Sony SCD-XA9000ES SACD player Measurements

Sidebar 3: Measurements

The Sony SCD-XA9000ES's maximum output level is specified as 2V, conforming to the CD standard, but it met this only from the multichannel outputs playing SACD. From the two-channel outputs playing CD, the output was an audible 0.57dB lower, at 1.875V; playing SACD, the output was 0.31dB lower, at 1.93V. These differences will be just audible, in favor of SACD in direct comparisons.

The output impedance from all outputs was a low 109 ohms across most of the band, this rising inconsequentially to 132 ohms at 20Hz. The Sony preserved absolute polarity, and its error correction of CD playback was among the best I have encountered. It coped with 1.5mm gaps in the data spiral on the Pierre Verany Test CD without sonic glitches.

The XA9000's frequency response for CD playback was effectively flat across the entire audioband with its reconstruction filter set to Standard (fig.1, top pair of traces). Setting the filter to Optional rolled off the response above 14kHz (fig.1, middle traces). There was no significant difference in response with pre-emphasized CD data (fig.1, bottom traces). With SACD data (fig.2), the output was down 1dB at 37kHz and 4dB at 50kHz, but the audioband response was still flat. In this respect, there was no difference between the two-channel and multichannel outputs. Channel separation between adjacent output pairs (not shown) was superb, at better than 120dB, at frequencies below 1kHz (with the exception of a slight "bump" in the crosstalk curve at 200Hz). Above 1kHz, the separation degraded at 6dB/octave due to the usual capacitive coupling. However, it was a still-excellent 95dB at 20kHz.

Fig.1 Sony SCD-XA9000ES, CD frequency response at -12dBFS into 100k ohms, with de-emphasis (bottom) and without (top), the latter with Standard filter (top) and Optional filter (middle). (Right channel dashed, 0.5dB/vertical div.)

Fig.2 Sony SCD-XA9000ES, SACD frequency response at -3dBFS into 100k ohms (right channel dashed, 1dB/vertical div.).

To compile fig.3, I performed 1/3-octave spectral analyses on the Sony's two-channel analog outputs while it decoded three different signals: 1) dithered 16-bit data from CD representing a 1kHz tone at -90dBFS; 2) DSD data from SACD, also representing a 1kHz tone at -90dBFS; and 3) DSD data from SACD representing a 1kHz tone at -120dBFS, this below the noise floor with CD. All three signals peak at the correct level in this graph, implying very low linearity error, and the traces are free from spuriae related to harmonic distortion. The increased dynamic range of SACD in the midrange and bass compared with CD unmasks a faint whisper of 60Hz hum, though at -126dBFS (left) and -129dBFS (right), this is negligible.

Fig.3 Sony SCD-XA9000ES, 1/3-octave spectrum of dithered 1kHz tone at -90dBFS, with noise and spuriae, 16-bit CD data (top) and DSD data (middle), with dithered 1kHz tone at -120dBFS, DSD data (bottom). (Right channel dashed.)

However, all three signals produced a peak at 200Hz in both channels in this graph, this not coincidentally the frequency of the bump in the crosstalk mentioned earlier. And SACD's lower-midrange noise floor reveals, with the -120dBFS data, peaks at the harmonically related frequencies of 400Hz, 600Hz, and 800Hz. (Analysis of the player's output while it decoded "digital black," not shown, found nothing at 1kHz, the fifth harmonic of the 200Hz tone, though it did reveal the presence of a trace of the sixth, at 1.2kHz.) Though the 200Hz tone peaks at just -107dBFS, which will be inaudible, it is something that should not be present at all.

Fig.4 shows the XA9000's two-channel output linearity error for CD playback, plotted as usual using a dithered 500Hz tone and a bandpass filter centered on this frequency. The Sony's noise floor is low enough that the plot is dominated by the recorded dither down to below -110dBFS, which is excellent. Repeating the test with the 1kHz spot tones on Sony's "Provisional" test SACD revealed just +1dB level error at -120dBFS and +6dB at -130dBFS, the latter basically dominated by the player's very low analog noise floor. The combination of low noise and low linearity error means that the XA9000's reproduction of an undithered tone at exactly -90.31dBFS is essentially perfect (fig.5). The three voltage levels comprising this signal can be clearly seen, and the waveform has excellent symmetry about the time axis.

Fig.4 Sony SCD-XA9000ES, right-channel departure from linearity, 16-bit CD data (2dB/vertical div.).

Fig.5 Sony SCD-XA9000ES, waveform of undithered 1kHz sinewave at -90.31dBFS, 16-bit CD data.

Sony specifies the SCD-XA9000ES into test loads of at least 10k ohms. Even so, the spectrum of its output into a load as low as 600 ohms differed from that into the lab 100k load only in having a little more high-order harmonics (not shown), though these were all still 100dB down. Into 8k ohms, probably the lowest load the player will see in real systems, the only distortion harmonics that can be seen above the noise floor are the second, third, and fourth (fig.6), with the third the highest in level in the left channel at -97dB (0.0014%). This harmonic lay at -107dB in the right channel, the same as the second harmonic in both channels. This is superb performance. Note, by the way, that the 200Hz idle tone can be seen in this graph, but peculiarly, only in the right channel (red trace).

Fig.6 Sony SCD-XA9000ES, spectrum of 1kHz sinewave, DC-1kHz, at 0dBFS into 8k ohms, SACD data (linear frequency scale).

Regarding intermodulation distortion, the Sony player produced an extraordinarily low amount of the second-order difference component into loads down to 600 ohms. Into 8k ohms, for example (fig.7), this lay at just -109dB (0.0004%). This graph was taken playing back CD data with the Standard filter. Switching to the Optional filter gave the spectrum shown in fig.8. This filter's early HF rolloff, seen in fig.1, drops the level of the tones, but because the ultimate rolloff rate is slower than the Standard filter, a strong reflection of the 20kHz tone appears at 24.1kHz; other aliased tones can be seen, albeit at very low levels, in the audible part of the spectrum. The tradeoff offered by this filter is better behavior in the time domain, and there is some evidence that the ear is more tolerant of small amounts of aliasing than it is of the time dispersion typical of very steep filters.

Fig.7 Sony SCD-XA9000ES, HF intermodulation spectrum, DC-25kHz, 19+20kHz at 0dBFS into 8k ohms, CD data, Standard filter (linear frequency scale).

Fig.8 Sony SCD-XA9000ES, HF intermodulation spectrum, DC-25kHz, 19+20kHz at 0dBFS into 8k ohms, CD data, Optional filter (linear frequency scale).

I used the Miller Audio Research Jitter Analyzer to look for evidence of inadequate word-clock jitter rejection in the XA9000's analog output. The measured level of jitter was superbly low at just 176 picoseconds peak-peak, while the spectrum of that jitter (fig.9) confirmed the low level of data-related sidebands (red numeric markers). The sideband pair highest in level were low in frequency, spaced 15.6Hz to the sides of the central 11.025kHz tone, and contributed 63ps to the total.

Fig.9 Sony SCD-XA9000ES, high-resolution jitter spectrum of analog output signal (11.025kHz at -6dBFS sampled at 44.1kHz with LSB toggled at 229Hz). Center frequency of trace, 11.025kHz; frequency range, ±3.5kHz. (Grayed-out trace is SACD playback of 11.025kHz tone.)

The noise floor in this graph is around 3dB higher than the lowest I have seen from CD playback. Though this is inconsequential, it does suggest that the Sony's dynamic range is a little less than is possible from CD. The grayed-out trace in fig.9 is a similar spectral analysis performed while the player decoded DSD data representing an 11.025kHz tone, without the low-frequency, low-level squarewave that the Miller Analyzer uses to diagnose word-clock jitter. The noise floor in this frequency region is around 4dB higher than with CD playback (presumably due to the noiseshaping used on SACD), and the sidebands at ±15.6dB are again evident, at the same level. But even allowing for the absence of data-related jitter sidebands, the SACD spectrum is noticeably cleaner than that of CD.

Overall, Sony's SCD-XA9000ES offers superb measured performance, as I have come to expect from this manufacturer. But even though there were no audible consequences that resulted from the presence of that spurious 200Hz tone, I would still have preferred not to see it.—John Atkinson

1 Sony Drive
Park Ridge, NJ 07656
(800) 222-7669