Kurre Posted October 6 Share Posted October 6 Are someone thinking/working on AI functions? Yes, it's going to be tough to develop for local software but there are several daw functions that could benefit from it. Audio alignment is one. It will likely come to daw's in a couple of years. 2 Link to comment Share on other sites More sharing options...
Noel Borthwick Posted October 6 Share Posted October 6 BandLab has a dedicated team working on AI music. SongStarter is one such product as well as stem separation. They are not yet available in Sonar but can be accessed via BandLab studio. Cakewalk Next does have access to Splitter. https://www.bandlab.com/songstarter https://www.bandlab.com/splitter But to answer your question yes we are thinking about fruitful uses for AI. 3 Link to comment Share on other sites More sharing options...
pwallie Posted October 6 Share Posted October 6 are bl using the user content as ai source? Link to comment Share on other sites More sharing options...
Noel Borthwick Posted October 6 Share Posted October 6 No its independent public domain data. Songstarter was a Google partnership. https://medium.com/platform-stream/bandlab-launches-songstarter-an-ai-powered-tool-built-to-fix-musical-writers-block-1581577ec02 The AI team also creates specialty training data using their own proprietary tools as well. There is a lot of investment in this. 1 Link to comment Share on other sites More sharing options...
pwallie Posted October 6 Share Posted October 6 cool, interested to see where it all goes Link to comment Share on other sites More sharing options...
Glenn Stanton Posted October 6 Share Posted October 6 not sure how much "public domain" hip hop and trap music for training is available given the length of time a copyright persists. it's one of those things that the legal system will end up deciding WRT AI training on copyrighted materials. Link to comment Share on other sites More sharing options...
Teegarden Posted October 7 Share Posted October 7 Glad to finally see some users understand the benefits! When I started a similar topic FR: AI tempo recognition and quantize two years ago the feedback about AI implementation was mainly negative, they clearly didn't understand the possibilities and benefits at the time. Link to comment Share on other sites More sharing options...
Alan Tubbs Posted October 7 Share Posted October 7 If AI writes the melody for you in BandLab, does google or ms get composer credit? And does cakewalk Sonar get mechanical rights for playing the performance? Asking for a friend. 1 Link to comment Share on other sites More sharing options...
Glenn Stanton Posted October 7 Share Posted October 7 and this years grammy for best mix goes to Microsoft! yay! Link to comment Share on other sites More sharing options...
murat k. Posted October 7 Share Posted October 7 2 hours ago, Teegarden said: Glad to finally see some users understand the benefits! When I started a similar topic FR: AI tempo recognition and quantize two years ago the feedback about AI implementation was mainly negative, they clearly didn't understand the possibilities and benefits at the time. Because you had been exposed to the meaningless comments of a group of status quo defenders. Don't worry. In time, they will all disappear one by one. Link to comment Share on other sites More sharing options...
Terry Kelley Posted October 7 Share Posted October 7 Or the defenders of art. Fake a song. It’s fine. Link to comment Share on other sites More sharing options...
Noel Borthwick Posted October 7 Share Posted October 7 While AI generation is new, algorithmic music generation has been around for 30 years or more. Products like Band In a Box, Jammer etc have been used on countless productions and there haven't been copyright issues, at least any I have heard about. While there is some legit concern about copyrighted data being used for training LLM's, for music often the training data itself is algorithmic. 3 Link to comment Share on other sites More sharing options...
Promidi Posted October 7 Share Posted October 7 5 hours ago, Noel Borthwick said: While AI generation is new, algorithmic music generation has been around for 30 years or more. Products like Band In a Box, Jammer etc have been used on countless productions and there haven't been copyright issues, at least any I have heard about. While there is some legit concern about copyrighted data being used for training LLM's, for music often the training data itself is algorithmic. Even though depreciated ages ago....even CAL scripts can be programmed to perform algorithmic music generation. I have written a few myself recently that utilises the Random command to generate usable musical and rhythmical phrases. Link to comment Share on other sites More sharing options...
murat k. Posted October 7 Share Posted October 7 5 hours ago, Terry Kelley said: Or the defenders of art. Fake a song. It’s fine. AI is a tool; art needs no defense. 1 Link to comment Share on other sites More sharing options...
Recommended Posts
Please sign in to comment
You will be able to leave a comment after signing in
Sign In Now