This AI thing is getting nuts (___ sings ____)

Damn this thread. I'm lost in a spiral of listening to these hilariously amazing (and shitty) Dio ai covers. Don't even care about these serious covers anymore.

 
Last edited:
If an AI can mimic pitch, tone, inflection, etc so well, what does a real human have to offer?

I predict live shows will take front stage (heh heh) to albums, as an AI "mixmaster" can produce virtually endless variations of songs. I can get it to make "good enough" music that it would be hard for a real band to stand out.

But a show, gonna need live humans for that, and if I insist on them playing and signing, not lip synching it, then they gotta have a base amount of talent and stage presence.

Or music will go in wacky directions as the humans try to out innovate the AI and we get the 80s all over again.

BRING IT!!!! :P
 
If an AI can mimic pitch, tone, inflection, etc so well, what does a real human have to offer?

I predict live shows will take front stage (heh heh) to albums, as an AI "mixmaster" can produce virtually endless variations of songs. I can get it to make "good enough" music that it would be hard for a real band to stand out.

But a show, gonna need live humans for that, and if I insist on them playing and signing, not lip synching it, then they gotta have a base amount of talent and stage presence.

Or music will go in wacky directions as the humans try to out innovate the AI and we get the 80s all over again.

BRING IT!!!! :p

Live shows could just end up as augmented reality.
 
Live shows could just end up as augmented reality.
Maybe some holograms or something, I just can't see folks gathering just to out on goggles though.

Now high Def VR recordings of concerts to watch at home, THAT would be awesome. Watched a few with my oculous and I sooooo wish they had good recordings of the bands I like back when they were young. Still gotta work on better audio but otherwise it's pretty good. So i hope bands are recording things even if they don't want to release them now. .
 
Aren't the survivng Beatles making an album with old George Harrison guitar samples from the 90s, and AI John Lennon singing?
 
Aren't the survivng Beatles making an album with old George Harrison guitar samples from the 90s, and AI John Lennon singing?
I don't think so, I think the use of AI here is about separating Lennon's voice from other recorded elements of unreleased material, so it's not generative in the way that people might think - or even in the way that people are mapping a voice to an existing performance as with these cover songs.

This tech has been around and widely available for a long time in professional software, I suspect that they were going to put some songs together anyway, but are calling what they've done "the work of AI" because it makes it more of an interesting story for the press.
 
Last edited:
Can you pick and artist and then Pick a song and have an AI generate it.

I would like BONO, Sting, and Sinatra to sing I get around by the beach boys.
 
If an AI can mimic pitch, tone, inflection, etc so well, what does a real human have to offer?

I predict live shows will take front stage (heh heh) to albums, as an AI "mixmaster" can produce virtually endless variations of songs. I can get it to make "good enough" music that it would be hard for a real band to stand out.

But a show, gonna need live humans for that, and if I insist on them playing and signing, not lip synching it, then they gotta have a base amount of talent and stage presence.

Or music will go in wacky directions as the humans try to out innovate the AI and we get the 80s all over again.

BRING IT!!!! :p
I'm thinkin' fuck paying for human effort, no more blood money, give robots every job, give me robots at the live show, let everyone relaxxxxxxxxxxxxxxx, this worlds already the matrix so may as well roll with it.
 
If an AI can mimic pitch, tone, inflection, etc so well, what does a real human have to offer?

I predict live shows will take front stage (heh heh) to albums, as an AI "mixmaster" can produce virtually endless variations of songs. I can get it to make "good enough" music that it would be hard for a real band to stand out.

But a show, gonna need live humans for that, and if I insist on them playing and signing, not lip synching it, then they gotta have a base amount of talent and stage presence.

Or music will go in wacky directions as the humans try to out innovate the AI and we get the 80s all over again.

BRING IT!!!! :p
For me, I generally dont give a shit if it's a real person or not. Of course as you said if it's concert time then ya you want to see the singers on stage. I'm not going to buy a ticket to stare at an HP laptop with a screensaver running music tracks.

But for music, I just like a good song. I dont care if it's a human who learned how to play guitar and met their bandmates at a high school and all the other grassroots stuff you read on a wiki page. Dont give a shit. Most singers I wouldnt even recognize on the street, never seen live and the only thing I have to go off are google image pics or an old YT rock video from the 80s.

If AI music can churn out great stuff fast with a fake voice, who cares. If it's good I'll listen to it. But right now it's so so. Some of these songs are clever, but some arent great. You can hear the cut off words or overlapping words from sentence to sentence which are impossible to do. So it's not perfect yet, but getting there. But then again, depending on the music you listen to from real singers they all do their crazy sound studio editing anyway so it's all improved upon regardless.

I think all this AI revolution for media and creative types comes down to whether someone supports it "I dont care, as long as it's pretty good", or if someone supports the human side "I want media done grassroots style by a human". Really no different than picking sides buying shit.... Buy local, or buy imports. Some people care a lot about the source/origins. Some people dont.
 
Last edited:
If AI music can churn out great stuff fast with a fake voice, who cares. If it's good I'll listen to it. But right now it's so so. Some of these songs are clever, but some arent great. You can hear the cut off words or overlapping words from sentence to sentence which are impossible to do. So it's not perfect yet, but getting there. But then again, depending on the music you listen to from real singers they all do their crazy sound studio editing anyway so it's all improved upon regardless.
I think some genres are more susceptible to this than others. Pop and EDM, for example, is at HIGH RISK, as it all feels very transient from an endless parade of very similar sounding artists performing stuff being churned out of the same source. Rap, though, probably still needs that street cred that an AI can't get. Or metal, I'd like to think the fans are keying in on the artist themselves, not just the sound. Jazz, I think, is the least vulnerable, it already sounds like cats being tortured :P
 
Someone should fix my Arnold Schwarzenegger autobiography audiobook. Arnold only reads the first chapter and then it's some random dude the rest of the time.

Also, no reason for Freddie Mercury not to fix his overbite now. The possibilities!
 
Last edited:
Continuing my Dio postings with Billie Jean. Hearing him tackle the MJ vocal ticks is great.



Ehmm ... Just like that other posting, I simply don't recognize Dio's voice here at all. Doesn't even sound remotely like Dio.

Just compare any live performance. The real voice is completely different.

 
Top Bottom