Netflix's bet on advanced video encoding

Overall sentiment on Netflix video quality

• Many commenters say Netflix’s encoding looks bad, especially:
  • 4K streams.
  • Dark scenes (banding, macroblocking, “pixel porridge”).
  • Fast motion, fine textures, and complex scenes (forests, waves, particles).
• Some report no problems on TVs/streaming boxes, suggesting tolerance or device differences.
• Several note that Netflix appears worse than Apple TV+, Blu‑ray, and some other services.

Bitrate, codecs, and device limitations

• Perceived decline since pandemic-era bitrate cuts; people doubt those were fully reversed.
• Apple TV+ is repeatedly cited as using much higher bitrates than Netflix for 4K.
• Netflix uses different maximum resolutions/bitrates depending on:
  • Browser, DRM level, and OS (desktop browsers often limited to 720p/1080p).
  • Device certification (many Android boxes/TVs get a web-app or lower quality).
  • Subscription tier (4K/highest bitrate restricted to premium plans).
• Auto-scaling often drops quality mid-stream; some suspect this is aggressive cost-saving, not just bandwidth adaptation.

“Advanced encoding” vs long-standing techniques

• Several people argue the article overhypes Netflix’s work:
  • Per-title and per-shot optimization is framed as new, but seen as an extension of long-known VBR/multipass ideas.
  • Some call decreasing color gamut or aggressive compression “ghetto encoding,” not “advanced.”
• Others note Netflix does use modern codecs (e.g., not 4K in H.264) and stress testing material, but say users rarely see those best encodes.

AI/ML-driven encoding

• Discussion of using object detection / attention models to allocate bits to important regions (faces, foreground).
• Some note that encoders already do perceptual weighting, but ML is being explored to improve predictions and decisions.
• Barriers include computational cost and model deployment at scale.

Business incentives, regulation, and “enshittification”

• Strong view that bandwidth cost optimization is prioritized over viewer quality.
• Debate over whether this is “monopolization” vs broader “market power” and shareholder incentives.
• Ideas floated:
  • Quality/bitrate standards for using labels like “4K”.
  • More regulation, though many see it as unlikely.
• Some see degradation as part of a wider pattern of services getting worse over time.

Piracy, local media, and user control

• Many say torrents, Blu‑ray rips, and personal servers deliver clearly better quality at similar or lower nominal resolutions.
• Technical discussion on how high-quality WEB-DLs are obtained (DRM key extraction, HDCP stripping, vulnerable devices).
• Frustration that users cannot force specific resolutions/bitrates or trade more buffering for stable high quality.
• Some prefer high-bitrate 1080p over low-bitrate 4K and wish industry had prioritized framerate/bitrate over resolution.