Thrax Audio Siren loudspeaker

Based in Bulgaria, European audio company Thrax has been active since 2009. Their ingenious and varied design approaches seen over several product lines have continued to intrigue me with their conceptual originality, innate musicality, and imaginative use of a broad spectrum of technologies. Their products range from valve (tube) amplification to digital audio and, more recently, loudspeakers. There are 16 products in the Thrax range.

Thrax founder and chief engineer Rumen Artarski packs a capacious scientific kitbag. His love of music and fine sound reproduction is self-evident. In the disputes that sometimes occur between measuring and listening, Rumen says, decisions made by the ear are final.

Writing for HiFiCritic in 2012, I supported Chris Bryant's evaluation of the all-tube Thrax Spartacus, a 40Wpc triode-based power amp. In 2014, Bryant examined the all-triode Orpheus phono preamp (footnote 1). You could say that we—Bryant and I—were both pretty much solid state diehards, even though we'd both enjoyed good experiences with tube designs over the years. In both the 2012 and 2014 reviews, we described how we were shaken up by these two thermionic designs, whose purist tube technology had been finessed using advanced electronic design principles to deliver a winning combination of tonal accuracy and transparency with the dynamics and musicality of low-feedback triode circuitry. We knew then that Artarski's Thrax was a force to be reckoned with. The range of distinctive high-end electronics has continued to expand to include a loudspeaker, the standmount Lyra, now joined by the smaller Siren ($13,600/pair), also a standmount and the subject of this Stereophile review.

Technology and engineering
Weighing a substantial 44lb (20kg), this standmount two-way marries an exotic, costly 6.5" (170mm) midwoofer to an unusual, semi-elliptical (biradial) horn of significant size. The horn "compression loads" a small, application-specific high-frequency driver, a 1" (25mm) ring-diaphragm design from pro audio specialist BMS.

Under the leather-covered central section, the construction is birch plywood reinforced by a cross brace. The walls of the 16 liter enclosure are not perfectly damped, but a knuckle rap is rapidly dissipated. Those substantial front and rear sections are milled from solid billet aluminum alloy. Fourteen stainless steel socket head bolts attach the rear panel. A filling of natural sheep's wool damps internal reflections. Artarski notes that the box/port system is tuned at 40Hz, with typical in-room extension to 33Hz at normal listening levels. The port's rear panel exit is mildly flared to reduce port noise at high air velocities.

Thrax prefers series topology for its crossovers, which is a little more awkward to implement than the more commonly used parallel form. In some designs, a series crossover may offer better acoustical blending of the driver outputs; simulation software has made the design task easier. As ever, crossover filter design is an interactive process involving theory and measurement, and must be partnered by meticulous listening.

Jantzen, a noted Danish manufacturer of audiophile-grade parts, was chosen to supply the costly Siren crossover components including three "zero saturation" wax-impregnated copper-foil, air-core inductors (no magnetic core to saturate!), plus several exotic film-and-foil capacitors (footnote 2) in each channel. The capacitors in the tweeter feed are the "Silver" type and enjoy pure silver lead-out wires. The whole crossover assembly is "sealed and potted" to damp vibrational disturbances from working drivers.

The horn is very efficient and imparts an unusually high intrinsic input impedance, which could otherwise add a trace of high-frequency excess with some tube amplifiers. So, an RC impedance-matching network placed at the input terminals helps maintain tonal balance when the Siren is used with tube amplifiers with high output impedance.

The Siren is single-wired with Aeco gold-plated copper alloy binding posts with 4mm plug, spade, and bare-wire connectivity. The DC input resistance measured precisely 4 ohms, a fair estimate of the "nominal" amplifier loading.

The deep-profiled custom flare of the BMS tweeter's horn was designed by Spherovox using electroacoustic modeling software. It blends seamlessly into the deeply milled solid alloy plate of the front baffle. The biradial transition is smoothly rendered and affords wider-than-usual horizontal directivity at higher frequencies. The driver element itself will never need to work hard, in view of its sensitivity: an exceptionally high 112dB/2.83V/m. That's some 25dB greater than the sensitivity of the completed Siren system (specified as 87dB/2.83V/m). Worked so lightly, it should remain free from compression overload at all sensible sound powers. It should also display low distortion and very good subjective dynamics and be free of dynamic sensitivity variations; some lower-sensitivity direct radiators, which draw more power, can compress due to thermal cycling during loud passages.

Partnering that custom horn is a 170mm, 6.5" midwoofer sourced from European "new-tech" company Purifi, affiliated with Bruno Putzeys of class-D amplifier and high-oversampling DAC fame. Purifi was founded in Denmark in 2019 by Putzeys and Lars Risbo with long-established audio entrepreneur Peter Lyngdorf. They spent several years on primary research. This Purifi driver has an unusual and highly distinctive cone design, with a circumferential brace and central reinforcement, while the assembly is supported on a distinctive, aggressive-looking, sectionally corrugated surround (see Sidebar 1).

This long-throw design incorporates an ultralinear moving coil motor system with deep excursion. The cone is fitted with a huge, sophisticated motor system built on a free-breathing die-cast alloy chassis. Purifi claims exceptionally low distortion, outstanding power handling, and a very low incidence of the usual menu of coloration-inducing mechanical misbehaviors. The driver's linearity and distortion are claimed to measure more like a good amplifier than a deep-excursion, spring force–limited, oscillating mechanical piston. It is easy to imagine the Purifi team planning their attack on the primary moving coil driver issues, then relentlessly advancing on less well-known problems, the process continuing to a practical vanishing point. Such complexity does not come cheap, as the Purifi driver costs perhaps four times as much as typically favored contenders.

The Purifi engineers have developed comprehensive mathematical models to simulate the key components of a drive-unit: the motor, inner suspension, cone, and surround suspension. Close examination of the motor section led to new insights about coil and magnet geometry, the selection of ferromagnetic materials, and their detailed configuration. Such complex mechanico-acoustic models allowed the engineers to analyze and control issues that previously were inexorably intertwined, facilitating the optimization of each source of error.

Purifi addressed distortion of all kinds arising from normal cone movement, in combination with excursion limiting so that the driver cannot easily be overloaded in the bass. The surround/suspension geometry is configured to radiate as little as possible, minimizing its unwanted (distorted) contribution. More subtle errors such as hysteresis distortion have been suppressed. There is no magic bullet but rather a systematic grinding down of well-known problems to a lowered baseline.

Typical moving coil drive units suffer 0.25% to 0.5% distortion at average sound levels, with better examples aiming for 0.15%; Purifi is aiming for a massive, 10–20dB reduction to just 0.05%, including intermodulation products. Over the important midband, the potentially whiney-sounding third harmonic reads just –76dB—0.015%, which is a remarkable achievement.

But how will this all fit together? Overall sound quality should be improved right away. The most obvious, immediate benefit would be an order of magnitude gain in clean bass power for drivers of this size.

Siren's song
After a few days' warm-up following a deep chill during their cross-Europe transit, and following trials with stands and placement, the Sirens were put to the test. At first hearing they proved to be well-balanced frequency-wise and not too critical of position. As expected, the best stereo imaging was with the speakers placed in a free-space location. Positioned about 6' apart, 4' to each sidewall, 4' 6" from the front wall, and 4' from my seat, they were toed in just enough that I could see down the inner sides of the enclosures. While optimal, this set up was not exacting: The fine imaging and overall quality proved relatively uncritical of their precise location.

At first, nothing in particular stood out; no special character drew my attention.

This was, in a sense, very promising, suggesting that the primary sound characteristics were in balance. Gradually I became aware of something special about them, firstly in the high frequency range. What I heard here did not sound like a tweeter of the usual kind; indeed, it was not aurally identifiable as a working mechanism but rather an open window into recordings. It's hard to explain. It sounded as if the tweeter was not present as a distinguishable entity but rather a blended extension of the fine midrange.

Footnote 1: HiFiCritic: Spartacus: Oct–December 2014; Orpheus Jan–Mar 2015 Vol.6 No.4.

Footnote 2: Not the usual metalized plastic film.

Thrax Audio
Bul. Kopenhagen, BL. 289
Druzhba 2, Sofia
(424) 344-0011

cognoscente's picture

Why not like with wines, and as German hifi reviewers do: "this is a 97 points speaker". Then we, the readers, your target group, have useful information. And we can compare this speaker with other 97 points speakers, which may be available for 6K or perhaps for 4K even.

JohnnyThunder2.0's picture

Stereophile obviously doesn't want to. I'm not saying I don't mind it in some of their sister pubs or other UK mags, but I think Stereophile's writers are ... better friggin' writers! What is not covered by numerical rankings is covered by their descriptive writing, their emotional response to the reproduction of music or a conclusion where they this is a class A etc. Their writers are also all different. Back in the day, what Art D liked was different than what Fremer liked, That is what they do for the recommended components rankings. it will never be qualified w a number.

Anton's picture

It's tough to take a subjective experience and then 'pretend rank' it as though it is something that can be reduced to two digits.

There are idiosyncratic things involved, as well: the room, the entire system, etc.

I take reviews as representing one person's opinion in one space during one period of time, so it wouldn't help me to associate a set numerical number to the whole thing.

I wine, a '97' rating doesn't mean that all '97' point wines taste the same, either.

The real rating is in the prose.

cognoscente's picture

Law of large numbers reflects something of objectivity.

Most buy wine based on indeed one wine rating and then say "this is a 97 point wine".

Indeed, reviewers have personal tastes and commercial interests. The more professional the less you can expect. The more objective they are, testing on objective standards. At least we, the readers, hope that is what they do, try. Same for hi-fi stereo reviewers.

I buy wines based on the average of at least 8 professional ratings and then I apply the figure skating method, I do not include the lowest and highest ratings in the calculation. Then I have a reasonable average. Then at least 6 reviewers say, the average between them I mean "this is a 97 points wine". Then you know, but what is sure in life, it is a 97 points wine. Then it is up to me to decide whether it is also my kind of wine, I mean my taste. And yes, you are right, 97 points wines taste different, but I know what my taste is and I'm looking for the cheapest 97 point wine within my taste. The best wine I like in my budget. Btw for most dinners a good selected 94 points wine is already excellent, however at special occassions I prefer a 97 points wine.

I wonder why so many other wine buyers buy much more expensive wines with a much lower rating? (yes yes we know, they buy prestige, not quality - I see the same in hi-fi stereo, or fashion or cars, actually in a lot - people buy prestige out of ignorance and insecurity).

Auditor's picture

Oh boy... Another case where the review sample has a problem! (In case you didn't read the Measurements section: the tweeter wasn't wired properly.) At $13,600, this is simply unacceptable. Based on the number of manufacturing problems that are discovered when Stereophile measures pieces of equipement, you wonder how many of us have faulty gear in our system.

Auditor's picture

And in this case, what a shame it is! These look to be amazing loudspeakers. But I would hesitate to buy a pair given that Thrax may have some quality control issues (at least that's unfortunate impression I get from this silly wiring problem).

Laphr's picture

Explain how this speaker is wired wrong.

Auditor's picture

Apparently your comment is directed to me.

I gathered the evidence using my eyes and brain. I read the following bit of text written by John Atkinson in the article:

"This sample had a severe suckout centered on 2.7kHz in its tweeter-axis response, while MC's measurements indicated that his Sirens didn't have this suckout. The loudspeaker's step response indicated that both drive units were connected in positive acoustic polarity. [...] My sample must therefore have had its tweeter miswired at the factory. I removed the Siren's rear panel and internal stuffing and rewired its tweeter in the correct inverted polarity."

Based on these statements, I arrived at the conclusion that the speaker was wired incorrectly. I hope this evidence is sufficient.

Laphr's picture

The tested sample had no assembly error. Amplitude and time responses are superb. I directed my comments to this speaker, not a prior example.

Rumen Artarski's picture

Sample sent form measurement was pulled out of the lab in a rush as the review pair got stuck in the UK. The error is evident and we don't try to hide it. It is my mistake picking the first sample I saw without clearing QC.

zimmer74's picture

JA measured a different sample of the speaker, due to logistical issues. The samples reviewed in the main text by MC were wired correctly.

John Atkinson's picture
Auditor wrote:
Oh boy... Another case where the review sample has a problem! (In case you didn't read the Measurements section: the tweeter wasn't wired properly.)

Incidents like this have happened a number of time over the years. I conjectured in an As We See It essay a long time ago that this might be because the review sample had been assembled, not by the skilled workers on the production line, but by the designer or his colleagues.

Auditor wrote:
At $13,600, this is simply unacceptable. Based on the number of manufacturing problems that are discovered when Stereophile measures pieces of equipment, you wonder how many of us have faulty gear in our system.

It should be noted by Martin's samples, which may well have been from regular production, were correctly assembled. And conscientious dealers do act as a further stage of quality control.

John Atkinson
Technical Editor, Stereophile

Auditor's picture

Thank you for your detailed reply.

I understand that Martin Colloms's pair was fine. And I'm sure most products that leave the Thrax factory are fine. But that doesn't change the fact that they shipped out a pair that was incorrectly wired. And this does look like the kind of problem that should have been picked up by a minimally diligent quality assurance process.

In the past, you might occasionally have received units that weren't assembled by the usual, skilled workers. However in this case, your unit apparently shipped at a later date than the one Martin Colloms got. A unit that shipped later more likely came from the normal production line than an earlier one. In other words, if a few "prototypical" units built by the designer left the company, they would be the earliest units produced and shipped.

Plus, I just had a look at the Thrax website and I saw the Sirens were launched in 2022. I should think a standard production process was in place by the time your unit shipped.

Glotz's picture

The last two paragraphs are rife with them. I think this is a forgivable error.

It's not a design flaw nor does this oversight negates the considerable high-level R&D and high-quality, custom parts this speaker sports. Martin was very thorough in how this product stands out in this section of the market and the considerable care that went into creating it.

I think I would care more about what you (auditor) would have to say if you put forth more productive discussion in this and other posts you have made. It would appear you haven't read the review, but just ran to the measurements with an axe to grind. Your communication does not extend to anything but this error.

It think most other understand this company is run by humans, not robots. Mistakes do happen. I don't think it negatively reflects on the company, as perhaps you intend.

Perhaps you see it as one mistake and the company should declare bankruptcy?

Auditor's picture

John suggested a possible explanation for the fact that situations like this are surprisingly frequent. If you go back through a couple years of Stereophile measurement sections, you'll find quite a few cases of review samples with manufacturing problems. John's explanation probably covers a few of these instances, but I was just saying that it seemed unlikely in this case. That's all. I wasn't indulging in idle assumptions for the fun of it.

I assure you I read the entire review carefully. It actually got me really interested in these speakers. They really do seem great. And I said it in one of my previous comments.

But they do cost $13,600. Which means this is definitely a high-end product. It's not a price a object to, though.

However, I would expect a product like this to be impeccably built. Mistakes can happen; I understand that. But when you build high-end products, you are supposed to have a rigorous verification and testing process to ensure that, if a mistake does happen, you'll catch it before the product goes to the customer. A lot of hi-fi manufacturers obviously have this kind of thorough quality control.

Lastly, each comment should be judged on its own merit. Be it negative or positive. No matter who wrote it.

Jazzlistener's picture

It seems we're not allowed to voice critical opinions, as they appear to trigger/offend sir Glotz. I think that was a fair comment re: a $14K pair of small bookshelf speakers should not have wiring or any other problems. Feel free to just ignore him - he trolls the comments section looking to criticize people for having an opinion.

Glotz's picture

Feel free to bitch and complain all you want, homey.


Anton's picture

This would be a great time for that.

I admit to finding it off putting. If JA1 can't get one wired right....

zimmer74's picture

in the print edition of the magazine. But even though he gives evidence of having read JA’s measurements, he chooses to completely ignore the wiring mistake, instead spending his time promoting his new factory, ongoing research, etc.

John Atkinson's picture
zimmer74 wrote:
There actually is a manufacturer’s comment in the print edition of the magazine. But even though he gives evidence of having read JA’s measurements, he chooses to completely ignore the wiring mistake, instead spending his time promoting his new factory, ongoing research, etc.

See this essay by Robert Deutsch from 1992: How To Write Manufacturers' Comments.

John Atkinson
Technical Editor, Stereophile

Auditor's picture

An amusing little article!

Based on Robert Deutsch's detailed typology, I would say we have here something akin to psychoanalytic denial.

teched58's picture

It's a shame that audio entrepreneurs don't understand the big competitive advantage they can achieve in a crowded, niche marketplace if they have competent media and public relations.

Competent PR isn't only about building a bigger foot. For startups, it can literally make the difference between surviving and folding.

You'd think more audio entrepreneurs would understand, given that most have been in the business previously and know the ecosystem. I'd guess that engineering types tend to see everything thru the prism of product ("If I build it, they will come. Because I built it. Me"). I don't know enough about marketing and sales guys to know whether they get the value of PR. Of course, it can be prohibitively expensive for very small companies. Decent, dedicated PR costs at least $3k/month.

Glotz's picture

They clearly will have rigorous verification and testing as the review attests. They will still make the occasional mistake despite ensuring 99% of every possible detail. And a lot of other hi-fi manufacturers are guilty of a mistake as well.

To assume Wilson or Magico or any other manufacturer hasn't made a mistake like this, is absurd. I won't offer conjecture as to why this issue happened. And I think it's offensive when someone pulls supposed answers out of thin air regardless of what JA used as past examples.

Your last statement is a bit leading, implying I want to censor you in some way. You here, rather are emphasizing to the deep degree this one minor oversight. It would seem to hold the utmost merit to you, but in reality would be reasonably forgivable for any reasonable human to understand Thrax's issue here.

I am quite sure the dealer and the manufacturer will be make it more than right. I have seen positive outcomes in many issues I have had in the past- PS Audio, NAD.. many others have risen well past their duty to give me something for my time, to the tune of thousands of dollars in my favor as well. That can also breed loyalty for that oversight or 'exception to perfection'.

I look forward to (perhaps) hearing these myself this year at AXPONA. I am interested in hearing where they lie in relation to the Magico A1's.

Kudos to Thrax in making one sexy speaker!

tenorman's picture

What manufacturer would be so incredibly lax in their quality control that they would send out their $13,000 ( $17,000 here in Canada ) speaker for review with the tweeter wired backwards ? Then , have the reviewer dismantle the speaker and rewire tweeter correctly so he could complete his review ? Definitely not a manufacturer I would ever want to deal with . If they don’t care enough to ensure their quality control is up to the highest standards for Stereophile , I doubt they would for their paying customers . I’ll pass

otaku's picture

It seems to me that every few issues there is a review item that malfunctions. One of my many favorites was a $39,000 CD player (no SACD's, thank you). The release lever for the AES/EBU input was missing so the unit had to be sent back with Jason's cable still attached. I am not sure how the manufacturer was able to run their QA on the unit without noticing.

Indydan's picture

Doesn't Robert Schryer live in Montreal? I was hoping he would cover the Montreal show again.

Auditor's picture

Yeah, I noticed that too. It's a shame. It's a good show and Robert always did a good job covering it.

Auditor's picture

Robert just posted his coverage!