Chord Electronics DAC64 D/A processor Page 2
In auditioning modern digital components, there are times when I begin to doubt whether there are any differences to be heard. This was not the case when I began auditioning the first sample of the Chord DAC64, using the Mark Levinson No.31.5 CD transport and with the Chord's RAM buffer switched out. Not only was the sound hard-edged and fatiguing, there was a noticeable sourness, as if the musicians were no longer playing in tune. Switching in the Chord's RAM buffer to its smaller size brought about a significant improvement in this respect, although the tonal balance remained bright. Peculiarly, when I set the RAM buffer to its larger setting the sound was, if anything, more fatiguing than with the smaller buffer—not less, as I'd expected.
I usually leave the measurements until after the auditioning, so my perceptions will not be affected by any expectations formed by the test results. But I couldn't believe the DAC64 was performing as its manufacturer intended, and so put it on the test bench to take a look at what it was doing. As I report in the "Measurements" Sidebar, the Chord's DAC was producing very high amounts of word-clock jitter without the RAM FIFO buffer and had a compromised dynamic range, but there was nothing to indicate that the unit was broken. I finished up my listening and had begun to write a negative review when I received an e-mail from Chord's John Franks, in the UK. Apparently, a manufacturing problem had affected some of the first batch of DAC64s, and Chord wanted to submit a second sample.
Franks had attached spectral analyses to his e-mail, showing the effect of jitter with and without the RAM buffer of both a faulty sample (with results identical to what I had obtained) and of a revised sample (with very much better results). I agreed, therefore, to put my review on hold while a new sample was shipped from the UK. The serial number of the first, substandard sample was USD013, that of the second sample was USD015. Following a comparison of the two samples, all of my subsequent listening was done with '015.
Soundwise, the two samples were chalk and cheese. Whereas the first was unlistenable without its FIFO engaged, having that peculiar sourness to its sonic character, the second DAC64 sounded warm, spacious, and smooth. Obviously, whatever the problem had been with USD013, it had been solved in USD015. I sat down to some serious listening.
There was no doubt in my mind that, without the RAM buffer, the DAC64 with CD sources was fairly ordinary. Tonally it sounded a bit dry, soundstage-wise a bit shallow. Switching the buffer in-circuit both sweetened the tonal balance and made the perceived stage deeper. (All my following comments refer to the sound with the RAM buffer engaged.) A/B comparisons of the two FIFO options were problematic because of the four-second time delay when I changed from one to the other. But if I had to swear on it, I would say that the larger FIFO made recorded piano sound slightly more "of a piece," the image slightly more fleshed-out and the high frequencies more silky-smooth. The different characters of all the cymbals Billy Drummond uses on Rendezvous (Stereophile STPH013-2) were superbly differentiated.
Tonally, the Chord's balance didn't favor one frequency range over another. But again, if I had to swear to it, I would say that the low frequencies were a tad more bloated than I'm used to in my system. The double-basses of the Rochester Philharmonic, as conducted by Christopher Seaman in Rachmaninoff's Piano Concerto 3 (Harmonia Mundi USA HMU 907286), sounded more gruff than I'd expected. But oh, how superbly real Jon Nakamatsu's Steinway sounded. There was a dynamic sweep to the $3000 Chord's presentation that was both addictive and not dissimilar to the way the $17,500 Mark Levinson No.30.6 presented the music's contrasts. And the DAC64 definitely got right the characteristic sound of brass instruments, with a raspy buzz to the leading edges that is too easily and incorrectly smoothed over by other components. But the nod would still have to go to the Levinson, for its better delineation of sonic objects within the image.