Microsoft's AI Generates Voices That Sing in Chinese and English (venturebeat.com) 32
Researchers at Zhejiang University and Microsoft claim they've developed an AI system -- DeepSinger -- that can generate singing voices in multiple languages by training on data from music websites. From a report: In a paper published on the preprint Arxiv.org, they describe the novel approach, which leverages a specially-designed component to capture the timbre of singers from noisy singing data. The work -- like OpenAI's music-generating Jukebox AI -- has obvious commercial implications. Music artists are often pulled in for pick-up sessions to address mistakes, changes, or additions after a recording finishes. AI-assisted voice synthesis could eliminate the need for these, saving time and money on the part of the singers' employers.
But there's a darker side: It could also be used to create deepfakes that stand in for musicians, making it seem as though they sang lyrics they never did (or put them out of work). In what could be a sign of legal battles to come, Jay-Z's Roc Nation label recently filed copyright notices against videos that used AI to make him rap Billy Joel's "We Didn't Start the Fire." As the researchers explain, singing voices have more complicated patterns and rhythms than normal speaking voices. Synthesizing them requires information to control the duration and the pitch, which makes the task challenging. Plus, there aren't many publicly available singing training data sets, and songs used in training must be manually analyzed at the lyrics and audio level.
But there's a darker side: It could also be used to create deepfakes that stand in for musicians, making it seem as though they sang lyrics they never did (or put them out of work). In what could be a sign of legal battles to come, Jay-Z's Roc Nation label recently filed copyright notices against videos that used AI to make him rap Billy Joel's "We Didn't Start the Fire." As the researchers explain, singing voices have more complicated patterns and rhythms than normal speaking voices. Synthesizing them requires information to control the duration and the pitch, which makes the task challenging. Plus, there aren't many publicly available singing training data sets, and songs used in training must be manually analyzed at the lyrics and audio level.
Alto Intelligence? (Score:2)
What's next, tenor?
Ckoe already tried that ... (Score:2)
... "I want to teach the world to sing ..."
Meh (Score:2)
This doesn't bring any Satisfaction.
Can it... (Score:2)
sing like Diva Plavalaguna?
Re: (Score:2)
Multi-pass
Started out English only... (Score:2)
AI only started singing in Chinese after they installed TikTok.
Artificial Art (Score:1)
Re: (Score:1)
President Xi deepfakes (Score:3)
I take it this stuff is already banned in China because well, subject line.
Though, if the deepfakes are any good, this might actually be something fun to play with.
President Xi singing the Winnie the Pooh theme song, anyone?
Re: (Score:2)
Xi's wife is a professional singer.
When they married, she was far more famous than he was.
For many years, he was referred to as "Peng Liyuan's husband".
Daisy, daisy, give me your answer do... (Score:2)
LOL, no, some things are human-specific (Score:2)
But there's a darker side: It could also be used to create deepfakes that stand in for musicians, making it seem as though they sang lyrics they never did (or put them out of work).
Oh, bullshit.
You want to tell me you can create a computer program that can 'sing'? Sure. Whatever. Maybe it'll even be technically accurate according to all standards of voice. But it'll never be a human singer. It'll never have the emotional depth that a human singer brings.
But Rick, autotune!
LOL 'autotune' just makes everyone into carbon-copies of everyone else. The 'flaws' in a singers' performance are just as important, as their technical skills as a vocalist -- and one could validly argue that the 'flaws' are sometimes
Re: (Score:1)
Re: (Score:2)
Re: (Score:1)
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
I don't think so-called 'AI' created 'music' is going to be any good, and an 'AI' 'singer' even less so. Not going out of my way to even hear it, not even clicking a link. Say what you want about that.
If and when they crack the code of the human brain and we can create *real* AI and not the ersatz they keep trotting out, and these general AI have actual personalities and are like real people, then maybe I'll give a damn. Otherwise they're just hyping garbage technology.
Re: (Score:2)
It'll never have the emotional depth that a human singer brings.
Why not? It seems to me that emotion should be easy to fake. For non-singing voice synthesis, adjusting intonation to simulation emotion is not difficult. This tech is advancing quickly.
Human singers win in the end.
Human singers expect to be paid. Software performs for free.
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
Pop singers have been faking for many years (Score:2)
Manufactured sound, voice enhancers.
They no longer need to sing in tune, the machine will do it for them. Even add a little vibrato, deepen it a bit, get it in properly time with the music.
Great video of what can be done (Score:2)
https://www.youtube.com/watch?... [youtube.com]
Re: (Score:2)
Ugh. That sounds like someone auto-tuned the crap out of it.
What until you see the translation what Tay sang (Score:2)
Sorry, the Slashdot guidelines prevent me from posting inflammatory and racist lyrics.
Songsmith Deluxe (Score:1)
Microsoft has finally completed Songsmith.
https://www.youtube.com/watch?... [youtube.com]
I'm gonna sell so many towels.