Not with the DAC your using.
Left manufactures specs for your Neoteck DAC, Right RMAA from my Asus ST sound card.
One thing to note here is that the ratings are dB vs dBA (A weighted, or human hearing weighted - mainly between 1-4kHz). This is a good example:
difference-between-db-dba think of the measurements weighted, compared to dB which aren't so simply comparing them without doing the maths is incorrect. I've not found a test where a calibrated wave form generator is used and measured in parallel with a spectrum analyser and the sound card. So perhaps I'll explain why some of this is moot. To convert dBA to dB you need to know what you're listing to as its frequency dependant.
I prefer dBV to consumer 1 volt reference. A lest significant bit for 24bit is ~59 nanovolt or 0.000059mV. That equates to -144dB dynamic range for 24bits. Hence the noise level (depending on the computer EMI) is going to be above the -144dB. So how is the noise floor below the dynamic range?
Human hearing A weighting (dBA) with noise shaping allows manufacturers to really mess with the numbers - you can be noisy as hell at low or high frequency but it will be blissfully low noise according to the dBA rating. dBA depends on the frequency being used to measure. Sneaky..
Intermodulation is the impact of your noise/harmonics on the sound itself during amplification - both as a product of the amplification and as the input into the amplification. Basically how much your sound is messed up by the noise/harmonics messing with the amplification process. A tube amp will be worse here. In the realworld any noise or harmonics OUTSIDE of the dBA weighting frequency range also plays a part in this figure unless they're only processing post dBA..
Solid state often use better active filters (opamp based for example) for harder cut offs.
I'll also highlight the stereo cross talk being below the 24bit range. How can you correctly report a measurement below your sampling bit resolution? The maths probably subtracts one channel from the other and I would hope that the floating point here is 64bit.. so this feels like a floating point conversion rounding error. I'd need to see the code.. or confirmation it's using 64bit (you'd assume it is as the programme supports 32bit?)
The DAC figures, although looking worse, provide more information w.r.t to the measurements than that of the Asus.
Also is the THD from the dBA figures? Which means the THD is also weighted and filtered.
TL;DR - Yes dBA is weighted for the perception of human hearing (of a child) as an end sound for a specific set of sounds, it shows you need to be careful in how you use it. Manufacturers are sneaky..
Lastly when you do your FFTs - ensure you document what windowing you're using. This shows how you're dealing with spectral leakage and this affects the figures resulting from the testing.
