What do you make of the FFT measurements showing the humps at 9kHz and 15kHz on the 12kHz tone chart for the D30 and the higher noise floor of the D30 in comparison to your measurements?
A bit hard to diagnose fully. I don't know what hardware he is using for measurements. But let me share some data regardless.
I tried to match his parameters as close as I could and ran the test against Topping D30 and Schiit Modi 2 Uber. First,
their results:
Topping D30:
And Modi 2 Uber:
Notice the broadening shoulders around the 12 kHz tone with Schiit Modi 2 which agrees with my review measurements. And this new set, run with just a 12 kHz tone (rather than J-Test):
The broadening of the shoulder due to jitter/noise is apparent again in Modi 2 Uber.
The Topping D30 has higher noise floor as I indicated in my last post over S/PDIF. Note that its output is 3 dB higher so you need to move the Schiit Modi 2 Uber up by the same amount to be apples vs apples comparison in that graph. When done, as the graph says, Schiit Modi 2 Uber has higher noise/jitter around the main signal tone of 12 kHz, but lower noise farther out. So a mixed result.
There are no humps though in Topping D30 that they are showing. Since my measurement is able to reveal the lower noise floor of the Schiit Modi 2 Uber, it is not a problem with my analyzer not being revealing enough. The noise is dominated by what the DAC is doing, and not the ADC of my Audio Precision Analyzer.
My only guess is that the higher levels of Topping D30 is overloading the the ADC in the soundcard he is using. He should run a test with another source that generates that higher output and see if the problem is in the source or his audio capture setup.
A side rant: this kind of discrepancy is why I use the (expensive) Audio Precision analyzer which is the #1 device for this type of testing in the world. This means that anyone can repeat my results by using the same gear. This is not true of sound card testing which the poster on the other forum seems to be using. So at some level, you need to ask him why his results don't agree with mine than the other way around.
One more note:
they are always showing jitter across a narrow range of frequencies. Here they cut off at 5 Khz on the low end in the measurement. It is important to keep going down to 20 Hz as that shows power supply problems and importantly includes distortions in the 2 to 5 kHz where our hearing is most sensitive. This is why my measurements show the full spectrum.
Their method leaves out data that is critical for determining audibility of distortion products.
Summary:
Their results agree with mine for Schiit Modi 2 Uber indicating random low frequency jitter/noise. I cannot reproduce their results for Topping D30 other than the fact that it does have higher noise floor over its S/PDIF input than USB. In comparison to Schiit Modi 2 Uber over S/PDIF, it outperforms it in low frequency random noise but loses in the wider range. Since my equipment and process is the industry standard, my suggestion is that they do some calibration of their measurement against the same so that they know whether their results are due to device being tested or the measurement gear itself.
Importantly, they need to show jitter spectrum that includes full audible band so that perceptual analysis can be performed on the results.