Audio redaction software
Audio redaction software, in Video Blackout, means covering a span of sound in the exported file so a spoken name, a phone number, or a radio call sign does not survive disclosure. The cover is applied at encode time: those samples are muted or replaced with a beep. The original file on disk is not overwritten.
Definition
Video redaction is usually discussed as pixels. Records officers still have a second channel: the soundtrack. A blurred face next to a clear “that’s Officer Ramirez” is still an identity. Audio redaction is the control for that channel. In this product it is manual only.
What is supported (and what is not)
| Capability | Video Blackout |
|---|---|
| Mark a start time and end time on the timeline | Yes |
| Multiple ranges on one clip | Yes |
| Mute or beep those ranges in the export | Yes |
| Automatic speech recognition / “find every name” | No |
| Speaker diarization or keyword spotting | No |
| Upload audio to this website | No — desktop only |
What buyers mean by audio redaction software
The commercial search is for a product that can withhold speech in a video export: a name, a phone number, a radio call sign. In Video Blackout that product is manual ranges plus mute or beep, in the same Windows app as the picture. Three things it is not: automatic speech recognition, a separate audio-only suite, and an upload site. If a shortlist assumes the software will “find every name,” it is scoring a feature this product refuses to ship.
Why we do not auto-spot speech
Automatic “redact every name” demos fail on bodycam radio, overlapping voices, non-English speech, and noisy school corridors. Worse, a false negative ships an identity; a false positive erases speech your memo required to keep. Manual ranges make the disclosure log boring in a good way: an operator chose 01:12–01:19. A second reviewer can replay the export and hear the gap without trusting a model score.
How operators actually use it
- Play the local file in the Windows app. Do not upload it.
- When speech must not be released, mark the range. Repeat for each span.
- Choose mute or beep per your house style.
- Export. Video redaction (faces, plates, other boxes) and audio ranges ship in the same MP4.
- Keep the original as the system of record. Circulate only the export.
This is slower than a fantasy “auto-redact audio” product and more defensible in a FOIA or DSAR log.
Offline audio redaction
Soundtrack work happens on the same PC as the picture. You do not upload WAV or MP4 to finish a mute. That keeps victim names and informant call signs inside the controller’s environment. See offline video redaction.
When audio is not enough
A muted name does not hide a face or a plate. Combine ranges with face redaction (auto-detect plus manual boxes) and license plate blur (same two modes). Other on-screen identifiers are operator-drawn: PII video redaction.
Bodycam radio traffic: bodycam redaction software. Access requests: GDPR video compliance, FOIA video redaction. Procurement wording: video redaction software. Research and ethics: video anonymization. All capabilities: features overview.
FAQ
- Does Video Blackout automatically detect speech or names in audio?
- No. There is no speech-to-text and no automatic mute. An operator marks start and end times. The export mutes or beeps those ranges.
- Can I redact audio without uploading the file?
- Yes. Audio ranges are marked in the Windows desktop app. The website does not accept audio or video uploads.
- Is audio redaction the same as face or plate redaction?
- No. Faces and plates have auto-detect plus manual boxes. Audio is manual time ranges only.
- Does mute leave the clear speech underneath?
- No. Selected ranges are rewritten in the export soundtrack. Keep the original file if you still need an unredacted master.
- Why not ship automatic name spotting?
- Speech systems invent confidence and miss accents, radio noise, and overlapping talk. For disclosure logs, a human-chosen timestamp is easier to defend than a model guess.