It's ironic that a few days ago Yoshiki was saying that he thinks that generative AI will be the death of music, and then he goes and uses it to create a version of his voice that can hold a tune so he can mime on TV...
That's how he's done this, by the way - just in case anyone has any doubt.
any proof for that claim?
I've been involved in music production for nearly 20 years and I spend most of my day job at the moment working with generative AI. I trust my ears!
It is masked to an extent in the verse by the breathy overtones of Yoshiki's voice (which the AI will have been trained on extensively) but when it goes into the chorus it becomes very 'uncanny valley' - like a robot singing with the transients from Yoshiki's voice.
Listen to this example of someone training AI on Freddie Mercury's voice and you might hear some similarities - it captures the rasp and timbres of his voice very well, but in some of the longer and higher notes it's... not quite right.
If you don't trust me at that, then fine - but ask yourself why Yoshiki has waited 40 years to magically show the world he can hold a perfect pitch?
Maybe it's because there is now technology than can reproduce his voice... or maybe it's just a big coincidence and Yoshiki has been shy this entire time.