Hacker News new | comments | show | ask | jobs | submit login

Anecdotal: there's a few different approaches to learning songwriting that seem to click for beginners. The "build up" approach is the most common and is what this link offers: It first teaches beats, then chords, then melodies and then, in theory, vocals etc. These lessons in this order make sense to many people, but not everyone.

If you're interested in learning to make music and the lessons in the link are confusing or overwhelming or boring, some students find a "peel back" approach to learning songwriting easier to grasp at first. A peel back approach just involves finding a song then teaching by stripping away each layer: start with stripping away vocals, then learn melodies, then chords, then finally learn about the drum beat underneath it all. A benefit of the peel back approach to learning is melodies and vocals are the memorable parts of a song and easiest to pick out when listening to the radio so a student can learn using songs they know and like. Either way, songwriting is hard and fun. Best of luck.

P.S. I think Ableton makes good software and I use it along with FL and Logic. They did a solid job with these intro lessons. But worth mentioning, there is free software out there (this includes Apple's Garageband) that offers key features a beginner just learning songwriting can practice on and mess around on before purchasing a more powerful DAW software like Ableton.

> there is free software out there

If anyone is interested in a Free/Libre/Open Source Software option (cross-platform Linux/Windows/Mac) I've really enjoyed producing with LMMS over the past 18 months or so: https://lmms.io/

It's definitely got room to grow in terms of functionality/interface but the development community is of such a size that it's possible to still make meaningful code contributions. I've contributed a couple of small patches to improve the Mac UI as a way to get familiar with the code base.

Of course, the downside is that I have to decide whether to write code or make music whenever I sit down to use it. :)

There's also a new project called Helio Workstation https://github.com/peterrudenko/helio-workstation it doesn't have much built-in instruments so you need plugins for everything, but the UI looks awesome

I like LMMS but it has enough vst issues that I just went with a cheap Reaper license. Plus I need an audio recorder as well.

I wouldn't say that "LMMS has vst issues". I would say that vst has a serious issue: it is not a open standard. Although it seems they are trying to improve that for the linux community: http://cdm.link/2017/03/steinberg-brings-vst-linux-good-thin...

Good to know. That said Reaper enabled me to use the 7-8 vsts I was really interested in using that didn't work (or worked poorly) in lmms. If I new c++ better I would contribute.

For recording, there's Ardour (among several other FLO options)

The Song Exploder podcast is awesome for both approaches, and I would really recommend anyone interested in writing and/or producing music to give it a go.

"A podcast where musicians take apart their songs, and piece by piece, tell the story of how they were made." @ http://songexploder.net/

Seconding this recommendation. The more electronic- or instrumental-oriented songwriters on the show tend to get more into the nitty-gritty details of production, layering sounds, etc. (the first two episodes with The Postal Service and The Album Leaf, for instance) while rock and singer-songwriter types tend more towards the songwriting aspect (the Long Winters is a personal favorite).

I sort of wish there were more technical details as a rule, but it's understandable given the relatively short format that they can only cover so much ground. I'd prefer longer episodes personally, but I suppose not everyone might, and there's tradeoffs in producing more content. I guess I'm just glad that the show caught on and is still going strong.

+2. Only podcast I listen to. The first few episodes are a little rough around the edges, bad questions, people not knowing exactly what they were getting into, but the rest are all incredible.

Protip: sample the guest's clips they put on his show :) I've gotten some really great material from this show sonically since most of them seem to be the individual instrument tracks.

+1 for SE. Also there is an amazing set of Motown tracks split into instrumentals and acapellas out there. I'm not sure if it can be legally obtained easily, but it's a great master class.

Is this a common distinction to acknowledge in general education environments? You pretty succinctly described the struggles I've tended to have in my education, and described it in a refreshing/revealing (for me) way.

I love looking at systems and peeling back the layers to find out what makes something tick. That's not an approach to learning that I really encountered until I entered the workforce and was met with complex systems that I needed to understand. And I loved it!

Interesting, I've never heard of the "peel back" approach, and I can totally understand why it would be instantly satisfying for a beginner in music to get started that way. Do you have any articles or books on the subject manner?

How would this approach apply to a more traditional instrument that doesn't have the advantages of having a "good" sounding sample already preloaded that can be easily layered into a song that you are composing? I grew up learning the violin and it was endless disjointed drills until it was put together in a classical song that I never heard before nor had the desire to play. 8 year old me just wanted to play the theme song to "Jurassic Park" and roar like a T-Rex.

I think there's a difference between learning composition, and learning to play an instrument.

In my view, learning an instrument has a lot in common with learning to code, in that some people take to it, and others don't. And we probably know some of the reasons, but not all of them. Of course teachers and teaching programs vary, as do kids and their family milieu. But nonetheless, music education has huge attrition.

For instance, by way of anecdata, I took string lessons as a kid and loved it, and my kids have gotten pretty serious on violin and cello. They actually like classical music, and it probably helped that both of their parents also enjoy it. So it definitely works for some people.

What you said about learning melodies and beats and chords kind of confused me. Do people actually learn how to make up music? I always thought it was just some natural ability that people have. For as long I can remember if somebody told me to write a song I would just spit it out after a while. Am I unuiqe in this respect?

I created an account just to reply to your comment. As someone who has played keyboard instruments for all my life, it's not crossed my mind lately that the idea that there is structure to music is not well known.

Just for fun: chords in scales are numbered from bottom to top in Roman numerals. I feels like home base, V feels like wanting to go home. If you want to create the feeling of going home but then not really go there you can go from V to VI instead of I. 'Sad but I have closure'-type ending? Major IV - Minor IV - I. Bluesy feeling? Add a minor seventh to your I, IV and V chords. Dreamy? Major seventh instead there, except on the V.

It's even entirely possible to learn to recognize all of these types of chord progressions and sounds instantly. I'm working on and off on an ear training app that randomly generates them that musicians can use to train their musical ear.

>I'm working on and off on an ear training app that randomly generates them that musicians can use to train their musical ear.

Sounds interesting. Please do a "Show HN" post when your app is ready for it.

Will do, thanks!

As someone who's recently tried to get into (basic) music theory, trying out the chord progressions you mentioned was fun! Do you happen to know of any resources that go through more of these well-known chord progressions?

I'm also wondering if these chord progressions work the same way for all scales, or if, for example, the 'sad but I have closure'-type ending only sounds that way in major scales? From experimenting I think it only works for major scales, but I'm not sure :)

You probably have learned most/all of your musical knowledge implicitly.

Some people have a great ear for music and can write solid songs without formal training in music. Other folks come at music from the more theoretical side, although usually with a lot of implicit knowledge of and experience with music as well.

For most people who are not formally trained in music, their songs can be improved upon on a technical level by someone who has deeper theoretical knowledge (learned either explicitly or implicitly).

For a good discussion of this, check out Tim Ferris' podcast interview with Derek Sivers. Derek talked about how he had learned a lot about music implicitly. In one summer, a teacher of his formalized that knowledge so efficiently that he was able to test out of lot of classes (1.5 years worth?) once he went to Berklee School of Music.

Songwriting can be taught, yes. In most music courses you start by analyzing the Bach Chorales, which (along with some Gregorian work from the middle ages) is what really kicked off contemporary music composition. By analyzing the Chorales and moving forward from there you learn how to manipulate chord progressions, harmony, point and counterpoint.

Composers classically trained this way tend (!) to have an easier time writing melodies, harmonies, and progressions in a consistent manner, ie not having to wait for "inspiration to strike". The composer, of course, still needs to develop an emotional connection in the music, but the point is that it can, and routinely is, taught.

My girlfriend is a trained classical singer whereas I'm a self taught musician. She doesn't really gravitate towards the rock music I like to write but because of her training she can easily jam, riff, or write anything far quicker than I can. Songwriting is a very technical skill indeed.

What I find difficult is that by the time I've got my DAW going and found some synths I like, the tune in my head has evaporated. Do all people find musical thoughts so insubstantial, or is it just me? If I imagine a picture or a paragraph of text, it'll stick around and I can remember it more or less indefinitely. I still recall snatches of crap poetry I thought up when I was a teenager, but any music I imagine just disappears before I can get it down.

The most successful tunes I made were more or less "discovered" from incrementally experimenting in the DAW, and not from any kind of original plan or idea. Maybe I'm just not a musician! (I'm an indie game dev who started making my own tunes for my games)

Imagine that you're a writer, and you have an idea, so you turn on your computer, wait for it to boot, log in, open up Word, and fiddle around with fonts for a bit... that's what you're doing.

Writers keep pens and notebooks by their bed so if they wake up in the middle of the night they can start writing right now. Or they have tape recorders. Anything works as long as it's immediately available. The iPhone has a "Music Memos" app, I'm sure there's something similar for Android. That's what I use.

Learning music theory and how to write music properly can come later. As long as you can sing, whistle, or hum a tune, you can record it.

you could consciously decide NOT to use synths to lay down the bones, always use a piano to begin with, once you have the tune idea down then you can move onto orchestration and picking synths and so on. Always keep the piano track as a guide and start adding tracks for all the other components until you have what you need.

From a remembering the tune perspective, I have the same issues, but I think it's more related to not applying musical lexicon and hearing skills the same way: you remember poetry or a paragraph of text because you remember the ideas and how to go from one to the other, if you are a musician and have something in your head and start thinking along the lines of "this is using a lydian mode, the progression is ii IV V I then it modulates to the relative minor and switches to dorian, also the theme is going down in thirds for two bars, then it will stay on the chord root for one and move to the dominant 7th" you are going to remember it a lot more easily than just by remembering the melody itself

It would be like comparing how easily you can remember poetry in English vs poetry in, say, Russian, where you only have the "sounds of the words" in your head to remember, but you don't have the syntax or the meanings to help you as well.

For me one of two ways works. Most often I start designing a patch on one of my synths and that ends up becoming a full song. Other times I start by noodling on the piano or organ and ending up with something I like. I suspect the more musically gifted do the latter more often, while the more technical ones like the process of patch creation, etc.

I evolved this way, though I'm far from gifted. Starting out, anything I made was driven by whatever sounds I was noodling with. Now, I almost start on the piano, compose the outline, and then pick the sounds that I think fit it.

The first approach has a sense of creative wonder to it, where your being guided by an outsider. As much fun as that is, it is very limiting and I suspect most people abandon that approach as their skill improves.

I found the same to be true. I've been trying lately to give up approaching music from the "I have this idea I want to get down" perspective. Instead, I set up my studio in such a way that I can easily "play around" and come up with ideas on the fly, and then elaborate on those.

Switching from a DAW to a mostly-hardware setup helped with this, as it's easier to "play" with knobs/sliders/keys/pads than virtual objects accessed via mouse/keyboard. Once you get things wired up, it's pretty straightforward: play around, find something you like, track it in, build more stuff over it.

Ever since making this switch, I found the parts that I used to practice/enjoy (like slicing and manipulating samples, for instance) feel much more tedious.

Another benefit is that it's easier to make mistakes, which often have more interesting results than the thing you originally intended. My guess is because this violates your internal "patterns" and forces you to think outside of your normal "music creation" schema, resulting in a more creative/unique outcome.

I've also tried to switch to "totally live" recording (i.e. minimal sequencing beyond loops and patterns, all automation and non-repeating parts done on the fly), and that's a bit more challenging, because you have to redo everything if you, say, screw up a little solo bit.

Improvising, screwing around, experimentation are great ways to generate ideas. I think most composers work this way, most generally don't start with some huge structure, they just find some germ cell ideas by improvising that then can be build upon afterwards with analytic techniques. I think there are two main approaches, the intuitive feelings and pure ear based approach which is what most people call the "talent" aspect and the analytic approach which is a lot like mathematics, it is about studying structure, and is a learned skill. The best composers will use both of these together. You should learn chord structures, scales and how to read sheet music. This will allow you to conceptualize a musical idea as a concrete mathematical object, and it will help you to not lose the idea. The reason you forget the music in your head is because you don't have enough reference points to define it in a memorable way.

You can understand a musical idea as a kind of memory impression, an echo that you can play back in your head, and also as a pattern of pitches and rhythmic structures . Having two reference points , sensory and abstract mathematical is very useful.

>by the time I've got my DAW going and found some synths I like, the tune in my head has evaporated

That's where music theory pays off. Learning to name chords, scales and arpeggios gives your brain a framework to reason about and remember musical ideas. It allows you to break the music into a more concise abstract representation, rather than holding it in your head as sound. If you understand the structure of music, it's far easier to make connections between different pieces of music.


Do you have much formal knowledge of music theory? If not, that might help.

When you "get a tune in your head", if you can describe it to yourself in abstraction, it will probably be easier to remember (or even just write down).

Check on this page on 12 bar blues for some examples of easy music notations. Similar types of notations and/or terms exist for different parts of a song.


I had exactly the same thing happen, losing the core ideas by the time I wired up the synth I wanted. So I dropped the DAW entirely. Now I create most of my music using Loopy[0] to layer the parts that I sing (or occasionally play). It's been fantastic for my creativity.

I'm starting to hit its limits for my workflow though. One of the really nice things about how easy it's getting to write software these days is that I can now fire up, say, a Swift playground, and after getting the fiddly basics of "how to record and loop audio buffers" with AudioKit, there are very few limits on what kind of idiosyncratic workflow tool I can design for myself. The UI looks and acts how I want it to, and since over the years I've trained myself to act like a human synthesizer, I can[1] compose a whole song without even worrying about having an instrument nearby.

[0]Loopy - Multitrack audio looping with very simple and expressive control https://itunes.apple.com/us/app/loopy/id300257824?mt=8

[1]The "can" is theoretical. This is my next big hobby project, and I'm still in the fiddly phase.

Maybe don't use synths for getting your ideas down.

If I have the beginnings of a song in my head, or I have been humming to myself, sometimes I just record the parts I have as vocals - humming or full-on beatboxing the bass/strings/lead/beats separately and as close as I can make them to my head-song (including filters with my mouth)- and then replace as I go, figuring out how to achieve the sounds that were in my head.

I have the same problem too - I solve it by humming/singing the melody and recording it on my phone. Afterwards, I'll find nice synths, put chords to melodies, write horn arrangements, tweak drum fills, etc.

My fiance who was a professional musician (she had record deals) always keeps a Tascam recorder. When she comes up with a melody or lyrics she puts them on that until we can get in the studio and record.

She hates music theory and trying to use her left brain for art. I'll say oh that is in F and she gets mad so, easier to just let her record it than try to notate it down.

I whistle into my phone. A few of my most complex pieces started that way.

Happens to me too. Ear training might help? There's an exercise where you take a song you know or a familiar recording and try to transcribe it by ear. (I find it's a pretty hard exercise.)

i spent 10 years building a project studio and optimizing the workflow and patch bay so that i can be recording almost any instrument within 30 seconds or so. that was a huge breakthrough and a huge commitment!

I think your natural composition skills are unusual, but not unique. I also begain to compose music at a very early age. My knack for picking melodies up out of the air and playing them on a piano when I was 7 years old was how my parents knew that I needed lessons. Naturally, I was surprised later in life to learn that other people had to learn basic things like pitch and rhythm; to me it had always been just as natural as speaking.

I believe the same is true with song writing, in a sense. You're still applying some parts of music theory, but most by-ear learners like ourselves simply grasp the concepts and have internalized them naturally, without needing to be taught. Music is little more than patterns at the end of the day, and our brains are very good at recognizing patterns. What you and I know intuitively, others can learn through training and repetition. Both approaches are valid, and yield interesting (and often different) observations.

I went going through Music Theory classes during my brief adventure with Liberal Arts Majors in college. I felt like I already "knew" the material in a way I couldn't quite put my finger on. It was like I was finally understanding what my brain had been doing all these years. I recommend it if you haven't yet had the experience.

>Do people actually learn how to make up music?

People have studied music and composition since at least ancient Babylon, so, well, yes?

>I always thought it was just some natural ability that people have.

With natural ability you can sing some melodies. For learning to play an instrument, adding chords to the melody, you need studying, even if you learn by yourself and by ear (as many folk musicians did). Song melody, one can have a natural feel for creating, but nobody just starts writing songs in full form "from natural ability".

>For as long I can remember if somebody told me to write a song I would just spit it out after a while.

What would that mean? You'd write a song on the guitar for example? If so, then you already know the chords. And not all of the theory, so how complex is your song? Just barebones songwriting (country/folk style)? Can you take it further? Can you write the parts for musicians to play on your song? Can you write different genres on spec?

There are more things in making music/songs than "spitting out" some melody.

When somebody asks me how to solve a particular database problem, or an IT problem, or how to write an algorithm to do something, I will think about it in the back of my mind and "just spit it out after a while," unless it's something difficult enough to warrant a literature review.

That doesn't mean that those subjects aren't covered in detail in textbooks and university courses, or that people cannot learn how to do it.

There are certainly people who have natural ability, and compose melodically, applying varying levels of knowledge in music theory.

There are other people who can't make heads or tails out of a keyboard, compose a tune in their head, or understand chordal progressions, but nevertheless compose music in layers and still do extraordinary work. They find what they like by playing with notes on the screen. Joel Zimmerman, a.k.a. Deadmau5, is an example of this.

I am an example of the former, with natural ability, bolstered by training in music theory. But I still use a layered approach when I am composing, generally starting with a beat or bassline, playing with melodic progressions in snippets, and eventually moving into a traditional composition process when I have something started that I like. Ableton makes this process extremely easy and productive.

Indeed. As a classically-trained musician, watching Joel's class on Masterclass and seeing him compose melodies by dragging notes around in Ableton until they "sound right to him" was eye-opening.

What surprised me was how he makes melody lines: Playing with chords until he likes the progression, and then pulling notes out of the chords to form a melody. And of course it makes sense on one level.

But I think melodically and tend to do a lot of counterpoint. Getting the chords out of my head and onto the screen is often the last thing I do. I don't know how well his approach would work with counterpoint, since counterpoint often creates and resolves dissonance using passing tones in double time.

Very few skills are just "natural ability". Music theory is an interesting and pretty important topic if you want to make music. Since people have been creating music for millennia, they have figured out many things that help composers.

Do you really mean to say you write songs without using any theory or explicitly sought knowledge whatsoever? Let's hear one.

I know a little rudimentary theory. Just enough to get mocked by someone with a real education. "without any theory" is sort of an impossible standard to satisfy, but I will say I never think about theory consciously, and go by how things sound. Anyway, this is what I'm working on:


Are the free instructions like this out somewhere (build-up and peel-back)?

Is Apple's Garageband free? I thought you need to own an OSX device to run it? (my understanding is OSX only runs on Apple hardware and also is not a free OS)

Yes, it's "free" (not open source). It's included with the purchase of a Mac.

The point is, that kind of free is marketing speak. More accurately, you can purchase Garageband as part of a package including Apple hardware. Or you could say Apple hardware is free with the purchase of Garageband.

Not any more.

So no, it's not even gratis, it's $500+ depending on how crappy hardware you want it to be tethered to.

That's a slippery slope to saying that OpenOffice for Windows isn't free software either because you have to buy a Windows box. This is not a useful definition of free you're using. GarageBand is not Libre software.

I dunno if it's a slippery slope starting at Garageband. I say the slippery slope starts at OpenOffice or GoogleDocs or something along those lines, given that OpenOffice could probably be run on a potato if you can find a way to install ubuntu on it and stick some RAM into it.

That's absurd, since OpenOffice runs in Linux, and is free, as in freedom.

when it comes to getting software, we've stopped including the price of the required computer since likeā€¦ 1995?

Garageband is completely non-free/libre/open which seems what you are saying

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact