Talk:Cross-Synthesis/@comment-24636659-20190124204106/@comment-53539-20190211145229

I'm adding to that, I know Yamaha wants Vocaloid to be always based on realism, but Vocaloid being able to produce less realistic results on call is not a downside. Re-reading this page you may get the impression XSY is entirely bad, but honestly? Chipspeech has really put things in perspective over the years, its nice to have a singer who can act as a replacement for a real singe, but at the same time... Sometimes producers just want a toy to play with in experiment music, seeing if something inspires a new song or creates that lovely "one off" result that they need for making a song have its own fingerprint. Vocaloid has never tried to really capable that effect, you can never create "mad" results like this and it can make Vocaloid boring at times.

For me, something really needed to be done to fix that and XSY became the answer at times. While still grounded in reality, since it uses two voicebanks based on it and an engine mimicking reality, it did have a chance of random failures that kicked normal results in the groin. It should have been kept for this reason no matter what and been in with the "style/colours" selection. "Styles/colours" is just so uniformed and makes every Vocaloid voicebank... kind the samey. I've listened to some results and honestly, its not fully a fair trade off even if the function works better.