Winner of Best Documentary at the 2019 Toronto Shorts International Film Festival and directed by Chasson Gracie.
(Learn more: https://www.musicandai.com)
As artificial intelligence continues to impact our human lives, many questions still remain, including: How will artificial intelligence impact one of the most human things on earth? Music. Will pop hits be created by algorithms? Will musicians disappear from the process? How important is the artist-fan connection?
"Gracie coalesces his roles as music documentarian and data science executive to explore the intersections of music and AI. He explores the implications for music, specifically targeting three affected areas: the industry (e.g. labels and marketing platforms), artists, and fans." - MacEagon Voyce
(Learn more: https://www.musicandai.com)
As artificial intelligence continues to impact our human lives, many questions still remain, including: How will artificial intelligence impact one of the most human things on earth? Music. Will pop hits be created by algorithms? Will musicians disappear from the process? How important is the artist-fan connection?
"Gracie coalesces his roles as music documentarian and data science executive to explore the intersections of music and AI. He explores the implications for music, specifically targeting three affected areas: the industry (e.g. labels and marketing platforms), artists, and fans." - MacEagon Voyce
Category
đ”
MusicTranscript
00:00Essentially, I don't think you can replicate a soul.
00:02I think you can replicate the impersonation of a soul,
00:05the impersonation of feelings based on learnings,
00:07but I don't think that automatic soul
00:09that we as humans cannot explain is there
00:12because if we can't explain it,
00:14how can we then create it?
00:19Best is when you, or when I play a song
00:24that's either new or a nostalgic song for people
00:27and they go crazy on the dance floor.
00:28I love the, oh!
00:30That's my favorite part of DJing.
00:34Just get somebody going like that is,
00:36you know, how does the computer gonna pick up
00:40that that guy just going absolutely crazy on the floor
00:45is, you know, that basically is telling me,
00:48okay, I'm heading in the right direction
00:51with my tunes at this point.
00:53Does the computer think the guy's just yelling at him?
00:56Or is he saying, you know, is he recognizing
01:00that the guy's going, this is the best track I've ever heard
01:04or something like that.
01:07I am a machine intelligence DJ,
01:10H-A-L, DJ 1200, MK3.
01:19Today I'm gonna DJ with my friend,
01:22no, back to back using a turntable and real vinyl records.
01:32Of course, I can control the pitch of turntable
01:35just like human DJ.
01:42Let me show you how it works.
01:49Tempo up.
02:23Let me show you how it works.
02:32The one thing I do know is that it's always moving
02:34and nothing's gonna stay the same, right?
02:36Like people talk about like the collapse
02:39of the music industry in like the digital age
02:41and Napster age, but that's not true.
02:43That was just the collapse of like
02:44the major record labels, right?
02:46And so it's like, sometimes in our conception,
02:49but as you said, before it was like sheet music
02:51and before that live performance was the main thing.
02:53So maybe we moved to a point where like
02:56the recorded music unit isn't the primary purpose
02:59or like the primary way.
03:00It's already not the primary to make money anymore, right?
03:03So maybe it's not the way we listen to it anymore.
03:06I don't know.
03:07That's kind of a bit out there, but I think it's exciting.
03:16So my name is Drew Silverstein.
03:18I am the co-founder and CEO of Amper Music.
03:22Amper is an artificial intelligence composer, performer
03:25and producer that creates unique professional music
03:28tailored to any content in a matter of seconds.
03:31We set out almost five years ago to build a technology
03:36that had never been successfully developed before.
03:39An artificial intelligence collaborator
03:42that would allow anyone from professional musician
03:45to complete amateur, to take their creative idea
03:50and turn it into the musical reality
03:53that they wanted to see come true.
03:54But then we said as composers, our job,
03:57our fundamental job is to translate emotion into music
03:59and music into emotion.
04:01So we suggested if we could build a creative AI
04:05that gives you the same collaborative experience
04:07of working with a composer within the time
04:10and economic framework that you need, would you want it?
04:13They said, yes.
04:15And so we got to work and four and a half, five years later,
04:19here we are today.
04:22âȘ Beyond what I can touch âȘ
04:26âȘ I wish I could feel beyond what isn't real âȘ
04:33âȘ I wish I could imagine, imagine, yeah âȘ
04:39âȘ There's more to who we are âȘ
04:43âȘ There's more to what we could be âȘ
04:46âȘ As I feel the weight of being âȘ
04:50âȘ I'm learning how to break free âȘ
04:54âȘ I'm breaking, I'm breaking free âȘ
05:01âȘ I'm breaking down, breaking free âȘ
05:09âȘ I wish I could feel beyond what isn't real âȘ
05:18The fundamental organizing principle of music
05:20is not a set of coherent sounds that fit together
05:24because God ordained them
05:26or because they naturally fit together,
05:28but rather that people all the time through their action,
05:31through making music and consuming it,
05:33through living their lives and shifting their preferences
05:36are reconfiguring the boundaries of those things.
05:39So these things are dynamic
05:40and they change because people are using them.
05:43So I would say instead of treating music
05:45as a set of sounds that are coherently organized
05:48into these little bundles we call genres,
05:51instead we should be thinking about communities
05:53as the fundamental organizing principle of music.
05:56What the most valuable thing for me with music
05:59is essentially experiencing
06:03and inviting new emotions in my life.
06:06Whether, you know, this is an emotion I've never felt before
06:10or it's an emotion that has come up
06:12from a moment in my past life
06:14that I'm now remembering through this music today.
06:17Yeah, I think it's the common thread,
06:19the fact that you're all experiencing
06:21the same emotion together
06:22and then you can, you know,
06:24you can talk about this experience afterwards.
06:26You don't even have to talk about it.
06:28It's a mutual feeling.
06:29They're going to feel differently about it,
06:31but at the end of the day
06:32it's a common thread between everyone.
06:34They're sharing this moment together.
06:35It's an intimate relationship now.
06:38So it's a really interesting phenomenon that I've noticed,
06:41especially as I've grown older
06:42and gone to a lot of different shows,
06:44a lot of different venues.
06:45I realized that no matter who the artists were seeing,
06:48everyone in the audience,
06:49whether you came with them or you met them
06:51or you're standing beside them,
06:52we're all one person at that point.
06:54We're all one entity enjoying the same thing
06:57and you automatically have connection to everyone.
06:59So that is a feeling that you can't replicate
07:01anywhere else other than I find in a musical venue,
07:03listening to someone that you're so in love with.
07:05There are no divides between everyone in that room.
07:08We're all appreciating the same thing at the same time
07:11and we're all really at peace, essentially.
07:13I'm very curious to see how AI
07:15is going to play a role in the creation of music
07:17only because there's a level of heart
07:19that I don't think you can replicate through algorithms.
07:22There's a soul, feeling, spiritual emotion
07:25that you feel when you hear music
07:27and I think that only transcends through the music
07:29because the creator also possesses that.
07:31So that'll be an interesting thing
07:33to see how well AI can replicate that type of feeling
07:38or if it'll still feel a little disconnected
07:41from the human element.
07:42It was interesting, like the way my set went
07:45is like I have an hour that I'm spinning
07:47from where doors open until like 8 o'clock
07:50and that's just kind of the rooms filling up
07:52and then after that, it's my job
07:54to like really get the crowd hyped.
07:56So for the first hour, I'm just kind of playing stuff,
07:58whatever, people are nodding their heads
08:00but then once it hits 8 o'clock,
08:02I really have to turn it up, you know what I mean?
08:04And that's when you grab the mic
08:05and you're like yelling at them
08:07and they're yelling back at you
08:08and then you're jumping in the crowd
08:10and handing out stuff and throwing water on them.
08:12So I definitely see an opportunity.
08:14I don't need to be DJing that first hour, for sure.
08:17That could be a playlist, that could be anything
08:19but those 15 minutes where it's really based
08:22off the interaction and the energy in the room,
08:25maybe someday that could be replicated
08:27but I don't know if it can right now.
08:29That's like a really intimate thing, you know?
08:38There have been examples where it's not clear
08:42how much human interference happened in the production
08:46and sometimes when you actually dig more deeply,
08:49there actually were humans making decisions.
08:52So there was an example, it was a music video
08:55from a few years ago that the advertising company
08:59Saatchi made, they actually exhibited it
09:01as part of the Cannes Festival.
09:05And they really used AI to do the casting,
09:09to make the drones, to write the story,
09:12to direct the drones, to film the video
09:15and there's actually a really interesting
09:18short video about how they did that.
09:22And that was an example of how artistic decisions
09:26were really made by AI.
09:29So I think that, and they're all different styles of music,
09:33sometimes, because it depends on what you teach the AI,
09:36what will come out the other end.
09:38So I think a lot of it has,
09:40but it's across different genres,
09:42I've heard ones that were Bach-based,
09:44I heard one that was someone trying to kind of
09:47make up a Christmas song, so yeah,
09:49there's a lot of different things.
09:52I guess I see myself more than anything
09:54just as a musician and no matter what I'm doing,
09:58I feel like I'm actually part of the creative process,
10:04whether it be producing for a vocalist,
10:08making my own music like I used to do a bit more,
10:12with the dance music and stuff,
10:14or engineering or mixing or mastering,
10:17which I've been doing a lot more of now
10:19because that's what I went to school for back in Winnipeg.
10:21It all really falls under the umbrella
10:23of just like I have,
10:25just like I work with people who I feel share
10:28similar taste and a similar vision that I do,
10:33and we're all working very closely,
10:36like I said, no matter what hat I'm wearing.
10:38And we can all just have our creative input,
10:42and it's nothing like,
10:43people aren't trying to steal the show,
10:45of like, oh, I'm the artist,
10:47so the producer, you just made the beat,
10:49kind of stuff like that.
10:51I feel like everybody has to work together
10:53if they want to actually do something unique.
10:55In music production already,
10:56I'm starting to see these tools
11:01that allow for random idea generation,
11:05and that's clearly being generated by AI.
11:08I mean, there's tools, for instance,
11:10in some groove production things
11:12where you just hit a button
11:14and it will randomize
11:15the types of samples you're using,
11:17and you'll get something
11:18that you would have never been able
11:19to come up with on your own.
11:21So I think that AI can definitely make
11:24sessions more creative for you,
11:26but I also think that it could be pushed too far,
11:29and it just will kind of like
11:30take away the creativity altogether.
11:33So it's like, the boundary's
11:36this extremely thin kind of like threshold,
11:39and it's like, I don't know where that is yet,
11:43and I think a lot of people
11:44are still trying to figure that out.
11:45So in terms of something not working for me,
11:48I think it's like replacing collaboration,
11:53because I think that a lot of
11:59what people are developing AI for
12:01in specifically music production tools
12:03is to make sessions more creative
12:05by doing things that you would never do,
12:06but I think that can be filled with
12:08collaboration with other artists,
12:09and I think we should be looking more towards that
12:11rather than trying to replace things completely
12:14with something computer generated,
12:16because at the end of the day,
12:18music is like human expression
12:21in an audible, invisible kind of way.
12:25So it's like, if you start filling that
12:28with a computer's random generation
12:30of what is programmed,
12:33what people think is good,
12:37I don't know, it kind of ruins it for me,
12:39but it can be used for good at the same time.
12:47One of the interesting things about me
12:49was I grew up in a household
12:51where my parents didn't really listen
12:54to a lot of music at all,
12:55so a lot of my process for discovering music
12:58and getting into music
13:00was just discovering it completely organically
13:03on my own, not through somebody introducing me
13:05to their favourite artist or something like that.
13:07So I never really got introduced
13:09to one particular genre of music
13:11that maybe my parents liked
13:12or relatives liked or something like that.
13:14So it was kind of just exploring in the world
13:17and just randomly stumbling across things
13:19that sounded good to me.
13:20So that was kind of how I got into music
13:23as a band for the first time.
13:24Basically, I started a company
13:25about two and a half years ago called Soda Tone,
13:28and we were helping record labels,
13:30management companies, booking agents,
13:32people in the music industry
13:33both find and evaluate talent,
13:36just understanding them from a data perspective.
13:38So one of the core elements of our system
13:41is machine learning and AI.
13:43So the ability to kind of predict
13:46for certain elements
13:47of what might make an artist successful,
13:50that's absolutely one big part of the system.
13:53AI has a lot of different applications,
13:56could have a lot of different applications in music,
13:59and some of them,
14:01I think the ones that are talked about the most
14:04are the ones where it infiltrates
14:07and affects the creative process
14:09in some way or another.
14:10And I think those are potentially
14:12the most controversial
14:13and most interesting to talk about.
14:15But honestly, there are so many other ways
14:19that AI can help the music industry
14:21that really doesn't relate
14:23to the creative process or infiltrate in any way.
14:26The one example I can give you
14:29is the way digital A&R used to be done,
14:35and to a large part still to this day
14:38is done in this way,
14:39is a team of different people,
14:42you know, every A&R going online
14:45and looking in various places,
14:47literally just manually clicking around.
14:49And that relies on a lot of kind of
14:53using your own mental power
14:55and literally the power of your fingers
14:57to try and click around
14:58and look through as much stuff as possible.
15:00Because if we're all playing this game
15:03where we're looking for music online,
15:06then it really just becomes a game of like,
15:10you know, a big part of it is
15:12who can kind of filter through
15:14the largest volume of music
15:16and who can do the best job manually filtering.
15:19So a lot of the basic filters
15:21of what A&Rs would want,
15:25it's actually very possible
15:27to do something through AI
15:29to help with that process
15:31and make it a lot more efficient.
15:37I think for that reason,
15:38the community reason,
15:39is why it won't go away,
15:43like the art side.
15:44Because I think that's a big part
15:45of why we listen to music.
15:46It's part of like your culture,
15:48your identity,
15:49and so that shared experience is like,
15:51it's a requirement almost, right?
15:55Yeah, like I said,
15:57it's an interesting time
15:58and I want to see how weird
16:00all these guys get with it.
16:01I feel like the human connection
16:02and the human experience
16:04that happens when you are DJing a party,
16:06I don't think that AI can replicate that.
16:09You have playlists,
16:10you have Spotify,
16:11you have all of these things
16:12that collect your data
16:13and like put together a playlist of songs
16:16that would be what you're supposed to like
16:18or what you might be interested in.
16:20But that doesn't necessarily work at a party.
16:22That doesn't necessarily work at a bar.
16:24That doesn't work at a wedding.
16:26That's not going to be
16:27the same type of interaction.
16:29So I don't really have any fear.
16:30I just, I want more knowledge
16:32and I'm excited to see
16:33kind of where it kind of takes our,
16:36like what journey it takes.
16:37That's the interesting part.
16:38So like that goes back to,
16:40I guess the first thing,
16:42which is Loud,
16:43it just wants to democratize
16:44the music industry.
16:45So if AI helps us do that,
16:47then I don't think
16:48there's anything to be fearful of.
16:50I just have a completely,
16:52I guess like a lot of people
16:53talk about music
16:54in a competitive way
16:55and I think that
16:56whatever makes great music
16:57is exciting, right?
16:59So I mean,
17:00I don't personally believe
17:01it's going to be
17:02artificial intelligence
17:04just because I know a little bit
17:06and I understand a little bit
17:07about machine learning
17:08and what that process actually is
17:10and I don't know
17:11if it's applied,
17:13if it can be applied
17:14in the same ways
17:15in actual music creation,
17:18like songwriting,
17:20composition.
17:22There's a lot about music
17:23that it's more about
17:24us communicating with each other
17:25than it is like
17:26who can do it better.
17:28So fears,
17:30I would say more excitement.
18:19I love you.
18:21I love you.
18:23I love you.
18:25I love you.
18:27I love you.
18:48I love you.
18:49I love you.