Space-Quest-V-Voice-Acting-Mod

Space-Quest-V-Voice-Acting-Mod

I modified Space Quest V to support full voice acting.

Sample video:

• https://www.youtube.com/watch?v=LB9o2LXhkkY

Installation

To install, get a copy of Space Quest V from the Space Quest Collection, and run the exe found in the Releases. Provide the exe with the path to your Space Quest 5 folder. The installer should remove all patch files, as they may cause issues running the game.

After installation, download the ~500MB resource.aud file, which contains the synthetized voices, and add it inside the Space Quest 5 folder.

Note that the game must be played with both text and audio, as the branching dialog requires text output from the Messager.

Runs with DOXBox-X and DOSBox (0.74-2 or earlier). ScummVM has an issue where loading a saved game loads with missing text boxes.

Pipeline

If you want to try adding voices to other old games, here is my general pipeline.

• Programming: Decompile and understand a game's code to figure out what and where changes are needed to get both text and speech working.
• Dataset Collection: Quality > Quantity. Think about a character and what voice would be appropriate and try to find or make a dataset that matches the character. Consider the trade-offs of using audio denoising on low quality audio.
• Fine-Tuning: Fine-tune voice models using Tortoise-TTS and RVC. If you want, you can also experiment with fine-tuning F5-TTS and E2-TTS. You can find some scattered tutorials here.
• Text Extraction: Extract text from a game and sort it by speaker if needed.
• Synthetic Voice Generation: Tortoise-TTS can take hours to generate hundreds of vocal lines on a GPU. Additionally, you may try generating multiple samples at once, as the quality and delivery may vary with random seed.
• Quality Control: Filter out bad audio samples and select the best ones. If a vocal line isn't working try generating more, correcting any issues with the text or rewriting it more phonetically. For example, rewriting "EVA" as "E.V.A." or "II" as "Two".

Synthetic Voices

As of the date this was published and updated, the use of "deep fakes" legally within my state and country has yet to be made into any sort of law. While this is a relatively innocent use of AI technology (as opposed to something, for example, commercial or pornographic), I am a US citizen and I want to continue abiding by all relevant laws and will remove all synthetic voice content and cease any further work of this nature, if and when needed, to comply with any relevant changes to laws.

Ordered by vocal appearance. Most are bit roles or don't sound that great because I tried limiting myself to Sierra voice actors and Sierra audio (I only made one exception).

Speech Synthesizers

• Tortoise-TTS
• RVC
• F5-TTS & E2-TTS
• A handful of free mouth noises from the Internet, most put through RVC.
• A handful of recordings of me delivering lines put through RVC.