Organic signal processing, or: How do I sound like Arca and Holly Herndon?
Hey y'all. I made a post over at /futurebeats
about the music that people like Arca and Holly Herndon make, check it out
if you have no clue what I'm talking about or
want to find out about who else makes similar music. And, even though this kind of music is often visually & conceptually interesting (artwork etc.) at the core of it there's just one thing: the sounds. And oh boy, the sounds are an array of different descriptive adjectives; like the real world and the digital realm fusing together into something that sounds like what a musical interpretation of Worlds.com
might sound like if done by a vaporwave-crazed field recording artist. That is: wonderful.
But how do they do it?
To be honest, I only have ideas. Clues. Things that might be helpful, that I know some artists use and that I think some artists use. Guidelines, color palettes. Nothing definitive
. But whenever you find a tidbit, or understand something about the music and the inner workings of it, don't hesitate to share it. Maybe we'll figure something
out. Okay, let's try.
Granular synthesis Maybe the most recognizable feature of this kind of music, granular synthesis is a method where audio is split into minuscule pieces (1-50 ms) and these pieces are then rearranged in what's called a cloud. Think of it as sampling, because that's what it essentially is, just on a very different scale than what you're used to. Given the right tools, this is actually a pretty easy way to create adequately interesting soundscapes. However, the rabbit hole is deep and granular synthesis does have quite a lot to offer. Here are some tools for granular synthesis:
The Mangle (PAID; 29.99 GBP)
Max (PAID; 399USD or 9.99USD / month)
A stand-alone visual programming language for audio (and other multimedia) manipulation. Holly Herndon recommends tutorials by Sam Tarakajian for learning, I asked (and she answered! now that was a fanboy moment!). The tool used by people who do more complicated digital signal processing.
Max for Live (PAID; inside Ableton Live 9 Suite which costs 599 EUR)
The Ableton version of Max. Robert Henke's free Granulator II is a great way to get familiar with the basics of granular synthesis and to make some wicked sounds, too.
Pure Data (FREE)
A visual programming language; an open-source Max alternative. The learning curve is kind of steep (I've tried), but I have seen some great things that have used Pure Data.
This one's more of a toy, but I've recorded some cool samples using this. You draw lines on an audio sample and magic happens.
More granular synthesis tools can be found at granularsynthesis.com's admittedly awe-inspiring list.
Spectral manipulation & synthesis This kind of music is a lot about audio manipulation as opposed to sound synthesis. Spectral manipulation is one way to do that. It's based on manipulating the frequencies of the sound to carve out new sounds and for example automated pitch correction plugins such as Melodyne are based on spectral manipulation.
Tools for spectral manipulation
GRM Tools (PAID; 600 EUR)
Info in English: Sound on Sound - INA GRM Tools 3.
iZotope Iris 2 (PAID; 149 USD)
Arca uses this (he's got a testimonial here) and based on what little I've seen of this at work, it's a formidable machine.
LinPlug Spectral (PAID; 149 USD (the site is kind of bad, but I think this is the correct price)) SoundMagic Spectral (FREE, MAC ONLY)
General production tips Automate, automate, automate
Seriously. A lot of this music sounds so "organic" and "moving" because it is moving. Automate filters, volume, cutoff, delay time, reverb length, sends, all that. You don't necessarily even have to draw those envelopes yourself; put that processing power of your computer to work. Here's a tidbit from M.E.S.H.:
Another thing I did on this record: with Ableton Live and with Max For Live there's lots of great ways to control the parameters using different types of LFOs and randomisers and envelope followers and stuff like that. There's a couple of tracks where I had a VST with like 30 parameters to determine what kind of sound I was making. I'd make all these LFOs that were sort of indeterminate, with a lot of jitter to them. And then I would assign that to different parameters on this preset, so when I start playing each parameter is slowly degrading in a certain direction. So the sound is sort of wobbling over time. When I'm working like that I just have to record everything. Sometimes the synthesiser will go in some really interesting direction and you'll have to record it really fast. Try separating sound design & sample-gathering from the actual music-making
M.E.S.H. to Resident Advisor in Machine Love: M.E.S.H.
This is another one from M.E.S.H... To be honest, I'm really bad in doing this kind of stuff, but it's probably worth a try:
For this album I really wanted to do something that felt like a single piece, even though it's divided into different tracks. So before I even started writing I spent a long time making sounds and collecting samples, messing about with things and saving presets. I made my own huge preset library with all the effects bundled together, just so everything's ready to go. Because I really, really hate starting with a blank slate. Of course, this doesn't have to be so binary (and probably won't be even if you tried), but just to prevent the aforementioned blank slate it might be good to think about this anyway.
M.E.S.H. to Resident Advisor in Machine Love: M.E.S.H.
[How I arrange and edit my tracks] varies of course, but often starts with a particular process, whether that be a vocal patch, or a percussion system, etc. Once there is an ember it just grows. I like to establish the vocal treatment and overall palette early on, but will re-write the primary rhythm section multiple times – like I often start a track at a different BPM with an entirely different drum pattern and swing – and that could change dramatically. Like: “...what would this sound like as a jungle track? Oh shit, that actually works way better than the swung hip hop beat I was using.” Sampling is crucial
Holly Herndon to Ableton in Holly Herndon: New Ways to Love
Sampling is one of the best ways to make sure that your music doesn't sound like a sterile DAW grid. Despite this kind of music being pretty high fidelity in a sense, it still avoids the pitfalls of sounding too clean by sampling anything and everything. YouTube clips, your voice, your cat's voice, field recordings, hip hop sample packs, whatever you get your hands on.
The voices on “Chorus” are from several different sources – online browsing, synthesis, my own processed voice, samples, etc. all recorded through my laptop. Resampling
Holly Herndon to Ableton in Input/Output: Holly Herndon
This is one of my favourite techniques and I'd assume that many of the people making this kinda music are utilising it as well. The gist of it is this: You take a synth sound, or really just any audio clip and you put into a sampler (e.g. Ableton's Sampler) and then you use that small bit of sound as a basis for a new synth. Often the "erratic" nature of the resulting synth makes it sound pretty unique. Since if you had, for example, a tempo-synced LFO modulating the VST you're using for the source sound, when resampling the LFO wouldn't be temposynced anymore. Instead, it would have a different speed depending on what note you're playing. So experiment with this!
Off the grid
As a counterpoint to the digital nature of this music you might want to move things out of the grid every once in a while. If you haven't got much latency and can kind of keep the rhythm, you can just try playing the parts yourself to give them a human touch. Otherwise, you can just manually move things so that they're not exactly aligned to the grid - this makes the music feel more varied.
Great VST's for this kind of stuff [does anybody know of any?]
Hopefully some of this helps you guys when you wanna try making android lullabies for your robot sweethearts or whatsoever. And, really, if you have any ideas on how I could expand this list (e.g. VSTs) or any tips of your own, do share! :)
And an itty bitty self-promotion: my Soundcloud is Taui, which unfortunately doesn't contain very much of this kind of music (...yet?) but is kind of dope in its own right anyway.
submitted by fiys