Using FMOD EQ to preview mix changes
Whilst working on Silt I had very specific ideas of how I wanted the game to sound but one of the difficulties in game audio is that by the point you are making final mix decisions during playtesting you are dealing with sound that has already been implemented. This presents a major issue. Your DAW where you generally make EQ/compression changes to the source audio is decoupled from what you're hearing; you're unable to hear changes in realtime.
To get around this issue I sometimes use a workflow that allowed me to mix in chunks using middleware and then apply those changes in the associated DAW projects.
The core idea here is very simple: you do 'mix sessions' in your middleware while the game is running and then at the end of the session, you bake your changes back into DAW projects and re-render the associated sounds. (For the purposes of this article I'll be talking about FMOD but I assume similar approaches can be adopted in WWise.)
What consitutes a mix session?
In Silt's case, a mix session would correspond with working on a defined set of associated scenes where the audio flowed from one scene to the next and where I felt that coherence was most important, both in terms of the and in the mood I was hoping to achieve narratively. So for example, the four to five minutes of game play prior to and including a major in game event would be a 'session'.
Making & baking mix changes
While you're mixing:
- define what scenes or parts of the game you're wanting to mix
- connect your middleware to the game so you can mix live whilst playing the game
- add EQ's (and in some cases compressors) to the channels or events that you need in your middleware
- tag which events you've added EQ to so you can find them again (more on this below)
- mix until you're getting what you want
Baking your mix changes:
- when you're happy with your mix you have to then apply the new EQ etc. to your audio files
- go through your new middleware EQ's and duplicate their settings to the associated DAW projects or audio files for that sound (more on this below)
- remove the new EQ's from the middleware (use the tags you set to find all instances of them)
- rerender the affected sounds so the changes are baked into the wav files
📷 The Crave EQ VST showing the same settings as the FMOD image above. Note the visual difference between the two due to the plugins different vertical scale.
In some cases it might be useful to mute any events/channels/busses that are not what you're working on, but for the most part you want to be hearing everything so you can tell how it's all interacting in the mix.
Losing sync with DAW projects
There have been instance in the final weeks of a project where I've gone from making changes in a sound's source DAW project to hard baking the changes directly into the audio file in my audio editor. The reason for doing this is simply time-constraints, especially with looping sounds where I often have to check or edit the loop after rendering from a DAW - just baking the change is faster.
I would only recommend doing this if you're really up against it timewise and most importantly, you're working on your own or in a pipeline that can accomodate these changes without throwing source DAW projects out of sync.
Tagging Modified Events
FMOD allows sound events to be tagged. When I'm in this mixing workflow I tag any events that have an EQ on them that needs to be baked so it's easy to get a complete list. I obviously then remove the tag once those EQ settings are baked into the source.
I also tag events in which sounds have had additional processing baked into them outside of the source DAW project e.g. where the 1:1 linkage with Reaper has been broken.
I'm sure I'm not the first to apply these types of methods and although they are a little clunky, the final mix pass where small 1 or 2 dB adjustments get made can have a dramatic culumative effect on the quality of the end result so it's worth the extra effort. Hopefully this approach proves useful or inspires positive changes to your mixing and listening pipeline.
TLDR: Hear something you don't like. Add an EQ in FMOD and tweak it until you do like it. Copy the changes From the FMOD EQ back to the audio file (if using an audio editor like Sound Forge) or source project (if using a DAW) and re-render it.