Patron Request - The Creative Process (Patreon)
Published:
2014-10-01 15:04:40
Imported:
2026-06
Content
A patron recently asked if I could explain a bit about my process for creating ambiences. I thought others might be interested as well so here it is. I’ve tried not to get too audio-geeky but if you need clarification on something just let me know. Alternatively, if you want me to get more geeky...then really it should be a conversation in person, with beer!
--
I like thinking about each ambience as a scene from a film. One idea and one location. The track should exist as a backdrop for the actions of the characters, it shouldn’t be the star of the show. If it is primarily musical, it shouldn’t be too distracting, it’s purpose is to serve and embellish the action, or reinforce the mood in the scene.
Once I've determined a location, scene, or mood for the ambience, I either pull from field recordings that I've made or heavily edit and layer various stock recordings in my DAW (digital audio workstation). If the location is a fictitious one or say, space, I’ll often start by creating a synthesizer or sampler patch. Or, a combination of all of the above. A starship background, for instance, might be a synthesized sub-bass with an LFO modifying the oscillator frequency layered on top of a recording of an industrial air-conditioning unit that has been pitched down 3 octaves. I make heavy use of pitching as it can change the character of everyday sounds in surprising ways.
Creating the background track is the foundation of any ambience, whether it’s a forest scene, an urban scene or a space scene. It’s the bedrock, like the rhythm section of a rock song. I’ll spend a long time making sure that it’s interesting, but not too interesting. If I’m going to repeat one of the background elements I check for signature sounds that will give away the fact that it’s being used again. If you hear the same bird tweet from the same location every 60 seconds, you’re going to pick up on it eventually and it will become annoying over 10 minutes. I’ll chop out any offending bits and stitch together the audio using crossfades, or by layering other tracks over the edit points to mask the splits. Finally I’ll automate volume and EQ envelopes over the whole thing to give a sense of change and motion over the course of a background. Using this technique I can layer a handful of tracks and clips and have it seem like 10 minutes of continuous evolving sound. Even if I’m using a full 10 minute clip from my field recordings I’ll still edit it pretty heavily. I do the same with stock audio. It almost never goes in as-is.
Once a basic background has been established I’ll select a convolution reverb, which is reverb based on impulse recordings of actual places or equipment. I try to match this to whatever the location of the background is. So if I’m doing outdoor ambiences, I’ll select an outdoor reverb type. I do this so that all the elements I add on top of the background will get routed, at some level, through this reverb and it will seem like they were all recorded in the same place. For instance, If I’m recording outside, and I want to add a branch breaking, I’ll route the branch clip to the reverb bus and adjust the reverb level (wet/dry mix) and pan so it will sound like it was breaking in a forest. I could use the same branch clip and route it through a different reverb impulse and it would work for a branch breaking in a cathedral, or an oil tank or inside a piano.
After the background and reverb are set up then I’ll start adding the incidental elements. These are subtle pieces of sound that will help further create a sense of place or action. Similar elements will get their own track or groups of tracks. Footsteps, for instance, will often get their own track group. Walla (unintelligible background vocals) will get another group. I’ll have left, right and center tracks in each group and vary the volume and eq for each specific element to push or pull their presence in the stereo field. Other elements will share tracks - I’ll have a track or two for miscellaneous effects. I’ll often use the same effect a few times but I’ll pitch it up or down a few cents and play with the panning and eq to make it appear to be different sounds.
Sometimes I’ll want to interact with the track in a more immediate way, rather than dragging individual clips into the timeline. For this I’ll build a sampler patch with lots of different incidental sounds attached to the individual keys of a midi keyboard. Then I’ll “play” the incidental sounds in real-time as the track records the midi data.
Occasionally I’ll listen to a finished ambience and decide that a slight musical bed will help to gel the track and add to the mood and emotion of the scene.
A typical project is between 15 and 30 tracks. I recently did some custom tracks for the Fragged Empire RPG which were 40 tracks.
Once everything is sounding good, levels are right, nothing leaps out as being incorrect or just wrong, I’ll render everything out to a 24bit/44.1khz uncompressed .wav file. This is then converted to 16 bit .wav in preparation for the .mp3. If needed, I’ll do some extra eq/compression and limiting at this stage, as well as a final dithering of the audio to compensate for the bit-depth change. Once I’ve got the final 16 bit .wav, I’ll convert the track to .mp3 for the web. I’m actually pretty happy with the quality of the mp3s that come from this process. Bandwidth costs being what they are I find that it’s an acceptable middle ground.
That’s basically it. Each track takes 2-4 days (some more) before I feel like it’s ready. And yes, sometimes I listen to things a month later and wish I’d spent more time on them! And that’s it, the creative process in a nutshell.