Skip to content

FEATURE_IDEAS.md

Latest commit

 

History

History
75 lines (65 loc) · 4.03 KB

FEATURE_IDEAS.md

File metadata and controls

75 lines (65 loc) · 4.03 KB
Raw

Feature proposals — better_native_video_player

Candidate features people ask video players for (sourced from this repo's issue themes, better_player/video_player issue trackers, and feed apps' usage), ranked by expected value for feed-style apps.

Status update (2026-06-11, feat/extractor-and-player-features): implemented — #2 analytics (PlaybackAnalytics), #3 startAt + checkpoints (load(startAt:) native on both platforms + PositionCheckpoints), #4 storyboards (StoryboardThumbnails: VTT + uniform sprite grids), #7 A-B loop (setPlaybackRange/clearPlaybackRange), #8 background toggle (BackgroundPlaybackGuard), #10 playlist (NativeVideoPlayerPlaylist). Extraction shipped as the separate packages/better_native_video_extractor package (Vimeo incl. referer, thumbnails, storyboard, expiry cache; YouTube) — the "app-side" placement argued for below, just maintained in-repo. Deliberately NOT implemented: #1 prefetch (perf Tier 3b, needs device A/B), #5 in-plugin Chromecast (app-level guidance: receiver subtitle tracks + native-controls routing), #6 offline downloads, #9 DRM token refresh (device/SDK-dependent, not verifiable here).

High value / moderate effort

  1. Prefetch API (precache(url) — roadmap Tier 3b). Warm the next feed items' manifests/first segments so scrolling to them starts instantly. Android: Media3 CacheDataSource/PreloadMediaSource; iOS: AVAsset preheating (MP4) or manifest-only warmup (HLS). Pairs with a feed's feed and the Vimeo HTTP extraction (resolve + prefetch together).
  2. Playback analytics events — a single analyticsStream emitting structured events (startup time, first-frame, stall start/end + count, bitrate/variant switches, watched-duration heartbeats, completion). Both platforms already observe everything needed (KVO / Player.Listener); this exposes it. Apps get engagement/QoE metrics nearly for free.
  3. Resume-position convenience — load(..., startAt: Duration) plus an optional onPositionCheckpoint callback (every N seconds, last value on dispose). Apps typically implement this app-side; plugin-level support removes the racy seek-after-load dance.
  4. Scrubbing thumbnail previews — accept a storyboard source (WebVTT storyboard / sprite sheet / BIF) and expose thumbnailAt(Duration) for overlay scrub bars. Vimeo's config even ships storyboard URLs. Pure Dart + one image fetch path; no native work.

High value / higher effort

  1. Chromecast support — the most-requested capability gap vs better_player ecosystems. Android: Media3 CastPlayer integration behind the same controller API; iOS: google_cast SDK. Big, stateful, needs real devices; consider a separate companion package.
  2. Offline downloads — Media3 DownloadManager + AVAssetDownloadTask behind a downloads API (queue, progress, license persistence for DRM). Large; only worth it with a concrete product need.

Nice-to-have / small

  1. A-B loop / clip range — setPlaybackRange(start, end, loop:). Trivial on both platforms (boundary timer + seek).
  2. Background audio-only toggle — keep audio when backgrounded without PiP (iOS audiovisualBackgroundPlaybackPolicy is already set; Android needs the media session flag we already hold). Mostly plumbing
    • docs.
  3. Per-source HTTP header refresh / DRM token renewal callback — lets apps hand the plugin a Future<String> Function() token provider so 401s on segment/license requests trigger a refresh instead of an error. Valuable for tenant-authenticated streams.
  4. Playlist/queue API — loadPlaylist([...]) with auto-advance and a currentIndexStream. Medium effort; interacts with the shared-player lifecycle, design carefully.

Explicitly not recommended

  • iOS-native sidecar subtitle injection (no sane AVPlayer API; already solved via the Flutter overlay + Android sideload).
  • In-plugin Vimeo/YouTube extractors — site extraction churns and belongs app-side or server-side (Vimeo HTTP approach), not in a player plugin.