Digital Processor Reviews

Sidebar 2: Measurements

The measured output impedances were pretty much to specification, at 27 ohms (Direct) and 495 ohms (Line), as was the frequency response, which measured –0.2dB at 4Hz and 20kHz, with only the merest hint of ripple in the top octave. Although the maximum output level from the direct sockets was 2.18V, as specified, that from the line-level sockets was a little higher at 1.05V. This is inconsequential, however, the 6dB reduction in output being sufficient to avoid overload problems with some older preamplifiers. As the analog output circuitry uses a DC servo to eliminate voltage offsets, I checked that, indeed, the outputs were at ground potential. They were.

Low-level linearity was among the best I have measured for a machine featuring the Philips 16-bit chip set, at –0.3dB/–0.4dB (L/R) and –4.7dB/–5.7dB (L/R) at the dithered –80.77dB and –90.31dB levels respectively. Listening to the fade of the dithered 500Hz tone from –60dB to –120dB on the CBS CD-1 test disc revealed a relatively pure tonal quality, which acquired a buzz of what sounded like mainly even-order harmonics as it approached, then went below –90dB. It was not quite as good as the Marantz CD-94 in this respect, which had a purer tone, though the more expensive machine did sound as if it had a slightly higher HF noise floor. Both sounded slightly cleaner than the ultra-expensive Sony DAS-R1, which had lower noise but more audible high-order harmonics.

The waveform of the undithered 1kHz tone (which is described by just the digital words for the levels –1, 0, and +1, ie, the equivalent of 3 LSBs) was rather asymmetrical, the positive-going half of the duty cycle lasting twice as long as the negative-going.

Looking at the 1/3-octave spectrum of the noise and spuriae present while the Black Box reproduces the –90.31dB dithered 1kHz tone on the CBS test disc (fig.1, footnote 1), the 1kHz tone can be seen at –96dB, reproduced approximately 5.5dB too low in level. Its second harmonic can be seen at 2kHz, and there is also a hint of fourth harmonic present, though higher-order harmonics, if present, are swamped by the general rising level of garbage in the top two audio octaves. (I assume that although this HF noise does contain a contribution from the higher distortion harmonics of the 1kHz tone, it is mainly due to the record dither used to ensure waveform purity when reproduced.) This ties in with the sound of the fading tone, which seemed to be even-order-harmonic–dominated in its tonality. The overall level of low-frequency noise is low, with only the full-wave-bridge power-supply 120Hz switching buzz noticeable at –103dB, its second harmonic lying at –106dB. This noise may have been measurable, but it was totally inaudible.—John Atkinson

Fig.1 Arcam Black Box, 1/3-octave spectrum of dithered 1kHz tone at –90dB.

---

Footnote 1: As the code representing this tone has had dither applied when the CD was cut, it should reproduce as a pure, if noisy, sinewave. Any distortion components present, therefore, can be laid at the door of the individual player's decoder and electronics. As this graphical representation of a player's low-level performance is more informative than a straight statement of by how many dB it compresses or expands the level at –90.31dB, I think that from now on we shall make this measurement (an average of an arbitrarily decided number of sample spectra, 11) a standard feature of Stereophile's CD player and decoder reviews.—John Atkinson