Sound Design Microtorial #3: Layered Wind

Designing wind can be tricky. And in my experience, different people mean different things when talking about wind. Some think of the hissing, high-pitched, noisy kind, others think of the rumbling, low-frequency kind that we hear when wind blows into our ears. Then there’s the howling, moaning, rustling, etc.

So first, let’s clear things up a little. Wind is a weather phenomenon. It’s the flow of air between high and low pressure areas. When looking at it from that angle, one could also say that wind IS sound, only of very low frequency.

Anyway, when analyzing the sound of wind, you can distinguish between three different phenomena:

(1) wind colliding directly with a membrane like a mic or eardrum, producing low frequency rumble.

(2) wind colliding with a fixed obstacle, creating twirls of air which produce broadband noise or even simple tones.

(3) wind colliding with movable objects that produce sounds themselves like rustling or rattling.

So when designing wind I usually take at least three layers: i.e. (1) a rumbling, (2) a mid-range noise and (3) some rustling or rattling. It’s a good idea to pick recordings that modulate in volume. Static sounds work too, but you probably have to apply some modulation yourself in order to make your wind sound interesting. Layer (3) is the most important one since it tells us about the materials and objects in the scene, and usually it’s the most noticeable.

Now, with those three layers you’re able to compose your wind. This is especially useful for “wind gust events”: Imagine a scene on an open field. For each wind gust you can start off with some rumbling, then blend into mid-range noise followed by rustling of grass. This will illustrate the movement of wind across the field quite adequately.

If your underlying audio files are long enough, you can simply duplicate such a composition and nudge the audio in the clips to different regions to create variants of the wind gust. This way you’re able to design a windy scene pretty quickly.

Be inspired by the objects found in the scene. Some may be of category (2) or (3) producing distinctive sounds. In a scrapyard scene for instance, layer (2) could be metallic moaning and layer (3) some rattling or soft banging.

I hope this was helpful. Have fun with your next windy scene!


Sound Design Microtorial #2: Explosions With A Crack

A common task of a sound designer is to create a fat sounding explosion. Wikipedia defines explosion as a “rapid increase in volume and release of energy in an extreme manner” (by that definition, gunshots are nothing different and the following applies for those, too).

There are many ingredients that make up a good explosion but today, I just want to talk about the very first part of an explosion: the initial bang. It’s quite hard to make it sound right and unfortunately, I don’t have a magic recipe. But there’s one thing that helps sometimes: giving it a crack.

Imagine you have a generic explosion sound. It should start with an extreme transient where the amplitude goes from 0% to 100% in a very short time. First, make that transient even more extreme by trimming the beginning of it. Cut to a high amplitude sample so that our explosion starts with 100% right away. Now it should sound something like this:

Explosion with trimmed beginning (tap here for iPhone/iPad version)

Now for the crack: chop off the first 5ms portion of the explosion and nudge it to the left by about 30ms (times may vary, use what suites you best). Make sure you chop during a maximum amplitude to get a maximum cracking effect.

Explosion with crack (tap here for iPhone/iPad version)

That’s it! Experiment with that a little and optimize it. It works best on explosions with strong low frequencies.

I came up with this by just fooling around, but I also have a theory about what happens here: by chopping up high amplitudes and creating a little silence in between, we basically simulate low-frequency distortion. Imagine a very strong, very low (like 60Hz) sine wave. Distort that very much and you get something like a 60Hz square wave. Now square waves are nothing more than no change of amplitude (silence) followed by extreme change of amplitude (slope). And one half of a 60Hz period is about 33ms. So we created a super-distorted half period of 60Hz in the example above.

So why does that sound fatter? My guess is this: we have an intuitive knowledge about distorted sounds. Distortion happens in every signal chain: in the air, in amps and speakers and even in our own ear. And usually it happens with strong signals. So that little bit of distortion we created implies a strong signal and makes us believe that the explosion is a big one.

That’s it for this time, hope you could benefit from it!

Sound Design Microtorial #1: Making a ringing cell phone sound realistic

Yesterday, I listened to the stuff I did for a scene of the movie I’m currently working on. In that scene, the main character pulls out her cell phone which was ringing. I made up a cute musical ringtone with Garage Band especially for her (I can’t recommend GB highly enough for those ‘I need a shred of music’ situations!).

But it still sounded crappy, dull and not believable. First I thought it was the ringtone itself, but then I realized what it was: it sounded static, not moved thru the air when she wielded it towards her ear! Back in film school, I had a similar problem once before and I remembered that I had found an easy solution:

Just apply a little flanger or delay to simulate the movement of the sound source. In reality, nearby surfaces produce echos with very short delay times. These times vary with the movement of the source, especially when it’s rotated. This of course applies to all moving sound sources.

So a Flanger or a time-modulated delay simulates exactly that. In fact, the perfect simulation is achieved by automating the ‘delay’ and ‘mix’ parameters of a mono delay. Delay time should be short (< 10ms). Feedback should be minimal, of course.

If you can’t bother to automate (I can’t) just make a flanged version and insert it into the clean clip during the heaviest movements, crossfading in and out.

Further reading suggestion: The famous concept of “Worldizing” is explained here.