Science Interviews


Tue, 8th Dec 2015

20 second music maker

Patrick Stobbs, Jukedeck

Listen Now    Download as mp3 from the show Music Technology: Do or Die?

Can technology and creativity exist together? When we think of music that creativity is often associated with the first sparks of a song and the initial Sheet Musiccomposition of melody, beats and pitch. But could a machine do that instead? Patrick Stobbs is a musician and co-founder of JukeDeck, a new system launching this week that uses software to make music...

Patrick - Weíre a London based startup.  We are building an artificially intelligent music composer.  A system that allows video creators, podcasters and any other content creators to create unique music at the touch of a button.  Weíre not launched yet, but weíre launching very, very soon.  For you, we will give you a little bit of a sneak preview but within weeks this will be available.

Connie - Fantastic.  Okay will you show me how it all works?

Patrick - Sure.  You simply come to our site, click new track and select your genre of music. We offer four or five genres, electronic, folk, rock, ambient.  Then you select your mood, so letís select and electronic genre, lets select a chilled mood.  You can select your speed, you can select your exact duration.  Letís just go for 27 seconds and click create and, in the next 20 seconds, our system will create a completely unique track of music based on your specification.

Connie - Electronic chilled.  Not bad for 20 seconds but what else could I make?  Five genres; electronic, folk, rock, orchestral and ambient and four moods; uplifting, dark Ė sounds intimidating, chilled and aggressive.  What do you see aggressive being used for?

Patrick - This has actually been a really key request from some of our users. Thereís a large community of people in the world who shoot videos of themselves playing computer games and theyíve requested slightly more hard, harsher sounding music because they think it really goes well with playing computer games.  So thatís why we created an aggressive genre.We are about to release at least ten, or twenty or so more moods.

Connie - Can I try making one myself?

Patrick - Please do.

Connie - Okay, so I want to make a whole new podcast.  Itís going to be about adventures in Siberia.  Okay, so I think Iím going to want some folk for my adventures in Siberia.  My adventures are going to be Ė theyíre going to be uplifting and for around 20 seconds.  Itís just going to be a short intro.  Hey, letís click and make itÖ And just another 20 seconds laterÖ  Nice I can really hear my podcast developing with this.  Iíve got a real picture of how itís going to workÖ

Patrick - Rustic SiberiaÖ

Connie - So Patrick.  This seems really easy but Iím assuming thereís something a lot more complicated going on behind it all to get out so many different types of music so quickly?  Can you explain to me whatís going on?

Patrick - Correct.  This sadly isnít a system that you can just throw together in a week.  Itís taken four years of hard technical development.  Broadly speaking we have two sections to our system.  One is a composition section which dictates what notes should come where - imagine itís writing a piece of sheet music.  Second is a production system which takes that sheet music and chooses how it should be performed.  It selects instruments, it selects how notes should be articulated and eventually it converts the piece into an mp3 file which is then given to you.

Connie - Doesnít it make you - I donít know Ė cry a little inside when you suggest that a computer can do what a highly trained musician could?

Patrick - Weíre composers, weíre musicians.  We totally get some peopleís instant reaction which is ďAh, thatís really weird.  Why are you getting a computer to write music?Ē  But I think if you dig a little deeper, it quickly becomes apparent that itís going to be a really useful and hopefully exciting thing.  Firstly, composition is a really restrictive practice, unless youíve had the really good fortune to have had a musical education or had the time to teach yourself, you basically canít create nice music. You can bang on a table and shriek and scream but composing a nice piece of music is really hard and most people canít do that.  So we think that by empowering non-musicians to create music at the touch of a button, that can only be a good thing.

The second thing to bear in mind is the possibilities this opens up.  When you have a system that can create music, it can necessarily do so very quickly because computers work incredibly fast.  And it leads to a situation where music can be composed in real time, i.e. on a second by second basis, dynamically reacting to what is going on in the world.  So this will, theoretically, be able to compose music that takes into account who you are, where you are, what youíre doing, how youíre feeling.  For example, say you are trying to get to sleep, a system like this could, theoretically, take into account your physiology, whatís going on inside your brain and compose you the perfect piece of music to lull you to sleep.

Connie - Do you think this is a tool that composers could use themselves in the end?

Patrick - Weíve already had musicians use this to create bigger and better works.  Weíve had various pop singers come along, create a backing track using our system, and then compose their own melody over the top of it and thatís really exciting to us as musicians.  So absolutely we want bake in a bunch more features to help musicians create bigger and better works of art using this system.



Subscribe Free

Related Content


Make a comment

See the whole discussion | Make a comment

Not working please enable javascript
Powered by UKfast
Genetics Society