The Elgar's linearity (fig.3) is indeed excellent, there being no deviation from perfect until -110dBFS, due to the unit's very low noise floor. The waveform of an undithered -90.31dBFS, 1kHz sinewave with 20-bit resolution (fig.4) looks
like a sinewave. Only in the otherwise superb noise-modulation plot (fig.5) is there a hint of what might be a spurious component at 2kHz. The Elgar's output spectrum while decoding 16-bit data representing an equal mix of 19kHz and 20kHz tones, each with a level of -6dBFS (not shown), was clean. While the 1kHz difference component was noticeable by its absence, there were a few high-order components visible above the noise floor. These should be inconsequential, however.
Fig.3 dCS Elgar, departure from linearity (right channel dashed, 2dB/vertical div.).
Fig.4 dCS Elgar, waveform of undithered 1kHz sinewave at -90.31dBFS (20-bit data).
Fig.5 dCS Elgar, noise modulation, -60dBFS to -100dBFS (10dB/vertical div.).
Finally, I wasn't able to perform any jitter tests. However, as Stereophile has just acquired one of Paul Miller's new test instruments, the Elgar's jitter measurements will be included in a Follow-Up. -- John Atkinson