Video footage can be an essential source of evidence in criminal investigations. Over 80% of criminal cases today are estimated to use video evidence, including an increasing amount of DVR surveillance video, body cams, and dash cams, but also video from doorbell cameras, home security solutions, or mobile phones. Recovering and analyzing evidence from these sources can present challenges without the right tools.
Built on the foundation of Magnet DVR Examiner, Witness enables you to acquire, review, and analyze video evidence, all with one advanced product. Witness recovers more video evidence than ever, including native and proprietary video formats, video from dashcams and body cams, plus cloud sources like Ring or Arlo. Review, analysis, and output are easier with new capabilities like synchronized review, sub-clipping, and more. Whether you’re in the lab or reviewing video evidence, Witness supports your entire workflow.
Quickly bring video clips together with video concatenation
DVRs record motion events triggered by people, vehicles, and other subjects in short clips, typically only a few seconds long. Viewing these motion events as separate files can be tedious and time consuming, both for investigators and in court.
Witness’ new video concatenation feature allows you to select and stitch multiple clips, or sub-clips, together so that they can be reviewed as a single video. With this feature, you can create timeline montage videos that follow specific subjects, simplify ingestion into other tools, and create more compelling court presentations.
Concatenation allows users to add clips to a playlist and then generate a concatenation job from that playlist. Users can select a particular encoder/codec and other configuration options, and a full report detailing the specific configuration and clips is included.
Automatically identify people and vehicles with video analytics
With the large volume of video evidence involved in the typical case, thoroughly reviewing it to identify subjects of interest can be a painstaking process. New analytics capabilities in Witness provide an easy and reliable method to reduce the amount of manual video acquisition and review that needs to be conducted during an investigation.
In the initial beta release of this feature, you can automatically identify video clips containing people or vehicles, greatly reducing your time to evidence by finding key videos of interest before processing.
For example, for a subject vehicle described as ‘white pickup truck’, a user can run a vehicle analytics job, select all clips containing vehicles and run a gallery preview to quickly scroll through thumbnails to find any white pickup trucks.
Analytics can be configured from the cliplist or playlists. Simply select the clips you wish to process for analytics and either right click or select Analytics from the toolbar. Users can select whether to find people, vehicles or both, and configure a confidence threshold applied to the identification process based on the sources/quality of the video to find the best results. Analytics jobs can also be configured as a post-scan job in the Case Wizard—this is helpful if you wish to run large analytics jobs across multiples sources in the background when building your case.
Once an analytics job has completed, clips containing persons or vehicles within the configured confidence threshold will receive a tag indicating the object type detected. The analytics job can be re-run to adjust confidence threshold settings, with tags added or removed based on the updated settings.