Exploring the Chord Genome, with Austin Brentley

Christopher: Welcome to the show, Austin. Thank you for joining us today.

Austin: Thanks for having me, Christopher.

Christopher: So I have been in enamored with The Chord Genome Project ever since our team came across it and I’m really looking forward to talking to you about that but I want to begin at the beginning and understand where this project came from and, for yourself, what what was the journey of learning music like?

Austin: Oh, okay. Well, the project originally began as just a tool to help me find easy three-chord songs. I was basically tired of playing “Three Little Birds,” and “Knock on Heaven’s Door” and the ones you keep singing again and again when you search online.

And then somewhere along the way, I think, Axis of Awesome came out with the four-chord song. It was a viral video where they stitched together all these pop tunes and so I realized that there are tons of tunes out there and it’s just a matter of finding them and so I started on creating a search engine that could help me easily find, at that time, three-chord songs.

Christopher: Nice. And let’s rewind the clock a little bit. What were you playing these three-chord songs on and what had you been doing up until this point to know that three-chord songs were even a thing?

Austin: Well, I’ve had a guitar pretty much my entire adult life. Every time I change countries I get a new guitar, well, old guitar from the pawn shop and consistently, it’s just languages there because I’d pick out a tune. I really did enjoy, like, bossa nova music or early rock and roll. Bossa nova was way out of my league, like, “One Note Samba,” do you know that?

Christopher: Mm-hm. By Ella Fitzgerald. Yeah.

Austin: Yeah. So that’s, you know, I was always punching above my weight but even with early rock and roll songs the progress just didn’t materialize because you’re trying to bring together all these unfamiliar chords all at once and if you stop early before you finish that song you don’t get anything, so a lot of wasted effort and then, you know, you feel, I don’t know, defeated, or I felt defeated and so, I’m sorry. Go ahead.

Christopher: I can certainly relate to that, you know, I think the traditional methods for learning guitar are very much “Let’s throw all of the chords at you as quickly as possible,” you know, whether you’re learning from a teacher or from a method book and I think that chord overwhelm hits a lot of early guitar players. Were you teaching yourself? Were you studying with a teacher? When did you begin guitar?

Austin: I probably began around 18. I’m 40 now. There weren’t a whole lot of online resources at the time but that really wasn’t the problem. Like, it, there were books and if I’d been more dedicated I could have done it but what often happens is that the songs that inspire you to rush out and get a guitar are usually out of your league or out of my league, anyway, and so it was the same cycle again and again and I said, “Okay. I’m gonna do it this time. I’m gonna sit down with a guitar,” and a week or two I’d give up in frustration.

Christopher: Gotcha. And what kind of family background did you grow up with? Were you someone who was expected to learn an instrument? Was this a total tangent for you compared with what your family were expecting, or…?

Austin: My family’s a very musical family. My parents met at conservatory. My grandparents taught violin, viola, piano for years in the DC metro area which is where I am from. My brother’s a performer. So everybody was expected to study an instrument. I studied flute and piano. You know, like, I did okay but I, it didn’t really click. But when I started to travel more the guitar just made a lot more sense and that really didn’t click.

Christopher: (Laughs)

Austin: I thought it would be easier because I’m not soloing, I’m just strumming chords. I was wrong.

Christopher: Mm-hm. Well, this is something I want to definitely dig into in this episode because I think it’s kind of crazy that piano and guitar are thought of as kind of the default instruments for people to learn in a lot of cases because they are two of the hardest instruments for a beginner to learn if you ask me and, you know, guitar, yes, you sit there and strum but that’s a lot of independent left-hand, right-hand stuff going on and, as we’ve touched on, there, often the method you’re following or the songs that you want to play are expecting an awful lot of expertise from you quite early on in the journey.

Austin: Right.

Christopher: So I’m sure a lot of our listeners who have tried guitar or even struggled through and become guitar players can relate to what you’ve been describing there.

Austin: Yeah. I think it’s, yeah. Like, there are some people who are just naturals and they really don’t understand what I’m talking about but I would suspect that the majority of people out there have tried the guitar, you know, just a friend’s guitar lying around, and the majority of those people are not playing a year later, so yeah.

Christopher: I’ve often thought that if learning guitar wasn’t about learning one new chord every week and one song for each new chord but instead was about learning three chords and then a hundred songs using those chords we might have a lot more continued guitar playing than we do and a lot fewer people giving up in the first six months.

Austin: Yeah.

Christopher: So on that note, how did things change for you? You were, it sounds like you were pretty consistent, you know, buying a guitar each place you moved and strumming away, having these bursts of enthusiasm. Were there any insights or breakthroughs that let you get away from that frustration?

Austin: Well, it, in building this platform, just, basically I kept discovering more and more three-chord songs, even songs that aren’t, like, normally considered three-chord tunes are playable with three simple chords, like, “Leaving on a Jet Plane,” or “I’m a Believer,” by the Monkees, which is, like, the first tune that I could call my own and basically I kept building a larger and larger repertoire and once you have those three chords it’s just a matter of polishing off these new tunes because the slightly different tempos and melodies but still fundamentally the same song, well, not the same songs but they use the same chords and it was just a much easier journey.

Christopher: Yeah. So let’s talk a little bit about that because the idea of three-chord songs has come up on the show several times before as a really neat way in to playing chords by ear, in particular and when we say three-chord song often we’re kind of assuming it’s gonna be the I, IV, and V chords but I believe with Chord Genome Project you’re not necessarily assuming that, right? When you say three-chord song you literally mean just a song that uses three different chords.

Austin: Correct. Like, the greatest number of songs are going to be, like, ones that use C, G and D or C, F and G, like, the same relationship but there are thousands of chords in the database. Just so we’re clear, like, when you click on a song in the search engine you’re directed to an off-site location, Ultimate Guitar, for example. So it’s really just indexing the songs that are out there in the world, you know, they’re not hosted locally but there are three-chord songs that use minors, seventh chords and the platform is flexible enough that any chords you type in you’re probably going to see results.

Christopher: So why was this important for your learning? I think people listening can probably understand that, you know, if it’s just three chords that’s a relatively easy song but was there more to it than that? Was there more to this idea of thinking really about the chords used in the songs and picking your next song to learn?

Austin: Well, actually, what kept me going is that the more people I told about this idea to, like, they were, like, “That actually makes a lot of sense. There should be a search engine like that,” and so that kept me going at a time when, you know, I wasn’t, you know, I was pouring money and time in but not getting anything back yet. In fact, every hour spent building this was an hour of not playing, so it really was kind of a big investment but that kept me going. Like, the feedback that I got from friends and family.

Christopher: Cool. And so let’s pause for a moment and think about that decision point. So you’ve been learning guitar, you’d been doing these bursts of enthusiasm, it had been a bit frustrating because it felt like every new song that you were kind of trying to take on the world and you gradually realized that actually you could leverage the fact that often there were chords in common across songs and you started to pick songs based on that, right?

Austin: Right.

Christopher: And at some point you thought, “I could build something to figure out these songs for me,” or at what point did you think, you know, this way of learning guitar is really working a lot better?

Austin: Well, so at the time it wasn’t actually for learning, it was just for growing my repertoire and then a buddy asked me, “Well, I’m starting from scratch. I don’t want to do three chords. If I just learned one chord what chord should that be?” and from that came a new feature. So, basically, we had all these songs indexed, organized by the chords they used and with that data you can determine what chord to add next no matter what your level is and that chord should be whatever one allows you to play the most new music, the most new songs and so we built this tool. It’s called “The Next Best Chord Feature” and it looks at what you know, the chords you know, and then it recommends. So if you know G-C-D, for example, it would recommend that you learn E minor next because adding that one chord allows you to play, I mean, we’re talking about going from about 4,000 songs to 12,000 songs just by adding a new chord.

Christopher: Amazing. And that comes back to what we were mentioning there, the 1-4-5 and the 1-4-5-6. You’re adding that six chord and because so many songs use these kinds of progressions your database knows easily 8,000 extra songs that suddenly you can play.

Austin: Exactly and if your starting point is C, F and G then A minor would almost certainly be the next best chord to learn so it’s adaptive, like, no matter what chords you know, including zero, it can tell you, “Well, if you’re gonna learn one chord, learn this one.”

Christopher: Very cool. So that is a really practical way to take what was, kind of, a search engine in the database and turn it into a learning tool, something that can really help people enjoy learning guitar, or, I guess, any instrument that plays chords, faster. Is that right?

Austin: Yes. That’s, basically, once we had these two components built that was the natural leap. Basically, starting from zero, you keep adding one chord at a time, which is much easier than trying to assemble five unfamiliar chords. It just, with that one chord, once you get it down, an entire universe of new songs become available. In fact, you use those new songs to get better at that one chord, which is tough in the very beginning but with practice gets easier and easier and then you’re ready for the next new chord.

Christopher: Nice. So I’m going to ask a couple of questions that I think might be on our listeners’ minds at this point and the first is, can you really play any songs with just one chord, when you were telling your friend what one chord to learn?

Austin: Yeah. Basically any song that you can, well, don’t quote me on this, but any song you can sing in the round, like, “Row Your Boat,” or “Frere Jacques,” probably is playable with one chord. So those are two songs that you can definitely play with a single chord and there is a song that they aired called, “The One Chord Song,” by Keith Irvin but I’ve never heard of it, all playable with any major chord that you want.

Christopher: Very cool, and so the other question they might have been thinking was at the other end of things, supposing you’ve got three or four chords under your belt or maybe half a dozen and this search engine is telling you thousands and thousands of songs you can play with those chords. That’s exciting but it’s also a bit overwhelming, right?

Austin: It is.

Christopher: So how do you turn that into something manageable and something that actually gives you a clear next step?

Austin: Okay. Well, so it’s not an elegant solution, partly because it’s so hard to find information out there, but the site comes with genre and decade filters so if you like rock music from the 50’s and 60’s you would apply those filters and you’d get a list of maybe 100 or so songs that use G, C and D from that. But it’s still highly experimental because songs get re-released every single year and so trying to find the earliest release date is not easy and trying to find, like, the most popular genre assigned, so, like, “Ring of Fire” by Johnny Cash, it’s been on soundtracks probably Christmas albums, pop albums, rock albums even though it’s undeniably a country song so anybody who uses the platform, like, understands that those filters are super, duper beta but they do help because thousands of songs is overwhelming.

Christopher: For sure, and, you know, from my background in digital music and kind of combining science and music I have friends who have worked for Last FM and Musicbrainz and the BBC and I’ve kind of seen a glimpse behind the curtains of how hellish that metadata problem can be.

Austin: Yes.

Christopher: You know, taking a single song and, you know, it’s, what’s obvious to a human that “Ring of Fire” is a country track by Johhny Cash actually can be incredibly hard to pull out from the haystack of data you find online. So —

Austin: Yeah. In fact, I’ve used both The Last FM and Musicbrainz and it’s tough to navigate but until I find a better solution those filters are here to stay.

Christopher: Nice. Well, it certainly does the trick and there were a couple of reasons I and the others on the Musical U team were really excited to discover this project and I think for me it was partly just that geeky appreciation of the challenge you’ve taken on and how simple you’ve made it for the user because behind the scenes I know there’s a crazy amount of implementation and analysis and, kind of, problem-solving that needs to happen but from the user’s perspective it’s as simple as you’ve described it. You know, you type in the chords, you hit search, you do some filtering if you want and you get back the answers.

Austin: Yeah.

Christopher: Super cool. So that geeky appreciation was definitely there but also it really scratched an itch for us because we teach a lot about how to recognize chords by ear and we talk about three-chord songs and four-chord songs and describe it as kind of the shortcut method where, you know, if you learn to recognize just a handful of chords or chord progressions by ear it opens up all of these songs to you compared with if you’re trying to take on all of the major and minor chords at once and so part of the challenge for us is always how can we give people practice tracks or practice examples to work with and so your website is a fantastic resource for that and I would certainly encourage any listeners aside from using it as intended and clicking through to see the chord chart on ultimateguitar.com or wherever it comes from, just that initial search can really be a useful training tool. So if you know you can play C, F and G, do the search, skim down the list of songs, find one you know and then just sit down and try and figure out the chord progression by ear, you know, where do the chords change? Which, is the change in the four, the five? That’s a really fantastic exercise and you’ve skipped that really, I was about to say, dangerous, that’s overstating it, but that potential pitfall of trying to figure out your random song on the radio and not realizing it’s using all kinds of weird chords that you’ve got no chance of recognizing by ear. So that selection process is a really helpful thing for ear training, I think.

Austin: Yeah. I can actually speak to that. Like, the more that I play, like, sometimes the song will come on the radio, like, I have a two-and-a-half-year-old and “Under the Sea,” the Disney tune from Little Mermaid, it’s like, “You know that, I think I can play that with three chords,” and it’s not an official version, like, Disney would not pay me money to perform it but it’s playable with the three chords and I’ve heard that song a thousand times. It never occurred to me and I think just as a result of playing more and more I’m making those connections that you’re talking about.

Christopher: That’s great, and you, we were talking a little bit before we started recording and you were saying how you’d kind of brushed up on music theory along the way to build this project and I think what you just described we should dwell on for a second because I think everyone knows the guitar player who has memorized 300 songs and played for a decade but can’t play by ear at all. So it’s certainly possible to immerse yourself and not develop your ear in that way but you, I think because you have the theory knowledge and this appreciation of what’s going on musically you have be able to kind of leverage that immersion to actually develop the ear skills, it sounds like.

Austin: It’s, you know, it’s an organic and slow process but I’m getting there. Like, I’m picking up music theory tidbits along the way and, like, you know, I wouldn’t hold me up as an example to the world but where I was and where I am now, for me, is a huge difference.

Christopher: And I had assumed, when encountering this project and discovering that you were the guy behind it, I had assumed you must be a programmer and amateur musician like myself who had kind of put this together in his spare time and it was growing into something bigger, but that’s not the case. Tell us a bit about what it took to build the Chord Genome Project.

Austin: Well, I mean, an endless search of developers around the globe, like, this is in year four now, but I’m not a coder at all and trying to explain what it is I want to do to the developer who, you know, is a genius at code but doesn’t know music theory, necessarily, it’s been a struggle but I’ve actually found one who admittedly doesn’t know much music theory, necessarily, but we’ve done some, we’ve covered a lot of ground together and so I explain what new feature I want to add and he’s able to implement it and I’m happy now.

Christopher: And have there been any particular challenges along the way apart from trying to communicate this to coders who maybe don’t get the music theory or quite understand where you’re coming from? Have there been any hurdles to overcome?

Austin: Every day. Every day there’s, like, a new challenge that I didn’t see in advance but, like, one of the big ones early on was the need to standardize chords in the back end. So you can spell, for example, A major many different ways. Just by itself, a capital A, or A with a big m, A major, A major with and without spaces but all those are fundamentally the same chord and so they have to be stored as one value so that a user who searches any of those spellings will see any of those songs use any of those spellings and then we have to do the same thing with all the other chord families to minors, diminished, etc. and also with harmonics.

So once we had indexed most of the songs we were going to use they were probably, I think, 50,000 different chords but after standardization there were 2,000 unique chords, 2,100 or so and as a result of standardization.

Christopher: Interesting. I guess that comes back a bit to that point that what seems obvious to a human is actually quite complex from a machine’s point of view, you know, a human who’s studied a bit of theory knows that a capital A means the same thing as A with a capital m next to it. That’s how chord charts work but to a machine those are two completely different things.

Austin: Yes. I have other challenges, like, I could fill up an entire podcast with — so we’re still in beta and probably for a while. Like, another thing that the songs that we index to, they’re on user-submitted websites like Ultimate Guitar which are fantastic resources but anybody can edit or remove songs so the, what we’ve indexed today won’t necessarily be how the song looks tomorrow, so it’s a constantly moving target and so I’d like to figure out a way to show users more accurate results and you can sometimes, I guess, resuscitate a song. Are you familiar with Archive.org?

Christopher: Mm-hm.

Austin: Yeah. It takes snapshots of the internet but I really don’t want to use that approach. I want to send the traffic to these websites because they put together amazing resources. So that’s another challenge right there.

Christopher: Fantastic. So you mentioned there how there are 2,000 or 2,100 unique chords that your database knows about and, again, I think the listener might be thinking that sounds crazy. (Laughs)

Austin: Yes.

Christopher: You know, “I’ve just learned my first three chords. How do I get from there to 2,000?” Can you speak to that at all, and, you know, if and how someone needs to learn those 2,000 chords?

Austin: Okay. So there are about 350,000 songs in the current index and you need 2,000 chords to be able to play all of those songs and we actually ran a process starting from the first chord and with each new chord you’re able to play a slightly higher percentage of the database. But what we’ve discovered is once you’re at chord number 23 you can play 50% of the songs in this database, which just blew my mind.

Christopher: Wow.

Austin: Thousands of songs, thousands of chords needed to play everything, but with just 23 chords you can play the majority of songs and, like, this index is no slouch. It’s a pretty representative cross-section of songs from genres all over the globe and users from around the world submitting songs they like. So there’s pop, rock, jazz and folk from all different decades. So by extension, in theory you could play roughly 50% of all English-language popular music with, it might not be exactly 23 chords but it would be absurdly low.

Christopher: Amazing. And, to be clear, we’re not even talking about transposing things into a convenient key, are we? Like, you’re talking about playing these songs in the key they were written in.

Austin: Well, yes. And, like, once you land on the page, as written, without any transposing, if you did transpose, I mean, it would probably be 80, 90% if you transposed all of those 23 chords up and down, well, I guess, 11 times. But with 23 you could do a lot of damage. Just 23 chords and so, like, I actually wrote a blog post, “The 23-Chord Challenge,” where, it’s ambitious if you learn one chord a week. I don’t know if I could do that one chord a week on the guitar but if you learn one chord a week then in less than six months you would be able to play roughly 50% of the songs out there, which is kind of a cool concept.

Christopher: Yeah. That’s super interesting and we’ll definitely put a link to that post in the show notes because I think whether or not you want to embark on this challenge or, indeed, use the database to guide you in it, I think that analysis alone is super-fascinating because, you know, we talked a bit about how guitar method books often just throw everything at you and I don’t want to overstate it, but I think often they’re a bit misguided in their choice of chords and, you know, a lot of musicians are left thinking, “Okay, first I have to learn all the major chords and then all the minor chords,” and that’s 24 chords, right there.

Austin: Right.

Christopher: And, you know, it’s not the case that the 23 chords you just mentioned are all of the major and minor chords, by any stretch, so I think if nothing else then everyone listening to this should go check out that blog post and see for yourself, what are those 23 chords that unlock half of the music to you and do a little bit of theory work yourself. If you’re familiar with the idea of 1-4-5 or 1-4-5-6, just look at that chart that Austin’s put together and ask yourself, why are they these chords? Why are these the ones that artists and musicians and songwriters have used again and again?

Austin: Yeah. There are a lot of seven chords. Yeah, if you just did majors and minors, which is an intuitive approach, but you’re leaving a lot of songs on the table.

Christopher: Mm-hm. So that’s a great exercise. Austin, you are a really interesting blend of music and tech with this project and I certainly am a big believer in how technology can make it fun and easy to learn music in a way that’s never been possible before. Aside from the Chord Genome Project, have you come across anything that can help our listeners learn faster or enjoy learning more compared with the traditional approaches?

Austin: Yeah, actually. Like, the purists out there, might hate this. My Dad, a flute teacher, definitely hates this idea but I use this, it’s a metronome. My phone is turned off for the podcast but it’s a metronome that speeds up imperceptibly but it’s, like, it slowly speeds up with time, forcing you to play faster and faster but, like, you don’t really notice it and then every now and then, like, I’ll stop the metronome and then go back 50% of the way, make it 50% slower than where I stopped and then I start it again so, like, because I’m taking a step back, I’m like, “Oh my gosh, this is unbelievably slow,” but it’s much faster than where I started so that has helped me a lot.

Again, my Dad hates it. He’s all about standard time and, but I would definitely check that out. There are apps like that for Apple and Android, as well.

Christopher: Teriffic. I love that tip. I’m kind of half-heartedly, well, enthusiastically but without a lot of practice time, learning drums at the moment. I have a pair of drumsticks on my desk and I’ve been following a few tutorials that are, like, “Try it at 60 bpm,” and “Try it at 70,” and blah, blah, blah, which is fine but it’s kind of tedious to stop and start like that. So I love that idea of, you know, speeding, both the idea of gradually speeding up and having a metronome that will do it automatically for you and that reset to 50% of the speed so you can enjoy that moment of, “Oh, this is kind of easy, now.”

Austin: Exactly. It kind of reminds me of the mood lighting that restaurants use where over the course of 15 minutes or half an hour the lights become a little bit brighter, a little bit brighter and then all of a sudden it gets dark and romantic, but, you know, it just goes on an infinite cycle, doing that again and again.

Christopher: Is that true?

Austin: Yes.

Christopher: Ha. I have never noticed. I mean, I’ve noticed it getting suddenly darker but I’ve never known they were doing a cycle.

Austin: Yep.

Christopher: That’s fascinating. I’m gonna watch for that next time I eat out. That’s very cool. Wonderful, Austin. It’s been such a pleasure to talk with you. Can you tell people where they should go to try out the Chord Genome Project?

Austin: Oh, okay. So thechordgenome.com or you can go to thechordgenomeproject.com. I think by the time this podcast airs we’ll have a search page up that doesn’t require any login. Just head over to the search page then if you’d like to see even more results there are paid versions, obviously, but try before you buy and, you know, I hope it helps the listeners out there.

Christopher: Amazing. I’m sure it will. I think, you know, we talked about checking out that blog post purely for a great insight into chord progression theory and the most useful chords. We talked about using it as a way to guide you in learning guitar or piano and what chords to learn next and also purely to get more fun and repertoire out of the chords you already know and I definitely recommend that ear training exercise, too, where you look up what songs use a certain set of chords and then you just sit down with your instrument and try and figure out for yourself how the progression goes.

So this is an amazing tool to have at your disposal. A big thank you and commendations to Austin for putting it together because I know it can’t have been easy and I think this is going to become a widely used and valuable tool for musicians.

Austin: If anybody does use this site, please do not be shy with feedback, like, if you ever face a problem others might, as well and so it helps me make it better and better, so.

Christopher: Wonderful. So Austin, a big thank you for coming on the show.

Austin: Thank you very much, Christopher. It was a pleasure.

Subscribe For Future Episodes!

Apple Podcasts | Android | YouTube | Google Play | Stitcher | RSS

Enjoying the show? Please consider rating and reviewing it!