Troubleshoot Redaction

Issues with video playback or redaction

Customers occasionally encounter problems with specific videos or video types during activities such as playback and redaction. In order to investigate these issues, customers need to share the original files with Axon's Engineering team.

If you're experiencing this type of issue, please email Axon Technical Support ([email protected]) with information about your issue and they will determine if your issue qualifies for investigation. If it does qualify, Axon Technical Support will email instructions to you explaining how to deliver this video to Axon. Please do not send Axon data from your Axon Evidence account via email, physical mail or file sharing services such as Box, Dropbox or OneDrive.

Additionally, by sharing a video with Axon you are authorizing Axon personnel to view the video. We advise you to consult with and get approval from your agency administrator or other appropriate management and counsel prior to sharing videos or other data.

When data is shared with Axon personnel using approved channels, we commit to the following:

  • Secure handling of data
  • Encryption of data in transit and at rest
  • Only Axon personnel that have undergone Axon's personnel security process, including CJIS background checks, will have access to the data
  • Viewing/playback of data only occurs at physically secure locations located in Axon offices in the United States
  • The shared data is retained for no longer than 90 days
  • Chain of custody reports will be added as 'Notes' on the evidence file

Invalid frames in Redaction Studio

What are invalid frames?

Redaction Studio decodes the video and audio packets to retrieve frames from them. A video packet can be made up of one or more multiple frames and an audio packet also typically usually contains one frame. A video or audio packet has invalid frames if Redaction Studio can’t decode a packet either because of incorrect input or because there’s no frame in the packet.

How does Redaction Studio identify invalid frames?

To scan the video and identify invalid frames, Redaction Studio creates a new decoder and then decode frames to find any errors. If Reaction Studio finds no errors, the frames are valid.

What causes invalid frames?

Several things can cause invalid frames:

  • If the recording device has a low battery level. This can either add data from an incorrect input to the video file or create an empty packet.

  • Video file issues such as picture artifacts or frozen frames.

What should you do if there are invalid frames in the video?

  • If the video was recorded on a device from Axon, check the device logs to see there was an issue with the video being captured and uploaded.

  • View the video with a local playback application, such as VLC player, to see if there are any issues with the video file such as picture artifacts or frozen frames. This applies to both videos recorded by Axon devices and video from other sources.

How does Redaction Studio handle invalid frames?

Several kinds of errors can occur during decoding.

Because invalid data input is the most common error, Redaction Studio applies a percentage calculation of invalid data input error in an invalid segment to determine if an invalid segment is either skippable or un-skippable:

  • A skippable invalid segment:

    • Has more than 30 consecutive errors

    • Invalid data input is greater than 50%

    • Has less than 30 consecutive errors. These are ignored.

  • For other cases, the invalid segment is un-skippable and the process stops.

How does Redaction Studio treat a skippable invalid segment?

  • Audio sampler: Redaction Studio replaces the invalid audio frames with zero (representing silence) when drawing the audio wave form.

  • AI and object trackers: Redaction Studio ignores them

  • Extraction:

    • If Redaction Studio finds the invalid segment in middle of a video, Redaction Studio replaces the invalid audio frame with silence, and the invalid video frame with a black frame.

    • If Redaction Studio finds the invalid segment at the end of a video, Redaction Studio picks the last valid frame found in the audio and video stream, and uses the last frame for the duration of the video. This duration could be different from either the duration of the metadata, or when examining the video). Redaction Studio then adds either silence or black frames if the last valid frame of the video or audio stream is less than the last valid frame of the file.

Why do we identify the last frame and use it as the duration of video?

An unexpected error can occur if there’re no video frame found while the audio frame is still active. The duration listed in the metadata isn’t totally true always accurate.

For example, a case with a 3-hour video with 3h duration but it’s may only play for 15 seconds when playing.