AI Eurovision Winner
'Beautiful the World' victorious in public online vote
Well the when is that from uncanny valley joins me now from Sydney Justin shave is the music producer Caroline grenades that innovation and strategic partnerships in Charleston il is head of music and innovation I mean it sounds great but I mean composing of a song normally is quite difficult but when it comes to using a I. does it make it even more so the sub you Justin.
But in a sense no I mean it the way we would normally approach production the writing of the song is you know the announcement come through to us to producers say can you take my my ideas in this song and.
That's that's amazing that you're looking to go off and in the same way this is G. X. rate our UPS account how do you how do you do that what what sounds you put it because you know we all know that even with traditional music composers use science the technology of music to to get the effects they won't there was a what did you put into the computer to achieve what we just heard a bit of.
Both 200 Eurovision tracks of the lyrics from the same amount of tracks we have had a similar idea wrestles a couple years ago so we had already collected all the lyrics from all the Eurovision songs at all and so we used 3 different neural networks in total 1 for lyrics 1 for music and then we had the idea for that the project that we've been working with the Google's credit left here in Sydney 23 the.
The neural net the thing you're only 8 weeks from the sounds from the same wildlife Cabarrus qualities to as many devils we sort of wanted to play how much to the devastating loss of a lot of luck in the straight during the bush fires off these devastating fires things on shelves and I could not she pick out the koalas I was of noise to the clothes make.
Just about to get to play on there that she's in the.
Who are the same because it's sort of it's generated a sound that otherwise doesn't really exist on any of our students a good noise that I mean did you have a do you have a talk about riff of you do this as part of the course will save the bit where you would where you just start the song the next it'll be after that is exactly the mixture of the wildlife sounds in the quality center as well.
No I don't I don't I I but we've missed the the 2 composite come believe it or not to a specific complaint but more in the second Justin when it when it does it doesn't take anything away from this is satisfaction though of creating I hits if you're using machine machine learning just just for you isn't worth of reviews individuals.
Hi I think the satisfaction probably more men in the studio I like I like happy accidents are like like ideas thrown at me and in this way with the computer came up with this quite well in some cases focus the lyrics and the melodies it's increased my level of satisfaction we all actually mold them into something that was quite catchy.
Give me an idea of the bunkers lyrics count on.
What's your favorite.
Open the door now we want to live in the dorms what is your reputation in 18 that I guess my.
Your average student there expects that it starts off with some pretty interesting pointed point flying from this world that is gonna pop.
I'm a lot of the things that's quite profound.
Things started coming out of but but but you know I mean I mean it's it is interesting it and perhaps some would say it's going to make your revision even even better the imp than in previous years but listen guys really well done I think we're just gonna hear a click there until until we know the kookaburras escaped and.
And trust with us for that but for the next time but well done to all of you thing winning via your vision the.