Board Thread:Off-Topic Discussion/@comment-25147000-20150320032731/@comment-53539-20150419223501

I almost snapped just now at Evil, I'm sorry, for that UTAU English sample. Because nobody understand what I'm trying to say. That no matter what, UTAU still sounds too same-y.

I'm just going to say it.... This isn't a rant... More like a problem I've observed over the course of 5 years.

You can improve the language, but if the the engine still sounds like Cher's singing results fromt he song "Believe", it has issues to resolve. Vocaloid has never sounded that bad, even back in V1, its always had a more human-esque toneeven at its worst. I've worked with Vocaloid and once gave UTAU a go and deleted it within day of installation.

I like UTAU, its existence is great, but Jez... Everyone trying to prove a point with UTAU always shows you examples that are PROVING EXACTLY the reasons UTAU is still behind Vocaloid. Just... If UTAU could get the human voice grain into its results... And remove the Cher-auto tune sound.... It will have caught up in leaps. Until then, improve the language all you want, it still sounds like auto-tuned samples... And Vocaloid doesn't sound like a "snap to sound" auto-tune sample set. Sure... It still as digital noise, but its not that bad. And at leas the "Twiki sound" has been lessened. There has been NO ATTEMPT to remove the "Twiki" sound from UTAU, hence why it sounds liek autotuned samples.

https://www.youtube.com/watch?v=yptoaCKZ0Tw

The sound I'm referring to is that "bebebebeb" sound he makes... If you listen to the translations between sounds/notes, you MIGHT get what I'm on at. Its when sounds don't smoothly go into one another because the ending of that sound doesn't match. So it sounds bitty. So this is UTAU's "Cher-auto tuned" problem in a nutshell.

Even if the other stuff is improved, Vocaloid4 still has large triphones capabilities in both English and Japanese and there are more layers and sounds being added... >_<