Country music legend Randy Travis, who’s largely been unable to speak or sing since a devastating 2013 stroke, is crooning again with help from artificial intelligence.
On Friday, Travis released a new studio-recorded song called “Where That Came From,†made using an AI version of his voice trained on past audio tracks. It’s his first new song since the stroke left him with severe aphasia, a disorder caused by damage to the side of the brain that controls language.
“Eleven years ago I never thought I would be able to have a hand in music production of any kind, but by God’s grace and the support of family, friends, fellow artists and fans, I’m able to create the music I so dearly love,†Travis shared in a Thursday Facebook post accompanying a teaser video for the tune about lost love.
“To hear his voice again is a miracle,†one wrote in the YouTube comments section for the music video. “What a beautiful song!†Wrote another, “It doesn’t matter if it’s AI, if it’s voice cloning, etc. This is Randy’s voice in one form or another; you can hear the authenticity regardless of how this was created and I love it.â€
As for Randy’s IA’d vocals: there’s really only two spots that made my ears go, That doesn’t sound exactly what I’d expect outta Randy Travis: the end of the second lines in the first two stanzas, there’s just something about them that seem more drawn out than what Randy would have done. Dunno if I can articulate exactly what I’m hearing, but that was my first impression.
Overall, I enjoyed the song: that’s what a country is supposed to aound like! If it take AI to get rid of tgis stupid-ass pop country bullshit, than I, for one, embrace our AI music overlords! … haha…
There’s so much auto-tune and pitch correction in popular music these days, I’m not sure this qualifies as much different. Presumably Travis is still writing the songs, he’s just using a synthesized voice, with “AI” (read - fancy computer algorithms) to duplicate previous recordings of his voice on the new lyrics? Artists/producers already use way more fake drums, guitars, keys, etc than people realize. This is just faking another “instrument.” And I think most people would be astounded at how much of their favorite singers’ performances are pitch corrected. It’s so common that lots of people don’t really know how good singing is supposed to sound anymore. The few songs I’ve heard that were actually written by AI programs have been horrible.
I look at it this way: the artist says “I have a song, but I can’t sing it, because … you know … can you help me out?”
Next: Imagine what Brian Wilson could do?
I’m not saying “if you can’t hit those notes anymore, use this” I’m saying, for artists who have -physically- lost the ability to sing, but can still write (Linda Ronstadt, Stevie Nicks, Joni Mitchell), this might be a path for them to continue to create
I’m saying, for artists who have -physically- lost the ability to sing, but can still write (Linda Ronstadt, Stevie Nicks, Joni Mitchell), this might be a path for them to continue to create
I agree. Of course, there has always been the opportunity to still create, and just have someone else sing your songs.
“I’m not saying “if you can’t hit those notes anymore, use this” I’m saying, for artists who have -physically- lost the ability to sing, but can still write (Linda Ronstadt, Stevie Nicks, Joni Mitchell), this might be a path for them to continue to createâ€
Mostly I agree for the same reasons. It must be a bittersweet moment to put that out after years of silence. I’m happy for him.
It’s also vastly different than someone trying to cash in on a famous voice by releasing “new†songs that the deceased artist had nothing to do with. As much as I’d like to hear some “new†Waylon or Marley songs it’s nothing I’m interested in if he had no hand in making the music.
i can’t remember who said it first, but it went something like:
" i want AI to clean the dishes and mop the floor while write songs and paint pictures, instead i’m cleaning the dishes and mopping the floor while AI writes songs and paints pictures."
.
The average listener would be shocked at how much of a “live” performance is prerecorded or enhanced. In the world that I play in (pun intended) there are a lot of bands that are using prerecorded backing tracks for nearly all their performance. The fact that there simply aren’t enough people or instruments on stage give them away, but no one seems to care.
AI is just a fancy, more easily controlled version, of prerecorded. Some of us might have seen a live performance of Jump by Van Halen where the backing track is dragging (they used a tape back then). Poor Eddie, in front of thousands of fans and while being recorded for a video, was trying to play the guitar and the notes he needed were not on the instrument. It is an entertaining train wreck to watch.
We use all sorts of processing equipment. The drums are not even the natural sound. Our sound engineer is correcting vocals in real time from the board in the back of the room. There are also automatic processing that is occurring on the vocals, including some minor pitch correction (You can’t tell).
My point is this is nothing new, just a new way of doing it, and a new application. The guy is still the creative “voice” for the product and owns the rights. He didn’t pull a Milli-Vanilli. I’m all for it and hope others follow suit.
And I think most people would be astounded at how much of their favorite singers’ performances are pitch corrected. .
I saw a band live once (Canadian band, Our Lady Peace), I had their album and liked it. Then I saw them live and dude literally couldn’t sing. He was so off key it was unlistenable. Ok, maybe he’s just having an off night, maybe he’s sick.
Nope, I saw them live on TV at some other point. Dude, just can’t sing. They clean him up in the studeo.
AI is just a fancy, more easily controlled version, of pre-recorded.
I get what you’re saying in general, but on this point, no. It’s 100% artificial; the artist never sang vocals to clean up or splice together or overlay or modify. It never happened. They added Randy vocal skin to another vocalist structure. Again, in this situation I think it’s the best way for the artist to put out his music and I applaud him for it, but I don’t think it’s akin to Vince Neil prerecording or autotuning that dogshit he pumps through his vocal cords on stage over the last half century.
That said, I’ll take AI Randy over autotuned Vince every day of the week.
AI is just a fancy, more easily controlled version, of pre-recorded… They added Randy vocal skin to another vocalist structure.
Not really.
AI does not work that way. AI listens to the recordings of the person and then synthetically creates the voice using the correct pitch, harmonics, inflection, grammar, etc. It is very advance machine learning. (Not really AI) There is not another person’s voice involved in this exercise. They are taking his voice and “processing it” to produce the new song. It is actually closer to a prerecorded then you might think.
Regardless, we agree on the main points. He’s an artist that is continuing to create, and he is not “hiding behind” the technology he is using. He is very up front about it. He has my respect.
AI is just a fancy, more easily controlled version, of pre-recorded… They added Randy vocal skin to another vocalist structure.
Not really.
AI does not work that way. AI listens to the recordings of the person and then synthetically creates the voice using the correct pitch, harmonics, inflection, grammar, etc. It is very advance machine learning. (Not really AI) There is not another person’s voice involved in this exercise.