
OVERVIEW
Vox is an adventure game that uses voice integration to allow the player to give verbal commands to the avatar character in order to solve puzzles and escape a mysterious mansion. Using the Hugging Face Unity API, the player’s speech is turned into text which is checked against a list of valid “points of interest” that the avatar can examine to gather clues.
Inspired by voice controlled games like Lifeline, Seaman, and DekaVoice, Vox has been my most challenging project to date. Despite the hardship posed by the unique control interface and other oddities that come with a voice controlled game, the development of Vox has left me with innumerable lessons that I will carry with me always.
DEVELOPMENT PROCESS
Development for Vox began shortly after Shoot First, Ask Questions Later was Finished. I had been toying with the idea of a game controlled with your voice for a good while, inspired mainly by the ps2 adventure game Lifeline. Before I even touched Unity, I spent my winter break going over the different options I had to enable voice recognition in Unity. Of the solutions available at the time, I had settled on Hugging Face, as they had a simple unity integration that I was able to have up and running in about 10 minutes on a free account. Their tools enabled the user to record their voice, which would be parsed into text. Using this, I was able to create a rudimentary system in which the player could say which object they would like the avatar to investigate. The player’s speech would be recorded, turned into text, and then the text would be parsed for keywords relating to the tags on the objects in the scene that could be examined.
Some VERY early test footage from when I had just gotten the navigation working. You can hear me working through a lot of the issues I had been facing the week before.
This project made LIBERAL use of the Pixel Crushers Dialogue System to display little “barks” of information when the player examines a point of interest.
While I was at work on the system of parsing speech to text, I had also begun working on an important bit of prototyping on the side. One aspect about the game that I grappled with a lot is verbiage. From verbs themselves to synonyms for different objects, I wanted to hear what words the players themselves would be using to identify the objects I was planning on using in the scene. So I spent a little bit creating an escape room style paper prototype that I asked some of my fellow students to participate in.