AMC Theatres will screen a Swedish movie 'visually dubbed' with the help of AI
Perceived quality of the AI visual dubbing
- Several commenters who’ve seen similar tech (Netflix series, Amazon’s Citadel spin‑off, YouTube demos) describe it as stiff, “gum‑flappy,” and sitting in the uncanny valley, especially when you focus on the actors’ mouths.
- Others watching the promo clips for this Swedish film think it looks “fine” or better than many conventional dubs, but note the company only shows very short, carefully chosen segments.
- Some say they’d find visible artifacts in a theatrical screening more distracting than traditional out‑of‑sync dubs or plain subtitles.
How this specific system works and its technical limits
- The actors re‑recorded their lines in English in a studio; the tool only alters lip movements to match that audio. It does not translate or synthesize voices.
- Commenters note that modeling tongues and inside‑mouth motion is still a major challenge; many systems seem to skip this, which contributes to the unnatural look.
- People question why, if actors are re‑recording audio, you wouldn’t also capture new visual references to drive a more accurate “deepfake‑style” lipsync.
Artistic and industry implications
- Some film lovers find the idea “gross,” but are open to it if it increases audiences for foreign films and reduces the push for inferior remakes.
- Others fear any success will accelerate displacement of professional dub/voice actors, translators, and related crafts, even if this specific project still uses the original cast.
- The practice is compared to colorizing black‑and‑white films and pan‑and‑scan cropping: tools that may broaden audiences but often degrade the original work.
Dubs vs subs and cultural habits
- A sizable group insists they’ll stick with original audio and subtitles; for them, even perfect lipsync can’t replace the nuances of the original voice and intonation.
- Others, especially from countries with strong dubbing traditions (e.g., much of Europe and Latin America), say they grew up with dubs, barely notice mismatched lips, and actively prefer dubbed content.
- Some argue the core problem is often poor-quality English dubs (celebrity voices, weak direction), not dubbing itself; high‑end regional dubs can be very good.
Dialects, nuance, and “AI” terminology
- There’s curiosity about how such systems handle regional dialects and whether dialectal flavor survives when actors switch to English.
- A side thread debates whether this should even be called “AI” or just specialized video processing, with some saying “AI” has become mostly a marketing label.