This review page is supported in part by the sponsor whose ad is displayed above
Reviewer: Kari Nevalainen
Source: EMT 928 and 930; Verdier Nouvelle Platine; Lumiere/Ortofon/Denon cartridges on Ortofon and EMT arms; Audiomeca/Sentec for digital
Preamp/Integrated: Shindo and Marco and various vintage designs; Behringer 1024 digital equalizer
Amp: Shindo Montille EL84 p/p; Marco 6V6 p/p; various vintage designs; Sun Audio 2A3 SET [on loan]; Densen B-330 [on loan]; 200wpc Sony receiver
Speakers: open baffle Phy, Fostex, Seas and Stentorian single-driver designs
Cables: silver-coated copper
Room size: 4.5m x 7m
Review component retail: Euros 1,400/pr

I could commence from autumnal Sweden. The day was bright and I was doing my normal tour de Stockholm, popping in and out of local HiFi and record shops. I had short-listed just one thing before coming over. Musta been something like ten years ago. The thing in question? A pair of Fostex FE103 Sigmas. I hardly need to introduce the FE103 here. Since the famous article by Gerard Chretien in L'Audiophile (Nr. 31), it has generated a small cult following. Virtually dozens of cabinets/enclosures have been proposed for this 4" wide bandwidth paper cone driver - small bass reflex cabinets, rear-loaded horns and everything in-between. Don't believe me? Simply surf the web.

The shop had one pair of FE103Zs left mounted in a closed box that wasn't much bigger than my fist. The sound was accordingly. Still, I was determined to have a pair so I didn't much care. I bought what may well have been the very last pair of FE103Zs in Scandinavia. This particular model was already discontinued and replaced by another apparently less in demand by the DIY savants. During the years that followed, I did my share of fitting the FE103 into different enclosures; from a small open back cabinet to an open pipe transmission line. I even once cut a mounting hole on cardboard boxes.

However, I really shouldn't start this story in Stockholm/ Sweden. It would make more sense to start it in Langenargen/Germany. That's where the European
Triode Festival 2004 was held. ETF is an annual event that gathers the cream of non-mainstream triode lovers from all over Europe and abroad. It was at ETF 2004 where I first heard Cornu Compact Spiralhorns (hereinafter referred to as Cornu CS). Cornu's pater familias, Daniel Ciesinger, was present and most willing to answer questions about his brainchild.

Having fooled around with FE103Z in the past I was -- like everybody else in my shoes would have
been -- immediately captivated by this new-to-me concept of how this small Fostex wide-bandwidth driver could be properly harnessed to serve the music. Equally noticeable was -- and I will come back to this later -- that the CS had the skill, not at all native to most speakers, of evoking a spontaneous smile on everybody's face who heard it, even those whose otherwise would not have called the Cornu
or anything like it their cuppa tea. People grin after having seen The Texas Chainsaw Massacre too - but something tells me that different mental mechanisms are triggered in the latter case.

But I won't start this story in Germany either. A show's a show, a festival's a festival. Even though I believe the conditions at ETF (with the exception of background noise perhaps) were clearly more favorable than they would have been at any other HiFi-Show (the Cornu CS was on display in a very big room), no reviewer with any self-respect would make conclusive judgments on the basis of one public event. The Cornu CSs sounded great there but I had no means to put its performance into any kind of familiar perspective. Couple that with high concept originality and you can appreciate why my interest was aroused. Products that launch with a similar surprise factor are less and less probable in this day'n'age.

So, finally, I kick off this story in mid-April of 2005 when the Cornu CSs arrived at my door steps in Helsinki/Finland. Quite a bit of time has elapsed since. Still, as surely as parts of the reasons for the delay were personal, another part was objective. I hope this becomes clear as we proceed. I was advised to make sure that the speakers were in good condition upon arrival. The Cornu CS is made of 4mm birch plywood and corners especially are in danger if impacted. The carton was, however, well protected against such accidents, sporting steel reinforcements in each corner. Hence both speakers were in excellent condition when pulled out of their packaging.

What are they?
To describe what sort of speaker the Cornu CS should properly be called becomes easy and tough at once. The easy answer is to call them rear-loaded hornspeakers with a single driver. That means no crossover, low cone excursions due to horn-loading, less impedance fluctuations, less phasing - as well as time delay between the driver and the horn, horn resonances, potential intermodulation distortion and so on, with other respective pros and cons of such driver loading lining up on either side of the fence.

Horn-loaded single-driver speakers are, of course, nothing new. They have a pedigree that traces back to at least the Voigt corner horns of the 1930s. And the concept has not lost its popularity. What's incontestably unique about the Cornu CS is their unconventionally flat, wide and square (70cm x 70cm) shape for a hornspeaker. The shape is due to the rare type of hornloading employed: a four-part spiral [see above]! Speakers with a spiral horn of various types are not unknown to DIY nuts but to the best of my knowledge, there are very few -- if any -- commercial variants based on the spiral horn principle apart from Cornu CS. Spiral horns complicate matters to the 5th power on many levels. Let's thus look at the easy one of frequency response first.

From a certain frequency range on upwards as with all single-driver speakers, the sound becomes solely attributable to the transducer. With the CS, that frequency divider sits somewhere between 1-2kHz. Mounted in the center of the front baffle, the paper cone is said to extend to 23kHz, exhibit good sensitivity (90dB/1m/1W) and a non-reactive 8-ohm nominal amplifier load. Fostex calls its FE108EZ a full-range driver but I prefer the more realistic wide-bandwidth descriptor. No single driver in this world is full-range, i.e. covers 10 octaves with equal sensitivity. Period.

The diminutive Fostex feature a hyperbolic parabolic diaphragm. According to Fostex, "... the 3rd-order curved surface structure of this technology results in the ability to reproduce faster-rising sounds more responsively, the minimization of self-resonating frequencies and the total lack of sound coloring characteristics." The claimed rise times are probably also a function of the driver's low moving mass of 2.7 grams.

Furthermore, the unit has what Fostex dubs a "tangential edge structure". Joining up-rolls and down-rolls at their tangential surfaces while utilizing sophisticated model analysis to determine the optimum bonding points creates the edge. This edge is claimed to have a major contribution to the overall performance.

Now comes the hard part. Because of its small size/cone excursion (0.28 mm), the FE108EZ is not capable of moving enough air to produce the lowest frequencies with equal sensitivity and output as the higher ones. Hence the obvious need to load the unit. Why horn-loading? After all, the choice is not obvious. Think Super 3s by Omega Loudspeakers or Gala Solo by Reflexion Acoustique,
designs that also make use of Fostex 4-inchers but pursue different boxes and loading schemes.

Cornu's website points at the driver and its very efficient motor (a large ferrite magnet with 5.2 Tesla in the gap) to explain why hornloading is required for proper bass reproduction. Hornloading is often considered ideal to increase efficiency in the lowest frequencies when a driver's resonant frequency (Fs) is fairly high. Here it is 77Hz.

Hornloading leaves many options. Fostex itself recommends a more conventional folded horn cabinet for their FE103 and FE108. The Horn Shoppe and zHorn [below] have similar ideas. Interestingly enough, one of the main reasons why Cornu opted for a spiral horn is that it allowed for a compact speaker that doesn't occupy much space in small rooms. Nevertheless, I doubt that spiral loading would have been chosen simply for the sake of size if either a folded or exponential horn would have produced a smoother response. There must have been more to it than size.

However, to further reduce size and truly become deserving of the compact moniker, the Cornu CS doesn't employ one long spiral horn but instead, four small spiral horns whose mouths vent sideways. Each horn is of a different length, the idea being that tuning the horns differently will cancel out resonances that small bass horns are claimed to suffer. Because the entire box is small and shallow, its horn paths naturally cannot be very long.

Cornu spiral? "The Cornu spiral, known also as clothoid or double spiral, is a curve whose curvature grows with the distance from the origin." Okay, I got that. "The radius of the curvature is inversely proportional to its arc measured from the origin." Fine, I think I can understand that too. "The parameter form consists of two equations with Fresnel's integrals..." Now I'm fading. The upshot? As shown, the Cornu spiral is dissimilar from both the logarithmic and linear spiral. Curves like these were most likely first studied by Johann Bernoulli in the 1690s.

What we have here in the end is a speaker that entrusts a single driver to cover sounds above the midrange and a compound spiral horn to boost its otherwise attenuated bass performance. On paper, that's an excellent idea. But to put two plus two together in practice? Well, like much else in this hobby, that's more art than science. Some say that single-driver speakers with hornloading are doomed to fail, period. Others are convinced that some of the best speakers ever designed followed this principle. Me? All I know is that if I had the requisite DIY attitude and competence, horns speakers would not be the choice for my first project and spiral horns even less so. (Cornu's DIY piped transmission lines at right).

Where to place them?
It's never trivial to find the best place for a speaker in one's listening room. As is all too well known, at times a few centimeters can make a world of difference. At other
times, there seems to be no ideal place at all. Our Cornu CS turns out to be particularly demanding in this regard. It took me quite a long time to understand the benefits of one placement versus another. After all these weeks and months, I'm still uncertain whether I truly hit upon the optimal site for them in my room. [German retail space below.]

Having listened to the Cornu CS for a while by placing them here and there, I asked Daniel Ciesinger how he would advise somebody without a master's degree in speaker/room interactions. This is what he gave me by way of an answer (typical setup):

  • The distance between his speakers should center on about 1.5m to 2.5m, less than the typical stereo triangle. Using vinyl/tube helps with stereo imaging.
  • Hang the spirals on the long wall.
  • Place the listening couch on the opposite wall. A wall carpet behind the sofa improves the sound.
  • For rooms of around 2.5m height, the spirals should be hung about 30cm below the ceiling.
  • There will be almost no bass in the middle of the room.
  • Typical multifilament cables can perform badly with certain amps. Solid core twisted pairs are better.
  • Too much/little bass mandates moving the speakers up or down.
  • Windows leak bass, especially the thin glass kind in old houses; hence the listening position and the speakers should have some distance to the nearest window.
  • High rooms absorb bass so place speakers near the floor and/or near corners for reinforcement.
  • 4 spiral speakers will result in powerful and deep bass everywhere in the listening room.
  • Transistor amps may exhibit midrange harshness and little bass with the spirals.
  • It is possible to place the spirals like conventional speakers. This produces better stereo imaging.

Many good points here to be sure. Let me start by sharing that I was quite happy with Daniel's reply. Just like my daughter who recently learned how to ride her bicycle without training wheels, I had covered many pointers on my own. I had the speakers on the long 7.5m wall -- with the nearest sidewall being 2.5 meters removed -- because the alternate short-wall option produced highly colored results. The window on one side was at a safe distance from the listener's ear. I first hung them up quite high on the wall but lowered them later so that the driver's center was 1 meter above the floor. (To show how critical this parameter was, 1.75m gave better results than 1.5m for example). The distance between the drivers was now 2 meters, having been 0.5 - 1.0 m wider before. The distance from the speakers to the opposite wall was 4 meters. Daniel's e-maiI made be confident that I was on a right track. [Prototypes of big spirals below.]

In his notes, Daniel seemed to be worried about the soundstaging and imaging capabilities of his speakers on the one hand and their adequacy of bass on the other. I wouldn't be. Explaining why could in one sense provide an overall theme for the entire review so let me proceed carefully.

An important confession first. I experimented with the Cornu CS leaning against the floor tilted towards the listener's ear. Why not? I then tried them on stands. Those were my own inventions, not very firm and cobbled together simply for testing purposes. Despite these short-lived experiments, it was clear to me from the beginning that whatever sonic merits there might be related to placing the Cornu CS somewhere other than on the wall, I would find it hard to do. Why? These speakers are made for walls (there's even a hook counterpart on their backs). They look like wall speakers, their form closely follows function and said function is to shine on the wall.
In the absence of sufficient justification, I just knew that I would end up listening to music with the Cornu CS hanging on the wall even if, for some reason, this setup should turn out less than ideal. So I did. Accordingly, all of my observations about their sound are on- rather than off-the-wall observations. Because other schemes have the potential of revealing something interesting about the Cornu's sound, let's look at their sonic merits first.

Daniel explains that if they're pulled away from the wall and slightly turned toward the listener, soundstaging and stereo imaging are improved. That's true but not the whole truth. The statement has to be qualified. First off, people who claim that proper soundstaging and imaging are unachievable with wall-hanging speakers often remain seated where they would have been had the speakers been removed from the front wall. No wonder there's no soundstaging or imaging then. To obtain either, the listening position should be moved closer to the stereo triangle to provide a 55-60° view of the speakers. (What the ideal stereo angle for the Cornu is becomes a very good question). Once listener distance is properly accounted for, the center image can be focused and musical events in the left-right domain delineated much in the same way as with conventional speakers. This even includes layered depth despite the wall's masking effect between the speakers. However, I won't argue for the truth of this last statement as this seems to be one of the eternally contested issues of HiFi discourse.

Secondly and more importantly, there are two different kinds of soundstaging and/or imaging. One is what happens when speakers -- preferably small two-way monitors -- are dragged into the middle of the room. In this setup, the sound is bound to float nicely between and behind the speakers, the presentation is airy,
all the instruments are in pinpoint order etc. We all know this type. The second kind of soundstaging is, paradoxically, not so much interested in the spatial aspects of the sound and more concerned over listening relief. Don't forget that stereophonic reproduction was originally meant for adding relief to monophonic sound reproduction, not for spatial reasons as such. [Another Cornu prototype above.]

It's not too difficult to see what I'm getting at. When speakers such as the Cornu CS hang on the wall, it's not that there is no soundstaging and/or imaging. There is but of a different kind. To my ears, the latter is not only more relaxing, it's also more realistic, natural and life-like. Unlike with the former, the musicians -- directly and without any ostensible interface -- present their music to the listener. The musicians themselves aren't first reproduced to then present their music to us. What's so charming about wall placement -- and about flush mounting or the corner position for that matter -- is exactly how it blots out the artificial HiFi-type soundstaging and imaging ideas from the listener's mind. Or so I hear it.

I don't want to, even for a moment, impose my attitudes on anyone else. The first kind of soundstaging and imaging are a viable reason to get on with this hobby and an important reason to many, particularly at the onset of their audiophile careers. If anybody with mini monitors as their ultimate yardstick -- or line arrays in the middle of the room -- complained that the Cornu CS's wall-performance was inadequate in this regard, I would accept that criticism without objection.

To me, however, such setups represent a degradation of the hobby to some extent. I find them musically phony, exemplified for instance by an unnatural (although sometimes appealing) presentation of a symphony orchestra in such systems. By mistaking spatial qualities for music, they can distract us from the hobby's original endeavors (and multi-channel home theater only capitalizes on that error). At least I personally don't long for the kind of soundstage and imaging that the textbooks recommend I should have. Whatever I missed in this respect while listening to Cornu CSs therefore didn't bother me at all.