- Written by John E. Johnson, Jr.
- Published on 25 October 2012
On the Bench
All distortion measurements were taken within an 80 kHz bandwidth. Unless otherwise stated, the Volume Control was set to 0.0. For digital signals, unless otherwise stated, the level was at minus 5 dB.
Let's begin with the coaxial digital input.
Using a 1 kHz signal at 0 dBFS, THD+N was 0.003%.
With the signal at - 5 dB, distortion was 0.004%. Note that specifications are usually stated within a 20 Hz - 20 kHz bandwidth, so the reason there is 0.001% more THD+Noise in this test is that the signal is lower, but the noise is not, so the noise is a slightly higher percentage of the total signal, and thus, the THD+N measurement is higher. I use an 80 kHz bandwidth because it will pick up any spurious signals in the out-of-audio-band region that are inaudible in themselves, but which can interact with signals in the audible band (music). THD by itself in the two graphs here is probably 0.001% as specified.
19 kHz and 20 kHz paired input signals showed no visible B-A peak at 1 kHz. Superb.
IMD measured 0.005%. Notice the lack of small peaks adjacent to the 7 kHz signal peak. This represents excellent performance, and a listening experience full of fine detail.
The measured Frequency Response was 20 Hz - 20 kHz, -0.3 dB.
Still using the coax digital input, but now with signals at 24/192 sampling, a 1 kHz sine wave produced 0.002% THD+N.
Using the 19 kHz, 20 kHz paired input signal did not yield any observable B-A peak at 1 kHz.
Measured IMD with 24/192 sampling was lower than it was with 16/44.1.
The Frequency Response was within 1 dB out to 42 kHz, after which it dropped precipitously at 42 kHz.
Now we take a look at the results using one of the HDMI inputs, fed from test discs played on an OPPO-BDP-95.
At 1 kHz, and 16/44.1 sampling, distortion was 0.004%.
As with the coax digital input, 1 kHz through HDMI did not result in any observable B-A peak at 1 kHz.
IMD through HDMI was only 0.005%.
The measured Frequency Response was the same as through coax.
Going to 24/192 sampling, THD+N using a 1 kHz sine wave was 0.002%.
Although there are some very small peaks in the 1 kHz range, I think the one peak there is about 900 Hz, not 1 kHz, so it is not a B-A peak. Therefore, as with the other tests, no B-A peak is observed.
IMD at 24/192 sampling was 0.001%. The left channel (yellow) has a slightly higher noise floor around 35 kHz - 40 kHz, and it is also present at 16/44.1 sampling, so it is something generated in the circuitry. It's inaudible in any case. I have found that signals below -120 dB can change from day to day. You never know what is going on to cause such artifacts.
So, with the HDMI v4.1 fully engaged with the SSP-800, HDMI is finally a true high fidelity input.
Lastly, we analyze the analog path. I used the XLR inputs and front left/right XLR outputs on the SSP-800, and generated analog signals with the Audio Precision.
At 1 kHz, THD+N was 0.002%.
Raising the volume control to + 5.0 lowered the input voltage necessary to produce a 2 volt output. (At a volume setting of 0.0, it took about 2.3 volts to produce 2 volts output.) The distortion was insignificantly higher.
With analog XLR input, and 19 kHz, 20 kHz sine wave input, there was a B-A peak 110 dB below the fundamentals. This is insignificant, but measurably different than when using the digital inputs.
IMD with the XLR connections was 0.002%, which is excellent.
THD+N vs. Frequency indicated a bit more distortion (and 0.0004% is definitely just a bit) in the left channel.
Output voltage, with the volume control set to 0.0, indicated a sharp knee at about 3 volts for both 100 kOhm loads and 600 Ohm loads. Clipping occurred at just slightly more than the knee voltage.
Turning the volume control up to 5.0 just about doubles the input voltage, and in use, I never went much above -10.0 when watching movies or listening to music.