This makes no sense.
A sound that is harsh in the original recording should sound harsh when played back.
Systems that make everything sound nice may be pleasing (and popular with audiophiles), but that doesn't mean they're high fidelity or high quality.
But how do
you know that it was "harsh in the original recording"? I presume you've listened to it via some playback mechanism, and judged it to be so - so, how did you allocate the harshness contributions of recording vs. playback chains: say, 100%, 0% or 50%, 50% or 0%, 100%?
Decades ago, I used to believe that there were "bad" recordings - but the more I worked at it the more I realised that the key faults
always lay on the reproduction side - if I could elevate the standard of playback to the highest level, then any perceived harshness in the sound would evaporate.
Recordings all sound different - it's not that they sound "nice" on a competent system, it's that they are fully convincing as a capture of real people, and instruments - and the illusion completely sustains.
I was just reminded of a compilation of Ike and Tina Turner tracks I have, some live in a club, dubbed onto a cheap tape recorder - you want harsh? The sound of these will sandblast the glass of the windows into translucent, on a conventional rig - and that's how it sounds to me also, if my setup is not optimised. But a 'miracle' happens when the system is good enough - it no longer slices my ears to shreds, but becomes an intense, 'raw', sweaty, power packed, emotional roller coaster .. the very reason for that type of music to be in existence.