Issue link: https://digital.copcomm.com/i/1546054
M OT I O N P I CTU R E S O U N D E D I TO R S 33 "The struggle itself towards the heights is enough to fill a man's heart. One must imagine Sisyphus happy." —A L B E R T C A M U S because it's often straightforward. States to adjust the mix between things like Exploration/Combat/ BossBattle, Gameplay/PauseMenu, Healthy/LowHealth, and AboveWater/ Underwater can all exist, overlap, and sometimes states can even be nested. Keep simplicity in mind as you develop states—you'll thank yourself later. Dynamic Mixing Dynamic systems run more-or-less automatically depending on game factors, including audio content. An adage about "enough rope" comes to mind. My advice: be choosy about what you implement, and build/test/tune each system individually before stacking them together. Standard Ducking Simple, heavy-handed sidechain ducking can quickly introduce "pumping" artifacts which make a mix sound unnatural. For full-signal ducking like this, I suggest combining it with mix states where possible. For example, for one game I worked on, there were lots of moments where narratively important dialogue would happen. For all of these, I had a mix state that would fire 0.3 seconds before the first word of that scene, and the mix would transition over that same amount of time. That mix state pushed most sounds (other than dialogue) down -3dB, and then I let the dialogue duck things another -3dB. In total, we were carving out 6dB for the dialogue, but the "pumping" was limited to a virtually imperceptible 3dB. Sidenote: For a future title, I expanded this idea to become a "Narrative Dialogue Enhancement" accessibility feature. The amount of change the state-based mix pushed was, by default, still -3dB, but that value was user adjustable down to -96dB. This allowed people with specific auditory needs to carve out as much or as little background as they'd like so they could hear the dialogue comfortably. Multiband Ducking Instead of lowering the overall volume of things when ducking, it is possible to set up multiband ducking. This has the nice advantage of pushing/ pulling sounds out of each other's way in a way that is far less prone to create noticeable artifacts. You can tweak the response for each band and really dial it in neatly. I've eked out a system for this in Wwise in the past using a cumbersome setup of cascading busses and aux sends (it was clunky, but it worked). Now with Wwise 2025's expanded sidechain features and parametric EQ dynamics, it's much easier to do. As such, I expect we'll see more people using multiband ducking techniques going forward. Bus Balance: A Sleek Alternative to HDR Several times I've evaluated HDR to determine whether to use it on a project. Each time, I've reluctantly decided not to. The promise of HDR is very attractive, but the heavy lift of tailoring and maintaining an entire project to serve the HDR system is not. In the piece I wrote for Wavelength last year, I suggested that upward compression might be an alternative to full-on HDR. Since then, I've refined and expanded the notion, developing it to a more complete system I call Bus Balance. The Bus Balance system uses explicit volume targets for each bus, combined with upward and downward dynamic adjustments beyond that range. That is, the system can apply program-dependent downward compression on a bus to ensure that when multiple sounds in that bus build up simultaneously, the bus gently compresses down toward the designated target range. Conversely, upward compression can gently elevate subtle elements when there's room for them (of course, the moment a prominent asset appears, the upward gain transparently backs off). In Wwise, this can be achieved by putting a meter on a bus and having that generate an RTPC that adjusts the Bus Output Volume on that same bus. I like to set my RTPC range from -96 to 0, as that lets me draw my curve relative to the full practical dynamic range of the bus, in dB. I'll start with the curve horizontal—0 all the way across—then make breakpoints at the designated range (Note: This may be different than your volume targets for sound design—those are typically in LUFS, but Wwise meter RTPCs are based on either Peak or RMS values). Here again, it might take some trial/ error to dial in the right values for your project. With the breakpoints in place around the target range, I can push the far left (lowest) value up and the far right (highest) value down, creating the boost below the target range and some resistance above it. I suggest something gentle in both cases, maybe ~2:1 to start, then adjust values/curves to taste. It's worth noting that, in addition to parallel application across different sound categories, these can also