Digital Processor Reviews

Harmonic distortion (fig.7) was very low, the third harmonic being the highest in level at -91dB (0.003%), even into 600 ohms. The only other harmonic to poke its head above the -100dB level was the seventh. Intermodulation distortion was also very low (fig.8), the 1kHz difference component lying just above -100dB (0.001%). However, note the spectral spreading of the primary 19kHz and 20kHz tones, which may be due to the Crystal sample-rate-converter chip running out of mathematical headroom with this demanding signal. (I originally suspected wordclock jitter as the culprit, but the P-3A has very low jitter, as you will see later.)

Fig.7 Perpetual Technologies P-3A, spectrum of 50Hz sinewave, DC-1kHz, at 0dBFS into 100k ohms (linear frequency scale).

Fig.8 Perpetual Technologies P-3A, HF intermodulation spectrum, DC-24kHz, 19+20kHz at 0dBFS into 100k ohms (linear frequency scale).

As supplied, the P-1A performs two functions: converting S/PDIF and AES/EBU datastreams to the lower-jitter I²S format, and applying Perpetual's proprietary Resolution Enhancement algorithm. But measuring the unit's effect was hardly straightforward. I started by connecting the P-1A to the P-3A with the supplied I²S cable, setting the P-3A to I²S Direct to bypass its Crystal SRC and data receiver chip, and setting the P-1A to increase the sample rate of an incoming 16-bit/44.1kHz datastream to 96kHz, and to increase the word length to 24 bits. I then fed the duo with various signals and tried to see what changes the P-1A caused.

Fig.9 shows a 1/3-octave spectral analysis of the P-3A's output decoding a dithered 16-bit/1kHz tone at -90.31dBFS, with the P-1A set to Bypass (bottom trace at 100Hz) and to Resolution Enhancement. You can see that, other than the region between 80Hz and 250Hz, the traces overlap exactly. However, as these traces are dominated by the dither content, it is possible that the increase in signal resolution is being obscured.

Fig.9 Perpetual Technologies P-3A, 1/3-octave spectrum of dithered 1kHz tone at -90.31dBFS, with noise and spuriae, 16-bit data, via P-1A in Resolution Enhancement mode (top at 100Hz) and in Bypass mode (bottom). (Right channel dashed.)

Accordingly, I switched off the dither and repeated the analysis (fig.10). At 1kHz and above, the traces again overlap exactly and show the harmonic content necessary to produce the waveform shown in fig.5. But without the dither noise, it can be seen that the P-1A is adding a significant amount of lower-frequency energy. Fig.11 shows the waveform of the enhanced analog signal. Yes, it looks more like a sinewave than fig.5, but not anything like a true 24-bit sinewave as output by the P-3A (fig.6).

Fig.10 Perpetual Technologies P-3A, 1/3-octave spectrum of undithered 1kHz tone at -90.31dBFS, with noise and spuriae, 16-bit data, via P-1A in Resolution Enhancement mode (top at 100Hz) and in Bypass mode (bottom). (Right channel dashed.)

Fig.11 Perpetual Technologies P-3A, waveform of undithered 1kHz sinewave at -90.31dBFS, 16-bit data, processed by P-1A in Resolution Enhancement mode.