AI made these movies sharper – critics say it ruined them
AI-driven “remastering” of classic films like True Lies and Aliens is drawing criticism for altering color grading, skin tones, grain and focus in ways many viewers find plasticky or soulless, even when based on high‑quality 4K scans. Commenters distinguish between legitimate restoration or upscaling and machine‑learning tools that invent or smooth over details, arguing this can undermine a director’s original visual intent and film history itself. Others counter that most audiences prioritize clarity and story over technical authenticity, and that better tools and standards for HDR, upscaling and labeling could eventually make such enhancements acceptable when applied with care.
Visual Changes and Aesthetic Complaints
- Many say the “AI-enhanced” frames look darker, flatter, greener/bluer, and “clay-like,” with ruined skin tones and plasticky smoothing.
- Critics argue the remasters often kill film grain, blur or incorrectly sharpen details, and alter depth-of-field and motion blur, changing the feel of scenes.
- Others note some AI remasters (e.g., Aliens 4K) look impressive and cleaner, though sometimes “too crisp” and de-grained for their taste.
HDR, Screenshots, and Comparison Issues
- Several point out the article’s comparison images are likely wrong: HDR frames grabbed and shown as SDR produce exactly the dark, green/blue look seen.
- There’s strong criticism of comparing highly compressed streaming copies to Blu-ray; any disc should look vastly better regardless of AI.
- HDR itself is called a mess: inconsistent player/TV handling, many sets too dim, and bad tone mapping making content appear “too dark.”
Artistic Intent vs. AI Alteration
- A major theme: AI remastering adds or “guesses” information that was never captured, making it no longer the original film.
- People worry about erasing the historical look of film (grain, color, softness, technical limitations) and overriding directors’ choices in lighting, grading, and texture.
- Others counter that remastering has always been interpretive (color timing, grain reduction, upscaling) and that AI is just another tool, good or bad depending on use.
- Strong calls that original versions must remain available; examples like original Star Wars show that often they are not, except via piracy.
Use Cases, Limits, and Technology Trajectory
- Some see value in careful AI usage: repairing damaged negatives, cleaning audio, limited upscaling, or fixing bad early CGI, provided it’s supervised and subtle.
- Others emphasize that AI “enhancement” is probabilistic: across billions of pixels/frames, visible artifacts and wrong guesses are inevitable if applied bluntly.
- Several liken the current wave to early CGI, the .com boom, or blockchain: overhyped, often tacky, but likely to mature; future tools may be far better.
Viewer Priorities and Nostalgia
- Split between viewers who prioritize story and are fine with modernized image quality, and those who value period-accurate texture, grain, and even VHS-era softness.
- Many note that “sharper” isn’t always better; some games and films look worse when over-cleaned or shown at unintended resolutions.