Oddly, the Theorem seemed to have as much extension in the lowest octave; the Cobalt's weightier presentation came more from the midbass rather than the lowest frequencies.
The combination of the Cobalt's hefty presentation and good dynamics teamed up to present a real sense of slam. Kick drum had a nice punch and drive, adding to the Cobalt's excellent pace. Moreover, the bass was taut and fast rather than slow and bloated. Some processors with weighty presentations tend to put a drag on the rhythm—the Cobalt achieved LF heft with speed and definition. As good as the Cobalt's bass was, it wasn't in the same league as Theta's Generation III in terms of drive, dynamics, or extension. Compared to the Theorem, it was no contest: the Cobalt had far better pace, weight, slam, and a more satisfying tonal balance.
Soundstaging performance—usually a Theta hallmark—was mixed. The Cobalt's resolution of image outlines and ability to throw a focused soundstage were excellent, but bettered by the Theorem's ability to convey a sense of depth and space. On naturally miked recordings, the Theorem was clearly superior at revealing low-level spatial cues that convey the impression of the listening room being replaced by the recorded acoustic. On Three-Way Mirror (Reference Recordings RR-24CD), for example, the stunning sense of space, depth, and the room's size were better resolved by the Theorem. The Cobalt's presentation was just a little flatter, lacking the three-dimensionality revealed by the Theorem. Moreover, the Theorem presented a nice bloom and "puff" of air around instrumental and vocal outlines not heard from the Cobalt. This quality better conveyed the impression of instruments hanging in space before the listener. Nevertheless, the Cobalt excelled at locating images at precise points in that space; the Theorem tended to be somewhat more diffuse, lacking the Cobalt's sharp delineation of individual outlines.
Despite the Cobalt's excellent image focus, it didn't do as well in resolving individual instrumental textures. There wasn't a clear sense that the presentation was made up of disparate and distinct instruments. Instead, the music tended to sound somewhat synthetic—like variations on the same tapestry rather than different tapestries. This was perhaps exacerbated by the Cobalt's slight layer of grain overlaying instrumental and vocal textures. One cannot expect miracles from a $599 processor, but the Theorem was better at portraying the varying tonal colors and nuances of different instruments.
In terms of transparency, the Cobalt was better than one would expect for its price, but didn't match the Theorem. The Cobalt seemed to put a very fine veil between me and the music, taking away that edge of realism. There was a slight thickness to the sound, in contrast to the Theorem's open, clear, and airy presentation. This is, however, one area in which the Theorem excels; the Sumo processor has a superb sense of immediacy, life, and direct contact with the music. The Robert Lucas CD Usin' Man Blues (AudioQuest AQ-CD1001) was a good example. The Cobalt tended to lack the Theorem's ability to put Robert and his guitar in my listening room. It was like taking a step back from real life, or, put another way, of the imposition of a fine veil in front of the music. I should add that this is the Theorem's best quality—many much more expensive converters don't approach its transparency—and is tough competition.
The Cobalt had a slightly softer treble than the Theorem, a quality I found appealing. Although the Theorem's treble was cleaner and had less grain, the Cobalt's presentation was less etched and analytical—my primary criticism of the Theorem. Sibilance tended to be more pronounced through the Theorem, and cymbals were more forward. The Cobalt's combination of a less aggressive treble and more bass heft gave it a better overall tonal balance. The Theorem, however, clearly excelled in revealing treble detail. There was much more of that filigree quality from the Theorem, presenting the listener with more information. I should reiterate, however, that the Theorem crosses the line from resolution of detail to becoming overly analytical, in my view.
Overall, the Cobalt had some excellent qualities for its modest price: powerful, punchy, and well-defined bass, an unaggressive treble, good image focus, and lack of glare. In the debit side of the column, the Cobalt tended to fuse instrumental textures, overlay the music with a slight coarseness, foreshorten soundstage depth, and obscure fine detail. All in all, however, quite impressive for a $599 product.
Conclusion
At $599, the Cobalt 307 DAC represents another price/performance breakthrough in digital processors. It is clearly better than previous low-priced DACs, and competitive with much more expensive units. On the plus side, the Cobalt has a good overall tonal balance, a lack of hardness in the treble, and a powerful, extended bass. I was also very impressed by the Cobalt's dynamics and punch, qualities that gave the music good drive and pace. On the downside, I found the Cobalt to sound a little thick, with a slight trace of soundstage opacity. The music didn't have quite the same immediacy and realism heard through the $799 Sumo Theorem. There was also a fine layer of grain overlaying the music in relation to the Sumo processor. Finally, the Cobalt was only fair at resolving differences between instrumental textures. These criticisms must be taken in the context of the Cobalt's cost; many imperfections can be overlooked at this price level.
The Cobalt's toughest competition, the Sumo Theorem, presents some interesting sonic tradeoffs. Although both are excellent for the money, they have very different presentations. Music in which bass drive and dynamics are important was better served by the Cobalt; selections of naturally miked acoustic music were portrayed with more realism by the Theorem. I should add that the Cobalt's $200 lower price is not insignificant at this level. Both are "must audition" products for those looking for an under–$1000, high-performance converter.
All things considered, I can recommend the Cobalt 307 DAC. It offers more than a taste of high-end performance at a price virtually everyone can live with.
At $599, the Cobalt 307 DAC represents another price/performance breakthrough in digital processors. It is clearly better than previous low-priced DACs, and competitive with much more expensive units. On the plus side, the Cobalt has a good overall tonal balance, a lack of hardness in the treble, and a powerful, extended bass. I was also very impressed by the Cobalt's dynamics and punch, qualities that gave the music good drive and pace. On the downside, I found the Cobalt to sound a little thick, with a slight trace of soundstage opacity. The music didn't have quite the same immediacy and realism heard through the $799 Sumo Theorem. There was also a fine layer of grain overlaying the music in relation to the Sumo processor. Finally, the Cobalt was only fair at resolving differences between instrumental textures. These criticisms must be taken in the context of the Cobalt's cost; many imperfections can be overlooked at this price level.
