To me it sounds exactly like they're playing the two instruments on top of each other, if not 'side by side' - which is to say, not like something I've never heard before.
Agreed, I was quite disappointed. I actually thought we would here new sounds humans have never heard before. This was more like two instruments picked from an orchestra or, playing two bank settings on a synth on a split keyboard.
I can replicate these sounds in a few minutes mixing on my 10 year old $120 Yamaha keyboard, if it's not in one of the built-in 999 x 999 sounds (because I can combine two sounds easily using the built in functionality), then I can mix it myself -- but it is already so close to the built-in sound.
Will be excited if an AI can make instruments that move me.
I'm asking myself if a synth is just a boring application for a model like this. But maybe it would be possible to implement a new kind of efficient sound compression algorithm.
It's a shame the article uses soundcloud clips, because I suspect the end sounds aren't the impressive bit. I think it's the user interface that's the interesting bit here.
> In addition to the NSynth “slider” that Engel recently demonstrated at Google headquarters, the team has also built a two-dimensional interface that lets you explore the audible space between four different instruments at once. And the team is intent on taking the idea further still, exploring the boundaries of artistic creation. A second neural network, for instance, could learn new ways of mimicking and combining the sounds from all those instruments. AI could work in tandem with AI.
A video of this would have been much more useful to me.
Looking forward to a more interactive demo of this tech. The headline is ridiculous though - I can use a RNG for a handful of frequencies and create a sound humans have never heard before.
oh wow. music producers invent sounds no one heard before all the time >.> ever listen to neurofunk? ;D
Why dont they make an AI that does something useful for a change instead of one that can look at cat pictures and produce jibberish!
17 comments
[ 3.1 ms ] story [ 57.4 ms ] threadI can replicate these sounds in a few minutes mixing on my 10 year old $120 Yamaha keyboard, if it's not in one of the built-in 999 x 999 sounds (because I can combine two sounds easily using the built in functionality), then I can mix it myself -- but it is already so close to the built-in sound.
Will be excited if an AI can make instruments that move me.
> In addition to the NSynth “slider” that Engel recently demonstrated at Google headquarters, the team has also built a two-dimensional interface that lets you explore the audible space between four different instruments at once. And the team is intent on taking the idea further still, exploring the boundaries of artistic creation. A second neural network, for instance, could learn new ways of mimicking and combining the sounds from all those instruments. AI could work in tandem with AI.
A video of this would have been much more useful to me.
The website for their project [1] is much more toned down, so you can blame the journalist or Google's PR team for pitching an interesting angle.
[1] https://magenta.tensorflow.org/nsynth
[Citation Needed]
http://www.moogfest.com/
Haven't seen anything like that before. Scrolling marques everywhere. I wonder if this will be a new trend.
http://brutalistwebsites.com/