Hello, Groundbreakers! My name is Will O’Neill. I’m credited as ‘Narrative’ on Techtonica, and that is more precise than you might think.

Unlike a larger (AA / AAA / AAA / AAAA / AAAAA?) studio with a dedicated team of writers in addition to specialists in story structure, quest design, casting, worldbuilding, voiceover production and beyond, my responsibility at Fire Hose Games is to wear all of those hats while collaborating closely with other many-hatted individuals. This kind of role isn’t for everyone, but I like it.

Best of all is that it affords me the credibility to walk you through one of the most exciting processes we go through: Adding a new character to the game!

Auditioning a New Voiceover Talent

A slice of the briefing provided to performers auditioning for MIRAGE

To ensure that we recruit the right performer, our process begins by utilizing an online platform built exclusively for connecting talent with a wide variety of voiceover jobs. As a small indie studio, it isn’t realistic for us to hire celebrities, hold live auditions, or spend months making decisions, so doing everything efficiently and online is crucial. 

For performers, auditioning for a new role is fairly straightforward – we ask them to record a brief monologue from the script for their character. However, this audition script is often a slightly altered version of what the actual lines will be, as we are also using the audition process to ascertain how a performer approaches specific challenges related to the role.

For example, because our new character (MIRAGE) for 0.5 was French, we inserted several extra lines into the audition that required them to utilize a French accent while speaking both French and English, as well as make use of several English expressions where it would sound natural for the character to ‘drop’ the accent somewhat. We’ll also move bits and pieces of a script around to get a feel for how performers transition between different tones and emotions that we know they’ll need for future character developments.

Once all of the auditions are received, we typically compose a shortlist of around five performers to discuss internally before making our final decision. In the case of a character like MIRAGE, part of that evaluation also involves consulting with a native French speaker to evaluate the authenticity of the accent.

One additional thing I will say about casting is that it is extremely competitive not only from a performance standpoint but from a baseline equipment perspective as well. Because we record our sessions remotely, the performer we hire must have access to professional gear – this often includes a high-quality microphone, preamp, isolation booth, and the software and hardware that puts it all together. 

The Recording Session

Excerpt from the September 2022 recording session script for SPARKS

Once a new performer has been selected and the script finalized, we are ready to begin! Typically, we record in sessions between one and two hours, as we are only doing incremental additions and revisions to our narrative throughout Early Access. Another nice thing about Techtonica is that it doesn’t require our performers to repeatedly undertake melodramatic reactions to various forms of violence, so our sessions are pretty light on vocal fatigue.

As a voiceover director, my background is more in corporate voiceover stuff, so I probably wasn’t raised right as a voiceover director for games. That said, I’d say my general principles are to always take things all the way through on the first go, trust the performers to find interesting possibilities that I hadn’t considered, and to be thankful that we are living in a digital age where it is easy to record as many options for a given line or scene as we want. There used to be people with razor blades and glue and gigantic reels of tape! I’m sure there still are.

I’ll say as well that performers also do a ton to prepare their minds, bodies and voices for a recording session, and that I appreciate it.

Selection and Implementation

Raw waveform from a section of the October 2023 Recording Session with PALADIN

A few hours after recording, the performers will send us a single file that contains the entire session. From here, we go straight into editing – splitting things up as needed, choosing the best takes, splicing them all together, and then implementing them into the game itself. 

Making careful selections and determining which individual reading of a line is the best one is, in my opinion, the most important part of this entire process. I think being both the writer and director makes a big difference here, as I have a seamless understanding of what a line sounded like in my mind as I wrote it, where to go looking for it while directing a performer, and how to bring those things together in an intangible kind of synthesis that makes my ears perk up at one tiny but meaningful variation of a line read over another. I know which takes to choose because I know what it should be, if that makes sense. Anyway, my point is that AI is trash and cannot do this.

Finally, rather than being ‘added’ to the game at this stage, the finished files replace existing placeholders that you may have listened to if you’ve ever played on our PTC branch. These synthesized ‘robovoice’ files allow us to put new narrative events into the game for testing and feedback before the final voice files are ready. It’s important to get a feel for the pacing and flow of the narrative events themselves – especially when they coincide with gameplay elements – so these placeholders are invaluable!

My final step is to hand the baton over to our musician / audio overlord / sound-hat-guy Brandon to ensure that everything vibes and sits right in the game. Post-processing and stuff like that is done in FMOD, which is that thing with the logo you see everywhere.

At that point, I’m done! I don’t have to work anymore! Until I have to start all over again!

Thanks!