Rotel RCD-955AX and RCD-965BX CD players RCD-955AX Measurements
Beginning with the multi-bit Rotel, the frequency response (fig.1) was virtually flat across the audible range, with only a small amount of ripple in the upper octavesripples characteristic of machines using Philips digital filters. The 1kHz squarewave in fig.2 is also typical of players using linear-phase digital filters, and although pre-emphasized CDs are in the distinct minority, the virtually flat de-emphasis response (not shown) guarantees that they will be properly balanced in playback.
The crosstalk from the left to the right channel is shown as the top curve in fig.3; the right-to-left crosstalk is the bottom curve. The separation indicated here is not exceptional, but unlikely to be audibly significant.
Fig.4, the spectrum of the RCD-955AX reproducing a 90.31dB dithered tone at 1kHz, shows some negative deviation from linearity but no significant artifacts other than a hint of third-harmonic at 3kHz. The linearity is more clearly shown in fig.5, where the deviation at 90dB is just under 2dB. Only the right channel is shown; the left was virtually the same. Listening to the fade-to-noise with dither signal from the RCD-955AX produced only one or two minor whistles at what I would estimate to be well under 80dB. They were well below the primary tone in level. The overall linearity of the RCD-955AX is very good for a multi-bit player in this price class.
The noise spectrum of the player's output when playing a "digital silence" track in fig.6 has no artifacts worth noting. It is worth noting, however, that we've recently become aware that some players are designed to actually mute the output of their D/A converters when presented with an all-zero digital signala practice which enhances their measured S/N ratio. That might be the case here; the measured spectrum (not shown) with the player in Pause was virtually identical to that in fig.6. In any event, this figure does show the noise contribution of the analog stages to the output to be negligible.
Fig.7 shows a 1kHz, undithered tone at 90.31dB from the 955. The desired stairstep sinewave is evident, though the nearly inevitable noise makes interpretation difficult. And in fig.8 the response of the 955 to a 19+20kHz, 0dB signal shows no significant IM artifacts. The aliasing tone at 24.1kHz is commonly seen in players using Philips technology; the relatively slow-slope filter in these machinesa plus for a number of reasonscannot completely eliminate this tone. It would only be a potential problem when used with an amplifier having marginal high-frequency stability.
The RCD-955AX is non-inverting at its analog outputs, neither of which had measurable DC offset. Its output impedance is just over 198 ohms on both channels.Thomas J. Norton