Just to be clear here, you are referring to a version of the NCores which have the SMPS integrated and on the same board as the amps, not a DIY version of the NC400 and separate SMPS1200A400 or the SMPS600N400, both of which document as producing amps with THD+N of .0005% or less (Sinad 105db).
And the paper you refer to lists the THD+N performance of the Pascal Modules as .05% at 10khz, .03% at 2.5 Khz, and .01% at 1kz which translates respectively into Sinads of 66db, 70db, and 80db. That performance seems to be in the same league as the Roland 535.
Right, you have to compare apples-to-apples. Measuring mono designs without/external SMPS yield to much lower noise. But that's nothing you will ever find in active speakers because of the packaging. Active speakers demand more channels and an integrated SMPS, also for the DSP.
You can look up the numbers at Hypex, so you have to compare NCxxxMP / NCASxxxxMP (p.26-36) to Pascal. You will find that the gap closes up:
https://www.hypex.nl/theme/Hypex2018/files/Hypex-Brochure-2020-web.pdf
This is not to say that the amplification isn't good enough and perfectly transparent, but I still believe it is a fair question to be asked.
Hearing threshold of THD are close to 0.1% @1kHz (most sensitive area) depending on level and Nth-order components. That's why it's common practice to give power numbers @0.1% THD for high quality audio. Of course you can and should go beyond, that's what Hypex and Pascal did.
You can go and hear it by yourself how sensitive you are to THD: http://www.klippel.de/listeningtest/?v=3 (also @amirm talked about THD hearing thresholds)
Again, we couldn't find any audible differences between Hypex and Pascal within our implementation.
Last edited: