Whats that coming over the hill?

Is it a monster? Or a cow with a ring modulator on?

So how exactly do you make the sound of something that doesn’t exist? How exactly do you record the sound of an intergalactic squid monster. Well, a squid monster has at least a basis in reality, but what about a squerch monster? What the hell does that sound like and how do you get the sound of one anyway?

Take the creature below as an example.

Creature 1 with no sound

This is obviously not a creature we will find in reality, so we can’t get near to one and record it. In fact, looking sat it, I wouldn’t want to go near one and especially not if I was carrying heavy recording equipment.

So we have to imagine what the creature might sound like. So we need to think about it for a bit. Is it friendly or unfriendly? Big or small? Loud or quiet? Aggressive or defensive? Intelligent or not?

To create this sound, I imagined that this was a big, unfriendly aggressive animal that is no more intelligent than a dog. So this gave me an idea as to the type of sound I could give this creature. As its an ‘animal’, it doesn’t need a language, so I don’t need to worry about words, or the creature giving instructions or describing something – its all about ‘noise’. This is why we need to be able to assess the intelligence of the thing we are making the sounds for. If something needs to speak in order to communicate something, then it needs to sound like words, or at least, the shape of the sound needs to follow what we know and understand of speech. For example, a question will rise in pitch at the end.

Looking at this creature though and the shape and size of it’s mouth, I think its fair to say it needs to be a ‘roar’ kind of sound. However, in reality, most animals that roar aren’t lizard like; and this animal is. The nearest animal that exists is probably a crocodile or alligator, and whilst they do make a very satisfying noise, it doesn’t fit this animal.

This is the final result. Read on and I’ll show you how I got to this.

Creature 1 with sound

I wanted something ‘shouty’ and ‘brash’ so started with an elephant. And yes, I know an elephant is a mammal. I know this creature looks nothing like an elephant, but bare with me on this.

This came across as a bit high pitched on its own and we know that big things sound lower in pitch than small things. Yes, I know elephants are big. I know, I’ve just said big things make low, deep noises and that the elephant’s sound is too high pitched…….. these ‘rules’ are generalisations. You can always find the exception to the rule and this particular elephant trumpet sound I’m using is that exception. There are probably better elephant sounds available and probably ones with a deeper pitch, but this is the one I have that has the brash edge I am wanting. In any case, I haven’t finished with the sound yet.

It needed more weight to the sound. So, I added a cow, but I used a pitch-shifter to lower the pitch a little bit.

Then I made the cow sound backwards and lowered it in pitch even further.

Then I added a walrus. They sound big and scary. Unlike the cow, there is a sharp attack to the walrus sound. It sounds like a shout or a bark.

And finally, so that the background animals had a sound as well as the voice of the one in the foreground, I added a stampede of wildebeest.

Listen to the finished sound again in the video above. Can you hear the separate sounds now you know what to listen for?

This is how it works. When we can’t record something as it is, we make it up by blending sounds together of things that do exist to create something new. You can clearly hear the different sounds in this now you know what to listen out for, but when put together, it sounds like one creature roaring away. Big, brash and quite scary.

These sound samples came from Sound Ideas XV.

There’s a procedure for everything

Imagine being on a rollercoaster. Every time you ride the rollercoaster, the experience is the same. It has been designed to provide a captive audience with a carefully planned and designed experience. This is an example of ‘ordinary media’, like films or TV. The director plans everything out for you and each time you watch it, you experience the same things in the same order along the same timeline.

Interactive media is different and so cannot be planned out by a director in the same way. Video games can loosely fall into two categories. Semi-linear and completely free. Semi-linear games are analogous to a maze; where the player makes choices at pre-determined junctions and is therefore relatively controlled by the game. The player has a limited amount of control and choice. Many adventure games, or narrative driven games fall into this category in broad terms. Each level begins and ends at the same place with the same cutscene, no matter how you choose to play the game. A completely free game, or ‘sandbox’ game, like the GTA series, offer different challenges to the sound designer. You can’t have a soundtrack for a level, or a prescribed length of time for a piece of music to play – the player may well not cooperate with your direction.

One answer to this issue is procedural sound design.

Imagine there are four Ace playing cards face down in front of you. You have to guess which suite the card belongs to before it’s turned over. After 6 or 7 goes, you are going to know the order of the cards. That game will become very boring very quickly. Now imagine you add a King, Queen and Jack into the game. Not only do you have to name the suite of the cards, but the order in which they come AND the order I mixed up at random. The guessing game now becomes impossible. There are a limited number of possibilities, but the number is so great that there is no discernible pattern. This is what we mean by procedural sound design.

Instead of playing cards, imagine instead we’ve chopped up five sounds of explosions falling into the sections of ‘crack’ ‘boom’ ‘tail’ and ‘thump’. If we stitch these back together in a completely random order, perhaps even randomising some filters, modulation and volume, then from the listeners point of view, we have an endless cycle of explosions without two full explosions ever repeating themselves. This has the benefit of sounding natural, yet we’ve managed to do it only a small amount of memory and space being used. To quote the Genie in Aladdin “phenomenal cosmic power in an iddy-biddy little space” (Disney 1992).

Here’s an example of it:

According to the Oxford Handbook of Interactive Audio,

In the visual and tangible realms, the apprehension of conditions and process is required, but it is usually the reĀ­sults of the process, not the process itself, that are of interest. (Andy Farnell, Edited by Karen Collins, Bill Kapralos, and Holly Tessler, 2014)

So, basically, how it works is of little interest to most people, but the results are important. When looking at a game that doesn’t use procedural sound design, the effect is obvious.

It can be argued that in a game like Minecraft, a more realistic or natural sound design policy wouldn’t fit with the aesthetics of the game, the point still stands that the same explosion sounds repeated does not sound realistic at all.

However, in games where procedural sound design does operate, the effect is a very natural result.

Gunfire, footsteps, shouts, explosions and most sounds feel natural and don’t instantly repeat on themselves.

In practice, however, it’s not necessarily easy to get right. If one is chopping up a sound that would normally be continuous, there are the issues around stitching sounds together that need to have similar enough timbres that they don’t jar on the ear. This is my first attempt and it suffers from not sounding properly ‘joined up’. This is obviously a skill tooth’s I need to learn.

So after a lot more work and (only a few) tears, I think I’m happy with this version. It feels more natural to me.