VSL Audio · Salvage Test

Same 60 seconds (your price section, ~35:00). The question: can we strip the music and keep your voice clean enough to use — or do you reshoot? Listen on headphones if you can.

1 · Original (what you heard)

your voice + the music bed, normalized

2 · Cleaned (music removed by AI)

demucs vocal isolation → de-noise → broadcast level
Listen for: is your voice natural, or is there underwater/warble/robotic artifacting where the music used to be?

3 · What got removed

the isolated music bed (for reference)
Your call: if #2 sounds clean enough → I run this over all 48 minutes, then we go back to the take-review workflow and you never reshoot. If #2 is too artifacty → you reshoot, but we fix the root cause first (mute desktop/music in OBS + mic closer), so the next take is clean from the start. Tell me which.