Marantz SA-11S2 Reference SACD/CD Player Measurements
The Marantz SA-11S2 Reference was assessed using Audio Precision's SYS2722 system (see www.ap.com and "As We See It" in the January 2008 issue), as well as, for some tests, our Audio Precision System One and Miller Audio Research Jitter Analyzer.
The Marantz's maximum output level at 1kHz varied according to whether CD or SACD was playing and which of the three filters had been selected. The level from the balanced XLR jacks was 4.4V with SACD and Filter 1 and 2, increasing to 5.23V with Filter 3. With CD, the maximum balanced level was 4.8V with Filter 1, 4.95V with Filter 2, and 5.0V with Filter 3. These differences in level will make direct comparisons of the sounds of the filters problematic. The unbalanced levels from the RCA jacks were exactly half the balanced figures, with Filter 3 almost 2dB higher than the CD standard's 2V RMS. The SA-11S2's output preserved absolute polarity; ie, was non-inverting from the unbalanced jacks, but inverted polarity from the balanced outputs, the XLR jacks being wired with pin 3 hot, the opposite of the AES standard. The unbalanced output impedance was a low 46 ohms at high and midrange frequencies, rising slightly to 52 ohms at 20Hz. The balanced impedances were exactly twice these figures, as expected.
The Marantz's error correction was superb, the output not suffering from any audible glitches or flagged errors in the digital output until the gaps in the data spiral on the Pierre Verany Test CD reached 2.4mm in length.
Playing SACDs, the SA-11S2's frequency response was identical with Filters 1 and 2 (fig.1, bottom two pairs of traces), rolling off by just 0.3dB at 20kHz, 3dB at 50kHz, and 10dB at 90kHz. Filter 3 gave a higher output overall (top pair of traces) and a slightly faster ultrasonic rolloff. Playing CDs (fig.2), Filter 1 gave the flattest top-octave response, with then a sharply defined rolloff (top pair of traces above 10kHz); Filter 2 gave a slightly higher midrange output and a slower HF rolloff starting around 6kHz (second-from-top pair of traces below 9kHz). Filter 3 gave an output 0.4dB higher than Filter 1 (top traces below 9kHz), but with then a more extreme HF rolloff reaching 3dB at 20kHz. The SA-11S2's response with a preemphasized CD showed a very slight positive error in the mid-treble (fig.2, bottom pair of traces below 9kHz). Channel separation (not shown) was better than 130dB in both directions below 800Hz, and a still superb 107dB at 20kHz.
Fig.1 Marantz SA-11S2 Reference, frequency response at 3dBFS into 100k ohms, SACD data, with Filter 3 (top) and Filters 1 & 2 (bottom). (Right channel dashed; 2dB/vertical div.)
Fig.2 Marantz SA-11S2 Reference, frequency response at 12dBFS into 100k ohms, CD data, with (from top to bottom): Filter 3, Filter 2, Filter 1 without deemphasis, Filter 1 with deemphasis (right channel dashed; 2dB/vertical div.).
For continuity with my published measurements of digital components (which go back to 1989), I investigate resolution by playing a dithered 1kHz tone at 90dBFS and sweeping a 1/3-octave bandpass filter from 20kHz down to 20Hz. The result with the Marantz playing CD data is the top pair of traces in fig.3; the peak just touches the 90dB line suggesting minimal linearity error, and the traces are free from any harmonic or power-supplyrelated spuriae. Repeating the spectral analysis with SACD data didn't drop the noise floor as much I would have expected (middle pair of traces), and the Marantz only just resolved a tone at 120dBFS, with 3dB of positive error evident (bottom traces). (The noiseshaping was turned off for these tests, as was the DC filter.)
Fig.3 Marantz SA-11S2 Reference, 1/3-octave spectrum with noise and spuriae of dithered 1kHz tone at 90dBFS with 16-bit CD data (top), SACD data (middle at 2kHz), and of dithered 1kHz tone at 120dBFS with SACD data (bottom at 1kHz). (Right channel dashed.)
Fig.4 was taken with the same test signala dithered 1kHz tone at 90dBFS from both CD (cyan and magenta traces) and SACD (blue and red traces)but was derived using a different kind of spectral analysis, FFT. Again the 1kHz tone peaks at exactly 90dBFS with both media. The noise floor is quite clean but about 6dB lower with SACD than with CD. I then confirmed the minimal linearity error with CD data by sweeping a dithered 500Hz tone from 60dBFS down to 120dBFS. Even though the SA-11S2's noise floor was a little higher than those of the best players I've tested, fig.5 really shows only the recorded dither noise and the player still easily resolved the three DC voltage levels that describe an undithered tone at exactly 90.31dBFS (fig.6). A dithered tone at 90dBFS played back from SACD was reproduced as a good, if slightly noisy, sinewave (fig.7).
Fig.4 Marantz SA-11S2 Reference, FFT-derived spectrum with noise and spuriae of dithered 1kHz tone at 90dBFS with CD data CD (left channel cyan, right channel magenta) and SACD (left blue, right red).
Fig.5 Marantz SA-11S2 Reference, linearity error, 16-bit CD data.
Fig.6 Marantz SA-11S2 Reference, waveform of undithered 1kHz sinewave at 90.31dBFS, CD data (left channel blue, right red).
Fig.7 Marantz SA-11S2 Reference, waveform of dithered 1kHz sinewave at 90dBFS, SACD data (left channel blue, right red).
The Marantz's output stage was unfazed by low impedances. Even with a full-scale 50Hz tone into 600 ohms (fig.8), the highest-level distortion harmonic, the third, lay at a very low 110dBFS (0.0003%). The demanding HF intermodulation test, during which I play an equal mix of 19 and 20kHz tones with a combined peak level of exactly 0dBFS, produced no significant levels of sum-and-difference tones, even into 600 ohms. However, it did reveal differences between the three filter settings for CD playback. Fig.9, for example, taken with Filter 1, shows fairly strong aliasing products between 20kHz and 24kHz. The picture was very similar with Filter 3 (fig.10), but with that filter's slower ultrasonic rolloff producing even higher levels of the two aliasing products. The ultrasonic rolloff of Filter 2, however, is apparently so slow that it dumps a large amount of aliasing products back into the audioband (fig.11). The 19+20kHz test signal is very much an extreme case; it is fair to note that this signal will never occur in music. The benefit of this kind of filter is that it sharpens the reproduction of audioband transient information; the downside is the high amount of aliasing it allows. If you like that sort of sound, Filter 2 is the one for you. But Filter 3 probably offers the best compromise between transient definition, top-octave resolution, and a clean noise floor.
Fig.8 Marantz SA-11S2 Reference, balanced output, spectrum of 50Hz sinewave at 0dBFS into 200k ohms, CD data (left channel blue, right red; linear frequency scale).
Fig.9 Marantz SA-11S2 Reference, balanced output, Filter 1, HF intermodulation spectrum, 19+20kHz at 0dBFS peak into 600 ohms, CD data (left channel blue, right red; linear frequency scale).
Fig.10 Marantz SA-11S2 Reference, balanced output, Filter 3, HF intermodulation spectrum, 19+20kHz at 0dBFS peak into 600 ohms, CD data (left channel blue, right red; linear frequency scale).
Fig.11 Marantz SA-11S2 Reference, balanced output, Filter 2, HF intermodulation spectrum, 19+20kHz at 0dBFS peak into 600 ohms, CD data (left channel blue, right red; linear frequency scale).
Finally, the Marantz SA-11S2 offered excellent rejection of word-clock jitter, producing just 251 picoseconds peakpeak of jitter with the J-Test signal on CD. Narrowband spectral analysis of the noise floor to either side of the 11.025kHz tone (fig.12) indicated that data-related sidebands were almost at the residual level. Repeating the test with the Marantz's word-clock input locked to a high-quality external source (a dCS 972 digital/digital converter) at either 44.1kHz or 88.2kHz increased the measured jitter slightly, to 271ps pp, but with an almost identical-looking spectrum, including the rise in the noise floor to either side of the central peak. I couldn't find any other measured differences resulting from externally clocking the player, but it could be that the dCS clock offers pretty much the same degree of stability as the Marantz's own.
Fig.12 Marantz SA-11S2 Reference, high-resolution jitter spectrum of analog output signal, 11.025kHz at 6dBFS, sampled at 44.1kHz with LSB toggled at 229Hz, CD data. Center frequency of trace, 11.025kHz; frequency range, ±3.5kHz.
As I have come to expect from Marantz products, the SA-11S2 Reference offers superb audio engineering. And while Filter 2's leaky behavior raises my eyebrows, there are two other choices of filter.John Atkinson