FAQs

Swear protects digital assets in real-time by mapping and fingerprinting every frame, pixel, and sound bite, while simultaneously watermarking nearly 100 dimensions of attribution data into each second of a recording. Swear then uses the fingerprints and watermarks to construct digital DNA that is permanently secured in a blockchain system. This all happens during recording, not after. The actual recording never leaves your device; only the digital DNA is stored in the blockchain.

When it comes time to verify the authenticity of a recording, Swear examines a digital recording by recreating the digital DNA and comparing the fingerprints and watermarks against the original recording’s DNA stored in the blockchain network. If a single pixel has been modified or a sub-second of sound changed or any of the dimensions of attribution data manipulated, Swear will flag that portion of the recording as inauthentic.

How accurate is Swear?

Swear can detect changes in a single pixel, a sub-second of sound, or any alterations in watermarked attribution data points.

What is the “Confidence Score”?

The Confidence Score is a summary of a frame-by-frame and second-by-second evaluation of a recording. The Confidence Score is comprised of an examination of four categories: 1) WHAT was recorded (visual elements and audio aspects), 2) HOW it was recorded (~100 dimensions of attribution data), 3) WHERE the recording occurred, and 4) WHEN the recording was created. The formula for the Swear Confidence Score is:

40% – Visual Accuracy
20% – Audio Accuracy
20% – Metadata Accuracy
10% – GPS/Location Accuracy
10% – Time Accuracy

Here is an example Confidence Score of a video a 1-minute video (1,800 frames) that has 6 seconds (180 frames) with changed visual elements and 10 seconds (300 frames) with a weak GPS signal:

40% * (1620/1800 frames) = 36% : Visual
20% * (1800/1800 frames) = 20% : Audio
20% * (1800/1800 frames) = 20% : Metadata
10% * (1500/1800 frames) = 08% : Location
10% * (1800/1800 frames) = 10% : Time
Confidence Score = 94%

It appears that location and time are important in the Confidence Score. Why?

Swear believes that protecting what, where, when, and how a recording is created proves its authenticity. For instance, if someone turns off their GPS, and we don’t know where a recording occurs, we believe viewers should be informed. Swear also doesn’t use the device’s time to validate the time, the Swear system uses the world clock to time stamp the recording. So, if the recording device isn’t connected to the internet, we technically don’t know when a recording occurred, so we inform viewers of the time challenge.

Why do you not put weight in anonymous VS non-anonymous recordings?

Although knowing WHO recorded something is helpful, we identified many environments where the person recording the system may need to remain anonymous for their protection.

If a confidence score is less than 100%, should I not trust the recording?

Examine the reason behind a Confidence Score less than 100%. It could be due to temporary GPS loss or other factors. However, non-100% Confidence Scores might indicate edits or manipulation in the recording.

When is my recording actually protected?

Your recording is protected in real-time during recording, not after. If you are four minutes into creating a recording, the prior three minutes and 59 seconds are already fingerprinted and protected within blockchain.

How do I know a video is protected?

Swear doesn’t just tell you a recording is authenticated; it shows you. Every recording is examined frame-by-frame to create a Confidence Score. The Score is recalculated every time you share your recording or publish it on your Swear.com channel.

If someone else sends me a recording, how do I evaluate it to see if it is authenticated?

Upload the media into your Swear app and it will be automatically examined frame-by-frame & and second-by-second.