Is it better to do a low cut of the dataset?

[static jasper]

If I make a low cut of about 60~80, will the noise be further reduced?

This is a dataset using the sound source of an actual singer.

Or what kind of work should be done with eq to further improve the quality of AI?

Since these songs have already been completed for commercial purposes, there is no need to touch them with EQ, right?

[lean zealot]

it won't do much, if you cut out those you will lose mic pops which can be better but may sound thin.

remember that most often, what happens in that frequency range, always stays in the frquency range