Making one voice sit next to another in a documentary or podcast mix: long-term average spectrum (LTAS) EQ matching with opendubstage ltasmatch, bridging a TTS or cloned voice into a location recording (band-limit, early reflections, room-tone floor), riding and placing an off-mic second speaker, phone and tape-sync matching, and when two voices should deliberately stay different.
安装
npx skills add gothicpixie/opendubstage --skill voice-matchingnpx skills add gothicpixie/opendubstage --skill voice-matching -a claude-code -gnpx skills add gothicpixie/opendubstage --skill voice-matching -a codex -ggemini skills install https://github.com/gothicpixie/opendubstage.git --path skills/voice-matchinggh skill install gothicpixie/opendubstage voice-matchinggit clone --depth 1 https://github.com/gothicpixie/opendubstage /tmp/gothicpixie__opendubstage
mkdir -p ~/.claude/skills ~/.agents/skills
cp -R /tmp/gothicpixie__opendubstage/skills/voice-matching ~/.claude/skills/voice-matching # Claude Code
cp -R /tmp/gothicpixie__opendubstage/skills/voice-matching ~/.agents/skills/voice-matching # Codex, Cursor, Gemini CLI, Copilot, OpenCode