As to infant steps towards "my" voice clone

Aside from celebrity soundalike industry…which I’m not interested in…there’s a teeny tiny amount of info as to how a guide singer sang a track and then Randy Travis’ actual voice syllables/inflections were plugged into/replace the guide…which…seems to at least be a glimpse into the path I would use to clone my own recorded voice…on my own…for my own uses.

In other words, I’ve not, until now, seen an actual description of how coder guys are approaching this…

“…Lacy tapped developers in London to create a proprietary AI model to begin the process. The result was two models: One with 12 vocal stems (or song samples), and another with 42 stems collected across Travis’ career — from 1985 to 2013, says Kyle Lehning, Travis’ longtime producer. Lacy and Lehning chose to use “Where That Came From,” a song written by Scotty Emerick and John Scott Sherrill that Lehning co-produced and held on to for years. He believed it could best articulate the humanity of Travis’ idiosyncratic vocal style…”


The infant steps are of course, the question of "how do you get from where we are…to the capability being a “module” in say, Spectralayers? :slight_smile:

I envision (I guess), we’re looking at least out to Spectralayers 20 :slight_smile:

edit: It appears audimee is in the game now for this type of thing…hadn’t heard of them until today.

1 Like

I think such a development will happen a lot sooner than SL 20. The evolution of GPU and AI technology are in J curve trajectory and Nvidia just became the most valuable company in the world at $3.335T. I don’t know about Steinberg’s internal thinking but I believe we are seeing the beginning of a golden age for DAWs. The manner in which music is created and the way it is performed is going to change radically and will demand an updated mindset. I can only hope Steinberg stays ahead of the curve.

Yeah, this is all beginning to point to golden era potential…which I wouldn’t have envisioned five years ago.

I, too, hope SB stay on top of the cutting edge, particularly between Spectralayers. wavelab. Cubendo capabilities as separate programs in relation to ongoing developments.

Much easier to divvy up previously-unimagineable-tech-breakthroughs between each program’s strengths…rather than say, Logic, or Reaper etc where devs have to figure how to stuff enormously complicated things into a singular daw-sandwich.

I think Nuendo-post guys are already screaming for dialogue-clone capability…which is the same general zip code as sing-cloning advancement requests.

As a side…but not completely…I dunno what Roland has struck as deals with riaa (it’s not yet well documented)…but I hope Yamaha is awake and doing similar.

Ultimately, whatever SB comes up with for tech breakthroughs is increasingly dependent on a whole lot of deep-pocket deals/goodwill with established music content owners throughout the publisher and sr-copyright holder universe imo.

These are exciting times and changing by the month.

1 Like

I saw the new Celine Dion doc. Not to spoil it but there is a scene where she discussed her illness, Stiff Person Syndrome and the paralyzing effect it’s had on her singing voice.

As was pitch correction once a new, controversial entry in live performance, so will real time voice cloning likely be met. Celine is a perfectionist and clearly lives to connect to her audience but her illness now limits her capabilities. I wondered, would she be willing to clone her own voice and use it in live performance? How would her audience feel? Would they think of it as chintzy as lip-syncing? Or would they love her even more?

Endless, exciting possibilities just around the next corner. :thinking:

IMHO the problem with this kind of feature is the end-user, since you don’t know were it’s going to be used… there’s already been some nasty cases of use of childrens voices and other scams… This just of one my views of the dangers of cloning voices