floor reflection. That one reflection is responsible for quite a lot of upper bass coloration, but to scotch such a reflection, you need a thick layer of acoustic foam. The carpet will damp the treble nasties, but not much else.
Speaker Set Up
Given this room's size and shape, I have found that the listener and speaker positions computed by the Rule of Thirds 29% Version tool to work very well with the previous speakers I've used in this room, so that is what I've started with using the M40.2s. With the room's 132" Main Wall width and 161" Side Wall length, that puts the center of the front baffle of the Harbeths 38 9/32" from the side walls and 46 11/16" from the wall behind the speakers. The listening position is 94 11/16 from the Main Wall behind the speakers, or 48" from the plane of the speakers. The speakers and listener form an equilateral triangle of about 55 7/16" on a side. This is near-field listening as I and most others define it. It keeps my head more than five feet away from the wall behind the listening position, which is important, in my experience, for the best spatial presentation. Bass room modes, while certainly both measurable and audible, are less than with most other arrangements and the spatial presentation is the best I've heard in this room.
On the Ton Trager or Something Solid stands, the tweeter center ends up about 40.25" above the floor (the tweeter is 26" above the bottom of the M40.2 cabinet). My velour-upholstered Drexel listening chair gives me a nominal very comfortable ear height of about 38.25". That puts my ears about 2" below the tweeter center. While Harbeth has long specified tweeter height as the proper listening axis, from the M40 on I've noted that 2" below the tweeter axis has always sounded best to me, especially for near-field listening. This puts the stage up a bit higher and provides significantly better height illusion, thus providing a superior sense of envelopment in the vertical dimension, as well as seeming to be the position where the tonal balance is most natural.
To get the Harbeths into position, I tape a strip of masking tape along the top front center edge of the flat part of the cabinet (the flat part begins about 3/8" behind the beveled front edge of the cabinet). I mark the center of the speaker on that masking tape and measure from the side and back walls to that mark with a tape measure and/or laser measuring device.
I adjust toe in so that the tweeters of left and right speakers point directly at their respective ears when I'm seated in the listening position. To do that, I temporarily remove the speaker grills and tape 2" circular flat mirrors directly below the tweeter guards. Then, with my head pointed straight forward, I look to the left at the left speaker with just my left eye and adjust toe in until I see my left ear's reflection centered in that mirror. For the right speaker with my head pointed straight forward, I look to the right at the right speaker with just my right eye and adjust toe in until I see my right ear's reflection centered in the mirror attached to the right speaker.
The trick to the positioning, and what takes awhile to accomplish, is to get three parameters—toe in, distance from side wall, and distance from the wall behind the speakers—all dialed in as closely as possible by moving the speaker stands just so, all without moving the speakers from their optimal position mounted on the stand itself as specified by Ton Trager's directions.
Once the speakers are thus positioned, I carefully put the grills back on the speakers, taking care not to move the speakers on their stands in doing so. This is more difficult than it sounds since the grill edges fit very tightly into a groove routed around the perimeter of the front baffle. Considerable pressure and a bit of grill bending must be done to get the grills back on. Removing the grills is at least equally problematic because of the tight fit. For both operations it helps to leave one arm atop the speaker cabinet and put some body weight on that arm while working the grill edges with my other hand.
Yes, I suppose that the speakers look better or at least as good with the grills off to show off more of the wonderful wood grain and to see the handsome Harbeth Monitor 40.2 badge. But, looks aside, sonically there is absolutely no contest. There are no sonic parameters improved by listening with the grills removed. Those who think the speakers sound better without their grills need to have their ears examined because they apparently are functionally deaf. The grills are an integral part of the design, making rather obvious contributions to both response smoothness and perhaps consequently the apparent detachment of the sound from the physical location of the speaker baffles.
Once the speakers are thus positioned, I also adjust the diffusers so that the center of each diffuser panel is at the first reflection spot of the nearest speaker front cabinet edge to that room boundary as seen from the listening position with a flat mirror attached to the relevant wall. Decades of experience with experimental placement of room damping and diffusing have shown me that these are the best spots to place the room treatment.
Sonic Evaluation
Bottom line at the Top: based on my early listening at home, to my ears, these are, far and away. the best yet of the M40 series from Harbeth. That makes the M40.2s overall the best speakers I've ever had in any of my home audio set ups. As I said at the very top of this thread, this is how I've always dreamed home speakers should sound.
Read Paul Seydor's review for The Absolute Sound. So far in my early experience with the M40.2s, I agree right on down the line with most of what he says. Below I'll try to mention some aspects not fully explored in Seydor's review.
Designer's Comments: Paul Seydor's review in TAS includes the transcript of an interview with Harbeth designer Alan Shaw. But I think if you want to get a fuller explanation of the evolution of the target frequency response and other aspects of the sound of the big M40.x speakers over the years, you should take a look at Shaw's comments on the Harbeth User Group. Start with post #60 in that thread, and also look at Shaw's subsequent comments in #87, 92, 94, and 146 in that thread.
It does appear that Alan Shaw pays a lot of attention to marketing feedback he gets from dealers, reviewers, and owners in terms of voicing his speakers. That could be good, in the sense of getting more information about how the product performs in a variety of real-world listening rooms other than recording studios. He does say that he designs in isolation with no one else having heard his designs before he finished them.
On the other hand, cynics may conclude that he may just want to grow his company and make some more money by giving his potential customers what he thinks they want. Having heard the evolution of the Harbeth sound over a couple of decades now, I'm inclined to side with those who find that evolution to be the designer's honest approach to making already fine products approach ever more closely the goal of having home speakers which sound like real music.
Measurements: Rather than show the effects of my room on low-end response, here is a link to the German AUDIO magazine's test report on the Harbeth Monitor 40.2. See the response curves in the left graph at the bottom of what is labeled page 24:
http://www.inputaudio.de/wp-content/uploads/2015/06/harbeth_m402_audio_test.pdf
To my knowledge, these are the only published frequency response measurements of the M40.2 to date. The axial response looks fine to me, like a target curve I would pick when applying an electronic equalizer to the sound measured from the listening position.
REG's response to seeing these measurements after having heard the M40.2s casually at Paul Seydor's house was, essentially:
Of course it looks nice--in a way. I am unperturbed by the bass rise. One assumes one has to control bass anyway and better a bit too much than too little to start with.
[F]urther up, if one looks carefully (compressed vertical scale as always in magazines) one sees a typical Harbeth recessed presence range. Between 1 and 2 kHz, there is what appears to be about a 2 dB shelf down. One might like this. The speaker sounds very easy on the ears. And the reviewer did like it[.] But I would probably pull the presence range up slightly just to get a little more brilliance(and truth) on violins. Still as speakers go it looks nice. (Sounds nice too along the lines suggested by the measurements--this is pretty much exactly what I would have expected from listening a bit)[.]
REG
PS: As such, speakers are not supposed to exhibit target curves above 500 Hz in their direct arrival. The roll off of highs etc in listening position measurement is generated by room effects (rolled down power response)[.] It is better to have a target curve sort of response than to have say a rising top but ideally one would like the speaker to be anechoic on axis neutral and the house curve to be generated by the house as it were. If you mess about with the on axis, one will hear it!
Near-Field Listening Coherence: First, never in my experience were large three-way speakers made which were so conducive to the type of near-field listening I do in my room (about 55 inches from the speaker drivers) as the big Harbeths. Frankly, very few large-ish multi-driver speakers have a degree of inter-driver coherence from this sort of listening distance which makes the speaker sound like a single driver.
The Stirling LS3/6 and Janszen Valentina Actives do okay from this sort of listening distance if one is very careful about getting the listening height correct and/or the speakers angled back just so. For best inter-driver coherence and tonal balance, the Stirling LS3/6s must be listened to with your ears level with the lower of the two tweeters. From close up on the recommended 400 mm-high stands, that requires either sitting on the floor or tilting the speakers back a bit. The Janszens are quite short and the middle of the electrostatic array is only about 22 inches above the floor, so it has to be tilted back a lot to get your ears on the proper axis from any reasonable listening height when you sit close to them.
The problem with tilting short-ish speakers back to get on the proper listening axis is that while the tilt back gets you properly balanced high frequencies, with most such speakers you will still be looking a bit "down" on the spatial presentation. The Janszens almost succeed in this respect despite their shortness since they have been designed to project images and a stage a bit above the top of the speakers and this design goal works pretty well even when listening closer than the recommended eight feet back.
One of the nice things about the big Harbeth speakers is that they are big. That implies that they are tall enough, when used on proper stands, to get the speakers high enough so that your ears will be in the right position with respect to the speakers for best coherence and tonal balance when sitting in a comfortable chair and with the speakers mounted vertically on their stands. Not having to tilt the speakers back, and the speakers having a traditional box shape makes for much less difficulty in positioning the speakers just so with a tape measure and/or laser measurement tool. A tilted-back speaker with an unusual shape like the Janszens makes getting the two speakers symmetrically positioned in the room with respect to walls and the listening position quite a bit more difficult.
As noted before, the small size of my listening room forces near-field listening. Speakers which do not perform optimally at listening distances of much less than 8 feet from the plane of the speakers—which is actually most speakers larger than mini-monitors—will not perform optimally in such a room. Yes, as explained above, I adjust things to maximize the inter-driver coherence from other speakers in this room, but there were still some audible compromises due to close-up listening which I was aware of but could mostly ignore, at least for awhile.
From past experience with the original M40, however, I KNOW that those speakers sound like ONE DRIVER from as close as 20 inches from the plane of the speakers, as long as the speakers are set up to point at my ears and the listening height is carefully chosen to be a couple of inches below the center of the tweeter. Four feet from the speaker plane is thus a piece of cake for the big Harbeths.
As coherent as the M40 was, the M40.1 was better yet. The front panel seemed to radiate sound all across its position, not at any driver locations at all.
But from the moment I put sound through the M40.2s, I knew that by further tweaking the crossover, designer Alan Shaw has wrought further clearly audible improvements in this seamless driver blending. Even from four feet from the speaker plane, the physical positions of the speakers truly disappear on well-recorded, and even much not-so-well-recorded, material, with sounds coming from various locations on the stage behind and in front of the speaker positions, depending on how close those sounds were to the microphones.
The inter-driver coherence in near-field listening like I use is at its quite considerable very best when the listening height puts your ears about two inches below the tweeter centers, as I described how to do in the set-up section. This height yields not only what I regard as the best overall subjective tonal balance, but also maximizes the smoothness of the transition from midrange to tweeter. From this listening height, the blend is seamless indeed. This blend was as audibly seamless as I'd ever heard even with the original 40 from such a listening position. The M40.1 was even better in this respect, and the M40.2 is—well—let's just call it truly seamless and truly undetectable.
No, as Paul Seydor's review mentions, the image height illusion is not the equal of true tall line sources, but from close up to the M40.2s, the speakers have a seeming line-source top-to-bottom radiation angle or space, creating lifelike image sizes, great depth, and wonderful immersion of the listener into the reproduced space of the performance. This quality supports and compliments the fabulous sense of "authority" for which the M40 series is justly famous.
Deep and Mid Bass: One of the first things I noticed compared to the M40.1 is that there is more bottom octave bass and less midbass with the M40.2. That is a very good thing for my smallish room. The specs say that the bass extends flat down another 5 Hz to 35 Hz. Subjectively, the bass warble tones on Stereophile's Test CD 2 sound pretty even down to 40 Hz, still strong at 30 Hz, but the 25 Hz and 20 Hz tones sound weak.
I know what true flat and even elevated 20-Hz-and-below bass extension sounds like in this room, having measured and equalized the response of the Stirling Broadcast LS3/6 plus AudioKinesis Swarm subwoofer array for flat or even elevated bass down to below 20 Hz. Yes, on a few pipe organ bass spectaculars, like the Dorian recording of Jean Guillou playing an organ transcription of Moussourgsky's Pictures at an Exhibition, the low notes in, for example, the Gnomus section are more awesomely rendered with superior room lock by great subwoofer support.
But absent that sort of material and direct comparison or good memory of how such sounds, you would never know anything was missing from the bottom octave with the M40.2. In contrast, the Janszen Valentina Active, while having good bass response in this room and never sounding thin, rolls off enough in the bottom octave to make the reproduction of such pipe organ spectaculars more polite and less than authoritative.
And Shaw has definitely tamed what many feel was truly excessive midbass for most real-world room set ups in the M40. In my opinion, without electronic equalization, very few domestic room set ups produce acceptably flat midbass with those speakers—REG is truly one of the fortunate few in this respect. The M40.1 was better, but still needed a favorable, bass-leaky room and ideal placement to avoid a peak of about 8 dB around 60 to 70 Hz.
The M40.2 has been further domesticated. Even in my small room, on most material the midbass is just concert-hall-naturally-full without electronic equalization, not overbearing in any way. I have not yet measured the bass response of the speakers as set up in this room, but I'd estimate a narrow band resonance in my room of about 4 dB in the midbass. This is most noticeable on closely miked acoustic bass solos with notes in that range. Outside that range, the walking bass notes are clearly differentiated and fairly evenly tempered as the bass playing moves up and down the scale.
The bass quality has a great combination of punch, detail, fullness, and warmth. It also will handle considerable power/SPL in this area cleanly. Bass quality is an area where the M40.2 clearly stands above its predecessors.
Authority: Several factors—the above-mentioned inter-driver coherence, the warm and rich frequency balance in the orchestral "power" range, the low distortion, great dynamic contrasts, the ability to play loudly without increased distortion, the generous depth and vertical height illusion—all contribute to reproduction which sounds naturally weighty and authoritative with all sorts of music, but especially acoustic unamplified instruments playing en masse, as with large-scale classical orchestral and choral pieces.
It is in this area where the contrast to the Janzen Valentina Actives I had in this room just before the Harbeths is most striking. With the Harbeths there is a lot of meat on the bone, a lot of gravitas where called for. Delicate sounds are suitably delicate, but with a fully-formed, lifelike balance which is in no way artificial, undersized, or lightweight. Glorious!
The Janszens are every bit as revealing about what is going on in the music both spatially and in terms of musical lines. And instruments sound natural enough. But there is something inherently small, light weight, and not forceful enough about their presentation, despite seemingly very high dynamic contrasts and plenty of bass. Many say that electrostatic speakers generally lack impact compared to good dynamic drivers. Maybe that's part of it. The Janszen presentation is exquisitely high on clarity and analysis without ever sounding analytical, but lower than it should be in terms of guts, heft, and authority.
The Midrange: Well, here it has all been said before. As Paul Seydor's review said, "how many variations can you ring upon 'beautiful,' luscious,' 'ravishing,' 'drop-dead gorgeous.'?" I would just add superbly natural and real sounding. If you know the sound of real acoustic instruments and voices in a favorable space, the Harbeth speakers, Harbeth M40 series in particular—and especially the M40.2—nails that sound as no other speakers I've heard do. Other speakers, like the Janszen Valentinas, Stirling LS3/6, or Gradient Revolution, which can for a time seem natural enough, will sound at least a bit artificial when compared to these M40.2s and the real thing. Most speakers, truth be told, sound QUITE artificial in comparison to these Harbeths and even the best of the rest just don't give you that disarming impression of "yes, this is the truth" about their sound. The impression on much material—not just a few audiophile specials—is overwhelmingly of the sound of actual players in front of you.
This is the part of the Harbeth sound which has been most important in repeatedly drawing me back to the M40 series year after year. It is in this area that they simply are not matched by any other speakers I've heard at any price.
As Seydor mentions, with the M40.2 this window on the all-important midrange is widened to include the upper bass/lower midrange all the way up through the presence range. And the perceived distortion is now lower than ever before, adding yet more realism. And, unlike the Quads, where later versions widened the magic midrange window at the expense of at least a bit of the magic the early ESL had in the heart of the midrange, nothing has been lost and much has been gained by the widening of this area of supreme naturalness in the M40.2.
Presence Range: The range from, say, 2 kHz to 5 kHz is largely responsible for how close we perceive the music as being to our listening position. The words "forward" and "recessed" are largely descriptive of the look of the frequency response graph of speakers in this frequency range. Elevated response in this area moves the apparent sound source forward, while depressed response in this area tends to back the apparent sound source off further from the listening position.
Because of the way most commercial recordings are miked, the traditional BBC speaker voicing had relaxed response through this region, with a depression typically measuring some two to five dB in this range or some part of it. Many classical music listeners believe that this response trough yields better overall realism from the majority of commercial recordings, the theory being that since most recordings involved miking from much closer than concert hall audience listening distances, the recordings have a bit of presence exaggeration built in and this response trough helps ameliorate that exaggeration.
With his earlier M40 and M40.1, Harbeth designer Alan Shaw definitely adhered to this "BBC dip" philosophy. With the M40.2, however, Shaw is on record as having decided to flatten out the presence range response as much as possible. Given the published test report measurements linked to above, there still seems to be a bit of presence-range recession/relaxation in the response. But this increased presence was the very first thing I noticed about the sound the new M40.2s in comparison to the M40.1 and original M40. Solo voices and instruments are more forward, not at all recessed.
But Shaw has managed a fine balancing act here between increased presence and flattering a wide range of recorded material. Through some legerdemain—or maybe it's just an upward extension of the "magic" which Harbeths have long had in the core of the midrange—this increased presence does not grate at all and does not in any way reduce the wonderful depth perspective which the M40 series has had from the start with classical music, large scale and otherwise. In other words, there is no sense of excess forwardness and no less depth of field apparent. Sound sources closer to the recording microphones appear closer to the listener on the sound stage, but sound at least as natural as before while instruments in the back of the ensemble are still waaaaay back there. There is thus actually an enhanced amount of depth information presented. Simply marvelous! This is the way things sound in a good hall from around audience Row 8 to 10: the violins are close but in no way screechy or unpleasantly close, while the woodwinds, brass, and tympani are way back in the next county.
I've noted the same positive effects on pop and jazz. Soloists pop with presence, but the ensemble depth is fully developed, making for a very involving and interesting presentation spatially.
Top Octaves: The M40.2s, like the earlier models in the M40 series, have truly extended, airy highs. There is no lack and no exaggeration here. The top two octaves are there in the proper proportions. The impact ting of sticks on cymbals and the following sheen and shimmer are there in the proper proportions in time, space, and frequency balance.
That was not the case, I'm afraid, either with the Stirling Broadcast LS3/6 or Janszen Valentina Actives. The stick impact was there, but the following sheen and shimmer were reduced, making percussive images that were lower in placement, not as spatially free, and sounding tonally a bit more like escaping steam from a valve. You'd think that electrostatics without any crossover between midrange and highs would get top octave air correct. You'd also think from REG's reviews and subsequent comments that the Stirling and Janszen both get this right. This just is not so, I'm now convinced, whether subjectively or measurably. Both the Janszen and Stirling measured as rolling off above 5 kHz from whatever distance or angle I placed the measuring microphone of the OmniMic V2 measuring system. Both measure down some 12 db by 20 kHz. I thought this was either a measuring artifact or a product of narrow dispersion in the highs.
But I knew that the same measuring system showed the M40.1 to have just a dB or two of gradual slope off above 5 kHz. And the highs of the M40.1 sounded realistically airy on everything from cymbals to violins, not to mention high trumpet note overtones. The M40.2 does just as well, if not better. See the linked published test results. The Harbeths get the balance between the stick impact and the following shimmer and sheen correct and generally have a very natural amount of airiness to their sound.
Judging by both ear and published measurements, many speakers designed in the past few years have excess air built into their design, showing a measured peak in response of 5 to even 10 dB somewhere above 5 kHz. Like REG (see his comments about the linked M40.2 test results above) I find this sort of response tailoring far more objectionable than the type of relatively benign smooth rolloff found in the Stirling LS3/6 and Janszen Valentina Active. A peak or roll-up in this area draws attention to the tweeter (the tweeter sound "sticks out") and in music with cymbals, draws them forward in the mix.
The Harbeth M40, M40.1, and M40.2 have each had a superb sense of blend and integration between the midrange and tweeter. You cannot hear the transition between the midrange and tweeter drivers and there is still the proper amount of high frequency air.
Low Distortion and Clarity: In this respect these Harbeths seem basically equal to the Janszens. While the Janszen Valentina Actives perhaps allow following musical lines in complex music to an even greater degree, these Harbeths are very close to that standard in that respect.
The Harbeths are superior to the Janszens and all other speakers I've owned or heard in allowing small details to emerge naturally without any frequency response peculiarities. Instruments sound both superbly natural/real and you can hear the small musical (and not-so-musical, such as breathing, chair noises, air handling roar and rumble, etc.) sounds with superb clarity. The clarity is not provided by any exaggeration of mids or highs or rolling off of bass or warmth ranges. The M40.2s fully match, and even sometimes exceed the level of detail audible through my Silver-Dragoned Audeze LCD-4 headphones, something I never thought any speakers could do until hearing the M40.2 in my room.
Distortion also seems at least as low as with the Janszens, which is as low as I've heard with speakers. Given decent program material, the sound is very, very clean, with no apparent distortion or noise. Part of this impression is probably the supremely low distortion and high signal to noise ratio of the combination of the Benchmark DAC-3 HGC and AHB2 amps when their gain structure is arranged to produce maximum signal-to-noise ratio and minimal distortion as I have done.
Low Level Listening & Correct SPL: Perhaps because of a combination of all the above factors, the Monitor 40.2s sound more realistic and satisfying at low volumes than any speakers I've ever used before. There is no need to listen at high volumes to achieve a natural low-frequency balance.
The proper or natural SPL for any given material is also very well defined for these speakers, more so than with any others I've used. That natural volume also seems to be at least a bit lower than with other speakers, allowing home listening to sound extremely well balanced tonally even without "blasting" the music. In this respect the speakers mimic the live unamplified concert hall experience better than other speakers. Unamplified acoustic music in concert rarely exceeds the 80 – 90 dB range from audience seats, even in loud portions. With the Harbeths, you get full concert hall realism at similar volumes.
This of course also means that neither the speakers or amps need be stressed by high SPL in order for the reproduction to sound "right." But, if you want to play your music very loud, the speakers are capable of delivering the goods quite well indeed.
High SPL Capability: Part of this is the Ton Trager stands, but even with the Something Solid stands it was quite obvious to me that the M40.2 will play at considerably higher SPLs without compressing or distorting on rock or large-scale jazz, for example, than either the M40.1 or M40. The M40.1 had definite limits which even my Lyngdorf SDA-2400 with its 200 watts per channel could begin to hit, not to mention the Sanders Magtech Monos with their 1600 watts per channel. Earlier speakers in the M40 speakers were natural sounding speakers at respectably high volumes up to and including the mid-90 dB range on peaks, but if pushed would start to complain at SPLs I'd classify as very loud where peaks measure about 100 dB or so.
I'm sure a bigger room would show the M40.2's limitations, but in my small room the M40.2s just get louder to higher SPLs than I care to listen for more than a few moments. They do this without any apparent increase in distortion. This applies from low bass to highest highs. They can do this without taxing my $6,000/pair Benchmark amps which are capable of clean power north of 400 watts per channel into the M40.2 load but are by no means the highest-power amps available.
This may not be important at all to classical music listeners. But if your tastes are eclectic like mine or are more in the rock or big-band jazz veins, the M40.2 can be immensely satisfying on such material at very high SPLs, at least in a smallish room like mine.
The Sum of the Parts: As good as the parts are when analyzed, the Harbeth Monitor 40.2s sound even more impressively natural and real when considered as a whole without any attempt at sonic analysis. That whole package tends to disarm analysis since it just sounds so real, so beautiful. The speaker does not favor certain music over other types; all sound incredibly life-like.
If your musical tastes are as eclectic as mine you will have found a willing partner for your musical explorations. These are speakers just made for today's world of millions of tracks available for internet streaming on demand. Even if you don't regard yourself as having eclectic musical tastes, I can guarantee that the M40.2s will encourage you to make new musical discoveries while it also brings yet deeper appreciation of old musical friends.
Best friends for your journey, that's what these are.
Sure, if Alan Shaw designs a Monitor 40.3 someday before he retires, I will investigate his latest thinking. But if retirement means I can no longer afford new speakers of this cost, as I said at the beginning, with the Monitor 40.2 I can sail happily through my Golden Years. They have the sound of music as I've always dreamed it would sound at home.
Back to the Future
Last December I began listening again seriously to my Harbeth Monitor 40.1s in my downstairs music room. The size of the presentation, the richness, the AUTHORITY and NATURALNESS of the big Harbeths! No, even dialed in to face the listening position they didn't image and stage as precisely as the Janszen Valentina Actives in my upstairs audio room. But they just sounded REAL, even spatially. It struck me that at some basic level, for all their clarity, precision, and seemingly natural tonality, the Janszens, by comparison, sounded artificial and small.
So, as has happened again and again since my acquisition of the original Harbeth Monitor 40s back so many years ago, speakers come, speakers go, but I keep coming back to the big Harbeths because they, and only they, seem to have the sound of real, live, unamplified music and have it in such an abundance, no matter the material. I bought my Monitor 40s around the year 2001, within a year or two after Robert E. Greene's (REG) seminal review of them in The Absolute Sound #116 (February/March 1999). I waited to buy mine until there was a North American distributor in the United States, and thus waited until Fidelis AV in New Hampshire (which remains the North American Harbeth distributor to this day) acquired distribution rights and established a dealer network so I could hear the Monitor 40 before making a purchase decision. Before that, Winter Tree Audio in Canada seemed to be the sole North American distributor/dealer for Harbeth, despite REG's walk-on-water review. Since I've owned the Harbeths, I've acquired and then sold:
AR-3a (vintage)
AR-5 (vintage)
AR-303a (vintage)
Rectilinear III (vintage)
KLH 12 (vintage)
Spendor SP1/2 (vintage)
EPI 100 (vintage)
Ohm Walsh 5 Mk III
Linkwitz Orion
Sanders 10C
Gradient 1.3
Gradient 1.5 Helsinki
Gradient Revolution Active + Gradient SW-T triple tower subwoofers
Stirling Broadcast LS3/6 + AudioKinesis Swarm
Now I've sold the Janszen Valentina Actives as well.
System Changes and Context
In my prior home, after my first wife passed away in 2010, I had up to nine complete serious audio systems at once in seven different rooms. Since I remarried in 2015, in our new home, I've had two rooms dedicated to audio for the past couple of years.
But now I've decided it's time to slim down further audio-wise. I have now consolidated to a single dedicated audio room. The lifestyle changes involved in now having a large family for the first time in my life (my wife is one of seven siblings, for example) means that even I need to limit the space dedicated to purely audio pursuits. And the rewards of having a large family are very much worth it. Thus, the downstairs Music Room where my Harbeth Monitor 40.1s resided is now a living room without any serious music system.
When the Monitor 40.1s were developed, I had the original pair which was making the rounds of reviewers (the same pair reviewed in Stereophile and TAS) in my home for a couple of weeks. As a result of that audition, I traded in my M40s through Fidelis AV for the Monitor 40.1s.
Some (REG, for instance) say that the original M40s are better than the subsequent 40.1s. REG objects to the Monitor 40.1's emphasis in response around 1 kHz which he thinks needs to be equalized out in order for the later speaker to be competitive in accurate reproduction with the earlier version. I disagree, finding this emphasis benefits the subjective results on most music even if the objective response measurement is a bit elevated in this region. One thing REG and I agree about as to the original M40 is that the tweeter guard on that model needs to be removed for best sound, sound competitive with the openness of the high frequencies of the later models. See my discussion "A Delicate Operation: Removing Tweeter Guards From Harbeth Monitor 40s."
But I certainly think that both the M40 and M40.1 are GREAT speakers, so who am I to argue with REG's preference? Even Paul Seydor, who reviewed the new Monitor 40.2 for The Absolute Sound says that REG’s superbly dialed in M40s yield a reproduction of symphonic music second to none in Seydor’s experience.
I'd heard the new Monitor 40.2 two years running at AXPONA in Chicago and was impressed at both hearings. Designer Alan Shaw definitely seems to be onto something in his new designs, something which allows his newest creations, the SLH5+ and now the M40.2, to sound simultaneously very accurate and yet forgiving enough to allow the 90% of recordings which are not sonic exemplars to still be deeply satisfying listening experiences. Now, with the help of Mike Kay at Harbeth dealer Audio Archon in Illinois, I've traded in my Monitor 40.1s for a new pair of Monitor 40.2s.
I have also now paired the M40.2s with the bespoke 14.25"-high Ton Trager Reference stands for the M40.2s. For the first couple of weeks with the M40.2s I was using my similarly sized tried-and-true Something Solid XF MkII stands I was using with the M40.1s. Power is from a brand new pair of Benchmark AHB2's run in bridged mono mode. I'm high on the Benchmark electronics right now, if you hadn't noticed.
Here's a rundown of the equipment I'm now using in my dedicated upstairs audio room (a converted bedroom):
Sources: Oppo UDP-205 for disc playback; Auralic Aries G2 for streaming Tidal, internet radio, and Airplay from my iPhone. Both are new to my room.
The update from my former EVS-modified Oppo BDP-105D means that I can no longer decode my Reference Recordings HDCD discs, but all of these are available decoded through Tidal. And since I exclusively use the HDMI output of the disc player so as to be able to input high-resolution disc sources (SACD, Blu-ray Audio) to my Benchmark DAC, I wanted the HDMI de-jittering feature which is new to the latest Oppo UDP-205.
Since I rely on streaming from the internet for a lot of my listening these days, I've chosen one of the top new internet streamers to replace my classic Logitech Squeezebox Touch and Apple Airport Express. See the linked Aries G2 thread for my comments on the improvements the Aries G2 brings about in streaming sound quality.
HDMI De-Embedder: Kanex Pro HDMI Audio De-Embedder (converts the HDMI output of the Oppo to coaxial digital for input the my Benchmark DAC).
DAC: Benchmark DAC-3 HGC (driving speaker amps) and Benchmark DAC-3 DX (driving headphone amp). One of the coax digital inputs of the HGC is configured by moving an internal jumper to be a digital pass-through output to the DX.
Speaker Amplifiers: A pair of Benchmark AHB2 amps, run in mono. I'll have more to say down the road about these amps in another thread.
Headphone Amp: SimAudio Moon Neo 430 HA, fed a balanced analog signal from the Benchmark DAC-3 DX. The DAC in the headphone amp is not used.
Headphones: Audeze LCD-4 with aftermarket balanced cabling in the form of the Moon Audio Silver Dragon premium cabling for the Audeze LCD series.
Speakers: Harbeth Monitor 40.2
Stands: The sources all sit centered atop an Ikea Lack table enhanced and leveled with felt pads under its legs plus Bright Star Little Rock damping weights (themselves with felt padding on the underside contacting the tabletop). The Oppo sits atop a large Little Rock which damps the top of the Ikea Lack table. A smaller Little Rock sits on the cover of the Oppo; it is small enough that it does not block all the ventilation holes of the Oppo Chassis. The Auralic Aries G2 sits atop that smaller Little Rock; that Little Rock is large enough for the footprint of the Aries G2. This position for the Auralic and Oppo provides the strongest wireless signal reception from the wireless Netgear Nighthawk X8 router (with Comcast Extreme 105 service) on the downstairs main floor almost directly under these units.
The headphone amp sits under the Ikea Lack table and atop a large Bright Star Little Rock which has three Bright Star Isonode sorbothane feet between it and the wood floor. The top of the headphone amp is damped with another Little Rock which also serves as the base for the two stacked Benchmark DACs. The top DAC is damped with an issue of The Absolute Sound magazine. The Kanex Pro HDMI De-Embedder sits atop this magazine on three small nylon button feet.
The Benchmark amplifiers each sit on separate Mapleshade 4"-thick maple platforms, themselves sitting on three Mapleshade Isoblock 1 rubber feet, two under the front two corners and one in the center rear. These small amps sit at the front of the maple platforms so that the cables attached to the rear are supported on the platforms and have enough room so as not to hit the acoustic panels behind the amps. The speaker cables and all other cables are also routed so as to avoid contact with the room's carpet to minimize electrostatic charges on the cabling.
The speakers now sit on the bespoke Ton Trager Reference stands for the Harbeth M40.2 speakers.
Cabling: With one exception, the cabling is either from Blue Jeans Cable or Benchmark. The exception is the Oyaide NEO d+ Class A Rev 2 USB cable connecting the Aries G2 to the Benchmark DAC-3 HGC.
The Oppo is connected to the DAC-3 HGC via the Oppo's audio-only HDMI 2 output by Blue Jeans Belden Series FE HDMI cable into the Kanex Pro HDMI Audio De-Embedder. The output of the Kanex to the DAC-3 HGC uses Blue Jeans Cable coaxial digital cable.
The two Benchmark DACs are connected by a Blue Jeans Cable coaxial digital cable. The balanced analog output cables from both DACs are by Benchmark.
The speakers cables are Benchmark, with Speak-On connectors at the amp end and locking banana plugs at the speaker ends, the configuration Benchmark recommends for its AHB2 amp when used with non-Benchmark speakers.
Gain Structure: To maximize signal-to-noise ratio and minimize distortion through optimized system gain structure, the Benchmark DAC-3 HCG is used without output padding so it is putting out a high pro-audio-level signal and the Benchmark amps are used at their low-gain setting.
Electrical: All electronics are driven from two dedicated 20-amp circuits (one for the amps, another for lower-draw equipment), each of which feeds a single quad of wall outlets behind the audio equipment. Both of these circuits are fed from the same phase of the home's 220-volt service.
Acoustic Room Treatment: A combination of 4"-thick female-pattern Sonex and Pi Audio AQD Diffusers are used. The diffusers are at the first reflection points of the speakers on the side walls and wall behind the speakers as seen from the listening seat. The Sonex is used for reflection control in the corners behind the speakers, on the wall behind the listening seat to cover the first reflection points, and on the ceiling first reflection points. Wooden blinds cover the room's window and the wooden furniture and CD racks in front of the speakers is arranged to be at least five feet from the speakers. The listening position is also at least five feet from any of the walls or furniture in front of the speakers; this minimizes "early" reflections audible from the listening position.
Other Tweaks: All non-soldered electrical contacts I can reach are treated with Caig Audio DeOxit Gold GL100, the brush-on liquid stuff, not any of the spray varieties. Most equipment (Squeezebox Touch and speaker terminals excepted) have EVS Ground Enhancers added.
Further details about the equipment and set-up are discussed in other threads in Tom's Corner, including:
Kanex Pro HDMI Audio De-Embedder
Sennheiser HD 800 S Headphones + SimAudio Moon Neo 430 HA Headphone Amplifier
Apple Airport Express as Internet Audio Streaming Receiver
Benchmark DAC3 DX
The Lowly Toslink
EVS Oppo BDP-105 Mods, Ground Enhancers, Black Discus & Mounting Tweaks
Electronic Visionary Systems (EVS) Ground Enhancers: Can You Spare $30?
Contact Cleaning: The Right Stuff
If It's Spring, It Must Be Time for New Speakers (and More): Janszen Valentina Active
Electronic Equalization
Sure, in my small (11' x 13') audio room I may need to use electronic equalization to best tame the big Harbeth low end. We'll see; so far on most program material there is no obnoxious flatulence. Designer Alan Shaw has further domesticated the low end of his big guys. But, hey, I've equalized the Harbeth low end many times before and I've long since become an expert at flattening the bass of the 40.x with such EQ, if needed. I've also owned a lot of equalizers in my time, most of them since I've had the Harbeths:
Cello Palette Preamp
Z-Systems rdp-1
Legacy Steradian (for Legacy Whisper speakers)
Rives PARC (the first equalizer I used with the M40s)
Rane DEQ-60L
TacT RCS 2.2XP AAA, both stock and fully Maui-modded versions
Audient ASP231
DSPeaker Anti-Mode 2.0 Dual Core (both 2012 and 2013 models)
Behringer DCX2496 + DEQ2496 (also provided stereo shuffling processing)
ART EQ355
RoomPerfect (in Lyngdorf TDAI-2170)
Z-Systems rdq-1
DSPeaker X4 (coming soon, I’ve again been recently promised)
If I decide that I do need or would just like to try electronic equalization, my trusty Z-Systems rdq-1 is waiting in the wings if the DSPeaker X4 does not soon materialize at my doorstep.
The Ton Trager Stands
Harbeth designer Alan Shaw is not in the habit of recommending speaker stands. To my knowledge, the Ton Tragers are the first speaker stands to get a solid recommendation from him, a video recommendation, no less, which can be viewed here.
My M40.2s are set atop the Ton Trager stands per Ton Trager's recommendations here. Following those directions as closely as I can seemed to make a positive difference.
Without meticulous adherence to those directions, the sound I was getting with my Something Solid XF MkII was at least the overall equal of that produced with the Ton Tragers. Those preferring a bit more high frequency zip and air and the largest possible stage presentation and greatest feeling of envelopment from near-field listening will probably still prefer the Something Solids even if the directions are meticulously followed.
But get the speakers positioned just so atop the Ton Tragers and give the stand feet a few days to really settle into the carpet and what you get is better depth, increased mid and high frequency smoothness and realism, as well as increased natural detail with all traces of excess tizz eliminated. Bass seems deeper and better defined. Most of all, at high playback levels, the speakers seem cleaner. It sounds like room and/or speaker box resonances or overload are reduced. The stage shape and size changes more from one recording to another.
At just under $1,400 a pair, the Ton Trager stands are expensive. The Something Solid XF MkII, the next best stand by far which I have used with any of the M40 series, is still a close second and is much less expensive, less than 1/3 the cost, even counting shipping charges from British dealer Deco Audio.
Appearance-wise, yes, I suppose the Ton Tragers are more handsome. Certainly my wife thinks so. But there is something to be said for the vestigial appearance of the Something Solids with their much more open framework. That more open framework also makes it very simple to approximately adjust the toe-in of the speakers to get each toed in equally by just looking at the frame from the listening position.
I should note that these comments are in the context of not using spikes with either of these speakers stands. The stands "float" atop the carpet and carpet pad atop the wood floor.
I generally do not like the sound of spiking the big Harbeths or any other speakers, for that matter, to the floor. Not using spikes allows the carpet and weight of the speakers to considerably damp a metal-frame stand like the Something Solids compared to the awful metallic ringing you get when the metal stand is spiked to the floor. With the Something Solids the entire lower rectangle of the framework of the stand contacts the carpet. With the Ton Trager, only the extended tenon "tone bed" makes contact with either the floor or the speaker. When plucked or tapped with a finger with the M40.2s on board, both the Ton Trager and Something Solid stands resonate a bit. The Ton Tragers sound like wood with an apparently lower-frequency resonance, the Something Solids like metal with a higher-frequency resonance, but neither has a long resonant "tail" to the excitation produced when tapped or plucked.
Even with heavy speakers on board, the Something Solid stands are fairly easy to move on the carpet by small amounts to get the speakers adjusted just so with respect to the walls and listener. The much smaller contact points of the Ton Trager stands makes such movements considerably more difficult, but not impossible. You just have to lift up a bit and concentrate on moving a particular corner at a time, rather than swiveling or sliding the entire stand. Obviously the whole stand does in fact move; it's just the technique that is different.
I also should mention that the Ton Trager stands apparently were designed to be mounted directly on hardwood floors or, even better yet, on slate rock atop the hardwood floor. That is not possible in my room given the room size and desired carpet damping of the listening room floor. Anyway, in my experience, you definitely want a thick carpet and pad at the first floor reflection between you and the speaker drivers as viewed from the listening seat. Thus, even if you mount the speaker stands on a wood floor, you'd better have carpet begin very close to the stand and the carpet should extend from there to and through the entire listening area.
Actually, Sonex damping of the floor sounds more wonderful yet, but is a dangerous tripping hazard, especially in a darkened room. But in my younger and even more idealistic years, I sometimes had not only 4" Sonex covering the first floor reflection areas, but as much as 12" of Sonex damping the floor reflection. That one reflection is responsible for quite a lot of upper bass coloration, but to scotch such a reflection, you need a thick layer of acoustic foam. The carpet will damp the treble nasties, but not much else.
Speaker Set Up
Given this room's size and shape, I have found that the listener and speaker positions computed by the Rule of Thirds 29% Version tool to work very well with the previous speakers I've used in this room, so that is what I've started with using the M40.2s. With the room's 132" Main Wall width and 161" Side Wall length, that puts the center of the front baffle of the Harbeths 38 9/32" from the side walls and 46 11/16" from the wall behind the speakers. The listening position is 94 11/16 from the Main Wall behind the speakers, or 48" from the plane of the speakers. The speakers and listener form an equilateral triangle of about 55 7/16" on a side. This is near-field listening as I and most others define it. It keeps my head more than five feet away from the wall behind the listening position, which is important, in my experience, for the best spatial presentation. Bass room modes, while certainly both measurable and audible, are less than with most other arrangements and the spatial presentation is the best I've heard in this room.
On the Ton Trager or Something Solid stands, the tweeter center ends up about 40.25" above the floor (the tweeter is 26" above the bottom of the M40.2 cabinet). My velour-upholstered Drexel listening chair gives me a nominal very comfortable ear height of about 38.25". That puts my ears about 2" below the tweeter center. While Harbeth has long specified tweeter height as the proper listening axis, from the M40 on I've noted that 2" below the tweeter axis has always sounded best to me, especially for near-field listening. This puts the stage up a bit higher and provides significantly better height illusion, thus providing a superior sense of envelopment in the vertical dimension, as well as seeming to be the position where the tonal balance is most natural.
To get the Harbeths into position, I tape a strip of masking tape along the top front center edge of the flat part of the cabinet (the flat part begins about 3/8" behind the beveled front edge of the cabinet). I mark the center of the speaker on that masking tape and measure from the side and back walls to that mark with a tape measure and/or laser measuring device.
I adjust toe in so that the tweeters of left and right speakers point directly at their respective ears when I'm seated in the listening position. To do that, I temporarily remove the speaker grills and tape 2" circular flat mirrors directly below the tweeter guards. Then, with my head pointed straight forward, I look to the left at the left speaker with just my left eye and adjust toe in until I see my left ear's reflection centered in that mirror. For the right speaker with my head pointed straight forward, I look to the right at the right speaker with just my right eye and adjust toe in until I see my right ear's reflection centered in the mirror attached to the right speaker.
The trick to the positioning, and what takes awhile to accomplish, is to get three parameters—toe in, distance from side wall, and distance from the wall behind the speakers—all dialed in as closely as possible by moving the speaker stands just so, all without moving the speakers from their optimal position mounted on the stand itself as specified by Ton Trager's directions.
Once the speakers are thus positioned, I carefully put the grills back on the speakers, taking care not to move the speakers on their stands in doing so. This is more difficult than it sounds since the grill edges fit very tightly into a groove routed around the perimeter of the front baffle. Considerable pressure and a bit of grill bending must be done to get the grills back on. Removing the grills is at least equally problematic because of the tight fit. For both operations it helps to leave one arm atop the speaker cabinet and put some body weight on that arm while working the grill edges with my other hand.
Yes, I suppose that the speakers look better or at least as good with the grills off to show off more of the wonderful wood grain and to see the handsome Harbeth Monitor 40.2 badge. But, looks aside, sonically there is absolutely no contest. There are no sonic parameters improved by listening with the grills removed. Those who think the speakers sound better without their grills need to have their ears examined because they apparently are functionally deaf. The grills are an integral part of the design, making rather obvious contributions to both response smoothness and perhaps consequently the apparent detachment of the sound from the physical location of the speaker baffles.
Once the speakers are thus positioned, I also adjust the diffusers so that the center of each diffuser panel is at the first reflection spot of the nearest speaker front cabinet edge to that room boundary as seen from the listening position with a flat mirror attached to the relevant wall. Decades of experience with experimental placement of room damping and diffusing have shown me that these are the best spots to place the room treatment.
Sonic Evaluation
Bottom line at the Top: based on my early listening at home, to my ears, these are, far and away. the best yet of the M40 series from Harbeth. That makes the M40.2s overall the best speakers I've ever had in any of my home audio set ups. As I said at the very top of this thread, this is how I've always dreamed home speakers should sound.
Read Paul Seydor's review for The Absolute Sound. So far in my early experience with the M40.2s, I agree right on down the line with most of what he says. Below I'll try to mention some aspects not fully explored in Seydor's review.
Designer's Comments: Paul Seydor's review in TAS includes the transcript of an interview with Harbeth designer Alan Shaw. But I think if you want to get a fuller explanation of the evolution of the target frequency response and other aspects of the sound of the big M40.x speakers over the years, you should take a look at Shaw's comments on the Harbeth User Group. Start with post #60 in that thread, and also look at Shaw's subsequent comments in #87, 92, 94, and 146 in that thread.
It does appear that Alan Shaw pays a lot of attention to marketing feedback he gets from dealers, reviewers, and owners in terms of voicing his speakers. That could be good, in the sense of getting more information about how the product performs in a variety of real-world listening rooms other than recording studios. He does say that he designs in isolation with no one else having heard his designs before he finished them.
On the other hand, cynics may conclude that he may just want to grow his company and make some more money by giving his potential customers what he thinks they want. Having heard the evolution of the Harbeth sound over a couple of decades now, I'm inclined to side with those who find that evolution to be the designer's honest approach to making already fine products approach ever more closely the goal of having home speakers which sound like real music.
Measurements: Rather than show the effects of my room on low-end response, here is a link to the German AUDIO magazine's test report on the Harbeth Monitor 40.2. See the response curves in the left graph at the bottom of what is labeled page 24:
http://www.inputaudio.de/wp-content/uploads/2015/06/harbeth_m402_audio_test.pdf
To my knowledge, these are the only published frequency response measurements of the M40.2 to date. The axial response looks fine to me, like a target curve I would pick when applying an electronic equalizer to the sound measured from the listening position.
REG's response to seeing these measurements after having heard the M40.2s casually at Paul Seydor's house was, essentially:
Of course it looks nice--in a way. I am unperturbed by the bass rise. One assumes one has to control bass anyway and better a bit too much than too little to start with.
[F]urther up, if one looks carefully (compressed vertical scale as always in magazines) one sees a typical Harbeth recessed presence range. Between 1 and 2 kHz, there is what appears to be about a 2 dB shelf down. One might like this. The speaker sounds very easy on the ears. And the reviewer did like it[.] But I would probably pull the presence range up slightly just to get a little more brilliance(and truth) on violins. Still as speakers go it looks nice. (Sounds nice too along the lines suggested by the measurements--this is pretty much exactly what I would have expected from listening a bit)[.]
REG
PS: As such, speakers are not supposed to exhibit target curves above 500 Hz in their direct arrival. The roll off of highs etc in listening position measurement is generated by room effects (rolled down power response)[.] It is better to have a target curve sort of response than to have say a rising top but ideally one would like the speaker to be anechoic on axis neutral and the house curve to be generated by the house as it were. If you mess about with the on axis, one will hear it!
Near-Field Listening Coherence: First, never in my experience were large three-way speakers made which were so conducive to the type of near-field listening I do in my room (about 55 inches from the speaker drivers) as the big Harbeths. Frankly, very few large-ish multi-driver speakers have a degree of inter-driver coherence from this sort of listening distance which makes the speaker sound like a single driver.
The Stirling LS3/6 and Janszen Valentina Actives do okay from this sort of listening distance if one is very careful about getting the listening height correct and/or the speakers angled back just so. For best inter-driver coherence and tonal balance, the Stirling LS3/6s must be listened to with your ears level with the lower of the two tweeters. From close up on the recommended 400 mm-high stands, that requires either sitting on the floor or tilting the speakers back a bit. The Janszens are quite short and the middle of the electrostatic array is only about 22 inches above the floor, so it has to be tilted back a lot to get your ears on the proper axis from any reasonable listening height when you sit close to them.
The problem with tilting short-ish speakers back to get on the proper listening axis is that while the tilt back gets you properly balanced high frequencies, with most such speakers you will still be looking a bit "down" on the spatial presentation. The Janszens almost succeed in this respect despite their shortness since they have been designed to project images and a stage a bit above the top of the speakers and this design goal works pretty well even when listening closer than the recommended eight feet back.
One of the nice things about the big Harbeth speakers is that they are big. That implies that they are tall enough, when used on proper stands, to get the speakers high enough so that your ears will be in the right position with respect to the speakers for best coherence and tonal balance when sitting in a comfortable chair and with the speakers mounted vertically on their stands. Not having to tilt the speakers back, and the speakers having a traditional box shape makes for much less difficulty in positioning the speakers just so with a tape measure and/or laser measurement tool. A tilted-back speaker with an unusual shape like the Janszens makes getting the two speakers symmetrically positioned in the room with respect to walls and the listening position quite a bit more difficult.
As noted before, the small size of my listening room forces near-field listening. Speakers which do not perform optimally at listening distances of much less than 8 feet from the plane of the speakers—which is actually most speakers larger than mini-monitors—will not perform optimally in such a room. Yes, as explained above, I adjust things to maximize the inter-driver coherence from other speakers in this room, but there were still some audible compromises due to close-up listening which I was aware of but could mostly ignore, at least for awhile.
From past experience with the original M40, however, I KNOW that those speakers sound like ONE DRIVER from as close as 20 inches from the plane of the speakers, as long as the speakers are set up to point at my ears and the listening height is carefully chosen to be a couple of inches below the center of the tweeter. Four feet from the speaker plane is thus a piece of cake for the big Harbeths.
As coherent as the M40 was, the M40.1 was better yet. The front panel seemed to radiate sound all across its position, not at any driver locations at all.
But from the moment I put sound through the M40.2s, I knew that by further tweaking the crossover, designer Alan Shaw has wrought further clearly audible improvements in this seamless driver blending. Even from four feet from the speaker plane, the physical positions of the speakers truly disappear on well-recorded, and even much not-so-well-recorded, material, with sounds coming from various locations on the stage behind and in front of the speaker positions, depending on how close those sounds were to the microphones.
The inter-driver coherence in near-field listening like I use is at its quite considerable very best when the listening height puts your ears about two inches below the tweeter centers, as I described how to do in the set-up section. This height yields not only what I regard as the best overall subjective tonal balance, but also maximizes the smoothness of the transition from midrange to tweeter. From this listening height, the blend is seamless indeed. This blend was as audibly seamless as I'd ever heard even with the original 40 from such a listening position. The M40.1 was even better in this respect, and the M40.2 is—well—let's just call it truly seamless and truly undetectable.
No, as Paul Seydor's review mentions, the image height illusion is not the equal of true tall line sources, but from close up to the M40.2s, the speakers have a seeming line-source top-to-bottom radiation angle or space, creating lifelike image sizes, great depth, and wonderful immersion of the listener into the reproduced space of the performance. This quality supports and compliments the fabulous sense of "authority" for which the M40 series is justly famous.
Deep and Mid Bass: One of the first things I noticed compared to the M40.1 is that there is more bottom octave bass and less midbass with the M40.2. That is a very good thing for my smallish room. The specs say that the bass extends flat down another 5 Hz to 35 Hz. Subjectively, the bass warble tones on Stereophile's Test CD 2 sound pretty even down to 40 Hz, still strong at 30 Hz, but the 25 Hz and 20 Hz tones sound weak.
I know what true flat and even elevated 20-Hz-and-below bass extension sounds like in this room, having measured and equalized the response of the Stirling Broadcast LS3/6 plus AudioKinesis Swarm subwoofer array for flat or even elevated bass down to below 20 Hz. Yes, on a few pipe organ bass spectaculars, like the Dorian recording of Jean Guillou playing an organ transcription of Moussourgsky's Pictures at an Exhibition, the low notes in, for example, the Gnomus section are more awesomely rendered with superior room lock by great subwoofer support.
But absent that sort of material and direct comparison or good memory of how such sounds, you would never know anything was missing from the bottom octave with the M40.2. In contrast, the Janszen Valentina Active, while having good bass response in this room and never sounding thin, rolls off enough in the bottom octave to make the reproduction of such pipe organ spectaculars more polite and less than authoritative.
And Shaw has definitely tamed what many feel was truly excessive midbass for most real-world room set ups in the M40. In my opinion, without electronic equalization, very few domestic room set ups produce acceptably flat midbass with those speakers—REG is truly one of the fortunate few in this respect. The M40.1 was better, but still needed a favorable, bass-leaky room and ideal placement to avoid a peak of about 8 dB around 60 to 70 Hz.
The M40.2 has been further domesticated. Even in my small room, on most material the midbass is just concert-hall-naturally-full without electronic equalization, not overbearing in any way. I have not yet measured the bass response of the speakers as set up in this room, but I'd estimate a narrow band resonance in my room of about 4 dB in the midbass. This is most noticeable on closely miked acoustic bass solos with notes in that range. Outside that range, the walking bass notes are clearly differentiated and fairly evenly tempered as the bass playing moves up and down the scale.
The bass quality has a great combination of punch, detail, fullness, and warmth. It also will handle considerable power/SPL in this area cleanly. Bass quality is an area where the M40.2 clearly stands above its predecessors.
Authority: Several factors—the above-mentioned inter-driver coherence, the warm and rich frequency balance in the orchestral "power" range, the low distortion, great dynamic contrasts, the ability to play loudly without increased distortion, the generous depth and vertical height illusion—all contribute to reproduction which sounds naturally weighty and authoritative with all sorts of music, but especially acoustic unamplified instruments playing en masse, as with large-scale classical orchestral and choral pieces.
It is in this area where the contrast to the Janzen Valentina Actives I had in this room just before the Harbeths is most striking. With the Harbeths there is a lot of meat on the bone, a lot of gravitas where called for. Delicate sounds are suitably delicate, but with a fully-formed, lifelike balance which is in no way artificial, undersized, or lightweight. Glorious!
The Janszens are every bit as revealing about what is going on in the music both spatially and in terms of musical lines. And instruments sound natural enough. But there is something inherently small, light weight, and not forceful enough about their presentation, despite seemingly very high dynamic contrasts and plenty of bass. Many say that electrostatic speakers generally lack impact compared to good dynamic drivers. Maybe that's part of it. The Janszen presentation is exquisitely high on clarity and analysis without ever sounding analytical, but lower than it should be in terms of guts, heft, and authority.
The Midrange: Well, here it has all been said before. As Paul Seydor's review said, "how many variations can you ring upon 'beautiful,' luscious,' 'ravishing,' 'drop-dead gorgeous.'?" I would just add superbly natural and real sounding. If you know the sound of real acoustic instruments and voices in a favorable space, the Harbeth speakers, Harbeth M40 series in particular—and especially the M40.2—nails that sound as no other speakers I've heard do. Other speakers, like the Janszen Valentinas, Stirling LS3/6, or Gradient Revolution, which can for a time seem natural enough, will sound at least a bit artificial when compared to these M40.2s and the real thing. Most speakers, truth be told, sound QUITE artificial in comparison to these Harbeths and even the best of the rest just don't give you that disarming impression of "yes, this is the truth" about their sound. The impression on much material—not just a few audiophile specials—is overwhelmingly of the sound of actual players in front of you.
This is the part of the Harbeth sound which has been most important in repeatedly drawing me back to the M40 series year after year. It is in this area that they simply are not matched by any other speakers I've heard at any price.
As Seydor mentions, with the M40.2 this window on the all-important midrange is widened to include the upper bass/lower midrange all the way up through the presence range. And the perceived distortion is now lower than ever before, adding yet more realism. And, unlike the Quads, where later versions widened the magic midrange window at the expense of at least a bit of the magic the early ESL had in the heart of the midrange, nothing has been lost and much has been gained by the widening of this area of supreme naturalness in the M40.2.
Presence Range: The range from, say, 2 kHz to 5 kHz is largely responsible for how close we perceive the music as being to our listening position. The words "forward" and "recessed" are largely descriptive of the look of the frequency response graph of speakers in this frequency range. Elevated response in this area moves the apparent sound source forward, while depressed response in this area tends to back the apparent sound source off further from the listening position.
Because of the way most commercial recordings are miked, the traditional BBC speaker voicing had relaxed response through this region, with a depression typically measuring some two to five dB in this range or some part of it. Many classical music listeners believe that this response trough yields better overall realism from the majority of commercial recordings, the theory being that since most recordings involved miking from much closer than concert hall audience listening distances, the recordings have a bit of presence exaggeration built in and this response trough helps ameliorate that exaggeration.
With his earlier M40 and M40.1, Harbeth designer Alan Shaw definitely adhered to this "BBC dip" philosophy. With the M40.2, however, Shaw is on record as having decided to flatten out the presence range response as much as possible. Given the published test report measurements linked to above, there still seems to be a bit of presence-range recession/relaxation in the response. But this increased presence was the very first thing I noticed about the sound the new M40.2s in comparison to the M40.1 and original M40. Solo voices and instruments are more forward, not at all recessed.
But Shaw has managed a fine balancing act here between increased presence and flattering a wide range of recorded material. Through some legerdemain—or maybe it's just an upward extension of the "magic" which Harbeths have long had in the core of the midrange—this increased presence does not grate at all and does not in any way reduce the wonderful depth perspective which the M40 series has had from the start with classical music, large scale and otherwise. In other words, there is no sense of excess forwardness and no less depth of field apparent. Sound sources closer to the recording microphones appear closer to the listener on the sound stage, but sound at least as natural as before while instruments in the back of the ensemble are still waaaaay back there. There is thus actually an enhanced amount of depth information presented. Simply marvelous! This is the way things sound in a good hall from around audience Row 8 to 10: the violins are close but in no way screechy or unpleasantly close, while the woodwinds, brass, and tympani are way back in the next county.
I've noted the same positive effects on pop and jazz. Soloists pop with presence, but the ensemble depth is fully developed, making for a very involving and interesting presentation spatially.
Top Octaves: The M40.2s, like the earlier models in the M40 series, have truly extended, airy highs. There is no lack and no exaggeration here. The top two octaves are there in the proper proportions. The impact ting of sticks on cymbals and the following sheen and shimmer are there in the proper proportions in time, space, and frequency balance.
That was not the case, I'm afraid, either with the Stirling Broadcast LS3/6 or Janszen Valentina Actives. The stick impact was there, but the following sheen and shimmer were reduced, making percussive images that were lower in placement, not as spatially free, and sounding tonally a bit more like escaping steam from a valve. You'd think that electrostatics without any crossover between midrange and highs would get top octave air correct. You'd also think from REG's reviews and subsequent comments that the Stirling and Janszen both get this right. This just is not so, I'm now convinced, whether subjectively or measurably. Both the Janszen and Stirling measured as rolling off above 5 kHz from whatever distance or angle I placed the measuring microphone of the OmniMic V2 measuring system. Both measure down some 12 db by 20 kHz. I thought this was either a measuring artifact or a product of narrow dispersion in the highs.
But I knew that the same measuring system showed the M40.1 to have just a dB or two of gradual slope off above 5 kHz. And the highs of the M40.1 sounded realistically airy on everything from cymbals to violins, not to mention high trumpet note overtones. The M40.2 does just as well, if not better. See the linked published test results. The Harbeths get the balance between the stick impact and the following shimmer and sheen correct and generally have a very natural amount of airiness to their sound.
Judging by both ear and published measurements, many speakers designed in the past few years have excess air built into their design, showing a measured peak in response of 5 to even 10 dB somewhere above 5 kHz. Like REG (see his comments about the linked M40.2 test results above) I find this sort of response tailoring far more objectionable than the type of relatively benign smooth rolloff found in the Stirling LS3/6 and Janszen Valentina Active. A peak or roll-up in this area draws attention to the tweeter (the tweeter sound "sticks out") and in music with cymbals, draws them forward in the mix.
The Harbeth M40, M40.1, and M40.2 have each had a superb sense of blend and integration between the midrange and tweeter. You cannot hear the transition between the midrange and tweeter drivers and there is still the proper amount of high frequency air.
Low Distortion and Clarity: In this respect these Harbeths seem basically equal to the Janszens. While the Janszen Valentina Actives perhaps allow following musical lines in complex music to an even greater degree, these Harbeths are very close to that standard in that respect.
The Harbeths are superior to the Janszens and all other speakers I've owned or heard in allowing small details to emerge naturally without any frequency response peculiarities. Instruments sound both superbly natural/real and you can hear the small musical (and not-so-musical, such as breathing, chair noises, air handling roar and rumble, etc.) sounds with superb clarity. The clarity is not provided by any exaggeration of mids or highs or rolling off of bass or warmth ranges. The M40.2s fully match, and even sometimes exceed the level of detail audible through my Silver-Dragoned Audeze LCD-4 headphones, something I never thought any speakers could do until hearing the M40.2 in my room.
Distortion also seems at least as low as with the Janszens, which is as low as I've heard with speakers. Given decent program material, the sound is very, very clean, with no apparent distortion or noise. Part of this impression is probably the supremely low distortion and high signal to noise ratio of the combination of the Benchmark DAC-3 HGC and AHB2 amps when their gain structure is arranged to produce maximum signal-to-noise ratio and minimal distortion as I have done.
Low Level Listening & Correct SPL: Perhaps because of a combination of all the above factors, the Monitor 40.2s sound more realistic and satisfying at low volumes than any speakers I've ever used before. There is no need to listen at high volumes to achieve a natural low-frequency balance.
The proper or natural SPL for any given material is also very well defined for these speakers, more so than with any others I've used. That natural volume also seems to be at least a bit lower than with other speakers, allowing home listening to sound extremely well balanced tonally even without "blasting" the music. In this respect the speakers mimic the live unamplified concert hall experience better than other speakers. Unamplified acoustic music in concert rarely exceeds the 80 – 90 dB range from audience seats, even in loud portions. With the Harbeths, you get full concert hall realism at similar volumes.
This of course also means that neither the speakers or amps need be stressed by high SPL in order for the reproduction to sound "right." But, if you want to play your music very loud, the speakers are capable of delivering the goods quite well indeed.
High SPL Capability: Part of this is the Ton Trager stands, but even with the Something Solid stands it was quite obvious to me that the M40.2 will play at considerably higher SPLs without compressing or distorting on rock or large-scale jazz, for example, than either the M40.1 or M40. The M40.1 had definite limits which even my Lyngdorf SDA-2400 with its 200 watts per channel could begin to hit, not to mention the Sanders Magtech Monos with their 1600 watts per channel. Earlier speakers in the M40 speakers were natural sounding speakers at respectably high volumes up to and including the mid-90 dB range on peaks, but if pushed would start to complain at SPLs I'd classify as very loud where peaks measure about 100 dB or so.
I'm sure a bigger room would show the M40.2's limitations, but in my small room the M40.2s just get louder to higher SPLs than I care to listen for more than a few moments. They do this without any apparent increase in distortion. This applies from low bass to highest highs. They can do this without taxing my $6,000/pair Benchmark amps which are capable of clean power north of 400 watts per channel into the M40.2 load but are by no means the highest-power amps available.
This may not be important at all to classical music listeners. But if your tastes are eclectic like mine or are more in the rock or big-band jazz veins, the M40.2 can be immensely satisfying on such material at very high SPLs, at least in a smallish room like mine.
The Sum of the Parts: As good as the parts are when analyzed, the Harbeth Monitor 40.2s sound even more impressively natural and real when considered as a whole without any attempt at sonic analysis. That whole package tends to disarm analysis since it just sounds so real, so beautiful. The speaker does not favor certain music over other types; all sound incredibly life-like.
If your musical tastes are as eclectic as mine you will have found a willing partner for your musical explorations. These are speakers just made for today's world of millions of tracks available for internet streaming on demand. Even if you don't regard yourself as having eclectic musical tastes, I can guarantee that the M40.2s will encourage you to make new musical discoveries while it also brings yet deeper appreciation of old musical friends.
Best friends for your journey, that's what these are.
Sure, if Alan Shaw designs a Monitor 40.3 someday before he retires, I will investigate his latest thinking. But if retirement means I can no longer afford new speakers of this cost, as I said at the beginning, with the Monitor 40.2 I can sail happily through my Golden Years. They have the sound of music as I've always dreamed it would sound at home.
