Google Get Stuck Into Deep Audio Morphing

US Machine learning from Google's Magenta team      04/03/21

Google's Magenta Studio is known for their innovative and creative uses of machine learning for audio. They've just released Tone Transfer which fuses two recordings into one, with an input sound (eg. birdsong) being generated in the style of the output sound (eg a saxophone). Not only can you load your own sounds to the input side, but each combo can be mixed, transposed and amplified. 

If you're not aware of Magenta Studio's work, I recommend the N-Synth Super which cross fades four sounds into one and the Blob Opera which delivers slapstick control over a choir of colourful singing blobs. 

Weird and wonderful results guaranteed. https://sites.research.google/tonetransfer

 

 

About the author [midierror]: midierror makes nifty Max For Live devices, innovative music hardware, award winning sample packs and hosts a podcast speaking to people in the music world.

< More From: GOOGLE

• MusicLM: Generating Music From Text 29-Jan-23
• The Blob Opera 25-Jul-21
• Google Launch Massive Electronic Music Museum Archive 11-Mar-21
• Google Have Unveiled an Experimental Synthesizer - NSynth Super 14-Mar-18
• Jam With Google's Neural Network Using A.I Duet 22-Feb-17

Even more news...