Audio Production Tools

The-Last-Of-Us-quality dialogue with a click of a button (vs months of manual labor). Streamlined audio assets review so designers could focus on the art.

The Last Of Us I and II had some of the most well-tuned dialogues. They were always clear and legible regardless of the surroundings, and the team left a satisfying amount of headroom to allow for different emotions to shine through within the characters’ performances.

It wasn’t until this 2022 GDC talk that everything clicked for me. The team at Naughty Dog had painstakingly massaged and leveled each audio stem by hand. More importantly, they were kind enough to share in detail the tools they developed to aid the process, which drastically minimized their workload.

It was such a rare glimpse into the kitchen that I immediately thought every studio would benefit from having such tools. If the same care were given to every performance in games, the dialogues (and by extension, the stories) would have had a much better chance to carry their intended weight. So I did the only logical thing—made them mine.

Tools:
Reascript, Lua, Python, ImGui.

Valuable resources:
The community sourced plug-in library Reapack
A better API documentation by Reascript guru X-Raym

1. Get all samples in each stem, divided by Cycle if using Peak mode, or by Window if using RMS mode.
2. Create envelope points at peak sample within each returned Cycle.
3. Adjust that envelope to the threshold value while applying slope between envelopes according to the Tension value.

1. Get all regions and return all markers within.
2. Add markers to Queue, select wait time between plays and how many times to play each marker.
3. All stems within markers will be randomly played back.