A feature of Suno v4 which I am coming to realise, is that v4 no longer models an uploaded audio voice clip accurately neither in cloning the voice or the original melodies. This is very disappointing. Probably something to do with copyright ,etc, in view of the current court cases.
I had been hoping to used a cloned model of my own voice, but I think I am giving up on this idea. I am also having to forget being able to use the original melodies and needing to use very strange punctuation to force the AI to sing my lyrics the way I intended.
I have probably tried harder with this song (an old favourite of mine) than any other of the nearly 200 songs I have woked on in Suno and loaded to Soundcloud. Finally, got something that is "acceptable" but very different from the song I wrote. - Maybe hearing it fresh without preconceptions you may like it.
