Sidebar 3: Measurements
Like the Naim CD5 XS, which Art Dudley reviewed in the November 2017 issue, the Rega Apollo has no digital inputs of any kind. That limited my measurement of its technical behavior to using 16-bit test files burned to a CD-R. I tested the Rega with my Audio Precision SYS2722 system (see the January 2008 As We See It").
The Rega's error correction was good—no interruptions were apparent in the player's output until the single gaps in the data spiral on the Pierre Verany Digital Test CD reached 1.5mm in length, when there were occasional glitches. (The Compact Disc standard, the so-called "Red Book," requires that a player cope with gaps of up to 0.2mm.) The maximum output level from the Apollo's unbalanced output was 2.18V, which is 0.8dB higher than the CD standard's 2V, and the output preserved absolute polarity (ie, was non-inverting). The output impedance was 600 ohms at 20 and 1kHz, rising slightly to 660 ohms at 20Hz, presumably due to the presence of a series capacitor in the signal path.
Fig.1 shows the Apollo's impulse response: It's typical of a linear-phase reconstruction filter, with symmetrical ringing either side of the single sample at 0dBFS. Tested with white noise sampled at 44.1kHz, the Rega's output rolled off quickly above 20kHz (fig.2, red and magenta traces), but hadn't reached full attenuation by the Nyquist frequency, 22.05kHz (fig.2, vertical green line). However, the aliased image at 25kHz of a full-scale 19.1kHz tone (blue and cyan traces) is suppressed by 100dB. The distortion harmonics of that tone are visible at 38.2 and 57.3kHz, at a respective –92dBFS (0.0025%) and –104dBFS (0.0006%). The blue and red traces in fig.3 show the Apollo's audioband response taken with spot tones; it is flat to 15kHz, and reveals excellent channel matching. The cyan and magenta traces in fig.3 show the response with preemphasized data: the output peaks by almost 1.5dB between 5 and 15kHz, which will make the sound a touch bright. Fortunately, pre-emphasized CDs are very rare these days.
Fig.1 Rega Apollo, impulse response (one sample at 0dBFS, 4ms time window).
Fig.2 Rega Apollo, wideband spectrum of white noise at –4dBFS (left channel red, right magenta) and 19.1kHz tone at 0dBFS (left blue, right cyan), with CD data (20dB/vertical div.).
Fig.3 Rega Apollo, frequency response at –12dBFS into 100k ohms: without deemphasis (left channel blue, right red), with deemphasis (left cyan, right magenta) (0.5dB/vertical div.).
Channel separation (not shown) was superb, at >100dB in both directions below 1kHz, but decreasing to 72dB at 20kHz. The analog noise floor was low in level, with no power-supply–related artifacts present. With dithered data representing a 1kHz tone at –90dBFS (fig.4), the graph actually shows the spectrum of the dither noise used to encode the 16-bit test signal. With undithered data representing a 1kHz tone at exactly –90.31dBFS, the three DC voltage levels described by the data are well defined, but with a very slight DC offset present in the right channel (fig.5).
Fig.4 Rega Apollo, spectrum with noise and spuriae of dithered 16-bit, 1kHz tone at –90dBFS (left channel blue, right red) (20dB/vertical div.).
Fig.5 Rega Apollo, waveform of undithered 16-bit, 1kHz sinewave at –90.31dBFS (left channel blue, right red).
The spectrum of a full-scale 50Hz tone into 600 ohms (fig.6) indicates that, even into this demanding impedance, the Apollo's distortion harmonics all lie at or below –99dB (0.001%). I then tested the Rega for intermodulation distortion with an equal mix of 19 and 20kHz tones. The resultant spectrum (fig.7) looks a little hashy in the audioband, but this is an artifact of the 16-bit encoding. Actual intermodulation products are very low in level even into 600 ohms, the second-order difference product at 1kHz lying close to –110dB (0.0003%). However, fig.7 shows an odd rise in the noise floor at the top of the audioband; this can also be just made out in fig.2.
Fig.6 Rega Apollo, spectrum of 50Hz sinewave, DC–1kHz, at 0dBFS into 600 ohms (left channel blue, right red; linear frequency scale).
Fig.7 Rega Apollo, HF intermodulation spectrum, DC–30kHz, 19+20kHz at 0dBFS into 600 ohms (left channel blue, right red; linear frequency scale).
Finally, when I tested the Rega Apollo with 16-bit J-Test data (fig.8), the spectral spike that represents the high-level tone at exactly one-quarter the sample rate is well defined, and most of the odd-order harmonics of the LSB-level, low-frequency squarewave are close to the correct levels (sloping green line). However, the harmonics closest to the 11.025kHz are too high in level, suggesting that the Apollo's rejection of word-clock jitter is not quite to the otherwise excellent standard of digital engineering revealed by the rest of its measured performance.—John Atkinson
Fig.8 Rega Apollo, high-resolution jitter spectrum of analog output signal, 11.025kHz at –6dBFS, sampled at 44.1kHz with LSB toggled at 229Hz: CD data (left channel blue, right red). Center frequency of trace, 11.025kHz; frequency range, ±3.5kHz.
