And all I'm saying is that the maximum analog output (or internal bus) level corresponds to 0 dB FS on the USB input. So when you target the nominal output level of 4dBu, you shouldn't drive the USB input to 0dB FS.
If I am playing some music through it, I have no control over the PCM samples. This thing will saturate not only at 0 dBFS, but even at -1 dBFS:
If all was well here, you could sort of hang your hat on this. But per above, performance is still horrible. This would not be acceptable in a $5 Aliexpress DAC. But we are supposed to accept it for a "professional" tool? We are talking distortion floor that is whopping 40 dB higher than CD spec
below 0dBFS! And again, remember, clipping behavior happens regardless of what you set the analog voltage level to.
Meanwhile people want me to go above 0 dBFS and test for "intersample overs."
When something is broken this way, we need to point it out so it is fixed. Zero benefit comes to any consumer for this type of overflow. It has no reason to exist. Literally hundreds of DACs I have tested pass this with flying colors regardless of price. Had there been a proper set of measurements used to verify the design, it would have been caught and fixed (hopefully).
So let's not attempt to paper over this. Company needs to be motivated to fix problems like this. It would have been a wonderful device and incredible value if it hadn't had such serious problems in its design.