I just scored a sweet pair of CS 2.4SE loudspeakers. Anyone else currently or previously owned this model? Owners of the CS 2.4 or CS 2.7 are free to chime in as well. Thiel are excellent w/ both tubed or solid-state gear!
All the technical discussions not withstanding, it's all about the music albeit first order, higher order, ... as long as you enjoy the music, it's all that matters.
Andy - Computers are a big part of the mix, and those tools are now easily accessible. To clarify: in Thiel's very beginning, computers were pretty limited. We used them from the beginning via Fred Collopy, another original collaborator who went on for his PhD in Decision Science - managing unimaginably large decisions with serious aid from computer modeling. We did Fourier Transforms (FFT) on impulse samples before 1980 and the number crunching took all night - literally we read the results the next morning. We used computers (Wang 600) before IBM or APPLE were in the personal computer business. So, we were ahead of winging it in the dark ages.
Your list addresses a few of the challenges. There are many more. Visionary unified engineering is rare and pure; I see Jim's work as that - one mind accumulating the questions and answers over tens of thousands of listening hours and iterations. That's a powerful thing.
One example of the contrary is the development of the CS2.7 by Warkwyn, among the best development engineering firms in the world with access to the Canadian National Resources Lab. It took them years of time and hundreds of thousands of dollars to develop the crossover for the 2.7 with the constant guidance and help of the team at Thiel. They have computers and resources to make your head spin. The tools are only part of the picture.
So I did my comparison today, the bryston 3bst and 7bst in parallel and serial on my Thiel 3.7. As both amps are off the same generation, they sound sonically identically. I used some jazz music, with organ and sax (domnerus, antiphone blues) and Brubeck, take five. First observation: the spl of the 3bst and 7bst in parallel were the same so comparing both in mono was very easy. Second observation: at 80 dB, what I consider at live loudness, both Amps were easily able to reproduce the music as I was seated 12 feet away. The only difference I could consistently detect was in the low bass where the 7bst was fuller. Any differences in the upper range, was undetectable by me, no signs of any distortion. 3. But then I played some female vocals, brendi carlille, and there it was, her voice was deeper on the 7bst and sharper on the 3bst. Enough of a difference to keep the 7bst.
Thanks to Tom for recapturing all the works went into the measurement phase. That seems like a lot of works. I supposed that was before the proliferation of personal computers?
I think today speaker designers probably have a lot easier time with speaker design with the advance of software and inexpensive access to computation. Some of the software are actually free and can be downloaded into your personal computers. I personally only spent about $270.00 and able to purchased a set of hardware that can perform all the measurements needed to speaker design. All the softwares that I am using are completely free and available on the internet. The rest is up to my own imagination.
As for measurement distant, I think the problem with making measuring at 8ft distant is that by the time the sound arrives at the microphone, you have a lot of sound reflection from the floor, the side walls of the measurement room. You could put a time bracket so you can only capture the sound at a very narrow time window before the reflections but that would compromise the low frequency portion. The longer the measurement distant, the narrower the time window, and hence the lower frequency response compromise. You could correct for that by measuring at closer distant such as 1 meter which is industrial standard. The problem is with first order speakers, at that distant, the mid and tweeter may not fully integrated so that is another challenge. Or you could build a large chamber so the floor and side walls distant is a lot longer than 8ft. Another problem is at low frequency, the wavelength is so long that you need a very large chamber so that you can accurately capture the one wavelength of low bass frequency before reflection. Tom mentioned that Thiel was trying to measure the speakers outdoor with the speakers mounted on a tree to eliminate reflection from the ground so I guess it had to be a pretty high tree :-). Eventually it may not be practical to build such a large chamber so there are a few methods of trying to merge the high frequency and low frequency response. There are software that could do this but they seem to have their own limitation.
But I think you could make it as complicated as you want, or you could simplify as much as you could to get a finish product.
Prof - I began my re-acquaintance with Thiel Audio when Jim died in 2009 which increased when the company sold in 2013 and more so when New Thiel ceased operations earlier this year. I agree with your assessment of appreciation. However, in my fairly extensive reading I am struck by how little the actual scope and accomplishment of the work is known.
Jim was very self-effacing and introverted and did not effuse, as I have been doing in this thread. And the company didn't spend effort on promotion or advertising and didn't develop liaisons which might have focused the understanding and value of the work more than the somewhat superficial understanding generated by reviews.
Flashing forward a few decades, I believe that phase-time will be revered in speakers much like it is today through the audio production chain from capture, right up to the speaker where it is commonly accepted that the ear-brain can descramble the signal well enough. Life is short and we leave very little mark. But what a trip!
That was an absolutely fascinating read, Tom. Thank you!
I'd followed Thiel for years, and especially when I was on the hunt for 3.7s in 2015 and afterward, I was fascinated with the story of Thiel's history...and it's demise at the hands of new owners - delving through all the Thiel articles and comments on Thiel at the time.
It's hard to think of any other high end figure who garnered so much sincere admiration and good will from writers, industry folk and audiophiles, as Jim Thiel.
JA - the whole thing was exciting. Beginnings are cleaner and purer and closer to the heart. We were exploring new territory from a first-principles perspective free from the prejudices that formal education or credentials would have imposed. Some form of progress was made every single day through decades of endeavor.
Prof - our measuring scheme was rigorous and thorough. But, being a bootstrap skunkworks, we created everything ourselves. We began with a rented HP dual trace scope and calibrated mic. We built a sandbox in the field - literally, sand - to bury the cabinet face-up to learn the 2pi, non diffracted, infinite baffle response. Later came hoisting the speaker into the walnut tree for free-field response. The mic was hung on a conduit, first 6' out and then a 10' joint - Hey! that's it! Outdoors we could use sine wave sweeps, but indoors the boundary reflections muddied the mix - so Jim designed and built 'the bleeper' an interrupted ⅓ octave stepped pulse automated signal generator: 3 cycles, one to accelerate, one to measure and one decelerate - advance ⅓ octave. The kids grew up 'singing' bleep, bleep. bleep from 20Hz to 20kHz. The breeze settles down after sunset for an hour or so of high quality measurements to analyze later in the night lab session. This scene is Georgetown Road, the country house where it began. The development lab stayed there for a couple of years after production expanded into Nandino Boulevard. At Nandino Jim's first lab there was 20 x 20 x 20'. Cubes are not good, but that corner could be walled off. The adjacent back parking lot was 150' x very long, perhaps 500' to first reflection. Outdoor ground-plane measurements materialized with live reference recording and test playback in the same environment. Our building was 100' x 300' and we built a roof access stairway to that 'infinite plane' with virtually no reflections 360°. Cable snakes dropped to the lab below for data recording. The final lab was pretty nice. Office / lab was 20' x 20' x 8' ceiling with a same-size balcony above, all connected to a soundproof room 17' high under 3' insulation x 20' x 32.5' (golden ratio) long plus the 20' balcony above the office. I would call the room quasi anechoic. We covered the walls and floor reflective points with layered sonic insulation and made a measuring tower 6.5' high (golden ratio) for a low reflection - highly intelligible environment. In other words, we knew exactly what the room was adding to the free-field measurements. Drivers were tested in an infinite baffle (flat wall) to identify diffraction effects in the cabinets. We compared everything to its base state. Time domain (diffraction, delays, internal reflections, etc.) must be engineered as temporal distortions. (Many designers treat those effects in the frequency domain, which is fundamentally incorrect.) We compared the incrementally improved 'bleeper', truncated noise, sine sweeps, etc. to commercial measurement devices and chose to keep Jim's stuff due to cost performance analysis in a small, frugal rapid growth environment. (New Thiel bought a Klippel System, which no one has used and Rob is now incubating.
Of critical importance is that Jim knew intimately the behavior of every aspect of the design and understood how they related to each other and the global perfection he sought. In this lab in the late 80s during the development of the CS5, which used best of form European drivers from Focal and MB, we judged that we could make a better tweeter than we could buy. The development of the CS5 aluminum dome tweeter (eventually also used in the 2.2 and 3.6) was driven by Jim's adoption of Finite Element Analysis. We took a pretty deep bite in cost and learning curve and, on the second try, took on FEI for all further developments. The dramatic reductions (10X+) in distortion due to motor subtleties - the copper shunt rings, pole and top plate shapes, cone geometries, etc. - were all facilitated by FEI. I know of no other company, large or small, that optimized such minute details.
All of these techniques are but half the picture. Every single exploration was co-developed by ear. Jim, Kathy and I each brought value to that equation and eventually we added a naive listener component for additional input. (I think I've talked about that.) Technical "improvements" were always vetted by listening, and often failed if they didn't sound at least as good after as before. Equally, options which sounded "better" were rejected if they compromised any aspect of technical performance. That bit separated Thiel from others. We required that every advance had to meet listening and measured criteria and if both were not simultaneously satisfied, then we went back to work. Our development cycle, even with successive learning, took years in the beginning and by the mid 90s was down to 9 months (like a baby), and that was a magnificent accomplishment, and orchestration, considering the extreme rigor involved.
Our listening room at the factory was in the lab, but nearly every night the product under development went home to the Georgetown Road farmhouse. We knew that room thoroughly well and it was a good room. I've mentioned the dimensions. Nice, not square, 10' ceiling, 45° bay windows, doors in the corners for bass vents, gypsum plaster on wood lath. Pretty sweet. We also used some other rooms including a plaster on brick with 12' ceiling in a downtown Victorian. Before 1990 we built the listening room at the factory. Soundproof. Quiet. Stiff walls and ceiling. Coves above lighting valence at the top edges. 12' high x 19.5' wide x 27' long - controlled decay dimensions, big enough that demure sound panels could control reflections. A truly lovely, neutral room which became the principle aural lab and demonstration room. We hosted many dealers, reviewers and other interested parties there. In 2012 I visited when the 2.7 final prototypes arrived. What a sweet reverie comparing them to 3.7s and 7.2s with the serious drive train in that room that I had built. I loved it. And I may not be participating here had that visit not happened.
I think that tells you much of the tenor of the undertaking. It was both scientifically serious as well as holistically grounded. It remained focused on music and its requirements while satisfying the demands of scientific engineering. For a small self-funded enterprise, it was genuine and rigorous beyond what any visitor imagined. And there were visitors from universities and large companies you would recognize. They invariably asked 'how can you do this'? And the answer was "because we love it'.
Such is the longer personal telling of Thiel's approach. The details would fill a book. I hope this narrative provides enough for you to appreciate some of what went into creating those speakers through which you enjoy your music. "For the Love of Music" was our first motto.
you guys nailed it when Jim optimized for an 8' - 12' listening distance. Agreed, the ear-brain does have much musical information to sort out under this type of testing condition(s). It would appear that team Thiel really had fun in those early days of research and development (R&D).
Beetle - indeed the 2 meter (80") distance is fairly valid. Ordinary speakers don't care much about distance since their drivers are not producing an integrated wavefront, but rather relying on the ear-brain to sort out the phase information. In the case of Thiel coherence, driver integration focuses into an integrated wavefront at about 8' where the measurements would be smoother and room-fill would be more even. 50" graphs are misleading, but 80" graphs are more than OK, certainly showing Thiel in the top tier. But 3 meter (10') are qualitatively better. We optimized for 8'- 12'.
I guess it's just personally disappointing to see all the effort that went into ruler-flat response being presented as less than its actual in-room / as-heard performance. Note that the coax drivers remove much of the mic-proximity degrade, since the more critical upper XO is fixed within the coax propagation geometry. Only the lower XO varies with distance and ear height. And that lower XO is more forgiving due to longer wavelengths and less directional specificity.
Point of History: In the O3 development in the late 70s, We mocked up a tri-ax with a 12"woofer with a huge diameter voice coil to allow a 4"x 1" upper driver coax in it. It was fantasy at that time; it took decades to develop real drivers moving toward that vision. In the mythical world, given resources, time & market, Jim would have developed a triaxial coincident driver. In today's world with Jim's wavy diaphragm, focused rare earth magnets and magnesium or carbon membranes, such a driver would be feasible. Youth has its potentials. _
For the record, I am also a fan of John Atkinson and what he has contributed to our industry. The 'fault' in the presentation of information goes to executive policy, which took its turns under multiple ownerships. Present ownership (Y2K+) deemed it politically incorrect to allow that phase coherence was a legitimate concern. Imagine the advertiser pressure if Stereophile continued to present phase / time as a real engineering benchmark! The arc of the journey began with an editorially open mind with JA collaborating with then-publisher Larry Archibald to explore the role of phase-time in the playback equation, and proceeded toward less respect.
Prof - please feel free if questions remain after consulting my previous Thiel measurement notes. Our measurement capability integrated with critical listening was at the very heart of our product and company development.
I enjoy JA's measurements, as well as, Paul Miller of HiFi News & Record Review. These gentlemen are class acts to the Audio press. Each leaves the reader with an impressive body of work.
Excellent discussion, as always, Guys. Tom- not vicious at all. Truthful. Jim's attention to the finer details and search for accuracy has certainly paid off in any of his loudspeaker models, IMO. I was fortunate in those early CS 2.4 auditioning days to have the Vandy 2Ce and 2Ci speakers plus a pair of Triangle Celius speakers to switch off from one another and shoot-out against the CS 2.4 same room. Thiel came out as the winner to my ears for its rich sound and presentation.
Thiel seems to have generally measured quite well, despite the limitations of the Stereophile methods.
A lot of manufacturers diss Stereophile and Atkinson for measurements that don't go their way. But nonetheless, I am grateful for the work JA has done. Even if flawed in ways, he's given us an amazing, wide-ranging body of work in terms of measurements of speakers and other audio gear, and attempts to correlate measurements to sound. I can't think of any gathering of data on high end gear that approaches what JA has provided to the paying public. (And even now, online, for free).
• Bad (compromised) information is in some ways more harmful than no
information. Real anechoic chamber or outdoor measurements are expensive
and Stereophile et al choose to side-step that expense without, in my
opinion, proper contextualization / education for their readers.
I agree that JA often fails to include a few simple statements that can help the reader interpret the graphs, including how the measurements might be misleading of actual performance. IMO, the quality of that publication has slipped over the years.
It is impolite for manufacturers to raise such issues in print, thereby becoming complicit in the misleading measurements.
I recall at least two examples where Stereophile printed measurements provided by the manufacturer (Vandersteen and Avalon) but, yeah, risky for manufacturers to push back.
Too bad they don't measure at distances that allow for proper driver integration, time accuracy or actual listening positions.
Most Thiels measured by Stereophile (at 50") show a suck out at the mid-tweeter XO point. Sometimes the text would explain that this was likely a problem of distance and sometimes not (but kudos to them for even trying!). Soundstage measures at 2 m, about 79". If the CS2.4 was disadvantaged in their test you can't tell by the "listening window" graph. +/- 2 dB from about 33-20K! I've only noticed one other speaker in their database that can match that! (A $$$$ Magico - no thanks)
The problem of measurement distance is well understood by those educated in the art. In other words, the measurers do not think that their 50" or 80" measurements present an accurate performance picture of a multi-driver phase coherent speaker. But their limitations are real. Reflections in real rooms overwhelm the actual signal, so they must bracket the time window of their quasi anechoic measurements to eliminate the reflection - noise. And they publish their results in the name of 'level playing field' - all products subjected to the same test, despite its known shortcomings.
Collateral damage includes:
• Bad (compromised) information is in some ways more harmful than no information. Real anechoic chamber or outdoor measurements are expensive and Stereophile et al choose to side-step that expense without, in my opinion, proper contextualization / education for their readers. • The normal reader does not have the education / information to extrapolate the real meaning from the compromised measurements. • It is impolite for manufacturers to raise such issues in print, thereby becoming complicit in the misleading measurements. • Many manufacturers design to measure well in the Stereophile-type quasi-anechoic measurements, rather than a justified standard. • There are no firm rules for record producers. They are second-guessing how a loudspeaker (without standards) might reproduce their mix. • A vagueness cycle (neither viscous nor virtuous) ensues.
And stuff like that. Note that the ear-brain, adept at synthesizing (remembering) how a real (insert instrument here) bass, etc. would sound in this playback room, (and should have sounded in that recording space - remember, we construct what we hear) can judge the more correct representation when given comparative choices. We at Thiel decided, at the beginning, that the only justifiable approach (to our understanding) was to design to anechoic-flat, just as a microphone is designed to anechoic-flat, except when it's not because Shure et al think that singers want to enhance their upper midrange formant. And the slippery slope gets slipperier and slipperier. I notice that there is more agreement now than 20-30 years ago about what is more correct. But, has there ever been an attempt by the Society of Audio Recording Engineers (and so forth) to standardize the design goal of the loudspeaker? Wouldn't that be a worthy undertaking? And the beat goes on - Amen.
^Kudos to them for using a neutral environment. Too bad they don't measure at distances that allow for proper driver integration, time accuracy or actual listening positions.
Jim does not add the customary underdamped bass hump, so Thiel bass is honest.
Teeheehee. I think I know which brand you're referring to ;^) Soundstage's loudspeaker measurements are done in a true anechoic chamber and, therefore, do not suffer the compromises of quasi-anechoic measurements done by Stereophile. Anyhow, if you look at their measurement library, the CS2.4 is down 6dB, relative the level at 1K, at ~31 cycles. This compares very favorably to other much more costly designs such as the Wilson W/P 8 (~38) and KEF Blade2 (~31). Nevertheless, some of the other designs with slower roll off can benefit from boundary reinforcement to increase low bass in an actual room.
When I got the CS2.4s, the first thing I noticed was improved bass definition relative to the otherwise well-accomplished Vandersteen 2Ce. I listened to many tracks before noticing the 2.4s don't go quite as low (specifically, a Tracy Chapman song with organ tones).
Correct - most designers shape the soundscape toward 'easier to handle'. Jim looked for honesty before all else. Most designers roll off the treble to match the bass roll-off. Jim kept the treble extension regardless of the bass extension. The 1.6 with folk music or a sub-woofer is voiced like the 3.6, etc. His approach was toward accurately reproducing the input signal.
^...Or a treble roll off. Which is why I think the smaller models sometimes can sound a bit tilted up compared to some of the competition. The more full range (and having the room to properly accommodate that can not be over emphasized) Thiel's are less likely (when properly amplified and set up) to sound bright.
1) Correct. A primary difference between the model 2 and model 3 is bass quantity and extension. The model 3 has a 10" woofer with longer excursion and larger passive to produce more than 1.75X the bass of the 2. 2) Jim does not add the customary underdamped bass hump, so Thiel bass is honest.
It depends on the size of your listening room. As good as the CS2.4, as far as the bass is concerned, it cannot compete with other speakers with 10in. or 2x10in bass drivers.
@andy2 The CS2.4 goes down to about 30 cycles in my room and I have them well away from walls. That’s enough to reproduce all but the left most one or two keys on a piano. Not much music down there. My previous speaker was Vandy 2CE Sig II, useful output into the mid 20s. My collection includes a single song wherein I noticed the deficit via the 2.4s. If you can afford Vandersteen 7s, go for it. They are fanatstic *and* have full output down to about 20 cycles. Or, if you really want those last couple of notes, you can add a subwoofer or pair to the 2.4s. I seriously considered subwoofers when I had the CS1.6 but that model quits at about 50 cycles. I have zero desire to add subwoofers to my 2.4s.
There seems to be some talks of what if you want to upgrade the CS2.4? Overall I still think its only real weakness is the lack of low bass. For clarity and detail and natural integration, it's probably still some of the best even for today. Some newer speakers now are using fancier tweeters such as be dome, ceramic which may offer better treble detail but I am not sure it's necessarily better. For speakers using higher order filter, they can afford to add like a 10in. driver such as the Wilson or Dynaudio and so on, but for the CS2.4 being first order time coherent, it's not that trivial and given the constraints I am not sure if it's even possible. So with larger bass driver such as 10in, you get a more fuller sounstage abeit with higher order type of sound. I guess it's all a compromise. You choose the CS2.4 for all its strengths, but have to accept its lack of low end grunt.
But can we add an active subwoofer with some type of room correction DSP? Probably - just like what Vadersteen did with their top of the line offerings that costing even above $50K, which is why I am a bit surprise why none of the Thiel speakers follow this strategy which would not only solve the low impedance issue but still retains the time coherent characteristics.
So the best upgrade for the CS2.4 would probably be some of the Vandersteen top of the line speakers (although a lot more expensive) with a built-in subwoofer.
Just sold my pair of CS 3.7’s to a gentleman in Southern Cal who has a a pair of 90 grand tube amps and similar pre. He’s loving it . Shipping from San Fran with special care was 1500.00. So glad he’s -happy. $8800.00 plus shipping.
Thank You for citing your amp upgrade and matching Cardas power cord (PC). There is much information in these Audio forums on B.A.T. gear and tube rolling to achieve a personal "sound". I do know that Cardas Clear and Beyond PC mate well with tubed gear. Keep me posted as you massage the VK-55 into your room, system.
If they need help, Rob will help you. If you want to take
them to the next performance level, we'll have an upgrade path for you.
IMNSHO, buying a nice pair of used Thiels might become the
most-cost-effective way to get near SOTA sound. For starters, the stock versions
are already excellent (as Tom noted), sonically competing with new models many
times their price. Buying a used Thiel is not worrisome given the available
service and parts from Coherent Source Service.
Tom Theil’s crossover upgrades should push sonic performance
to the next tier. We’ll soon know. If that’s true, you can buy a nice used pair
and upgrade the passive parts to get something that approaches SOTA sound. As a bonus,
you have speakers that look really good in your listening room.
jon - I know it's thin ice for a manufacturer to speak up. But I'm out of the game, so here goes. Over the years, our products were competitively evaluated by some big name as well as audiophile manufacturers. Their 'business formula' would assign a sell price of up to 10x our sell price, and then back out some 'extraneous costs', which is what they call giving the customer more than he knows. Point is that for the customer whose values are aligned with ours, Thiel prices are extremely slim.
Schubert, have a great time with your 3.6s. If they need help, Rob will help you. If you want to take them to the next performance level, we'll have an upgrade path for you. Welcome to the Thiel corner of the world.
Pick a dozen brands / products at $20K and compare your new
3.6s toe to toe. I bet you'll be thrilled.
Yes! There are 2, maybe 3, speakers south of $20K that I would *consider* trading my CS2.4SEs for. And that's before I finish my XO upgrade . . .
the relationship between price and performance is extremely
weak
Indeed, I can think think of a few models as high as $40K that I would not trade my 2.4s for. But I can also think of a handful north of $40K that I would pick over any Thiel (maybe I would change my mind after a Tom Thiel XO upgrade?). But they damn well should sound better given the chasm in pricing. The bottom line is an audiophile can be *really* happy with a pair of used Thiels and not have to worry about how to get some incremental SQ improvement . . . until after winning the lottery. :)
Tom is right about the value here. It's one of the reasons I have such a dim view of audiophiles in general. They'll spend ridiculous amounts of money for stuff that simply doesn't perform that well. Thiels are a bargain at MSRP and they historically haven't held value that well so you can get an obscene bargain for the performance if you buy used. When I bought used 2 2s it really opened my eyes to this. They replaced a pair of 3k msrp speakers from a larger company in a second system. They were so obviously, laughably superior that I don't think there's a single person on earth that wouldn't have recognized it. It took a while for me to accept it but I eventually concluded that the relationship between price and performance is extremely weak. I don't know what "audiophiles" want but it's clearly not the truth.
Viewed another way, you got an unbelievable bargain. In the realm of passive speakers, which is still the norm in home systems, their design-engineering equates to products selling in 5 figures. I know factually that multiple other manufacturers would claim $15,000 as cabinet value alone.
I understand the desire to buy stuff as cheap as possible. But, I also encourage you to consider the intrinsic value of the goods. Pick a dozen brands / products at $20K and compare your new 3.6s toe to toe. I bet you'll be thrilled.
Good to see you again. $800 would have been a fair offer. I usually see this Thiel model in the $1000-1500 neighborhood. Were the Serial Numbers listed?
You must have a verified phone number and physical address in order to post in the Audiogon Forums. Please return to Audiogon.com and complete this step. If you have any questions please contact Support.
