PrimaLuna ProLogue Eight CD player Measurements
I examined the PrimaLuna ProLogue Eight's measured behavior mainly using Audio Precision's top-model SYS2722 system (see www.ap.com and "As We See It" in the January 2008 issue), as well as our Audio Precision System One and the Miller Audio Research Jitter Analyzer for some tests. As always, I experimented with the grounding between the player under test and the test set to give the lowest level of measured hum (see later).
The ProLogue Eight's error correction was one of the best I have encountered. While its digital output flagged errors when the gap in the data spiral on the Pierre Verany Test CD reached 2.5mm in length, the player's output didn't mute, and the sound of the 500Hz tone remained continuous. It finally muted when the data gaps reached 3mm in length. This is extraordinarily good performance.
The PrimaLuna's maximum output level at 1kHz was 2.13V, this 0.55dB above the CD standard's 2V. The output preserved absolute polarity; ie, was non-inverting. The player had a significantly higher output impedance than the norm, ranging from 2.7k ohms in the treble and midrange to an extraordinary 12k ohms at 20Hz. This player really does need to be used with an amplifier having an input of 50k ohms or more if the bass is not to sound a little lightweight. Many solid-state preamplifiers will not really be compatible with the ProLogue Eight; fortunately, PrimaLuna's ProLogue Three preamp has an input impedance of around 100k ohms in the bass and midrange. However, I believe that this high output impedance was the reason I felt the ProLogue sounded lean with the Parasound Halo JC 2 preamplifier, which has an input impedance at low and midrange frequencies of 27k ohms. By contrast, Fred Kaplan's Krell FBI amplifier has an unbalanced input impedance of 200k ohms, which is perfect for use with the PrimaLuna CD player.
Even so, with the Audio Precision System One's 100k ohm input impedance, the PrimaLuna's low-frequency response began to gently shelve down below 200Hz, reaching 0.5dB at 45Hz and 1dB at 18Hz (fig.1, top pair of traces). This will not be enough to make the player sound lightweight or lean, but it might lead to a feeling of improved low-frequency definition. At the other end of the spectrum is a slight rise in top-octave output. With pre-emphasized CD data (fig.1, bottom traces), the ProLogue Eight offers the not uncommon depression in the treble. As a result, the player will sound a little mellow on the few pre-emphasized discs that have been released. Channel separation (not shown) was better than 100dB in both directions in the upper midrange, but dropped a little at the frequency extremes.
Fig.1 PrimaLuna ProLogue Eight, frequency response at 12dBFS into 200k ohms, CD data (blue left, red right; 2dB/vertical div.).
To maintain historical continuity with past Stereophile reviews of digital components, I assess resolution by sweeping a 1/3-octave bandpass filter from 20kHz to 20Hz while the product under test decodes dithered data representing a 1kHz tone at 90dBFS. The resultant spectrum for the ProLogue Eight is shown in fig.2; while the left- and right-channel traces peak at 90dB, large peaks can also be seen centered at 60 and 180Hz, and a small one at 300Hz. As mentioned earlier, I experimented with the grounding between the PrimaLuna player and the Audio Precision test sets to give the lowest level of hum. However, the frequencies of the peaks in fig.2 suggest that they are due not to a ground loop but to magnetic coupling between the power-supply transformer and the audio circuitry. Fig.3 shows the same spectrum, but plotted on a linear frequency scale and derived in a different manner: by applying a Fast Fourier Transform to time-domain data. The hum components can again be seen, and though harmonic spuriae are absent, some stray, very-low-level tones are present.
Fig.2 PrimaLuna ProLogue Eight, 1/3-octave spectrum with noise and spuriae of dithered 1kHz tone at 90dBFS with 16-bit CD data (right channel dashed).
Fig.3 PrimaLuna ProLogue Eight, FFT-derived spectrum of 1kHz sinewave at 90dBFS into 200k ohms (blue left, red right; linear frequency scale).
Plotting the PrimaLuna's linearity error with a 500Hz tone swept from 60dBFS down to 120dBFS gave the traces shown in fig.4. The increasingly positive error below 90dB is due to analog noise. The ProLogue Eight's reproduction of an undithered 1kHz tone at exactly 90.31dBFS was good, with the three DC voltage levels clearly visible (fig.5). However, some higher-frequency analog noise is present, leading to more jagged-looking traces than usual. In addition, the low-frequency power-supply noise overlays the left- and right-channel traces and moves them apart. It is fair to note that I couldn't hear this AC supply hum from my listening chair at normal playback levels, but I could hear it when I turned up the volume to rock-out levels with recordings having wide dynamic range, such as Attention Screen's Live at Merkin Hall (Stereophile STPH018-2). By contrast, the Benchmark DAC 1 was deathly quiet at the same volume setting.
Fig.4 PrimaLuna ProLogue Eight, linearity error.
Fig.5 PrimaLuna ProLogue Eight, waveform of undithered 1kHz sinewave at 90.31dBFS, 16-bit data (blue left, red right).
I had been warned that the PrimaLuna ProLogue Eight would show an increase in distortion with continuous tones approaching 0dBFS. PrimaLuna's design engineer in Holland, Marcel Crouse, says that this is due to saturation in the 12AX7 input tube. This "may seem odd in this time and age of vanishing low distortions and noise figures," he explained, "but we think that the real gain in sound quality is largely obtained elsewhere. Not that we purposely neglect harmonics and static noise, but we think that, when they are so benign as to be essentially inaudible because music itself is full of lower-order harmonics and hiss (such as tape hiss or ambient hall noise), we'd better concentrate on distortions that generate entirely new frequencies that are not present in the musical instrument at all in the first placedistortions that shift harmonics in phase, which degrade attacks and dynamics; and noise modulation that impairs transparency, focus, and imagingall artifacts that we believe are far more seriously destroying the quality of the reproduction and the fun of listening to well-recorded good music."
Fair enough. However, I measured fairly high distortion with a maximum-level 1kHz tone driven into 100k ohms, the second harmonic lying at 36dB and the third at 43dB, with many higher-order spuriae visible. The situation was similar at low frequencies. Reducing the signal level to 3dBFS dropped these harmonics by 6dB or so (fig.6), but not until I reduced the signal level to 10dBS did all the harmonics drop to 50dB (0.1%) or below (fig.7). The 60Hz and 180Hz hum components can also be seen in figs. 5 and 6.
Fig.6 PrimaLuna ProLogue Eight, spectrum of 1kHz sinewave at 3dBFS into 200k ohms (blue left, red right; linear frequency scale).
Fig.7 PrimaLuna ProLogue Eight, spectrum of 1kHz sinewave at 10dBFS into 200k ohms (blue left, red right; linear frequency scale).
Concerned that we might have been sent a faulty sample, I contacted Marcel Crouse, who reassured me that my measurements were typical. "You saw the same thing in the PrimaLuna ProLogue Three preamplifier [reviewed in December 2006]. The tube stage of the ProLogue Eight is the same [as that of the Three] save for the volume control. . . . Also, the 60Hz hum seems to be consistent; I measure 86dB at 50Hz. (It seems to be the power transformer's stray field.) So we may conclude that your review sample is not faulty."
Looking at intermodulation distortion at maximum level, using my usual equal mix of 19 and 20kHz tones, each at 6dBFS, was, as expected from the earlier results, disappointing, with the 1kHz difference tone at 40dB (1%) and many other higher-order products present (not shown). Dropping the signal level by 10dB gave the spectrum shown in fig.8. (Ignore the rise in the noise floor above 15kHz, which is due to the noiseshaping I used to generate a 16-bit signal from a 24-bit original.) The difference tone lies at 54dB (0.2%) ref. the signal (64dB in absolute terms), which is respectable. More important, all the other spuriae are well down in level, though a product can be seen at 23.9kHz, this due to a rather "leaky" anti-aliasing filter.
Fig.8 PrimaLuna ProLogue Eight, HF intermodulation spectrum, 19+20kHz at 10dBFS peak into 200k ohms (blue left, red right; linear frequency scale).
I used both the Miller Jitter Analyzer and the Audio Precision SYS2722 to examine the ProLogue Eight's rejection of word-clock jitter. The two systems gave similar results; the spectrum produced by the Audio Precision is shown in fig.9. The jitter level was a fairly low 406.5 picoseconds peakpeak. More significant, the highest-level sidebands that can be seen either side of the central 11.025kHz tone in fig.9 lie at power-supplyrelated frequencies of ±60Hz and ±120Hz. Actual data-related sidebands were much lower in level; if it weren't for the supply-related spuriae, the ProLogue Eight would have very good jitter rejection. I suspect that it is these supply-induced sidebands that led Fred and me to feel the ProLogue Eight's upper bass sounded a little soft.
Fig.9 PrimaLuna ProLogue Eight, high-resolution jitter spectrum of analog output signal, 11.025kHz at 6dBFS, sampled at 44.1kHz with LSB toggled at 229Hz, 16-bit data. Center frequency of trace, 11.025kHz; frequency range, ±3.5kHz (blue left, red right).
Summing up the PrimaLuna ProLogue Eight's measured performance is difficult, as some of the things it does wrong, such as the increasing distortion at levels above 10dBFS, will not be that audible with typical musica distorted snare-drum peak sounds the same as a clean one. But I was concerned about the effect of the transformer-radiated hum. I was also bothered by the very high output impedance at low frequencies, which will make system matching more difficult than usual.John Atkinson