During and after Munich. At the Munich High End 2016, Gilles shared a room with Swissonor. Perhaps alternating demonstrations between Swissonor and Leedh speakers wasn't the best strategy for broad recognition during an event where visitors are bombarded by solicitations. Nonetheless, I honestly thought that the demo of the Leedh E2 Glass prototype was best sound of show. Gilles Milot spent a full night to set up and fine-tune the system. The decision to use two subwoofers placed outside the main speakers had a huge impact on the overall result. To my ears, this was the only room capable of throwing the immersive sonic bubble I had previously seen during my very best multi-channel experiences. This soundstage was close to perfection. Considering how it was achieved using Gilles' personal DIY amplification and digital source merely heightened the key contribution of the Leedh E2 Glass prototypes.

So that was something quite unique. Milot succeeded in creating a top performance system which was at once sonically and physically invisible. I cannot say if on a trade show basis this exhibit ended positively or negatively for our maker when most visitors here were completely convinced that this unusual performance was due to the "bigger" far more visible Swissonor speakers. They remained entirely ignorant of the other pair of speakers present. Most of the times when someone from the Leedh staff informed them that the sound in fact was emitted by these alien-looking speakers, they were so surprised that many decided to leave immediately. Quelle horreur. It seemed they were confused, aghast and unprepared for such a strange phenomenon. So much for the negative effect. The positive fact was that no other loudspeaker at the MOC could create this quite perfect holography. Of course, holography isn't the sole and unique criterion whereby to assess a loudspeaker's performance. Other competitors undoubtedly achieved better tonal balance, greater dynamics and higher resolution. But my opinion about the overall performance remains unchanged: this year Acoustical Beauty had the best sound at the MOC. And I truly think that this level of soundstage quality remains out of reach for any electrostatic, ribbon or planar effort, period.
This Munich demo put an end to the glass prototype adventure. Two months later Gilles Milot came to me with his first pair fully kitted out in glass diaphragms. It took some time to finalize the manufacturing process of the tweeter membranes and it shall remain a very labour-intensive manual process. Each membrane is produced then baked at a high temperature one at a time. Initially 16 diaphragms went into the furnace. After eight hours, only 4 were usable. The current process requires less baking time but can only produce one membrane per cycle. Perhaps a more sophisticated industrial oven might achieve higher quantities but the cost for such equipment would be completely unsuited for such a limited production endeavour. Which gets us to the real point: how did the final product compare to the prototype?


This time my enthusiasm didn't compare to my previous experience because I more or less knew what to expect. I was convinced that the new tweeter would increase subjective bandwidth, dynamics and detail in the low end. It would also create even greater cohesion since all membranes would be identical - nothing mysterious there considering the effect of the glass membrane on the other drivers. But mere confirmation of my expectations was a superb achievement given how very ambitious the initial promise had been. Let's briefly revisit the cornerstones of the design. Milot's specific goals were to reduce as much as possible all parasitic vibrations and to assert full control over his directivity. A byproduct of these goals is a substantial lowering of harmonic distortion (which usually leads to a loss of timbre, unpleasant saturation and listening fatigue) to approach levels comparable to that of the best amplifiers.

His sealed tubes exploit air compression for acoustic suspension and set the resonant frequency of the driver at 90Hz. Due to very high magnetic damping, this works at a very shallow slope of 3dB/octave at 100Hz up to 12dB/octave below 40Hz to insure the membrane's constant excursion. The Leedh tweeter is a more conventional dome diaphragm but still a 100-micron glass membrane. It too uses an iron-free motor whereas the ferrofluid joint and piston tube are no longer applicable to such a small driver. Another important goal for the Leedh E2 was, again, directivity. At low frequencies where the wavelength is greater than the distance of the speaker from the walls, boundary proximity is an advantage because it strengthens the sound wave. Conversely, at middle and high frequencies where the wavelength is less than the distance of the speaker to the walls, this proximity becomes a disadvantage because it tends to excite the room with unwanted colorations and parasitic reverberations and resonances. As loudspeakers are not made to be used inside an anechoic environment, working on progressive directivity is a very relevant achievement.


Unlike usual loudspeakers which rely on the geometry of their baffle, the spatial organization of the full set of HPAB drivers allows the Leedh E2 to be omnidirectional up to 200Hz, then gradually decrease its directionality with frequency, reaching a solid 90° (±45° around the listening axis) for a constant response of about 400Hz to 20kHz with a progressive fade over the listening angle (-3dB at ±30° and - 5dB at ±45°). This combined with a perfect impulse response allows a natural reproduction of tones and total stability of the sound space without requiring one to be perfectly positioned in the sweet spot. Furthermore, on many floorstanding speakers where woofers are classically arranged at ear height (about 1m), there is usually a cancellation between 100Hz and 1kHz. This phenomenon—the result of the summation of direct and ground-reflected wave—cancels certain frequencies and creates nonlinear issues over this part of the bandwidth. On the Leedh E2, the set of low-frequency drivers is close to the ground to avoid this cancellation effect, then further avoids nonlinearities due to their progressive filtering between 100Hz and 1kHz.


The Leedh E2 Glass structure was designed for minimal vibrations. Parasitic resonances are a key issue when any diaphragm excursion causes mechanical stress that extends to the whole structure. Enclosure vibrations also were significantly reduced thanks to very small dimensions made with thick walls of ceramic composite and extensive internal bracing. Then there is the force-cancelling effect of back-to-back drivers in a push/push configuration. The same concept applies to the midrange module whose internal passive drivers absorbs spurious vibrations. Finally the resultant structure with its tubular supports is light and low mass to not store energies. The 20.1 subwoofer acts as a natural extension of the Leedh E2 speakers. Its goal consists of providing a wider soundstage and greater bass energy especially for electronic music. In fact, 20.1 means 20Hz at -1dB compared to -8dB for the Leedh E2 solo. The Leedh subwoofer is based on the isobaric principle and includes two internal 23cm Athom drivers in a push-push setup within a 16 litre enclosure; and two external professional Audax 38cm woofers in reverse mode to face the smaller ones also in push-push mode. Two 2-channel 300wpc Athom plate amplifiers drive all 4 woofers. Both pairs of 38+23cm woofers operate in push-pull so the volume of air between them remains unchanged.

The ideal subwoofer would use the biggest possible woofer within the widest possible acoustic load. To this end, here the bigger woofers working in open air are acoustically loaded by smaller heavier drivers of longer excursions loaded within a relatively small enclosure. The smaller Athom woofers exhibit exactly the same amplitude and phase response curves as the big Audax. Thus the latter operate exactly as though loaded by an infinite acoustic load whilst in reality their enclosure is terminated by the enclosure of the smaller internal drivers. The internal woofers create 15 times higher kinetic energy than the visible woofers. They require an excursion of ±10mm to deliver the same maximum SPL as the Leedh E2 drivers at 50Hz. Here the subwoofer cancels most of the negative impact of these high acoustic pressures released by the Athom drivers with the opposing force of the external woofers. Given the 50Hz resonant frequency and damping of the 38cm woofers, the power feeding the 4 drivers is filtered to achieve -1dB at 20Hz. Each woofer pair sees 300w/4Ω. This double amplification pursues minimum distortion for the bigger woofers which need 15 times less power than the inner ones. Gilles Milot considers his design to be very close to extremist Japanese audiophile systems based on in-wall 38cm woofers loaded by the volume of the room behind their wall. Of course the Leedh system offers rather higher flexibility than the type of audiophile madness which designs an entire house around the playback chain. Very low distortion levels allow a relatively high subwoofer cut-off at 70Hz with a shallow 12dB/octave slope. Thus the 20.1 subwoofer can reinforce the upper bass without adding too much coloration of its own. Toward this end, Gilles Milot proposes an optional 50Hz@6dB/oct. filter capacitor connected to the amplifier's output feeding the main speakers. The goal is to lighten their bandwidth burden when run with the sub/s.