ONLINE - ISSUE 21
Wyetech Pearl - a search for
the "right" preamplifier, Part 2
This review is part of an ongoing evaluation of line stages (see Part 1 http://www.positive-feedback.com/Issue17/soniceuphoria.htm). When I began this adventure, I was pleased with my Audible Illusions Modulus 3A, but had undertaken to review the small but potent Silvaweld SWC 450BFA. Exposure to the Silvaweld, plus the acquisition of a passive unit and a used Sphinx Project 4, led to a five-way comparison (which included a Jolida I already had). I learned from that comparison that a preamp can make profound differences in the sound of an audio system. I found that the little Silvaweld delivered the best overall sound, at a relatively modest price, but it had to be returned to the distributor and could not be a part of my evaluation of the Wyetech Pearl. However, I was able to obtain its big brother, the Silvaweld SWC 1000, which offers a more appropriate price comparison. This led to the four-way, top-of-the-food-chain shootout that constitutes part of this review.
Something very interesting goes on when a preamp handles a low-level signal (or not so low in the case of CD). In theory, a preamp should be unnecessary when the source is capable of 2 volts (standard maximum output for a CD player), but experience has shown me that doing without an active preamp does not lead to improved sound. One possible explanation for this is that most CD players have relatively weak output stages. In most cases, they consist of a few op amps and/or some discrete transistors, driven by a power supply that can barely be deemed adequate by high-end audio standards. For the purpose of driving an amp, most are inadequate, and the result is often a bit harsh and/or harmonically "bleached." There is resolution but not refinement, suggesting a system under stress at musical peaks. Also, the sound often lacks dynamics. Many internet contributors claim that this is the sound that was actually recorded, and that if you don't like an aggressive, flat soundstage and lifeless dynamics, you simply can't handle the truth. Some even conclude that an active preamp, which can provide effortless dynamics, a deep, wide soundstage, and palpable, 3-D imaging, is in fact generating artifacts or "enhancements" that are not on the recording. I don't believe that these effects are artifacts, but information that is lost when those recordings are played on lesser systems. Making classical recordings in a real space has led me to believe that these spacial characteristics are real, and can be either captured or added in the mix.
Another argument for eliminating an active line stage is it does not in fact amplify anything, but acts "merely" as an attenuator 90 percent of the time. While it is true that the input is often of a higher voltage than the output, an active preamp also gives the signal a deep reservoir of current drive and voltage stability from its own power supply. Think for a moment about how a preamp works. First, the signal from the source arrives and is attenuated, then the signal is amplified again—replicated, if you will—by a circuit that has a larger and more stable power reserve than the original source. Since the gain of most preamps is fixed, the active part of the line stage is always amplifying the signal, not attenuating it. Since it is only the amount of signal to be amplified from the volume attenuator that changes, the concept that an active line stage is only attenuating the signal is false. In fact, the preamp first reduces the incoming signal, then beefs it back up with (ideally) the same amount of information, but greater drive and control.
The fact that preamp designers have recognized the need for a great deal of stable power is shown by the fact that more than a few top-of-the-line preamps are now being designed with exotic power supplies. These preamps SHOULD have greater stability under dynamic conditions than the source components, and should respond with a greater sense of ease (and less voltage fluctuation) when demands are made on their power supplies. This may or may not be the case, as not all preamps are created equal. It seems possible, then, that letting a bigger engine do the pulling maintains dynamics, and the resulting preservation of low-level information benefits a preamp's imaging and soundstaging abilities.
What about circuit topology? In my opinion, the best circuit is the one that delivers what the power supply has to offer with the least interference. All circuitry introduces some of its own character, but in my experience, relatively simple circuits, backed by the best power supplies, result in great (but not equivalent) sound. Obviously, the circuit design is responsible for most of the sound (Allen Wright of VacuumState suggests a 50/50 split in importance between power supply and circuit design). If implemented incorrectly, the circuit can be a major source of coloration and a killer of dynamics and low-level resolution. Likewise, an inferior power supply can ruin the best circuits. At lower price points, preamps are typically compromised in both their circuit design and power supply. As you move up in price, there is usually a substantial increase in the quality of the circuit, the quality of the power supply, and the quality of the parts. Nevertheless, I have found that there is generally some compromise that prevents many preamp designs from being truly great. Once you climb into the higher price reaches, the compromises (or design flaws) in each aspect of the design tend to become much smaller, and you begin to see fundamental limitations in the circuit and power supply topologies chosen by the designer.
Although the Pearl is only the second-most-expensive preamp made by Wyetech Labs, at $5300 it must be considered a design of minimal compromise. It is described in the company literature as a single-box design that is sonically as close as possible to the top-of-the-line Opal. Many of the parts in the Pearl are the best available, and the circuit design is the best that the designer knows how to execute. The Pearl (in fact, every Wyetech product) is the result of the singular vision of Roger Herbert. I say that Herbert's vision is singular because he believes that there is only one correct way to build a preamp. This includes designing a large power supply around choke filtering and passive shunt voltage regulation, no use of printed circuit boards, zero negative feedback, and above all, a circuit topology called the grounded grid. I know of only one other preamp that uses the grounded grid, that being the Transcendent Sound Grounded Grid (used in this review for comparison). Using this topology gives the Pearl a stated bandwidth of up to 500kHz (-1db). An interesting feature of this design is that it uses a cathode follower on both the input and output, with the GG stage between them.
While Wyetech makes four line stages, all of which use the grounded grid topology, the cheapest one makes a few concessions—it has a printed circuit board and a less ambitious power supply. The upper three models (the Jade, the Pearl, and the Opal) use hand wiring on vector board (perforated plastic board) and top-grade components. The Jade, Pearl, and Opal all use three 6SN7 dual triodes in the grounded grid circuit because Herbert believes that this tube has better linearity than the smaller nine-pin tubes, and is capable of greater drive.
Images courtesy of Srajan Ebaen of 6Moons.com
As you can see from the photos, the Pearl is beautifully constructed, inside and out. It sports Wyetech's trademark lavender automotive paint casework and gold knobs. I like it because it dares to be different, but friends and colleagues have made mixed remarks about its appearance. Wyetech's use of top-quality parts extends to the switches and attenuators. In the case of the Pearl, this means Shallco military switches on the source selector and balance control and an Elma 24-position attenuator with one percent, surface-mount, metal film resistors for the volume control. Toggle switches for the power, mute, and source/tape selection are also industrial quality, from NKK. The toggle switches are very industrial in feel, as are the Shallco selector and balance switches. Selecting a source results in a solid "thunk." Oddly, the volume control has a somewhat loose and vague feel. The Transcendent Sound Grounded Grid preamp has the exact same switch, and it has a much more solid feel. Strange, but no matter—the control works perfectly well. I never felt that I couldn't find the right volume, so I think that the Pearl's gain structure and resistor steps have been well judged.
Image courtesy of Srajan Ebaen of 6Moons.com
Inside, the circuit and power supply are built on their own vector boards for better isolation. The construction is very tidy, and the wire routing is reminiscent of a PCB—not the spider's web one normally sees in point-to-point wiring. This type of wiring also results in a short signal path compared to a PCB. Resistors are metal film and capacitors are metallized polypropylene. The power supply has a fairly beefy toroidal transformer, and a very large (for a preamp) electrolytic capacitor for further smoothing. There are also two chokes for additional filtering, and a large number of high-power zener diodes provide voltage regulation for the high-voltage supply. Herbert states that his power supply has low impedance and no feedback, allowing it to respond very quickly to sudden demands.
While this preamp has not been significantly compromised in its execution or parts, this does not a $5300 preamp make. The sound determines whether or not the unit is worth its asking price. As far as I am concerned, it could be made from toothpicks if it sounded correct, but then, I would not be prepared to pay thousands of dollars for a preamp made out of toothpicks. Like most people, I like to see high value in a piece of gear that sells for a high price. I think the Pearl delivers on this. My response upon opening the shipping box was, "Ah, so this is what a $5000 preamp looks like," rather than, "Is that all there is?"
I did most of my listening with the Pearl in my main system, where it replaced a Transcendent Sound Grounded Grid preamp built from a kit (great fun and easy, too!). I left the Pearl on for an hour before my initial listening, as Wyetech's U.S. distributor had told me that it was fully broken in—so I felt that this was sufficient warm-up time. Upon turning it on, a red light comes on to indicate that the preamp is muted. The turn-on sequence lasts about 45 seconds, at which time the red light goes out and (unless you have the mute switch engaged) the green one above it comes on, telling you that the preamp is ready. The first thing I noticed about the Pearl, even before playing music, was that it does not get really warm. I know that a preamp is not supposed to get as hot as an amp, but several of the preamps that have come through here got rather toasty. I then remembered reading that the tubes in the Pearl are run rather conservatively, contributing to a long tube life. I mention this because I suspect that the tube bias has an effect on the Pearl's sound.
After the one-hour warm-up, I gave the Wyetech my undivided attention for three hours. My initial response was that it didn't sound like any of the other preamps I had heard in my system. There was a very high level of transparency and precise imaging. There was also a smoothness to the sound, a texture that reminded me, in some strange way, of white, translucent silk. There was a great sense of coherence, and no hint of congestion or of instruments running over each other. Music made sense, and had definite order. Bass was tight and rhythmic, pushing the pace of the music along nicely. There was nothing—and I mean nothing—sluggish about the sound of the Pearl. In fact, if anything, it sounded slightly lean and mean.
The Wyetech clearly allowed the timing relationships between instruments to pass unfettered. This is a common attribute in gear that eschews the use of negative feedback. Dynamics, both micro and macro, were superb, up there with that of other top preamps, and again, this is a trait common to zero-feedback gear. The Pearl's low-level detail and preservation of subtle dynamic shifts were excellent, among the best I had heard in my system. There was no sense of strain at any level, and the continuity of sound from top to bottom was excellent. This is something the much less expensive Transcendent Sound cannot quite manage.
Nevertheless, I was not wholly satisfied with the Pearl. I found that I was not connecting emotionally with the music, and was getting mildly fatigued while listening. I then reconnected the Transcendent Sound and left the Pearl on continuously for a week. When I returned to the Wyetech, I found that things had not changed significantly. I was again impressed by the Pearl's transparency and dynamics, but again decided after some time that I had had enough of it. This continued for several weeks. I tried other amp and speaker combinations, but always came away with the impression of a component that seemed nearly perfect, yet somehow failed to engage me. The harmonic envelope of the music was somewhat thin, much as it is with many solid state preamps, even very good ones. I do not mean to say that the overall sound was thin. It did not lack midbass foundation, but the harmonic development of each instrument in the sound field was somewhat incomplete.
For example, when listening to the excellent Vivaldi Concertos for Three Violins with soloist Carmignola on Divox Antiqua, I seemed to hear the strings more than the body of the violin. This made it easy to discern the way in which the piece was being played, but lessened the impact of what was being played. All instruments were treated similarly—it was quite easy to hear the notes, even in a complex arrangement like a chamber orchestra, but I didn't quite get the sense that I was hearing complete, solid instruments. Images seemed to be less dimensional and more "see-through." I am accustomed to tube preamps creating very dimensional sound, not just in space but in tonal quality, with images having a tonal density or body they did not quite have with the Pearl. "Nearly like a top solid state preamp," I could hear myself thinking, and yet not. The sound was still flowing in a way that only tubes seem to be able to carry off, and it had a smooth texture—"liquidity" would be an appropriate term. This texture was seamless, owing to the Pearl's high level of coherence, but it was noticeable. While most tube preamps display a darkness of character, the Pearl displayed a transparent whiteness that was unique in my experience.
This preamp had a "character," which does not mean that it was not transparent, or that it did not provide a largely truthful portrayal of music. Where did this character come from? The best explanation I have heard was put forward by Harry Pearson, who posited that a large part of the character of an electronic component does not come from its steady-state frequency response (practically all preamps measure flat in the audio band), but from its dynamic response throughout the frequency range. If, for example, a preamp has a better dynamic response in the bass than in the treble, it will have a somewhat dark character. In the best case, the bass is not smearing or obscuring the upper frequencies, but is merely creating more powerful transients in the bass relative to the upper frequencies. Transparency is preserved, but the overall character of the preamp is shifted to the lower frequencies. This is the sensation that I got from the Wyetech, except that it had better dynamics in the upper/mid frequencies relative to the lower ones, resulting in a slightly whitish character.
What were some of the other consequences of this apparent shift in dynamics? One was a flattening of the soundstage. I had encountered this with the Sphinx Project 4. The Pearl's soundstage was not as flat as the Sphinx's, but it had a similar overall tonal balance, especially in the high frequencies, which didn't seem to be as natural as the rest of the spectrum. The high frequencies in the Pearl seemed to be projected slightly ahead of the rest of the spectrum. Since the upper harmonics and overtones are in this region, they affect the sense of space, so the Pearl was foreshortening the sensation of depth. This was apparent on all recordings, but most noticeable on those with a deep soundstage. With studio-made recordings in which excessive reverb had not been added, the effect was essentially unnoticeable. These recordings sound flat with all preamps.
About a month into the review, a friend came over to hear my system with the Pearl in place. He greatly enjoyed the system's resolution and power, but was bothered by a lack of emotional involvement. He attributed this, at least partially, to his feeling that the soundstage was too flat. When I switched back to the Transcendent Sound preamp, he immediately said that he could now "feel" the music, because the sense of depth, spaciousness, and body that was missing had now returned.
I had pretty much made up my mind about the Wyetech, deciding that while it had attributes that made it near-state-of-the-art, it was lacking in some critical areas in which even the inexpensive (but modified) Transcendent Sound sounded better. Nevertheless, I was not ready to give up on the Pearl so easily. I decided not to listen to it for some time, but I left it on continuously, thinking that it might need further break-in after all. During this time, I acquired the Silvaweld SWC1000 and VacuumState FVP5a preamps for comparison, and spent the next few weeks becoming acquainted with them in preparation for comparison to the Wyetech.
As I said earlier, I heard no significant change in the character of the Pearl after it had been on continuously for nearly three weeks. I am not usually a believer in huge sound differences resulting from break-in, but after the Pearl had been on continuously for at least two months, it improved significantly. To be sure, it still had the same basic character. It was still fast and dynamic rather than rich and lush, but now the high frequencies seem to be much better proportioned. Either they had taken a step back into the soundstage or the lower frequencies had taken one step forward. As a result, there was a sense of depth that had not been there before. The soundstage depth seemed more natural, as if it was really in the recording as the Pearl presented it. If not, there was not a false sense of depth, unlike the Transcendent, which slightly exaggerates this effect. In addition, the quality of the high frequencies had improved. I could hear the harmonic development of sound in a cymbal strike with some of the complexity of tone and decay that this instrument creates. Before, the sound was quite splashy, and lacked that distinct impact, bloom, and decay. I still didn't hear all the complexity that I know was there, but the shaping up of the high frequencies had greatly lessened two sources of irritation.
I was not sure whether the improvement in the high frequencies had also affected the preamp's texture, but something had, because I could barely sense the "translucent white silk" that I had heard before The Pearl was now largely without discernible texture, a good thing in my book. Those not accustomed to this level of performance would probably not notice what texture there was. Was the Pearl now perfect? Not quite, as I still felt that the tonal density of images was less than it should be, and the sense of thinness persisted, leading me to believe that this was how the designer wanted his preamp to sound. I believe that this is related to how the tubes are biased. Perhaps it is overly simplistic to assume that a lower running current will lead to a less rich sound, but more complete-sounding preamps run pretty much full out in the bias department. A friend who is an electrical engineer (and a former tube electronics designer) suggested that my pet theory was at least plausible. Otherwise, I cannot explain why none of the other tube preamps I have tested (or heard, for that matter) sounded this way.
Despite all this, the Pearl no longer sounded fatiguing, and I found its presence in my system rather enjoyable. I was now much more enthusiastic about its performance, and no longer questioned whether it was performing at a high level. This brings up the question, "At a high level compared to what?" The answer is complex. While I felt that the Pearl had some deficiencies, I also felt that it should be compared to several real-world contenders, all of which would have their own strengths and weaknesses. I once again managed to wrangle up a few competitors that fell within shouting distance of the Pearl's retail price. These included the Silvaweld SWC 1000 (6000 Euro—approximately $7500), the VacuumState FVP-5a (4000 Euro for the kit —approximately $5000—and 7500 Euro built), and the Transcendent Sound Grounded Grid ($650). The Silvaweld and the VacuumState are reasonably close in price to the Pearl (given that the VacuumState includes a phono stage), and can be seen as direct competitors. The Transcendent Sound is a low-cost kit, but at approximately a tenth of the price, it would be instructive to see how it compared to the other three preamps. The Transcendent Sound also uses a similar grounded grid circuit to the Pearl, which made it a unique addition. Plus, I just happened to have it around.
The Silvaweld SWC 1000 is the big brother of the SWC 450 BFA that I reviewed a while back, and is effectively the top-of-the-range line stage from this company. (There is an even more expensive one, the Aphrodite, but only two exist.) The SWC1000 is much more impressively built than the SWC 450. It is a two-chassis design, with thick aluminum plates bolted to corner posts—the same as the Wyetech, but with even thicker plates. The power supply is choke filtered, tube rectified, and tube regulated. In fact, there are no transistors to be found in this preamp. Most unusually, it uses a 300B tube as the high-voltage regulator, which mean that it runs very warm. The circuit chassis has two 6922 and four 417a tubes in a true balanced configuration. I asked several knowledgeable people about the unusual use of the 417a, and they told me that it was an excellent tube but difficult to use as it is susceptible to microphonics and high-frequency oscillation. The volume controls are Silvaweld's own 31-step silver attenuators, and all wiring is Teflon-jacketed solid-core silver. Source selection is by switched relays. Balanced and unbalanced inputs and outputs are provided with transformer coupling to convert unbalanced signals to balanced and vice versa. This preamp is a true exotic, with impeccable build quality and beautiful esthetics—as it should be for the price.
The VacuumState FVP5a, on loan from tube guru Allen Wright, was not an elegant piece of equipment as it turned to be his prototype unit, which had been built into an old chassis. Because the preamp uses quite a few tubes, it runs hot. I mean, it was cookin'! Wright told me that the chassis was too small for the number of tubes, given their bias, but he assured me that this unit was up to their current production specs. Interestingly, it was the only full-function preamp in the trial, and is capable of handling both MM and MC cartridges. I was informed that the MC stage is a hybrid cascode design, using a bipolar transistor to provide the extra boost needed for low-output phono cartridges. Wright referred to the output stage as a "super-linear" cathode follower. The high voltages are fully regulated by a custom-designed shunt regulator known as a SuperReg, which gives the preamp an extremely stable power supply. According to Wright, the only good regulator is a shunt regulator, which is used in both the Wyetech and Silvaweld preamps. From reading Wright's Preamp Cookbook, I gathered that this is a mature design that has evolved over the last twenty years or so.
I own the Transcendent Sound Grounded Grid preamp, and bought it because: (1) it was cheap, (2) it had a good reputation and innovative circuitry, (3) it looked easy to build, and (4) I could easily experiment with it. Its total cost was $685 including shipping from Switzerland and a 24-step attenuator from Elma (the same attenuator that is in the Wyetech). I assembled the preamp in about seven hours and commenced testing almost immediately. It sounded tight, fast, rhythmic, and too harmonically lean—similar to what I heard at first from the Wyetech. I suspected that the reason for the leanness was the 7dB of negative feedback, and to test this theory, I took the preamp to an electrical engineer who swapped the 100kOhm feedback resistor for a 500kOhm resistor. The result was a much nicer harmonic presentation, with little or no loss in speed, along with a distinct loss in bass power. This wouldn't do, so we went to a 350kOhm resistor that proved better. The bass had punch again, but the harmonic presentation had changed very little—a good thing. Still, I continued to hear something wrong in the bass. My electrical engineer friend suggested increasing the value of the coupling capacitors from 1μF to 10μF and the use of top-quality Mundorf capacitors. This took care of most of the bass coherence problems, and it was in this configuration, along with quite a lot of chassis damping, that this preamp was used for comparisons.
I set the output of each preamp to within 0.5dB using pink noise and a calibrated SPL meter. I used tracks from the following recordings: Chopin, Etudes/Preludes/Polonaises with Marizio Pollini on Deutsche Grammophone LP. This is a fantastic recording of solo piano from one of the great pianists of the last thirty years. It is recorded somewhat up-close, but not as badly as some piano recordings I have heard. The dynamics are astounding, with approximately 30dB of dynamic range. Playing this record will indicate if the system has any hint of strain. It also tests the quality of low-level dynamic shifts, and is a good test of transients, tone, and note decay. Cannonball Adderly and Milt Jackson, Things Are Getting Better on Riverside LP is one of my favorite jazz albums because the music is superb and the recording is nearly holographic. The vibraphone is a great test for tonal density and 3D imaging, as is Cannonball's sax. Donald Fagen's The Nightfly, on CBS LP, is a great pop recording. The title track has a complex mix of instruments, all of them pushed up in the mix, which makes it difficult to follow the threads of the song. This means that the track is a great test of image focus and overall coherence. I again used recordings I made of my girlfriend Anna playing solo violin and violin with cello. I know these recordings the best because I made them, and because they are recorded in the real space of a music hall. Another recording that I used at the end of the survey is a minimalist recording on CD of Prokofiev's Romeo and Juliet, made by Russell Dawkins, a recording engineer new to me. Dawkins used only two ribbon microphones in Blumlein configuration to capture an entire symphony orchestra (the Ukranian Radio and TV orchestra). Only 4dB of compression was applied, along with a subtle, 0.75dB lift above 4kHz. This recording is extraordinary in its dynamic range and instrumental tone, and I have no better recording for these two attributes.
On Things are Getting Better, the Wyetech showed great instrument separation and transient response, but the highs were a bit washed out compared to the other three preamps. The bass was clearer than the Transcendent's, equal to the Silvaweld's, and had a bit less drive than the VacuumState's. Vibraphone transients were very good, again the equal of the Silvaweld's, but slightly superior to the VacuumState's and the Transcendent's, both of which seemed to round off the attack of the mallet by comparison. However, the decay of the instrument with the Pearl seemed to occur a little too quickly, while all of the other preamps held on longer. Image density was best with the Silvaweld, followed closely by the VacuumState and the Transcendent. The Pearl showed its usual leanness on the sax and vibes, but its coherence was up there with the Silvaweld's and the VacuumState's. The biggest drawback of the Transcendent was its somewhat murky and disconnected bass, in contrast to its stellar midrange and high-frequency performance. The VacuumState preamp was very good on this recording, with only a slight bit of warmth that I perceived as being related to a slight rounding of transients. Of the four preamps, the Silvaweld seemed to intrude less of its own character. Interestingly, the Transcendent and the VacuumState were the best at generating a sense of space, but the Silvaweld was close behind in this regard. The Wyetech created a flatter perspective.
On Chopin's Polonaise, Op. 26, No. 1, from Etudes/Preludes/Polonaises, the Pearl again showed superb transient response, making the attack of the hammer on the strings clearly audible. However, the piano sounded slightly smaller and less full than it should. This was more evident in the low notes, when the sound of the piano had more string and less soundboard. Imaging was superb, allowing me to actually sense the length of the piano strings. The Silvaweld and the VacuumState were equal in this regard, but they retained the attack and gave more of the body of the instrument, giving a decent impression of a grand piano in the room. The VacuumState retained the harmonic richness, impact, and sense of scale of a real piano, but the transient attack was the slightest bit rounded and lacked that sudden thwack when Polini really pounded on the keyboard. Still, the VacuumState did an excellent job of conveying the drive and power of the instrument, and was the best of the group in this regard. While the Transcendent shared this slight rounding of transients to about the same degree, I wouldn't call any of these preamps slow or muted, except in comparison with each other. Notes on the Transcendent were clean and clear, and even high-speed runs remained easily discernible.
In dynamics, where this record will separate the men from the boys, all four preamps did a convincing job, but the VacuumState was the best at conveying the big dynamic swings, with the Wyetech and the Silvaweld close behind. It was in this (and in bass control) that I could hear the limitations of the Transcendent's power supply. All of the more expensive preamps have had much design effort put into their power supplies. The Wyetech and the Silvaweld could both accurately portray the subtle shifts in emphasis and expression in quiet passages. The VacuumState was very close to them in conveying convincing microdynamics. The Transcendent was good in this regard, but not quite as good as the other preamps. The winner in the "It's a real piano in the room" sweepstake, however, was the Silvaweld, which blended all the elements I have described into the most coherent whole.
On Donald Fagen's The Nightfly, the tonal density issues with the Pearl were much less prominent. The Wyetech seemed to deliver the whole package—dynamics, transparency, tone, separation of instruments, and tight, driving bass. The Silvaweld was also excellent with this track, but possessed that little bit "extra" that, quite frankly, wouldn't be noticeable without direct comparison. The Transcendent and the VacuumState were extremely close. Surprisingly, the Transcendent's bass seemed quite tight and punchy, belying some of the coherence problems I had noted with the classical and jazz works. I suspect that The Nightfly, like most pop/rock recordings, does not make big transient demands on the electronics because its dynamic range is so compressed, and that the power supply of the Transcendent could cope more easily with its near-steady demand. In fact, if I mostly listened to pop/rock recordings, I would buy the Transcendent and not worry about missing anything. It gives up little ground to the big dogs with this type of music, and costs so much less.
It was clear that the character of each preamp was the same for all recordings—sometimes less prominent, sometimes more, but consistent. The lone exception was the bass performance of the Transcendent, which seemed to do better with rock/pop music. The only preamp that seemed to have little discernible character was the Silvaweld, which had the speed, transparency, and coherence of the Wyetech but a more completely developed harmonic envelope. One aspect of the Silvaweld's sound that could not be equaled by the other preamps in this (or the previous) survey was the tonal differentiation of instruments. What do I mean by this? If two violins or two oboes play the same note at the same time, it is possible to distinguish that the two instruments have different tonalities. You can hear that one musician is playing a Stradivarius and the other a modern instrument. The Silvaweld preserved these harmonic nuances better than the other preamps in the survey. The VacuumState could give some sense of this with top recordings, but lacked the relaxed ease that the Silvaweld showed. I felt that it was slightly undone in this regard by its subtle amount of added warmth. The Wyetech and Transcendent units hinted at this ability, but could not carry it off convincingly.
The preamps in my last roundup didn't get it either, and in fact, only the little Silvaweld suggested that it was missing. It wasn't until I heard its big brother, the SWC1000, pull off this trick that I realized its importance in creating musical realism. It also goes to show just how much even very good preamps homogenize music with their own colorations. The recording that best showed off this trait was the Prokofiev Romeo and Juliet, which sounded extremely close to a live performance in terms of tonal color. I had heard this music live just one week before, with Zürich's Tonhalle Orchestra in Zürich's Tonhalle, so my memory of the sound was quite fresh. Obviously, my system cannot reproduce the full dynamic impact of a large orchestra, but when I focused on the tone of instruments (mainly in quiet passages), I realized that this recording is largely correct, especially with the Silvaweld line stage in the system.
So is the Silvaweld a perfect preamp? No. It has one flaw that is evident when using the unbalanced inputs. Using the balanced inputs and outputs bypasses the coupling transformers that convert the unbalanced connections to balanced ones. These transformers are not totally transparent, and they color the sound. The effect is subtle, but noticeable with the high-resolution Caliper Signature and Stax ELS F81 speakers. It gave this essentially characterless preamp a hint of warmth that was more than the warmth inherent to the VacuumState and Transcendent preamps. As a result, I conducted all of my listening with the Silvaweld's balanced inputs and balanced outputs, even if the source was unbalanced.
This was not an easy survey to conduct because the four preamps all performed at a much higher level than any of the ones I had previously tested, with the exception of the Silvaweld SWC450. All had fewer colorations, greater transparency, and better dynamics than the ones in the last shootout. The Wyetech in particular impressed me in all but one area—the harmonic completeness of instruments. Once broken in, the Pearl had great coherence, transparency, micro- and macrodynamics, low-level resolution, presence, and imaging. Without comparing it to the other preamps, or to live, un-amplified music, I probably would not have noticed the deficiency, but I feel that in a highly resolving system, with other gear capable of preserving such subtleties, this characteristic of the Pearl will be clearly audible. The other, more severe problems that I heard earlier in the review process—the flattened soundstage, less-than-stellar high frequencies, and subtle, yet smooth texture—were all but eliminated with two months of continual on-time. While the Pearl's soundstage depth was not as believable as the Silvaweld's or the VacuumState's, and its highs were not at the same level, these problems were greatly attenuated after the full break-in period, and were no longer a drawback to me.
The Transcendent Sound Grounded Grid preamp is an outstanding performer for the money. It was competitive with the much more expensive competition in some areas, but was mainly let down by its bass integration, which will probably not be noticed by owners of non-time-coherent speakers. It was less noticeable, for instance, on the Apogees than on the Stax speakers. It also has a slightly exaggerated sense of depth and very slight rounded-off transients. Nevertheless, it is still a fast-sounding preamp, with good tonal density and fine transparency.
The VacuumState FVP-5a is a lot of preamp for a not-insubstantial amount of money, though the price includes a very good MM/MC phono stage for around $5000 (as a kit). I only focused on its line stage, but can report that the phono stage is good enough to give some quite expensive stand-alone units a run for their money. It is clearly better than the highly regarded Audible Illusions M3A (with John Curl gold board for MC) in all aspects of vinyl reproduction. As a line stage, it offers great dynamics, transparency, and musicality. Only a slightly warm character keeps it from being a top contender.
The Silvaweld SWC 1000 was the best overall preamp of the four I compared. It had less of its own character. It preserved detail, dynamics, and tonal subtleties better, and had equal or better macrodynamic capabilities than any preamp I have heard in my home. In my opinion, it is a reference-level preamp, on par with the top Kondo, Jadis, and Conrad Johnson preamps. (I would also put Allen Wright's top model, the RTP-3c, in that elite category, confirmed recently by a second lengthy listening session.) Only the small coloration I heard when using the unbalanced inputs and outputs marred its otherwise stellar performance.
This brings me to the final ranking of the preamps I have reviewed in the last year:
1. Silvaweld SWC 1000
2. VacuumState FVP-5a
3. Wyetech Pearl
4. Silvaweld SWC 450
5. Transcendent Sound Grounded Grid
6. Sphinx Project 4 MkIII
7. Audible Illusions Modulus M3A
8. Sonic Euphoria PLC
9. Jolida JD3000b
I dithered between putting the Wyetech second or third on the list. The problem was that the Wyetech and the VacuumState were extremely close in overall performance, yet their characters were so different that I had a difficult time deciding which was more correct. The main problem with the Wyetech was primarily subtractive, and was not really obvious unless the Pearl was directly compared to live music or to other preamps that are more developed in this regard. The VacuumState, for all its formidable strengths, added a subtle bit of warmth to the sound, but in the end I felt that it presented music somewhat more realistically, and this gave it a slight advantage. The Silvaweld SWC 450, discussed in my last review, also has a hint of added warmth and is slightly less transparent. I felt that these colorations, though subtle, were more distracting than the flaws of the Wyetech.
The best values, by far, are the Transcendent Sound Grounded Grid and the Silvaweld SWC 450, with the nod going to the Transcendent because it is less than half the price of the Silvaweld. If I had less than $3000 to spend on a preamp, I would take either of these two and not look back. Also note that even though the Jolida was ranked last on this list, it is more transparent and has better high-frequency reproduction than the Audible Illusions or the Sphinx. Only a definite, relatively large weakness in the midbass keeps this preamp from excelling. You can usually find the JD3000b used for just a couple hundred dollars. As such, it is much better than any of the cheap solid-state preamps I have heard.
The Wyetech Pearl is expensive, but performs as an expensive preamp should. It is not perfect, and has its own unique voice, but most of my early reservations about it are now gone. Although it performed at a high level, I am not going to give it a "go out and buy this now" recommendation because I think that it has stiff competition in its price range, and that it is missing that last bit of "rightness" possessed by (for me at least) the best preamp in this survey. Still, the Pearl (as well as its big brother the Opal, which would also fit price-wise in this competition) deserves serious consideration. No preamp is perfect, and the Pearl does so much right that it should be on any serious audiophile's short list.