Skip to content

A simple R&D player, entirely created by AI without human involvement.

License

Notifications You must be signed in to change notification settings

OpenIPC/rnd-player

Repository files navigation

R&D Player

A browser-based video stream analyzer for DASH and HLS. Think of it as a lightweight, instant-on alternative to tools like StreamEye — no install, no desktop app, just paste a manifest URL and start inspecting.

Live demo

R&D Player — quality comparison with filmstrip, GOP tooltip, and bitrate graph

What it does

  • Split-screen quality comparison — put two renditions side by side with a draggable divider, synchronized down to the frame. Spot compression artifacts, banding, and detail loss at a glance.
  • Filmstrip timeline with frame-level zoom — scroll through the stream as a visual strip of thumbnails. Zoom all the way in to see individual frames with color-coded borders: red for I-frames, blue for P, green for B.
  • GOP structure inspector — hover over any segment in the bitrate graph to see a per-frame size breakdown with I/P/B classification. Understand how the encoder distributes bits across the GOP without leaving the browser.
  • Per-segment bitrate graph — measured from actual network responses (not just manifest estimates). Instantly see bitrate spikes, CBR consistency, and how the encoder reacts to scene changes.
  • Save any frame at full resolution — right-click a filmstrip thumbnail and export the exact frame as a PNG, decoded from the active rendition (not the thumbnail stream). Frame-accurate, position-based capture that works across different GOP structures.
  • ClearKey DRM support — auto-detects cenc:default_KID from DASH manifests and prompts for the decryption key. Filmstrip and frame export work on encrypted content too.
  • Real-time stats panel — codecs, resolution, dropped frames, buffer health, network throughput, color gamut, and more. One right-click away.
  • Audio level meters — per-channel dB meters with peak hold, supporting mono through 5.1.
  • Clip export — set in/out points with I/O keys, pick a rendition, and download the segment as MP4.
  • Keyboard-driven workflow — J/K/L shuttle, frame-step with arrow keys or comma/period, volume, fullscreen, zoom — all without touching the mouse.
  • Sleep/wake recovery — detects system sleep via timer-gap analysis and restores playback position automatically.

Works with any public DASH (.mpd) or HLS (.m3u8) stream. No server required — runs entirely in the browser.

Browser Support

Automatically tested on every commit:

Chromium · Linux Firefox · macOS WebKit · Linux
Chromium on Linux Firefox on macOS WebKit on Linux
Safari · macOS Edge · Windows
Safari on macOS Edge on Windows

Getting Started

npm install
npm run dev

Open the app, paste a manifest URL, and click Load. If the stream is ClearKey-encrypted, you'll be prompted for the hex decryption key (or pass it via ?key=).

Build

npm run build

Stack

License

MIT

About

A simple R&D player, entirely created by AI without human involvement.

Topics

Resources

License

Code of conduct

Stars

Watchers

Forks

Releases

No releases published

Sponsor this project

Contributors 4

  •  
  •  
  •  
  •