Strip away every content pack type in The Choicer Voicer and you’re left with the same basic unit repeated everywhere: an audio clip. Voice packs are built from them, dub packs sync recordings against them, judges react with them, and Twitch chat triggers them by keyword. The rules aren’t identical across all four contexts, though, and mixing them up is a common source of packs that load but don’t behave right.
Across every context that uses audio — voice clips, dub samples, studio music — the game reads three formats: WAV, MP3, and OGG. There’s no meaningful advantage to picking one over the others for most purposes; use whichever your editor exports most conveniently.
This is where people trip up, because the limit changes depending on what the clip is for:
Here’s the part that catches people off guard: the game doesn’t check any of this at the pack-selection screen. Validation only happens once it actually builds a batch for gameplay. That means a corrupted file, a wrong format, or an oversized clip can sit in your pack list looking completely normal right up until the moment the game tries to use it.
Format and length determine whether a clip loads at all. Loudness determines whether it plays fair once it does. The scoring system reads waveforms, and small, quiet waveforms are genuinely harder for it to evaluate accurately — which means a pack full of soft-spoken clips doesn’t just sound weaker, it produces scores that feel arbitrary and frustrating to the person performing. Normalizing your audio before finalizing a pack isn’t a nice-to-have; it’s the single change most likely to make a pack feel good to actually play.
Different pack types treat filenames very differently, because they’re solving different problems:
01_IntroLine, 02_PunchlineOne, and so on. Skip the numbering and the sequence becomes unmanageable fast.scoreblip1 through scoreblip5, tied to the order votes happen in — not to which judge cast them.judgeX_voice, but tied to a specific judge rather than vote order.config_chatter file, where they’re mapped to the keywords that trigger them.Since version 0.5.0, captions and tags for voice clips are handled through Extras > Edit Voice Packs & Clip Metadata rather than separate accompanying files. From there, you can write a caption describing what’s actually being said, apply tags so players can filter for the kind of content they want, and — worth knowing — nested subfolders inside a pack are automatically treated as their own tags unless you turn that behavior off. The trade-off is load time: because clips need to preload to support this filtering, a pack with a large total file size can take a few extra seconds to fully appear.
Beyond the numbering convention, dub samples hold three extra pieces of data that a standard voice clip never needs:
If you’re only going to remember two things: check the length limit for whichever pack type you’re building (60 seconds for voice, 6 for dub), and normalize your audio before you call anything finished. Everything else — naming, tagging, metadata — is about organization and polish. Those two rules are about whether the clip actually works the way it’s supposed to.