Skip to content

Lip Sync (Phonemes) #13

@LeadKiller

Description

@LeadKiller

Team Fortress 2 introduced MP3 compression for voice lines with the Mar 12, 2015 Update, which also removed phoneme expressions from the VCD files and replaced it with script files.

Solutions

Using the original .wav files
This works well, but has three issues:

  • It is 418 MB
  • It does not work for new lines
  • Lip Sync Data is out-of-date

Manually recreate Lip Sync Data
This works for any line, but has two issues:

  • It takes a while to do
  • It must be done for every line

Use AddPhonemeFile
Exposing engine->AddPhonemeFile to Lua would allow developers to load phoneme data from scripts.
This would allow all lines to effortlessly be up-to-date with Team Fortress 2's lipsync data.
Currently the phoneme data for Team Fortress 2 resides within scripts/game_sounds_vo_phonemes.txt.
This could work perfectly, but has one issue:

  • Only Facepunch can do this.

Manually face pose characters for voice lines
Reading the scripts/game_sounds_vo_phonemes.txt file through Lua and manually recreating the lip sync system works would function similarly to above, but has four issues:

  • Timing can be incorrect
  • Words don't work properly
  • Very tedious to get working
  • Takes 1-6 FPS depending on how lines are cached

Metadata

Metadata

Assignees

No one assigned

    Labels

    currentcurrent branches (dev/master)enhancementNew feature or requestremakeremake branch

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions