The Sound in Tonal Terms
Let me start with the bass. This is the least distinctive part of the speaker, but it is in fact extremely good. “Gnomus” from Mussorgsky’s Pictures at an Exhibition arranged for pipe organ and played by Jean Guillou [Dorian] had not only the required power but also excellent definition. Pedal tones that often are undifferentiated rumbles became precise musical (and mechanical) items. These speakers supply full-range bass with superlative precision on their own and so do not need subwoofers. (The nominal -3dB point is 30Hz, but given room gain and the slow roll-off of sealed enclosures, these speakers are full-range.)
What’s needed here is power. The nominal sensitivity of the PX1s is 82dB—quite low! The otherwise excellent Benchmark AHD2 amplifier, with 100 watts into 8 ohms, 190 into 4 ohms, did not enable the PX1s to give their best; there was clipping on orchestral climaxes at even moderate levels. So I brought over my big Bryston 14 ST, which can pretty much drive anything, and it drove the PX1s without any fuss or bother. (The PX1s are rated to accept 1000W peak power so you are not likely to over-drive them, but, say, 250 watts on bass transients won’t be as loud as one might think.) Both the Benchmark and the Bryston have clipping indicators, so it was clear what was happening in metered as well as listening terms. The Sanders Magtech would have been another obvious choice, but I was using it back at my place. (The Muraudio review samples were in fact in Paul Seydor’s home, as I did not have the space for them at that moment and as PS was partnering in the review in any case. I am very familiar with PS’s listening room and system—we live not far apart and often listen together at his place—so this was not an issue. And PS was out of town during part of the review period, so I could indulge myself in listening without being a nuisance.)
One just has to face the fact that the PX1s are not terribly sensitive speakers, and you must give them the kind of amplification they need. In this price range, this does not seem a major issue since suitable amplifiers—suitable and then some!—are available at prices far lower than the speakers themselves.
The speakers’ maximum SPL output is rated at 105dB at 2 meters (this would be around a 200-watt input for a sensitivity of 82dB, if one takes these things at face value). This is loud, especially with an omni speaker. In practice, I never felt any need at all to play them louder than where they seemed happy playing. Due to the speakers’ omni pattern and resultant “direct arrival” loudness, this SPL is actually louder than its number suggests. In any case, 105 dB is plenty loud! Still, you do need a powerful amplifier to get out of them what they can offer. (Big bass notes can amount to a surprising lot of power for short times.)
Beyond the bass, two things were immediately striking. One was that the sound was extraordinarily smooth in both in-room and perceived response. In-room measurements were also much smoother than one usually finds. Every speaker is pushed around a little bit below around 300Hz by room effects (though the PX1s looked good from there down, as such things go), but from there on up the PX1s’ RTA measurement (1/6th-octave smoothing) was uncanny in its smoothness. It matched within 1dB a very gently sloping target curve, say a 2dB droop by 10kHz. There was a very subtle broadband lift at 1kHz and a small dip at 2kHz but effectively the speakers were in-room flat up to the usual (and desirable) roll-off of the very high treble. This is as good an in-room performance as I can recall ever seeing without DSP, and far, far better than most. Moreover, it was very stable over a variety of listener positions. Usually such super-smooth curves (e.g., regonaudio.com/Harbeth%20Monitor%2040.html) are obtainable only at a particular sweet spot, but the PX1s did the trick over a variety of listening positions. Large displacements vertically from the center of the electrostatic unit caused some irregularities and extra treble roll-off. But otherwise stability was the rule.
The slight tendency to relax a bit around 2kHz brings up another point: Above 1kHz the PX1s generate more diffuse than direct sound compared to most speakers, and the ear’s response to a diffuse field is quite different than the response to frontal arrival. This means that a speaker with more diffuse field will in fact sound different, other things being equal, than one with relatively more direct arrival and a less diffuse field. The nature of this difference is known, with the main distinction being is that around 3kHz there is a dip of about 5dB in the ear/brain’s diffuse-field response compared to frontal-arrival response. Physically both responses rise, but the diffuse field rises considerably less. So diffuse sound will have in effect an audible suck-out around 3kHz. This works the other way, too—when the response of a microphone that is picking up a diffuse field is played back frontally, it will have an apparent peak of about 5dB at 3kHz. This is the reason many speaker designers have found that a dip at 3kHz makes things “sound better” (cf. Siegfried Linkwitz’s website).
The effect of all this is that the PX1s, which are quite close to truly flat in in-room RTA measurement, sound pleasantly non-aggressive and natural in the 3kHz range compared to flat speakers with primarily frontal radiation. This is, in the case of recordings where a lot of diffuse field was recorded, a kind of higher truth—the playback resembles the sound the microphone picked up and hence sounds more natural.
This is not to say that the PX1s lack top-end sparkle. Sibilance in speech, for instance, is not lost. Nor is high percussion dulled. If anything, the speakers are a bit extroverted on top though not in any displeasing way. Note also that many speakers roll the RTA response off considerably earlier than the PX1s, as a result of their tweeters becoming beamy. Played at natural levels the PX1s give front-row-center sound in that sense.