Philips SACD1000 SACD/DVD-Video player Measurements
I did an uncharacteristic thing before measuring the Philips SACD1000: I read the manual. And it's just as well, as doing the obvious thing and plugging a pair of interconnects into the pair of analog outputs marked "Stereo" would have led me astray—these jacks output a signal band-limited to 20kHz even when a wide-bandwidth SACD is being played. It appears that they're supplied solely for connecting the player to a home-theater system for stereo playback of DVD-Video discs. I therefore did as Chip did and assessed the SACD1000's performance from its Left and Right Front Surround outputs.
Another point to note: The digital outputs source only a Dolby Digital or DTS bitstream from DVD-Vs. They do not provide a PCM output even from CDs, so those who want to use an external high-end DAC for CD playback need be warned: The SACD1000 is a standalone player for SACDs (as expected) and CDs (not expected). It must be used with an external surround processor or receiver for DVD-V surround playback, as the analog surround outputs work only with surround-encoded SACDs.
The maximum output level (MOL) was a lowish 1.86V from the discrete L/R surround outputs, and just above the 2V CD standard, at 2.1V from the DVD "Stereo" outputs. This was with the '1000 playing CDs; playing the "tentative" Sony Test SACD, the maximum level at 1kHz from the discrete L/R jacks was 2.1V, or 1dB higher than the nominal 0dBFS level from CD, which will bias direct SACD/CD comparisons in favor of SACD. Signal polarity was non-inverting from all analog outputs, and the source impedance varied from a low 34 ohms across most of the audioband to 40 ohms at 20Hz.
Playing back CDs, the frequency response (fig.1, top traces) was flat over almost all the band, with just a very slight shelving visible above 5kHz. This shelving was accentuated with pre-emphasized data (fig.1, lower traces), though not nearly to the point where sound quality would be adversely affected. Playing back Sony's test SACD, the ultrasonic response rolled off differently, depending on the setting of the rear-panel "filter" switch. With this set to "50kHz" (fig.2, top traces), the output was exactly 3dB down at 50kHz. At "40kHz," the -3dB point was exactly 40kHz. In each case, the response was down by at least 20dB at 90kHz. Crosstalk between the discrete Left and Right channels was beneath the noise floor below 1kHz (fig.3), rising above the audioband to reach a still-superb -85dB at 80kHz.
Fig.1 Philips SACD1000, CD data, frequency response at -12dBFS, without emphasis (top) and with emphasis (bottom). (Right channel dashed, 0.5dB/vertical div.)
Fig.2 Philips SACD1000, SACD data, frequency response at -3dBFS with filter set to "50kHz" (top) and "40kHz." (Right channel dashed, 2dB/vertical div.)
Fig.3 Philips SACD1000, SACD data, channel separation ref. 0dBFS: L-R (top), R-L (bottom). (10dB/vertical div.)
Fig.4 shows spectral analyses of the SACD1000's output while playing CD data representing a dithered 1kHz tone at -90dBFS (top traces below 8kHz) and SACD data (bottom traces below 8kHz). (The 1dB difference in MOLs for the two media has been compensated for.) The DSD-encoded data on the SACD has about a 12dB greater dynamic range in the low treble and bass regions than the 16-bit PCM data on the CD, at least as far as the left channel (solid traces) is concerned. The right channel has some suspicious-looking but still low-level peaks at 60Hz, 120Hz, and 240Hz with both CD and SACD playback. These are, of course, power-supply-related, and might have been affected by the grounding between the player and the Audio Precision System One test set. Above 8kHz, the aggressive noiseshaping used by SACD's DSD encoding makes its presence felt, and SACD actually has less dynamic range than CD in the top audio octave.
Fig.4 Philips SACD1000, 1/3-octave spectrum of dithered 1kHz tone at -90dBFS, with noise and spuriae, 16-bit CD data (top below 8kHz) and DSD SACD data (bottom below 8kHz). (Right channel dashed.)
The Sony test SACD doesn't appear to have the equivalent of CD's "digital black" signal, but it does have 1kHz tones encoded down to a roots-of-the-universe -160dBFS. To compare the SACD1000's noise spectra with no signal playing but with the mute lifted, I therefore performed wideband analyses of its analog output playing "digital black" on a test CD and the -160dBFS tone on the Sony SACD. The results are shown in fig.5. With CD replay, the 16-bit noise floor is, again, about 12dB higher than the SACD noise floor, at least in the midrange and bass. The two are equivalent in the mid-treble, and while the CD floor features a rise in energy above the audioband, presumably due to the player making use of a bitstream DAC, this rise is dwarfed by that featured in the SACD spectrum.
Fig.5 Philips SACD1000, 1/3-octave spectrum of "digital black" (16-bit CD data) with noise and spuriae (top below 4kHz), and of dithered 1kHz tone at -160dBFS (DSD SACD data, bottom below 4kHz). (Right channel dashed.)