Sampling

As much as I like to use and write about synthesis, I must admit simply sampling sounds is highly effective and in many ways more efficient. In my daily work as a film sound designer, I use sample libraries and samples of sounds I’ve previously made almost as often as my synthesizers to create new sounds I need (though I do tend to process said samples with synthesizers, so categorise that as you will). But, I wanted to delve into the many, MANY ways you can use sampling, and maybe give some tips on recording the right sound for the job at hand.

Sample-Based Synthesis

First, I want to briefly touch on sample-based synthesis. This is a synthesis method which employs samples as the oscillator(s). This means you get all the filter, modulation, effects, etc you get with a synth, but with a greatly expanded soundset from wavetables or other synthesis methods. A few synths that have used this to great effect are the Korg M1 and wavestation, the Roland D-50, and my personal favourite, the Yamaha SY77, which can use up to 4 layers of 2 FM voices and 2 sample voices at once. Very cool.

You also have straight ROMplers, like various E-mu instruments and the venerable MPCs, which are basically just sample players but with keyboards so you can re-pitch the sample. Some of them also go into multi-samples, which is standard for instrument libraries today, where each note or set of notes has its own sample. Modern library tools like you see for orchestral work utilise this extensively, playing multiple notes the exact same way to capture variations in the sound, as well as sampling different velocities and techniques to create a truly playable instrument. Very cool.

Here in the sound design world, we mostly more use the first category, where we take samples, often from libraries but also self-recorded, and manipulate them in various fun ways. Stretch them, pitch them, apply various effects, all par for the course when designing a sound. You can even completely transform a sound if you wanted, but typically the magic comes from layering along with manipulation. A sample-based synthesizer, or sampler with a long list of VST effects processing it, are typically what you’ll see in a sound designer’s studio.

The Process

I should cover this in more detail, I think. The typical process for creating a custom sound from samples starts with choosing the right sound for a given end result. If you’re making a creature sound, you’ll probably start with samples of real world creatures rather than, say, a door being slammed shut (though depending on the end result that might be warranted). Say you want to make a low passive grunt of this creature you’re designing for. You’d probably start with existing grunt sounds, like a cow or walrus:

Already pretty close with either of those two, and that’s just raw field recordings! Now, we’ll throw that sample into Phaseplant’s sample module. I use Phaseplant for a lot of my work with layering and stuff.

Lot going on now! So I have the walrus sample going through a lowpass to get only the bass part of it, and I’m mixing that with the cow sample (layers r ur frens). That’s being slightly overdriven, then run through a modulated formant filter going from i to the point you see the dot at. This will give the grunt/growl a little more of a iiiooouuu sound. Then it’s EQ’d to boost the bass cuz bass is good, then phase distorted with a saw wave. This adds a tiny bit of grime to the sound (you may recall from previous posts that phase distortion is effectively remapping the phase of an input signal by another signal, in this case the saw, which is at the same frequency of the samples so we’re just adding the tiniest non-linearity to the sound). We then run that through a tape stop, which will handle the pitch shifting of the sample up and down as the sound progresses. Convolution for giving the sound some space, wide highpass to reduce the lows from the reverb, and a limiter to make sure we don’t clip. End result? This:

Different notes played will give you different results, both in timbre and pitch, giving you lots of variations very quickly. You can also adjust modulations, effects, and more to get the desired results. Some of those sound kinda weird, what with the tape stop effect and the loud noise at the end of each, which you can further edit out in your DAW or other sample editor, along with adding fades, reversing, etc. Point is, you can do cool stuff to cool noises to get cool results.

Hwo 2 record gud

How you get the initial cool sounds varies. You can do what I do and just use your phone with a Rode VideoMic on it, or use a Zoom H1N, or you can do what other sound designers I know do and invest entirely too much money into a million microphones and accessories and set up dedicated recording sessions. I spent all my money on synths, so I go the cheap route. Good results can be had either way (the above raw sounds were recorded with a phone).

Being a field recordist is a bit like being a photographer (which I also dabble in). You need to frame your subject (point your mic at it), it has to be in focus (minimal other noises), your subject needs to have the right exposure (basically using a proper mic and not something with highly limited range), you need to be there at the right time, you need to be ready to record any sound anywhere, and you need to get it right the first time (or take a lot of shots because your subject may move and sound different each time you record it). As with photography and many other skills, all you can do is practise, learn from your mistakes, and have fun with it!

It helps to have a goal in mind when you go out to record. If you’re aiming for creature sounds, that’s going to be a very different technique and location compared to ambience. It’s ok to get a little of both as you go out, of course. A macro photographer will inevitably get a few landscape, animal, or architectural shots in, as will a trip to the zoo to record animals result in nice walla, footsteps, and other general noises. But if you’re going to be recording animals in a zoo, you’re going to want a shotgun mic, possibly on a boom if allowed, to stay away from the animals but still capture them clearly. This is like using a telephoto lens to shoot wildlife pictures. If you’re capturing ambience, you’ll want more omnidirectional mics, possibly even employing stereo or XY mics. This is like using a nifty fifty lens to capture the world around you. There’s also mics for EMI, underwater, ground tremors, and more, all of which are worth exploring – much like a photographer having multiple lenses in their bag.

One thing I do want to mention is the proximity effect. This is how distance from a sound affects it. The further you are from the source of the sound, the more of the environment you will capture, be it in the form of reverb, ambient sounds, delay, etc. What this means is you can use proximity creatively to create variations of a sound. This is where multiple mics helps, if you can manage it. Repeating a sound will invariably not be identical to the previous sound, so you can’t effectively use proximity to your advantage with only 1 mic.

Legality of Sampling

We’ve likely all heard of how a musician sampled another song in their song and had to buy rights to do that. Albums have been delayed due to this, artists have been sued over this, and frankly it’s a problem for you as well, as you go about your day recording whatever takes your fancy.

You need the permission of the person who’s work your sampling. This can mean using library samples, recording a sound someone else made including things like a car’s beeps, or recording people, either individually or as a group. Your local laws may have more or less strict endorsements to this concept.

Now, you may ask, if we’re out here re-pitching, stretching, warping, effecting, re-synthesizing, and otherwise mangling samples to suit our needs, why does any of this matter? In my experience, it doesn’t. At least, in the 16 or so years I’ve been doing sound design professionally, no one has ever asked for sources of samples, or sounds in general (except on discord but that’s less important). That said, check your sources, and ensure legal rights to use sounds, especially in paid work.

Libraries

I realise most of this article has been about recording your own sounds and mangling them, but a lot of sound designers get by entirely using library sounds. These are sample packs, usually paid, usually organised in some fashion like UCS, which are designed either for building sounds with or using the sounds as-is. Most sound designers will use libraries as placeholders, but sometimes those placeholders end up working so they stay in. You can hear an example of this in the jotnar shrine opening sound in God of War 2018 and the CoD Zombies mystery box opening sound: they both use a metal latch sample likely from boom mechanicals (shoutout to PapiPallanes on the sound design server I moderate for that find).

Most sample library users will make use of layering tools like S-Layer and sound combination tools like Kontakt to get the sounds and variations they need. Many sample library creators will also offer such tools for sale alongside their libraries. It’s definitely the better, if more expensive, option for extremely fast turnaround, as is required for TV, YouTube, and other fast-paced media.

Konkloojun

In conclusion, sampling is dope and you should definitely make use of it. Get yourself a decent mic and start recording the world around you. Grab some libraries and start combining sounds to create your own sounds. Or, y’know, just stick, with good old synthesis. Your call 🙂