digital subtractive synth artifact test

[Shy]

For your reference, here is a simple test I made using Abakos Pro, a discontinued software synth that uses brute force to tackle this issue, and not one of the fancy new methods (minBLEP, minBLAMP, whatever), to showcase the importance of a "virtual analog" synthesizer's operating sample rate (and/or oversampling).

This was rendered at 192kHz with 16x oversampling (and converted to 44.1kHz).
To emphasize how CPU intensive this is: with my 3.4ghz AMD Phenom II 965 CPU, with a single note, the synth's performance fluctuates from 0.9 to 1.1x real-time (hence not playable).
FYI: the artifacts and overall sound were much worse at 96kHz and 128kHz, and just worse at 176kHz.

The attached zip file also contains the MIDI file that you could use if you'd like to compare things yourself (set to 140bpm). The synth settings are simple: sine, saw, square or triangle waveform for oscillator 1. Set oscillator 2 (for ring modulation) to the same waveform as oscillator 1 and set it to not follow notes (set to C3 when not transposed, irrespective of note input). Set oscillator 2 semitones to - or + 12, 24, 36, 48. A slight volume envelope and that's all.

Obligatory "why" and closure:
Audio samples to provide a clear reference and showcase the amount of difference it makes. I would expect something like comments about why it all doesn't matter so much after all, followed by philosophical debates and theories, followed by criticism of why this is not an ultimate, all-inclusive test (as if it was meant to be), followed by some idiotic personal comments. So I'll say right now that this is a main reason I made this thread, to let people listen, compare and decide for themselves.

Additionally, I'll emphasize that I'm not asking anything and would probably have nothing more to say here. Finally, if you're a troll or just someone who doesn't like the fact I shared something I care about or doesn't like my attitude, go somewhere else, I have no time for that.