How do you make your songs? // Workflow discussion

[Skrew]

How do you make your songs? Well most of them. This is pretty much my method. Well for when I'm making my R&B, Pop, Hip-Hop stuff.

Play/listen some classic piano songs to get my mind thinking. I really like relaxing to Moonlight Sonata.
Spend hours/days messing around with Massive, zeta+, Sytrus, and reFX Nexus making patches
Make various loops with the patches and save them all (either midi or audio if I already EQ'd it)
Listen to them and find a good starting point
Build on an intro, several basslines, and a starting structure.
Sort through my loops to find the perfect one. (Or just make a new one)
Start to layer out my song (verse, chorus, bridge, ect.) Longest part.
Choose my drums/percs I want to use and EQ/compress them.
Go through the song adding drums/percs. I never make loops, I just listen through the song and place them where they fit. Although I do make 1 or 2(they get repeated) because I get lazy.
Add any extra sounds, samples, ect.
Master/mix the song.

After this point, I have either listened to it long enough that I grow tired of it and throw away or gave to this guy Tori I knew who moved to Oklahoma(Joined the Air Force) and started his own small label. I never asked him to pay me but it felt good when I heard him or one of his friends using it in their songs. Other than that, nobody has ever really heard my music other than family/close friends.

I haven't done anything in months really except re-sampling and stacking up my library.

[Augment]

I already have a template that includes drops n everything, so it's easy for me to get going when I'm in the mood.
But usually I spend lots of time making sounds/beats, and then when I've got lots of samples, I'll make a song. Not always though

[hudson]

I usually start of with a sample that inspires me and take it from there. Beats come first most of the time, then bass, then leads then everything else.

[wub]

1) get more samples than you know what to do with. Spend hours sampling everything you can get your hands on. I don't think we've ever had a CD that we couldn't get at least one sample off. Try and avoid stuff you would naturally gravitate towards. Get unusual stuff.

2) When sampling, check especially the start and ends of things. Loads of good stuff there. Middle 8's / breakdowns too

3) Once you have 100000000's of samples, find one that you like and get a loop of some sort going. It doesn't matter if you get rid of this sample later. You just need something to give you a key etc.. to work around.
HINT - you should like this sample so much that you can listen to it for 7 hours straight without getting sick of it. If you can't do that, it's probably not the right one.

4) Once it's looping, audtion millions of your other samples over the top of it. If you stumble across something that you think might work, load it up and try tuning it. If you don't get any joy in a couple of minutes BIN IT. If it you can get it in tune do some surface fucking around with it. e.g. Try it an octave up or octave down / reverse it - that sort of thing, just to see what happens. You can do more complex stuff later. Don't bother right now.
If it more or less works carry on with the process until you have more stuff than you need all layered together.

5) Now you have loads of stuff that all more or less works together you can carry on with the track and start ditching the less good stuff.

6) Let the core sample (whatever it is you've ended up with) dicate the rhythm / bassline / feel of the tune. For this reason it's better to start with a sample of some sort, rather than beats etc...


The key to this is working very fast. If something's not happening straight away ditch it and move on. DO NOT spend hours tweaking something you're not 100% convinced about - and bear in mind that even when you're 100% convinced about something, you'll mostly be 75% wrong.
If you're overworking a sample because you don't have anything else that might work better, you need to stop writing the track and go and do more sampling.

^^ That, basically.

[oprs]

i agrees with what wub agrees on.

and thats the first time ive read it, but 95% of the time thats a guideline im following

[Dystinkt]

I usually do beats first, but lately i start with melodies and basslines and iv been a hell of a lot more productive in both quality and quantity of tunes made. I dporbably put it down to an increase in synthesis skills in terms of creating bass sounds, its the part i enjoy the most tbh.

[wub]

Slightly off topic, but love this;

[wayoftheworld]

interesting video...

[IC0N]

I usually start with a bassline or melody that I randomly think up. Build the drums off that. Build intro. Put all that together until I have the first half of the song solid. Then work I'll work on the second half (which is where I almost always get writers block). Then mixdown and master.

[meltusriddler]

Get high....see what comes out

[FSTZ]

I usually start of with a sample that inspires me and take it from there. Beats come first most of the time, then bass, then leads then everything else.

same here

people have told me I am crazy for building whole tunes around a 1 second movie sample LOL

[Sine69]

I don't have an exact process that I go through, but I typically work on the intro to a song first. Lately though I've been starting with the bassline

[ToxicBass]

interesting video...

What's every bodies thoughts on this method of writing?
I end up with tonnes of unfinished tracks and was wondering if anyone could vouch for this method.
I currently mix as I go working from writing the intro first to the outro last.

[skyh]

1)drums
2)bass or lead
3)additional melodies
4)repeat

sometimes I focus more on sound quality as opposed to arrangements or vice versa
I pretty much just make beats whenever I have a clear mind or when I feel inspired.

A whole sit down for me could be spent tweaking levels and EQing separate tracks within my DAW.

[snypadub]

Drums drums and more drums. spend loads of time getting percussion spot on then I build around that.
Just sort of flows, I like to go into a track with no real idea what it will sound like and let the sound find itself as I work through it.

[ambinate]

interesting video...

What's every bodies thoughts on this method of writing?
I end up with tonnes of unfinished tracks and was wondering if anyone could vouch for this method.
I currently mix as I go working from writing the intro first to the outro last.

that seems a little too strict for me. i definitely see his point and i think there's some validity to it, but i'd go fucking nuts if i couldn't pan instruments or apply some EQ or effects while working on the musical material in a song. leaving the major mixing 'til after the majority of the song has been written can definitely be helpful, but for me, at least, i think it's better to just address what needs to be addressed as it comes up. something like panning or cutting out certain frequencies of a synth can be crucial to getting a song to work on a musical level. when i'm working on a song, it's sort of a dynamic process, and i like to just work on what i think needs work at that moment, and i try to keep myself in check as best as i can. if i've spent a few hours trying to get a kick to sound perfect before i've gotten anything melodic down, i'll probably back up and start working on something else. but if a lead and a pad are conflicting while i'm working on the melodic stuff, i'll try to fix that then and there.

this post just started to ramble but i think the point is that, in my opinion, there's no perfect way to go about writing/producing a track - it's a little different for everyone. i'd try out a few different methods and just see what starts clicking for you.

[wub]

Drums drums and more drums. spend loads of time getting percussion spot on then I build around that.
Just sort of flows, I like to go into a track with no real idea what it will sound like and let the sound find itself as I work through it.

Won't the flow always be dictated by the drum programming that you've already laid down though?

[meltusriddler]

I agree with the whole thing of sketching out a track have been working this sorta way for about a year now and always try to think what a sound could be like ahead but actually leave the fine tuning till later. Alot of stuff can get changed or added like a fx and shit in places but having the whole thing being pretty much layed out arrangement wise before getting more into mixing i think definitely works.

[zerbaman]

Wub's method is mine. Well lately at any rate, I've been struggling to start tunes totally from scratch, always some sort of sample in there

[Toric]

Hey folks. Torbjørn here,

I wanted to add to this discussion. Who knows, maybe I'll inspire someone and help them with work flow. Either way, this is how I do it.

Setup

I use Logic Pro 9. I built myself a preset to open the program with right off the bat. The setup is; a drum kit, one of each synth you like to use, Sampler, a piano (this is a must, I'll explain why later), 2-7 aux channels set-up with Delay, Reverb, Multi output channels for Ultrabeat. It's so much easier to make 7 aux channels and have 5 of them already setup to Ultrabeat multi outputs. It' a lot of tedious work to be doing while you're making a song and it WILL get in the way of my creative work flow. Not too many aux channels on the preset, only the things I will use every time I make a song.

Brostep/Pop/Radio

Usually I start with the drums. I find that it's really easy to get a song to sounds good and radio friendly if I have it arranged in drum form beforehand. I label everything with markers, and use a different color for the drum regions. If I change a region to have something that's different from another region, I make a lighter or darker color, so I can easily see what drum beat is what. Secondary drum beats are usually on the opposite color spectrum for me, so I don't have to do any guess work. I also never use the looping function on drums. I copy paste seperate regions. The reason being, is so I can swap out drum parts quickly and effectively. No more chopping up looped parts and dealing with 4 new regions to delete. Once I've gotten the drums into a few sections and markers set-up and everything colored, I start working on the notes I've left myself (Markers will include things like Main melody. Main melody +1, Dubstep part. ETC.)

So basically the order of work I usually do is.

Drums/Markers/Colors
Melody
Bass
Drops
Samples
Stabs, cuts, warping effects
Intros/outros
EQ/Doubling/Layers
Mixing
Mastering
Drive around, turn volume to 100, pop in CD.

Here's an example of what my workspace looks like for the song "Servo Machine," and the song so you can compare.


Soundcloud

Now to explain the piano. First thing is, when you're sampling, you generally end up having to retune a sample to get it to play the correct pitch. This is the first, probably most obvious reason to always have a piano out. You can always hide it and then unhide it when you need it. I also think the piano is the best way to work on melody ideas when you're coming up with the core Idea. This includes basslines as well. It's so much easier to hear what the melody should sound like on a piano. It takes a little bit of tweaking to get it right, but it's well worth the extra effort.

Last thing I do always is make sure the song is catchy and listenable.

Experimental/Dirty/Wankstep

In short? Throw everything I just said out the window and get to it. Usually, I do start with a drum beat, for practical purposes, but I don't really do any arranging. For the most part I like to drag things on for WAY too long, while modifying the basslines for hours and hours. I usually cut that up (generally I bounce the regions in place so I know it will sound like it should, as automation tends to get sticky sometimes) and arrange it how it sounds best. Since usually I'll play the keyboard and mess with sliders at the same time, I generally have to spend some time looking for what sounds best for what I want to make, and throw out the rest. I know, throwing things out isn't really a good idea, but at this point it's such an experimental process that it doesn't matter, as whatever the sound is, it's by accident. If I find something particularly cool, I'll spend some time figuring out what I did at that point in time, and add that to my knowledge of stuff.

Make sure your knowledge of stuff is constantly growing
This is probably the best way to keep my workflow going. If I didn't learn anything new in making a song, I probably failed to create something exciting and original. Doesn't mean it's garbage, I'm just not satisfied with myself at that point.

End
I may edit or add to this in the future, and maybe even have Thrasher tack on his creative advice. I hope this helps someone. I know I'm not super specific here, so maybe I'll make a tutorial in the future, if requested.

-T