Uncover how lyrics are driving music innovation with our guest, Darryl Ballantyne, CEO and founder of LyricFind.
This episode dives deep into the magic of LyricIQ, a tool that uses AI to expose the concealed meanings in songs, proving that even the most cheerful tunes can have a darker undertone—you’ll never hear MMMBop the same way again! We'll also explore how LyricFind's innovative technology enables services to create child-friendly streaming platforms through content filtration.
Darryl discusses how synchronized lyrics, lyric translations, and content filtration can amplify your driving experience. Imagine the convenience of quickly accessing your preferred songs through voice search and AI without the need to remember the title or artist.
Finally, Tristra and Darryl examine the role of blockchain and AI in the music industry. Understanding the intricacies of publishing rights can be overwhelming, but we'll break it down for you, explaining how AI can streamline processes and maximize efficiency, leading to greater content availability and increased revenue for songwriters and publishers.
Looking for Rock Paper Scanner, the newsletter of music tech news curated by the Rock Paper Scissors PR team? Subscribe here to get it in your inbox every Friday!
Listen to the full episode here on this page, or wherever you pod your favorite casts.
Listen wherever you pod your casts:
Episode Transcript
Machine transcribed
0:00:10 - Tristra
Welcome back to Music Tectonics, where we go beneath the surface of music and tech. I'm your somewhat irregular host, Tristra Newer-Yeager, chief Strategy Officer at Rock Paper Scissors, the PR firm that specializes in music innovation and music technology. I recently got to catch up with a longtime friend and supporter of Music Tectonics, Darryl Ballantyne, the CEO and founder of LyricFind. Since co-founding LyricFind in 2004, Darryl has grown LyricFind to be the largest licensed Lyric database in the world, having established decade-long relationships with the majors and over 6,000 music publishers from around the world. Lyricfind is trusted as a premium provider to top industry clients, including Amazon, google, youtube, deezer and many, many more.
Now, for those of you who listen regularly, you know Darryl has been on the podcast before to talk about how LyricFind pioneered lyrics as a revenue stream for creators and rights holders and about all the innovative tools LyricFind builds for an evolving music industry. This time, Darryl and I got to talk about filtering explicit content you know you want to hear about that how people interact with music differently while they're driving the dark side of Hansen, and so much more. So stick around and come check out these surprising stories about lyrics and music tech. Hey, Darryl, welcome back to the Music Tectonics podcast.
0:01:38 - Darryl
Thanks, Happy to be back, and this time talking to you instead of Dmitri.
0:01:44 - Tristra
I know, I know we wanted to mix it up a little bit, just so you know you get to be subjected to the other hosts. This time, exactly, I'm an equal opportunity guest. I love it. Well, lyrics can do a lot of really cool things, and one really important aspect of lyrics is how they can help us find and sort songs effectively, so basically be able to curate not just by genre or BPM, but by emotion or even subject matter and a bunch of other parameters. So I was wondering if you could tell me a bit about the state of lyric-based curation and how you see lyrics shaping the way people are finding and putting songs together for specific purposes, whether that's for, like, a fitness class or for some kind of brand campaign. How do things look from LyricFind’s perspective?
0:02:36 - Darryl
Yeah, that's definitely a key area that we've been working a lot on and we think that it's extremely powerful and extremely important. We have a product called LyricIQ that is analyzing music through the lyrics, not looking at things like BPM and that, but looking at just what the song is actually about and what is the content of the song, because there's a gap there.
If you think about songs that sound happy and fun, then that doesn't necessarily align with the lyrics and what the song is actually about. One of my favorite examples that we discovered was MMMBop by Hansen. Everybody thinks about that song myself included being like a happy, fun, upbeat song, but then when you actually look at the lyrics of it, it's about getting old, losing your hair, losing your friends and that and like. The MMMBop is gone, excitement of life is gone and it's actually kind of depressing.
And if you're trying to find the right song for an ad campaign or for a workout or that, then and you want something that's happy and fun, or for a TV show or a movie or any of these six different sync licensing opportunities, that might not be the right song.
Or pumped up kicks is another one where it sounds happy and fun, but it's been a school shooting, so it's not.
Yeah, it's not the right lyrics. The audio sounds one way and the lyrics are actually another way, so it's really important to be able to differentiate between those and identify those to get the right song for for the moment. So, using the LyricIQ and both the emotion analysis, sentiment analysis, content filtration like that, we're helping a lot of services with that kind of problem and helping them to identify the, the emotions and sentiment and and topics that are in the song and what it's actually about. To be able to filter that in or out or make better decisions in in playlists or sync licensing or any of those types of areas. In store music, for example, it's key to have the right type of of lyrics, not just the right type of music, in that scenario. So it's it's actually a really, really interesting area and it's been fun to kind of explore and make those discoveries that blew my mind and a really good laugh Really realize that I thought of that song entirely wrong my entire life.
0:05:28 - Tristra
It's great that you know by using AI you're able to do a deeper literary analysis of the work of Hansen.
0:05:35 - Darryl
Yes, that's really what the people have been clamoring for.
0:05:41 - Tristra
So I was curious. You know we mentioned things. You mentioned things like filtering out content and that could be things like profanity or language that talks about, about sexual activities, that kind of thing, drugs, that sort of stuff. Are there any sort of positive examples that you've seen? Have people worked with LyricIQ and found you know, a set of songs based on their lyrics that really underlined their message in a positive way, like, have you seen that kind of example as well, in addition to just trying to weed out the pumped up kicks.
0:06:14 - Darryl
You know it's an interesting question and one of our weird realities of our business is that we provide the data but we don't necessarily see the granular information about the use.
0:06:24 - Speaker 1
Got it.
0:06:25 - Darryl
So we license some fitness platforms, for example, that use LyricIQ to filter things out or in based on emotion and sentiment and topics and things like that. You know we're one of my favorite LyricIQ clients. It's a company called Gabb Wireless that does a streaming service for kids. So you know that's I've got two small kids, so seeing that type of use and filtration to make sure that things don't make it into a service that aren't appropriate for them is really great, because there's so many different categories of content, there's so many different types of things that go beyond, you know, the traditional George Carlin seven dirty words.
So true, of course, language filtered or like the parental advisory sticker that doesn't really have any rules about when it should be on. You've got. And they're one of the. Again another Canadian example: "Life is a Highway" by Tom Cochrane isn't flagged as an explicit song, but they use the word shit in it. There's a line "after all the shit we've been through" and it doesn't even get bleeped half the time, I guess, because there's only one shit? so it's not.
I don't know what Canadian like broadcast law involves, but If you're a parent and you don't necessarily want your kids to hear that shit, that it doesn't matter if it's one or if I get me.
I actually it's not true. It does matter if it's one or if it's if it's 15 15, but it's still there. So there's no real rules around it. Then it's hard to measure. Or there may be scenarios where something is highly sexual but they never say a specific, explicit word that would flag flag it where it might be innuendo, or it might be slang that doesn't get picked up on or that. So you want to be able to Recognize and evaluate that on a pretty cut and dried system. You know, as a parent you want to make sure I don't have that stuff playing with my kids, for example. So it's yeah, it's a long overdue Set of data.
0:09:04 - Tristra
I Think that's so important to point out that you know before, I believe, a lot of the explicit you know markers etc. They were flagged by who was ever you know that was uploading the stuff themselves, or or you know Contributing it to a distributor. You know Putting the sticker on the front of the CD or the record or whatever, and that was all kind of like self-reported right. And Sometimes you'll see stuff marked explicit on Spotify for example, that you're like well, that's kind of I mean, I guess that's a little little naughty. But then there's stuff that you're like this isn't marked explicit what you know. So I think it's great to have that consistency using actual data that doesn't require humans to listen to like millions of hours of yeah, it'll be impossible to do it on a Manual basis, and a lot of older stuff too.
0:09:52 - Darryl
The explicit label would get applied to an album, so then it would be slapped on that's right track and, you know, maybe that's what contributed to Tom Cochrane, maybe the rest of that album had Nothing on it that was at all objective and they, they didn't want to label all 14 tracks as explicit for that one word.
0:10:17 - Tristra
Tom could have just said stop, anyway, we won't. We won't try to correct great songwriters here, but this is. This is really really exciting. So I'm also I'm curious. You know you're talking a bit about, you know you know a family oriented service. You're talking about being with kids. One of the places that I know people spend a lot of time with their families, sometimes in enclosed and inescapable spaces, is in a vehicle, and lyric find has done some really interesting deals with Vehicle manufacturers, which may be somewhat unexpected to some of our listeners. But I'm wondering if, if you could talk a little bit about some of that past and why lyrics are important To the company's building in car entertainment kind of mean. We might also need a little background on what in car entertainment means right now. So maybe let's start there. What are we talking about? We talk about in car entertainment takes. It's not just like a radio with a CD slot or a like way you can connect to Bluetooth.
0:11:19 - Darryl
Yeah lyrics in cars. Right, who would have ever thought that? That?
Life's a rich Not me, if you'd asked me 10 years ago, like are you gonna license lyrics to be used in cars? I would have said are you insane? I don't want to kill people, exactly, but the car experience has changed dramatically over the last little while with connected cars and and the opportunities that that opens up. So now you've got A cars that have a full-time data connection all the time you're, you're streaming music directly Into the car and they're they're connected in to the car without even having to to use your phone. They've got video services and, like, you can watch Netflix on a Tesla if you want.
0:12:14 - Tristra
And that explains the a self-driving mode and it's failings, perhaps Anyway no, it's not silly.
0:12:21 - Darryl
Actually, that's exactly one of the main motivations between the the rise of automotive infotainment systems is autonomous driving, mm-hmm. The automotive manufacturers are all looking to the future of an autonomous driving world where you don't actually need to be paying attention and suddenly you have a captive audience in the car and they have nothing else to do. So it becomes a little bit like being stuck on a plane, except More room, less people and you actually know who you're gonna be sitting next to. So that opens up all these possibilities of having a connected system and having full entertainment in there, because now people don't have to pay attention as much and you you've got a Captive audience for entertainment purposes in the car and you've got a high-speed internet connection. That opens up all of these possibilities.
And lyrics fit in perfectly as a lightweight data solution as well with what has always been the primary activity in the car and that's listening to music. So being able to have that visual aspect and on these large screens that are in cars now and have synchronized lyrics, lyric translations and everything like that in there, to really enrich that experience Right now, not for the driver but for everybody else that's in the car, creates huge value and huge entertainment value for everybody that's there. So we're able to now provide lyrics to the car over its own connection, have the, the translations in there, have the synchronization in there and you know, if you go and buy a new Mercedes right now, you'll get lyrics with that that car. That's coming from lyric find and we have four or five other oems launching with with lyrics over the next year. So it's going to start to be really everywhere Over the next short period of time as a key component to the connected infotainment system.
But when you talk about families in the car, it is a Really key part of what that automotive experience can be. When we tie that back into that content filtration that we were talking about earlier, because if you're in a car and you're listening to Spotify coming through the car and you've got a random playlist or or it's just generating a Playlist on its own, oh, and all of a sudden Cardi B comes on and you have a five year old in the backseat, maybe that's not appropriate for them, maybe you don't know if it's appropriate.
0:15:24 - Tristra
So many questions. So many questions will be asked. Yes, yeah, you don't want to answer those questions.
0:15:31 - Darryl
Not at that point, maybe later.
Push that off as long as you can, and as parents, we can't be expected to know what the content of every single song is that is out there.
There's 100,000, 200,000 songs being released every day. It's impossible to keep up, but we can keep up with that with technology, and we can keep up with that through lyrics and content filtration and being able to flag for a parent that's driving the car with their kids in the back that, oh hey, this song that just came on the radio or that you are now playing off of Spotify, has highly sexual content. It's really violent, it's racist, it's discriminatory in one of our way, and we LyricIQ breaks that down into 31 different categories, but really we can boil it down into one is this appropriate for kids or not? And then you can make a decision Do I want to keep playing this song and, you know, let them hear it. Maybe they're asleep, maybe maybe they're pretending to be asleep and you want to Put the station to a different one or skip the song and create a more family friendly environment in the car when you need to.
0:16:56 - Tristra
Yeah, there's. I mean, I could also imagine the a beautiful future where there is a filter on things like songs involving, let us say, certain bodily functions, where if you have certain kids, like my kids, if I could just block every song involving involving certain certain things, it would make our car rides less, less like battles.
0:17:19 - Darryl
Yeah, exactly.
0:17:20 - Tristra
There's a whole world of content that's just like customized for, for like nine year old boys searching.
0:17:27 - Darryl
Oh yeah, sometimes they don't even realize. Like I have two younger sisters and the older one of them, lara, if she listens to this show, kill me, but I remember her as a little kid running around the house singing liquid dreams by O Towne, having no idea what any of that meant, and my parents too didn't really catch on because it was just on the radio and it's like that's, yeah, she's, she's like seven right now, maybe not give it 10 years, little sis, oh, amazing.
0:18:10 - Tristra
Well, one other area that was really hot for a minute there about, you know, probably three or four years ago, four years ago now was voice and I'm wondering, Darryl, if you could give us like this is a little bit of a curveball, but your mention of lyrics in car made me think of not just the passengers but the driver and the poor soul sort of yelling at their phone or another device trying to get it to load the song that they want and they don't remember the title because it's spelled really weird or whatever. I'm curious where we are now with, like you know, voice search lyrics, that whole world. I know there's been a lot of sort of soul searching in you know, the Alexa and smart speaker world of late and I'm curious if you know how, how you're, how you're thinking about that as someone whose lyrics could really make that function, you know, make them those smart, smart speakers and voice search a lot more functional.
0:19:04 - Darryl
Yeah, certainly, voice search in the car is a big area. One of our clients, soundhound, is doing a ton of work in that, that space too, and they're they're doing a really great job with it. But you pretty much all automotive OEMs are implementing voice controls at one level or another. Because it just makes so much sense was when you need to search through all the music in the world. You can't do that typing on a screen. When you're driving you can't do that, like you know. Imagine using like T nine type of texting, and that's kind of what a lot of automotive interfaces are right now.
So voice just circumvents all of that distraction and makes it it's super easy.
So being able to search not just on artists and song name but also on on lyrics when you want to play a song, or asking for playlist by theme you play me songs about Toronto, play me songs that are happy about driving, or it's like that all that can be enabled through voice search and metadata around the music to make that much, much easier to do and much safer to do.
Like a lot of things in automotive really come down to safety, can you do this task without causing somebody to run off the road and crashing the tree. So that's the biggest decision maker factor in a lot of these implementations is is it safe, can we make it safe and how can we make it safer and enable more functionality and features there? So definitely, voice is going to be more and more important in automotive going forward. I would expect that in the not too distant future pretty much everything that is done will be through voice or will be able to be done through voice in the car to minimize those distractions where you're not turning and looking at a screen and taking your eyes off the road.
0:21:24 - Tristra
That's exciting. I can't wait to yell, yell lyrics at my car and have it play music.
0:21:29 - Darryl
I can't test the song that goes this yeah, exactly, dumb did it.
0:21:34 - Tristra
Dumb did it. Oh yeah, oh yeah. Who knows, it'll come up. It'll be exciting. All right, I want to. You know we've talked a little bit about the future, about all sorts of cool things that are just emerging, but I wanted to now take a little trip down memory lane with you, Darryl, because you started Lyric Find at a time when things were really different than the way they are now, in a very specific moment of both the internet and lyrics and the music industry, and you've seen a lot of changes to music and tech. So a lot of our listeners are, you know, have a startup or are working, maybe in a more entrepreneurial way, within a larger company, and so I thought it would be really fun to talk to you a bit about that journey and your experiences. Hope you don't mind. So, when you think about that journey, what lessons did you learn that you'd like to share? What would you say to yourself? You know, many years ago, as you were getting started, that you think might be helpful to other founders who are getting started right now.
0:22:41 - Darryl
Oh man, well I'm, you know it's a big one.
0:22:44 - Tristra
I'm so old now I've forgotten most of it. Just blocked it all out.
0:22:52 - Darryl
So long ago. Yeah, there's definitely a lot of things that I learned along the way. There are a lot of mistakes that I made along the way and sometimes we made the right decision and sometimes we made the wrong one. Some of the right decisions that we made, you find people smarter and more experienced than you and get them involved as soon as possible. For us, having people like Ted Cohen, andrew Stess, melta Holman and them involved from an early stage saved us so much time, so much money, so many headaches and being able to leverage their experience and knowledge and connections, because there was a ton that we did not know. We were coming into this industry straight out of university, had no institutional knowledge ourselves, very minimal connections.
I was fortunate enough to meet Ted at a conference early on and have him kind of adopt me and you know, having people like that around it prevented us from making a lot of bad decisions early on and kept us going and saved the company in many, many ways. Like every person there, multiple times did something that stopped the company from being bankrupt and disappearing. So I can't stress enough how important it is to have experienced people and advisors around you from an early stage to help steer you in the right direction, but I also did stuff the wrong way. So when we were trying to raise money early on, I was terrible at it. Well, my biggest shortcoming as a founder is my inability to fund raise. I never did it the right way. Obviously, we'd never succeeded in raising any money. The only person I was able to raise money from when we were really trying was my mother, and she had that obligation as soon as she had me anyway, so she don't signed up for that decades ago, but not focusing on that.
In the end it worked out fine.
Obviously, we're still here and I'm happy that we didn't take a bunch of money from a VC early on and be kind of beholden to that. But it was hard. It was tough that we couldn't do everything we wanted to do early on because we had to bootstrap everything. And if you're going to be raising money, then you pretty much need to focus on just raising money and do it in an organized way. And there's lots of stuff that's written out there now about how to organize that and how to be efficient at it, but I never followed any of that and I never succeeded in raising money.
0:26:06 - Tristra
So uh, you it, it, that's. It's interesting because you know your biggest mistake may also be one of your greatest strengths, because, while it's, you know it's wonderful to be able to raise money and that can be the right thing for certain kinds of businesses. Being forced to build slowly and, you know, methodically is also a kind of a blessing in disguise. Perhaps it's kind of amazing that that you persisted and, you know, kept building a business that's remained viable. That's pretty.
0:26:38 - Darryl
Yeah, I mean, you're right that it ended up being, uh, one of our biggest strengths, in the end, is our independence and, you know, not being stuck on the schedule of a VC fund that needs a liquidity event within, you know, five to 10 years, because we didn't really hit our stride until more than 10 years into it. So if we've been forced into a liquidity event or earlier, uh, I wouldn't be doing this now. We wouldn't have reached the success that we, that we have uh, and it would be a very different uh world and a very different life for for me and and for Moe, that's for sure. And because, because of that, you know, we've gotten to this point and, you know, maintained full control, we've maintained, uh, that, that ownership and that freedom to do the things that we think are right for the business and right for the our uh, our, our staff and our clients, and not necessarily worried about, uh, you know, hitting, hitting the next uh quarterly number or things like that.
0:27:55 - Tristra
You're in an interesting position in the business and that you sort of you know, you, you came into this and created a new market. I mean there wasn't prior to Lyric find, there wasn't really Lyric licensing per se as an independent and you know income stream, um, and you know, because you're kind of new to this, you have this whole new area. Um, I'm sure you had to make a lot of complicated decisions about technology, because there's really nothing that's sort of ready made for what you were doing. I'm wondering if there were sort of you know overall lessons you learned about deciding how to you know what technologies to to to use, how to, um, how to sort of sort out in a more like big picture way what was hype versus what was like could be a really fundamental part of your business. So you went with AI and you've been, you've been working with, you know, machine learning and AI, I think for what like four years now I'm thinking about, when LyricIQ launched.
0:28:55 - Darryl
Um, yeah, it's been years, I don't know. It all comes up, exactly, exactly.
0:29:03 - Tristra
That was just like like before it, before it started, before times. Um. However, I've never heard Lyric Lyric find breathe a word about things like blockchain, which is so how did you make those decisions as as a leader? I mean, that's fine. People can like blockchain all they want. It's not a note, no, no, no shade thrown at blockchain, but you've made very clear decisions about technology and I'm wondering how you made those decisions and if you have anything to share with, uh, with with emerging founders, um, on how to make those decisions.
0:29:31 - Darryl
So a lot of that comes down to my co-founder and CTO, mo. He keeps an eye on a lot of the emerging technologies and tries to figure out what's real and what's not and what's useful, and he's never been a fan of blockchain. Uh, for the music industry and for publishing rights in particular, uh, just because it doesn't actually solve an issue that that we have. You know, we don't have a problem keeping track of ownership information that we have. We have a problem getting ownership information in the first place, and yeah, I'll mend to that.
I've always used the analogy that like garbage and equals garbage out, regardless of how nice that garbage can is that you store it in. And and that's really what a lot of music publishing ownership information is, because people, uh and it happens on both ends of the spectrum whether it's the indie artist who just doesn't know any better, or huge publishers that just are so antiquated that they haven't caught up. Now, most of them have. Now they're much, much better down than they used to be, uh, in the years past. But, uh, you know, we have publishers that we've dealt with that can't tell us what they own. And if they can't tell us what they own, then what good is blockchain going to be for that life? People who actually own it can't do it.
Or people that are creating a song You've got a guy in a garage that uh is recording a cover of you know, don't stop me now, or something and they release it. They don't know that. Oh, I have to register that this is this song, this is this is queen. This must be owned by Sony ATV. Then, because it's like they don't know and you can't reasonably expect them to know, so they're not going to register it because, with the amount of music that's being released now. It's impossible to manage that on the long tail. So blockchain has never solved anything for us about how we manage data or how we we store that ownership. But a regular database works just fine. It's getting that information really at the point of creation is a problem and you're never going to to be able to do that with the scale and long tail of of creation that's happening now.
So there isn't really a technology solution to having proper ownership and tracking of every song that's released now, because it's just the wild west. So you can't. You can't use something like blockchain. You have to do things like the MLC, and the MLC has a good step in the right direction where it forces people. If you're going to try to go after somebody for statutory damages for copyright infringement, then you need to have registered it there. That makes sense, right To let people know. It would be interesting if DSPs started only accepting content if all of that ownership was properly tracked and recorded and registered when it was uploaded. I don't think that anybody's going to do that because they would put them out in massive competitive disadvantage compared to the other services. So that's unlikely but could help solve the problem. But blockchain doesn't solve any of that, at least not in the music industry.
0:33:19 - Tristra
But AI has proven, at least for organizing information, to be really a benefit, and you have such a huge data set that AI can be machine learning and large models can do lots of really cool stuff.
0:33:33 - Darryl
Yeah, absolutely, and we're using AI for a lot of different things.
On our back end, we're pretty much always pairing that with human curation, though, because it's not perfect, it's not something that you can set it and forget it and never have to look at it again, because it's going to spit out exactly what you need For the foreseeable future.
You're going to need that level of human curation and moderation to make sure that things are actually correct and not embarrassingly wrong, but it's getting there, and there's a lot of things where AI is a really useful tool. Translation is one of those where you can make translators so much more efficient by using AI. Word by word synchronization, which is super tedious and time consuming and expensive to do manually with a human, but you can use AI to get pretty close and then fix it from there and be much more efficient and cost effective. So there's a lot of stuff like that where the AI tools that are becoming available now can really make humans much more efficient and effective, and it enables us to make so much more content available and, as a result, generate more revenue for the songwriters and publishers that own that content, because so much more can be covered, so much more opportunity is unlocked by that efficiency that really grows the market overall.
0:35:23 - Tristra
On that cherry note, thanks so much for talking to me today, Darryl. This is really really cool. I love how we got to talk about everything from cars to translations, to all that good stuff. So thanks so much.
0:35:56 - Speaker 1
And, while you're there, look for the latest about our annual conference and sign up for our newsletter to get updates. Everything we do explores the seismic shifts that shake up music and technology the way the Earth's tectonic plates cause quakes and make mountains. Connect with Music Tectonics on Twitter, Instagram and LinkedIn. That's my favorite platform. Connect with me, Dmitri Vietze. If you can spell it, We'll be back again next week, if not sooner.
Let us know what you think! Tweet @MusicTectonics, find us on LinkedIn, Facebook and Instagram, or connect with podcast host Dmitri Vietze on LinkedIn, Twitter, and Facebook.
The Music Tectonics podcast goes beneath the surface of the music industry to explore how technology is changing the way business gets done. Weekly episodes include interviews with music tech movers & shakers, deep dives into seismic shifts, and more.
Comments