Never give Art Baker the aux cord because he will straight up play the saddest country songs you've ever heard in your life.
*Three Wooden Crosses by Randy Travis plays out*
Art is completely fine and vibing, already searching for the next song. Garraty is fucking dead in his seat with tears streaming down his face, McVries is staring out the window and contemplating death, Stebbins is nodding along, he can dig it!
I want to talk about the new Randy Travis song, because I feel like there are a lot of misconceptions about the whole thing (which certainly isn’t helped the lack of information on Warner Bros's part), so I'm going to be going over whats happening and give my own perspective.
For those that don’t know, Randy Travis is a country singer that suffered from a stroke in 2013 and hasn’t had the ability to sing since. He partnered with Warner Bros to release a song, “Where That Came From”, using AI to replicate his voice, using songs he recorded before his stroke. To be clear, this is not an AI generated song. This song was written by Scotty Emerick and John Scott Sherrill. Calling this song AI generated discredits Randy and his team.
As for how the vocals were generated, I can’t seem to find a solid answer. Warner says they used “AI models”, and according to The Tennesseean, months were spent tuning the vocals. To me it sounds a lot like using a vocal synth, like Vocaloid or SynthesizerV, the differences being that instead of having a vocalist record a series of syllables to be put into a database (referred to as a voicebank), I think the AI actually pulls syllables from the songs provided to it, like Vocoflex, a program Dreamtonics announced a few weeks ago (see their Youtube channel for the announcement and further information). It also seems that AI generated the vocals based on a demo Randy's producer recorded, and it was fine tuned from there. Like I said though, not much information was given, so I can’t say for sure, this is just based on my knowledge of vocal synthesizers.
I don’t actually understand why Warner hasn’t released more information on the software they used. They said two different programs were used, but it’s not clear if Warner developed both of them, what they are, how they work, etc. It’s all really vague, and I don’t understand why. I wonder if the program(s) will stay private or eventually be put on the market, and they’re worried about competitors? I really can’t say.
I feel like where this goes from an ethical perspective really depends on what Warner does next. I'm curious to hear other people’s thoughts on this, especially folks in the vocal synth community, since I haven’t seen anyone there discussing this yet.
If you wonder how long I'll be faithful
Well, just listen to how this song ends
I'm gonna love you forever and ever, forever and ever, amen
I'm gonna love you forever and ever, forever and ever, forever and ever, forever and ever, amen