I have no complaints about Solarias English except maybe the missing screamo mode 
However, I tried myself at a Hikaru Utada song and it just sounds way too harsh, I was checking my de-esser and compression settings twice but no chance. Setting the strength of the consonants works halfway but she tends to get a lisp from that.
Maybe I’m using it wrong, any tips?
You may want to go through other parameters such as tension, breathiness, etc. During the post-process workflow you may want to put the voice through an equalizer or a multiband compressor. These may help with the problem.
By the way, you can try to export aspiration to separate files and process it separately, and finally mix it with the other part.
Thank you very much, using the isolated aspiration track did work well although it does leave a few undeseried duck-like sounding noises here and there.
Edit: It works but its impossible to get anything crisp and clean above ~6khz
Isolated aspiration splits the output into two audio channels or files when rendered. After applying your different EQ and other effects to the individual parts you need to recombine them in your DAW to get the final result.
To be clear, if you don’t process/mix the two components differently they will sound the exact same after being recombined as if you had rendered them as a single output in the first place.
I just found a nice and easy solution, additionally to the normal de-essing settings (from your multiband compressor or whatever) you can also use TDR Nova (free version) and adjust the tiny range to duck with the de-essing preset. I’s alot easier than drawing all the automation by hand.