How to check a video editor’s export before you subscribe
Will a free editor deliver the file you promised a client? We took the same sentence out of the same four-second recording in two free cutters. Both exports kept the original 800 × 400 picture. One file was nearly ten times larger. You can download both and inspect the difference yourself.
See the measured exports · Check your own file · Find a spoken line after a cut
One recording, the same cut, two MP4s
On November 15, 2026, we removed source seconds 0.200–1.450 from a four-second H.264 recording in Valmera’s browser cutter and Ezgif’s online cutter. That leaves 0.200 seconds of lead-in before the remaining footage and a target running time of 2.750 seconds. Both runs used the removal mode with its default encoding settings. A person selected the cut in each tool.
| File | File end | Picture | Audio | Bytes |
|---|---|---|---|---|
| Original MP4 | 4.000 s | 800 × 400 · H.264 | AAC · 16 kHz | 74,834 |
| Valmera MP4 | 2.750 s | 800 × 400 · H.264 | 16-bit PCM · 48 kHz | 587,214 |
| Ezgif MP4 | 2.888 s | 800 × 400 · H.264 | AAC · 16 kHz | 59,073 |
Why the size changed. Valmera’s stereo PCM audio occupied 528,000 bytes of its file. Ezgif’s compressed AAC audio occupied 23,511 bytes. Picture coding and container overhead account for the rest. If you have an upload cap, weigh the finished file. Its dimensions alone tell you little about its size.
Why the displayed lengths can differ. The Ezgif file’s video starts at 0.080 seconds and ends at 2.840. Its audio starts at zero and ends at 2.888. A player rounding the running time can show “0:03” for either export. Compare the actual track times when a cut must fit a precise slot. These measurements describe this recording and these settings.
For a dialogue check, listen for “Um, use the green one for print.” The blue-label sentence should be gone. With Valmera’s companion SRT, the surviving line starts at 0.310 seconds. Review the first word, the last word and the join in the player where you plan to publish.
Where did the same line land after the cut?
The four-second source says “Um, use the green one for print.” Take source audio from 2.560 to 3.280 seconds and look for the same sound in each downloaded edit. Removing 1.250 seconds earlier in the recording should move this excerpt to 1.310 seconds. A picture-only cutaway should leave it at 2.560.
| Edit | Expected | Found | Difference |
|---|---|---|---|
| Valmera sentence cut | 1.310 s | 1.310 s | 0 ms |
| Ezgif sentence cut | 1.310 s | 1.430 s | 120 ms later |
| Valmera MCP picture cutaway | 2.560 s | 2.560 s | 0 ms |
The first two rows use the manual cuts above. The third uses an edit made through the AI connector. Choose a row below to repeat its check. To look for the sentence that was removed, change the source range to 0.440–1.200 seconds. The two sentence cuts give no confident match for that range.
For your own files, select two short recordings with the same voice and choose a distinctive source excerpt of 0.30–2.00 seconds. The browser reads files locally, up to 90 seconds and 32 MiB each. WAV works across browsers. Other formats depend on browser decoding support.
For the examples we decoded the first audio track to mono at 8 kHz with FFmpeg. The finder searches for similar loudness patterns and compares the underlying waveforms around the strongest candidates. It reports uncertain or repeated matches. A dubbed, time-stretched or heavily remixed voice can prevent alignment. The measured offset locates sound within the audio. To check lip sync, also inspect a visible consonant or clap on the picture timeline.
Check your downloaded video
Choose an exported MP4, MOV, WebM or MKV. Add the source or a second export to compare the files side by side. Your browser reads both on your device.
“File end” is the last media timestamp. Audio and video can start or end at different times. A negative audio start can reflect encoder priming. Play the file to check the actual sound, captions and watermark.
Check what the file properties can’t show
- Branding. Watch through the final second and look in all four corners. A template can add an ending card even when the rest of the clip is clean.
- Speech. Listen on headphones across the cut. An audio track can contain silence, a clipped consonant or the wrong language. Its presence only tells you that audio was packaged.
- Caption delivery. Turn the player’s captions off. If the words stay visible, they are part of the picture. For a separate SRT, open the actual download and match a later cue to speech after the cut.
- The next export. Repeat the edit with the AI feature and assets you plan to use. Record whether the second export needs more credits, a higher plan or a different format.
To choose a plan, compare the free editors’ export limits. For a conversational edit, bring the original recording and describe the change in Valmera. Hosted AI editing requires an eligible subscription.