Use video crop when the same rectangular area of every frame needs to be the only thing that plays back — static letterbox bars, fixed watermarks in the corners, or unwanted edges that never move through the clip. That single decision drives the rest of the workflow: a one-rectangle-every-frame operation, applied with exact pixel coordinates, producing a new file whose visible frame region is smaller than the source's but whose playback duration matches the source. Cropping is not a substitute for trimming a clip to a shorter span, resizing it to a smaller resolution, compressing it to a smaller file, or extracting its audio — each of those targets a different property of the file. Knowing when to reach for crop means recognizing whether the visible frame, the duration, the pixel dimensions, or the audio is what actually needs to change. Once that answer is "the visible frame, everywhere, the same way," crop is almost always the right call; once it is anything else, another tool will get you there with less risk.
How Cropping Differs From Trimming, Resizing, and Compressing
The decision to crop becomes clearer when you set it next to the four edits that often get confused with it. Cropping touches the visible frame region on every frame. Trimming touches the time range that plays back. Resizing touches the output pixel dimensions. Compressing touches the bitrate and the resulting file size. Video to audio touches only the audio track.
A practical way to phrase the question: what would be wrong if you left the source alone? If the answer is "the visible area still has those black bars, that watermark, or that logo strip," crop fixes it. If the answer is "the clip is too long," trim fixes it. If the answer is "the resolution is too high for the platform," resize fixes it. If the answer is "the file is too big to upload," compress fixes it. If the answer is "I just want the soundtrack," video to audio fixes it.
This distinction matters because the tools and their limits differ. A cropper takes exact pixel coordinates and re-records only the requested rectangle. A trimmer takes in and out timestamps. A resizer takes target width and height. A compressor takes a target bitrate or preset. Reaching for the wrong one usually means either the operation fails outright, or you end up doing extra work to undo the side effects of the wrong pass.
Six Triggers That Mean "Use Video Crop Now"
Several recurring situations make video crop the right answer. Each one is a fixed-rectangle case — the unwanted area does not move during playback.
Static letterbox or pillarbox bars. The source has black bars on the top and bottom (16:9 inside a wider container) or on the left and right (4:3 inside a 16:9 frame). The bars never move and you want them gone in the output.
A fixed watermark, logo, or timestamp. A channel bug, lower-third logo, or "CAM 01" overlay sits in the same corner of every frame. Cropping it out keeps the rest of the frame.
Platform-specific frame shapes. The destination expects a specific rectangle — 1080 × 1920 for vertical short-form video, 1080 × 1080 for a square, or 1920 × 1080 for a 16:9 upload — and your source has more around the edges than the destination wants.
Broadcast or color safe area cleanup. A known rectangle outside the safe area carries content that does not need to ship. Cropping to the safe rectangle trims it once and for every frame.
Screen recording with a fixed chrome. The browser frame, app sidebar, or system clock around the recording is the same in every frame and is not part of the content.
Composite or edited footage that needs re-framing. A camera was offset in a multicam shoot and the unused side of the frame carries crew reflections or set pieces. Cropping to the active area cleans it up.
If the unwanted area shifts during the clip, this approach will not work — the rectangle is locked to source coordinates, so any moving intrusion will pass through the crop window at some point.
When Crop Is the Wrong Tool
A fixed-rectangle crop only handles a fixed problem. If the issue moves, or if it changes over time, crop will not track it. The crop rectangle is applied to every frame at the same source coordinates; it does not follow a moving subject or grow over time. For anything that requires per-frame attention, you need a tracker, a blur, or a redact tool, not a rectangular crop.
Cropping also does not trim time. A one-minute source that needs to be a 12-second highlight still needs a trim operation, even after you have cropped the visible area. Cropping also does not change the file size by itself — the bitrate the recorder picks is derived from the cropped pixel area at 30 frames per second, bounded from 180 kbps through 8 Mbps, but that does not turn crop into a compression tool. If the goal is a smaller upload, run the cropped output through a compressor afterward.
Cropping does not remove objects, blur faces, replace a background, or restore an aspect ratio by scaling. It only takes a rectangle from every frame. If you find yourself wanting a partial-frame effect that is not the same shape throughout the clip, crop is the wrong starting point.
How to Crop With Exact Pixel Coordinates in Your Browser
The Video Cropper applies one rectangle to every frame and re-records it as a WebM. Follow these steps to keep the operation pixel-accurate.
- Choose a browser-decodable video (MP4, WebM, MOV, M4V, or Ogg) and wait for its source dimensions to load. The decoded width and height populate the form once the browser finishes reading the file.
- Enter whole-pixel X, Y, width, and height values. X is the offset in pixels from the top-left corner, moving right. Y is the offset moving down. Width is the horizontal extent. Height is the vertical extent. Every value must be an untrimmed nonnegative whole number, and width and height must each be at least two pixels.
- Confirm the rectangle stays inside every source frame: X plus width must not exceed the source width, and Y plus height must not exceed the source height. The output area must also stay at or below 3840 × 2160 pixels. The defaults select the full frame, so you only need to override them when you want a smaller rectangle.
- Select Crop video. The tool plays the source locally, draws only the requested region into a Canvas, and records that Canvas stream as WebM. Keep the tab open during processing — a one-minute source takes about one minute, because the recording happens in real time.
- Download the resulting WebM. The tool appends audio when the browser exposes a capturable track; verify that both sound and picture are present in the download. Then play the file back in the destination player and confirm duration, rectangle, and compatibility before you delete anything.
If the rectangle is invalid — out of bounds, smaller than two pixels in either dimension, or larger than 3840 × 2160 — the crop fails without shifting, clipping, rounding, or producing a partial file. Fix the coordinates and try again.
Limits That Decide Whether a Browser Cropper Can Finish the Job
Before you crop, check the source and the requested output against the tool's combined safety budget. The input and the rectangle share these constraints:
| Limit | Source | Output rectangle |
|---|---|---|
| File size | Up to 500 MiB | n/a |
| Duration | Above zero, no more than five minutes | Matches source duration |
| Source sides | No side above 4096 pixels | X + width ≤ source width, Y + height ≤ source height |
| Source area | No source area above 3840 × 2160 | Output area ≤ 3840 × 2160 |
| Codec | Browser-decodable (extension accepted but not guaranteed) | VP9, VP8, or generic WebM |
A container extension alone does not guarantee codec support. The browser has to be able to decode the source; if it cannot, the form will not populate and no crop will start.
The recorded WebM has a finite-duration container. The recorder writes a Matroska Duration in timestamp-scale ticks, so players report a real duration instead of an open-ended stream. Reference for that container format is the IETF Matroska Media Container Format draft, and the Canvas drawing step follows the WHATWG HTML — Canvas drawImage section. Audio is appended when the browser exposes a capturable track through media capture; support varies, and the source may be audible during processing. Subtitles, chapters, rotation tags, HDR signaling, color metadata, and similar container features are not preserved — verify the output if any of those matter for your destination.
Verifying the Output Before You Delete the Original
A successful download is not the same as a verified result. Before deleting the source, replay the WebM in the destination player and check five things.
Frame region: the visible rectangle matches what you typed into X, Y, width, and height, with no off-by-one pixel shift and no leftover bars from the source.
Duration: the player reports a finite duration that matches the source. Streaming MediaRecorder output sometimes ships without one, so the tool's Matroska duration writer is what turns the file into something seekable.
Audio: the soundtrack plays back, at the right sync, when the source had audio. If the source had no capturable audio track, the output will be silent by design.
Compatibility: the file opens in the app or browser where it will actually be used. Some players are stricter about WebM variants than others.
Container features you need: subtitles, chapters, rotation, HDR — none of these are preserved by the crop, so if any were present in the source and your destination needs them, reattach them through a different step.
If any of those checks fail, keep the original, adjust the rectangle or the source, and re-crop. For more decision help on whether crop is the right call for a specific file, the guide on how to decide if you need to crop a video walks through the same triggers in more depth.
For a deeper look, see When Should I Use Video Crop Instead of Doing It Manually.