A documented video crop procedure is a written set of pixel-precise instructions that takes a browser-decodable source video, defines an exact rectangular region inside every frame, and produces a finite-duration WebM that preserves the chosen area while discarding the rest. The procedure has four fixed checkpoints: confirm the source dimensions before drawing, enter whole-pixel X, Y, width, and height values that stay inside every decoded frame, run the crop in real time while keeping the tab open, and review the downloaded WebM for duration, crop accuracy, audio, and destination-player compatibility before deleting the original. Because every crop rectangle must be specified in source pixels rather than in percentages or aspect-ratio presets, the documented procedure also records the source's decoded width and height, the exact four coordinates used, and the relationships X plus width must stay within the source width and Y plus height must stay within the source height, so the same crop can be reproduced or audited later.

What a Documented Video Crop Procedure Covers
A documented video crop procedure is more than a list of button presses. It records the inputs the cropper needs, the exact rectangle applied to every frame, the limits the tool enforces, and the verification steps that confirm the output matches the documented intent. For Video Cropper, which decodes a browser-decodable source, draws one fixed rectangle into an opaque Canvas, records that Canvas as WebM, and patches a finite Matroska duration entirely in the browser tab, the documented procedure has to capture the source format (MP4, WebM, MOV, M4V, or Ogg), the decoded width and height of the source in pixels, the four crop coordinates (X, Y, width, height), the expected real-time processing duration (roughly equal to the source duration), the output container (WebM via VP9, VP8, or generic WebM), and the verification list: duration, crop placement, audio presence, and compatibility with the destination player. Without each of those notes, the procedure cannot be reproduced, audited, or handed to another person without guesswork.
Why Video Cropper Matches a Documented Workflow
Video Cropper is built around reproducibility. It accepts one fixed rectangle, applies it to every frame, records the cropped region into a Canvas, and writes a finite-duration WebM without uploading the source or the output. That deterministic behavior is what a written procedure needs: given the same four pixel values, the tool produces the same crop on the same source. The tool also exposes its limits in plain numbers, including 500 MiB maximum file size, no more than five minutes of decoded duration, no source side above 4096 pixels, no source area above 3840 × 2160, output area no larger than 3840 × 2160, and rectangle dimensions of at least 2 pixels, so the documented procedure can include a pre-flight check that fails fast instead of producing a partial file. Because decoding, cropping, recording, and duration patching all happen in the current tab, the workflow can be repeated once the page is loaded, and the documented steps do not depend on a remote service being online.
Documented Steps to Crop a Video With Video Cropper
- Open the Video Cropper page in a desktop browser and click the file picker. Choose a browser-decodable source video in MP4, WebM, MOV, M4V, or Ogg format, then wait for the decoded source dimensions to appear in the tool before continuing.
- Record the decoded source width and height in the documentation. Every later step depends on these two numbers.
- Compute the crop rectangle in source pixels. Set X (the horizontal offset from the top-left corner, in pixels) and Y (the vertical offset from the top-left corner, in pixels), then set width and height (each at least 2 pixels). Confirm X plus width stays within the source width and Y plus height stays within the source height, and confirm width multiplied by height does not exceed 3840 × 2160.
- Enter the four values into the tool's X, Y, width, and height fields. The defaults already select the full frame, so replace them with the computed rectangle.
- Select Crop video. Keep the browser tab open and in the foreground. Processing happens in real time, so a one-minute source takes about one minute; do not close, reload, or navigate away.
- When the tool offers the download, save the WebM to a known location. Open it in the destination player and review the entire file.
- Run the verification checklist: confirm the duration matches the source, confirm the crop rectangle is correct on multiple frames, confirm the audio is present (when expected) or absent (when not), and confirm the file plays in the destination player without artifacts.
- Keep the original source file until the entire output has been reviewed and accepted.
These eight steps form the documented procedure. Steps 1 and 2 collect inputs, step 3 produces a valid rectangle, steps 4 and 5 execute the crop, step 6 captures the output, step 7 verifies it, and step 8 protects the original. Each step is a single decision, which is what makes the procedure writable and auditable.
Input Limits and Output Properties
Before opening the tool, write down the limits so a reader of the documented procedure can pre-validate the source and the rectangle without trial and error. The table below lists the values the tool enforces and the output properties the writer of the procedure should record.
| Property | Value or behavior |
|---|---|
| Accepted source extensions | MP4, WebM, MOV, M4V, Ogg (codec support depends on the browser) |
| Maximum source file size | 500 MiB |
| Maximum decoded duration | Five minutes, above zero |
| Maximum source side | 4096 pixels |
| Maximum source area | 3840 × 2160 pixels |
| Maximum output area | 3840 × 2160 pixels |
| Minimum width and height | 2 pixels each |
| Coordinate origin | Top-left corner of the source frame (0, 0) |
| Coordinate type | Untrimmed, nonnegative whole pixels |
| Output container | WebM via VP9, VP8, or generic WebM (first reported by the browser) |
| Output bitrate | Derived from cropped pixel area at 30 fps, bounded 180 kbps to 8 Mbps |
| Processing location | Current browser tab; no upload |
| Audio handling | Appended when the browser exposes a capturable track; may be audible during processing |
| Duration handling | Finite Matroska Duration inserted in TimestampScale ticks after recording |
| Preserved across crop | Visible rectangular region of every frame |
| Not preserved across crop | Subtitles, chapters, attachments, rotation tags, HDR signaling, color metadata |
Invalid rectangles, including out-of-bounds, non-integer, below 2 pixels in width or height, or producing an area larger than 3840 × 2160, fail without shifting, clipping, rounding, or producing a partial file, which is the intended behavior for a written procedure that needs predictable failures. The output container is WebM with a finite Matroska Duration value inserted in TimestampScale ticks (per the IETF Matroska specification), and the crop is rendered through the standard HTML Canvas drawImage path. A documented procedure that cites both specifications gives a reviewer enough context to confirm that the output container and the rendering step match the documented intent.
Recording the Procedure for Repeat Use
A documented procedure is only useful if it can be repeated. To make the Video Cropper workflow repeatable, record the source filename, the decoded source width and height, the exact four coordinates used, the bitrate floor and ceiling applied, the output filename, and the verification result. Save those notes next to the output WebM so the next person reading the folder can reconstruct the rectangle without re-decoding the source. When the rectangle is reused on multiple sources, store the coordinates relative to a reference resolution (for example, X equals 100 pixels and Y equals 60 pixels on a 1920 × 1080 frame) and rescale them per source. When the rectangle must match a platform's frame, whether 9:16 for Shorts or Reels, 1:1 for a square post, or 4:5 for a vertical feed, record the aspect ratio and the exact pixel dimensions used, because aspect-ratio presets are not built into the tool and the procedure must supply them.
Verifying the Cropped WebM After Export
Verification is the step that turns a crop into a documented result. Play the downloaded WebM in the destination player from start to finish, not just the first few seconds. Confirm the duration matches the source within the recorder's tolerance, confirm the rectangle is in the right place on at least three frames (opening, middle, end), and confirm the audio is intact when the browser exposed a capturable track. If the destination is a social platform, upload the WebM to a private or draft slot before publishing, because container features that are not preserved (subtitles, chapters, attachments, rotation tags, HDR signaling, and color metadata) will not survive the crop and may also be re-encoded by the platform. For a deeper review checklist, follow the step-by-step verification workflow for a local video crop. Keep the original source until the entire output, including the platform-side upload, has been accepted.
Documenting a video crop procedure is the act of writing down four numbers, three input limits, and one verification list. With Video Cropper, those numbers come from the source's decoded dimensions, the limits are the values listed in the table above, and the verification list is the eight-step procedure applied in order. Once the procedure is written and the output reviewed, the same crop can be re-run, audited, or handed to another editor without re-decoding the source or guessing at coordinates.
If you're weighing options, Fix a Cropped Video Result That Looks Wrong covers this in detail.