Recording
VXG supports 3 types of video recording:
Cloud recording
On-prem Cloud NVR recording.
Camera SD recording.
Regardless of the recording type, video can be watched and managed from a single user interface.
Video can be recorded either continuously or by-even, such as motion or object detection. In the case of event recording, pre and post-event intervals can be defined.
The cloud and on-prem storage are S3 compatible which makes VXG interoperable with the major storage providers. VXG partners with AWS and Seagate for cloud storage. Cloud NVR installs an S3 driver converting on-prem storage into S3 compatible object storage.
Video is recorded in mp4 chunks making the video format HTML5 compatible.
The depth of the video recording can be configured per camera. For example, one camera can have 90 days of recording, and another having 3 days.
Playback
VXG supports multi-format streaming to make the video compatible with any video player.
A low-latency HTML5 compatible format (WebRTC) for applications when you need to control your camera using PTZ or cameras with bi-directional audio.
The most reliable HTML5 format (HLS) that provides the best user experience even in any network.
Image polling (JPEG polling) for low bandwidth networks or when you need to watch many streams simultaneously (grid view).
Non-HTML5 standard formats (RTMP and RTSP) for interfacing with VA and AI engines.
Broadcasting using Content Delivery Networks (CDN) allows sharing video with an unlimited number of viewers.
As well, VXG offers an HTML5 video player, and native Android and iOS players with the following functions:
Live playback with a selection of formats.
Timeline, calendar, and playback of recorded video.
Scrubber (dragging timeline with immediate response).
Backward and forward multiple speed playback.
Digital zoom and full-screen mode.
De-warping for fish-eye cameras.
Bi-directional audio.
Time-lapse playback of recorded data.
Processing
VXG performs video pre-processing for Video Analytics and Artificial Intelligence. Using video pre-processing saves up to 100 times of the operating costs as well as saves significant bandwidth. For example, instead of ingesting video into an AI engine, VXG will extract images when motion is detected on a camera and will only send extracted images to the AI engine.
The platform supports the following operations:
Extracting images from the live video at any sample rate or by the event.
Extracting batches of images from recorded video.
Extracting clips on-demand and by the event.
All operations are fully auto-scaling and very fast. For example, if you need to extract an image from 10,000 cameras it can be done with a single API call.