Video Trimmer & Cutter

Digital video containers (such as MP4, WebM, and MOV) organize compressed video frames into sequential structures known as Groups of Pictures (GOP). Within each GOP, different frame types dictate how video data is reconstructed:

• Intra-Coded Frames (I-Frames / Keyframes): Self-contained full-image frames that do not rely on previous or future frames for decoding. Keyframes serve as anchor points across the timeline.
• Predicted Frames (P-Frames): Contain motion vectors and difference data based on the preceding frame.
• Bi-directional Frames (B-Frames): Interpolate motion using both previous and subsequent frames for high compression efficiency.

1. 1Zero-Latency Timeline Scrubbing: When a video is selected, the browser allocates video frame decoders to extract equidistant frame thumbnails across the duration, rendering a visual filmstrip.
2. 2Draggable Range Selection: The user defines $T_{\text{start}}$ and $T_{\text{end}}$, producing the target duration $\Delta T = T_{\text{end}} - T_{\text{start}}$.
3. 3GPU Hardware Stream Capture: The video engine seeks precisely to $T_{\text{start}}$ and captures high-resolution video and synchronized audio streams directly through the browser's hardware MediaRecorder pipeline until $T_{\text{end}}$ is reached.
4. 4Instant Container Assembly: Raw encoded frames are encapsulated into a standard .mp4 or .webm container and provided as an immediate download Blob without any server mediation.