Firstly to Steinberg, thank you so much for including Omnivocal with the Cubase 15 release.
As a songwriter / lyricist / producer not gifted with a singer’s voice, I have sometimes struggled with my musical direction in the past as I couldn’t fully visualise the finished song. I had even toyed with giving up writing lyrics completely and sticking to instrumentals only. However, lyric writing brings me so much pleasure that I really don’t want to restrict my music in this way. Omnivocal gives me the ability to hear how my songs might sound with vocals at last.
I have played around with the instrument today and, considering the app is in Beta, I am quietly impressed. Most of my songs are in the late-night, smooth jazz and quiet storm genres. By using the controls I have be able to create a soulful voice to a degree, enough for me to imagine what a human singer could do with it.
The biggest areas that need improvement for me are expanding the AI singer’s vocabulary and to have more control over the expressiveness of the voice. There are plenty of English words not recognized by the AI singer, I’ve been able to mitigate this to some degree by noticing the phonetics that are displayed word-by-word and using those when needed.
An example of what I mean comes from the first line of one of my songs, to quote:
No one ever warned me Cupid’s arrow has a poisoned tip
The best I can get from Omnivocal at the moment is:
No one ever warned me Dupid’s arrow has a poisoned tip even though a K is shown in the phonetics, rather than a D.
Also, the AI singer seems to have a slight accent, possibly German ![]()
Overall though, I can’t stress how joyful it is to finally imagine how my songs might sound in the hands of a professional singer. Thank-you Steinberg, I look forward to the app’s development.
