After you run a video through Video Compressor, you should review the input byte size, the output byte size, the percentage change between them, the output dimensions, the playback of the downloaded file from start to finish, and whether the destination accepts WebM with the codec the browser produced. Compression runs locally in your browser tab at real time, so the moment the source finishes playing is also the moment the tool has actual numbers to show. What you review at that point is not a promise about compression; it is a report of what happened to your specific file. Output dimensions can shrink to fit a preset's maximum width while preserving the source's aspect ratio, but the byte change is content dependent and may go in either direction. The review therefore begins with reading the report the tool actually gives you, then continues with playback verification on the device or platform where the file will be used.
The Three Numbers the Tool Reports
Video Compressor reports three concrete values when the source finishes playing: the input bytes of the file you selected, the output bytes of the WebM it produced, and the signed percentage change between them. The percentage is computed only from those two actual Blob sizes, not from a preset name, so a larger output is shown as a positive number rather than as a reduction. The tool does not relabel a grown file as a saved file. Alongside the bytes, the page also lists the output dimensions, which reflect the preset's maximum width, the source's aspect ratio, and a downward round to even pixel values for codec compatibility. Read these numbers together, not in isolation. A 70% reduction on a Small preset at 640 pixels and 24 fps tells one story; a 10% reduction on a Quality preset at 1920 pixels and 30 fps tells another. The percentages only carry meaning once you connect them to the preset you picked, the source dimensions, and the kind of content on screen.
Preset Limits You Compare the Report Against
| Preset | Maximum output width | Target frame rate | Recorder bitrate range |
|---|---|---|---|
| Small | 640 px | 24 fps | 180 kbps – 8 Mbps, clamped by a bits-per-pixel heuristic |
| Balanced | 1280 px | 30 fps | 180 kbps – 8 Mbps, clamped by a bits-per-pixel heuristic |
| Quality | 1920 px | 30 fps | 180 kbps – 8 Mbps, clamped by a bits-per-pixel heuristic |
Each preset preserves the source aspect ratio, never enlarges the source, and rounds output dimensions down to even pixel values. A source already at 640 pixels or below stays at its native size on Small; a source at 1280 stays at its native size on Balanced. When you review the output, ask whether the dimensions you see match the preset you picked and the source you fed in. If the source was 1920 × 1080 and the report shows 1920 × 1080 on Quality, the canvas path did not downscale. If the same source on Small shows 640 × 360, the width was capped and the height followed the aspect ratio.
Run the Compression and Walk Through the Review
- Choose a browser-decodable video no larger than 500 MiB, five minutes, or 3840 × 2160 pixels. Files may carry MP4, WebM, MOV, M4V, or Ogg extensions, but successful decoding still depends on the codecs installed in the current browser, so a valid extension cannot make an unsupported codec decodable.
- Select Small, Balanced, or Quality and start compression. The page plays the source locally, draws each displayed frame into a bounded Canvas, and records that Canvas stream with a WebM codec supplied by the current browser through MediaRecorder.
- Keep the tab open while the source plays in real time. A two-minute source therefore takes approximately two minutes to process rather than completing instantly, because the recorder observes the canvas stream at playback speed.
- Review the actual dimensions, input bytes, output bytes, and the signed percentage change that the page displays when the source ends.
- Download the WebM and play it from beginning to end before deleting the original. Watch for scenes with motion, dark areas, faces, and on-screen text, because compression problems are easiest to spot in those passages.
- Decide whether the result meets the destination's requirements. If it does not, return to the preset list and try another pass on the original, since re-encoding an already-compressed WebM usually loses more quality than re-encoding the source.
Verify Picture, Sound, and Duration in a Full Playback
The numbers on the page are a starting point, not the final review. Play the downloaded file on the device or platform where it will actually be used. Skim the file once to confirm picture clarity, then play it again with audio to confirm that the sound is present, in sync, and at a usable level. Open it in a different browser if you plan to share it widely, because MediaRecorder support varies; the page tries VP9, then VP8, then generic WebM, and the codec your browser picks affects how other software decodes the same file.
Audio presence is its own checkpoint. The recorder adds a track only when the browser exposes an audio track through HTMLMediaElement captureStream. If your source had sound but the output is silent, the issue is browser exposure rather than the tool's intent. The page never normalizes volume or mixes tracks; it either captures the track the browser makes available or it does not. The source plays audibly during real-time processing when it contains audio, so use device volume if you need to mute the room during a long encode, but do not assume that muting the tab has no effect on a captured track. Verify both picture and sound before deleting anything.
What to Review Beyond the Tool's Report
A short list of container features is not preserved by the re-encode, and you should check the destination's expectations for any of them. Subtitles, chapters, attachments, rotation tags, color metadata, and HDR signaling do not survive the Canvas-and-MediaRecorder path. WebM is the only output container because browser MediaRecorder support for generated MP4 remains inconsistent. If your destination needs an MP4 with subtitles, chapter markers, or HDR metadata, the local re-encode is the wrong tool for that final delivery, and a maintained desktop encoder is the right next step.
Confirm duration and seeking behavior as well. The page inserts a Matroska Duration float in timestamp-scale ticks so downloaded playback has a finite timeline, but downstream players still read that value through their own WebM demuxer. If a player reports an unknown duration, that is a container-metadata consequence of the local re-encode path rather than a missing frame. If a player refuses to seek past a certain point, the same explanation usually applies and the file may need to be remuxed with a desktop tool.
When the Result Needs Another Pass
Three common review outcomes should send you back to the preset list. First, the output is similar in size or larger than the source: low-motion or already-efficient material is the usual cause, and a stricter preset usually shrinks it further. Second, the picture shows blockiness, banding, or motion smearing during fast scenes: the preset pushed the bitrate too low for the content, and a higher preset or a different tool is appropriate. Third, audio is missing or out of sync: the browser did not expose a capturable track, or the capturable track drifted, and you may need a different browser, a different source file, or a tool that handles audio mixing explicitly. In each case, keep the original until the next run has been played from start to end, and consult a guide on fixing compressed results that look wrong when the issue is picture quality rather than size.
Keep the Original Until the Review Is Complete
Every value the tool shows is a report, not a guarantee that the file is ready for the destination. The page keeps source bytes, decoded frames, and output bytes inside the current tab; nothing is uploaded and no server ever sees your video. Use that local-only property to your advantage by replaying the output as many times as you need before deleting the source. When playback, byte change, output dimensions, audio, and codec all meet the destination's requirements, only then is the original safe to remove.
Related reading: When to Use Video Compression Instead of Doing It Manually.