Members bookumdano4 Posted November 10, 2016 Members Posted November 10, 2016 Well, here we go. I thought it would be another 20 years for this. We ain't in Vocaloid territory any more. Load in an isolated vocal of say, Sinatra singing something or other, type YOUR lyrics in as text, they replace FRANK"S lyrics, and there you go. Cut, paste, a little Elastique, Melodyne here and there for re-phrasing,, re-load into your head-banger self-written instrumental track and you've got Sinatra (or Taylor Swift etc) singing about slashing the dudes and riding Harley's after midnight. The possibilities (and lawsuits) are endless. So cool. If it works half well.
Members RoadRanger Posted November 10, 2016 Members Posted November 10, 2016 The lawyers are gonna LOVE this LOL.
Phil O'Keefe Posted November 10, 2016 Posted November 10, 2016 Here's an article I found about it... looks like it needs 20 minutes of source material (speaking?) to create the model from... and no, it's not available - it's an experimental project they were demonstrating... https://techcrunch.com/2016/11/03/adobes-project-voco-lets-you-edit-speech-as-easily-as-text/
Members RoadRanger Posted November 10, 2016 Members Posted November 10, 2016 A similar technology (probably many years away) will input all film footage of several actors, a script, and some "direction" - and turn out a new movie with anyone you want (dead or alive) in the roles .
Members sharkbait Posted November 10, 2016 Members Posted November 10, 2016 What's the ostensible use?I immediately see the potential to forge false incriminating statements.
CMS Author MikeRivers Posted November 10, 2016 CMS Author Posted November 10, 2016 [video=youtube;I3l4XLZ59iw]
Members 1001gear Posted November 11, 2016 Members Posted November 11, 2016 Some cop show already used the premise to fake a persona and of course they stung the mark cue music roll credits... This new tech is more machine and none people.
Phil O'Keefe Posted November 11, 2016 Posted November 11, 2016 What's the ostensible use? I immediately see the potential to forge false incriminating statements. Give me (or any expert audio editor ) twenty minutes of random conversation, a DAW and a bit of time, and we can do the same thing.
Members AlamoJoe Posted November 11, 2016 Members Posted November 11, 2016 Well.....I've got a tune I always wanted Robert Plant to sing....
Members bookumdano4 Posted November 12, 2016 Author Members Posted November 12, 2016 looks like it needs 20 minutes of source material (speaking?) to create the model from... and no, it's not available - it's an experimental project they were demonstrating... Yeah, Adobe was demoing with spoken voices. I'd look to guys like Peter whats-his-name from Melodyne to unravel the current Adobe approach and make it workable for sung notes. Which no doubt several super-tech brains will figure out sooner than later (everyone can't just keep coming up with new virtual mic pre plugs for the next seven years). Adobe themselves may not really care much. I see this as cutting edge stuff with mindblowing possibilities (and for sure, lawsuits) that will arrive relatively soon. I have to revisit how the statutes work when you trademark a particular voice. I vaguely remember tv coffee commercials etc getting in hot water for impersonating known voices in the background. Dunno what will happen when some kid starts posting videos in five years of Robert Plant singing "Stitches". For now, I'd be vewwy cawful about letting any of my isolated sung vocal tracks up there on youtube.
Recommended Posts
Archived
This topic is now archived and is closed to further replies.