Skip to main content

Cameras overview

Cameras in Toposync are more than a video URL. A camera can have multiple video sources, ONVIF discovery and metadata, live transmissions, physical placement in a composition, PTZ controls, and pipelines that turn images into events.

Start from the camera source, not from a manual pipeline. For normal live viewing, Toposync can generate the transmission and the continuous pipeline for you.

The camera model​

The main pieces are:

  • Camera: the device or logical camera entry.
  • Source: one video feed from that camera. A camera can have more than one source.
  • Role: the intended use of that source: main, sub, zoom, or custom.
  • Publication: the choice to expose a source as a live transmission.
  • Transmission: the generated live video artifact used by the player.
  • Implicit pipeline: the continuous pipeline that keeps a published camera source feeding its transmission.

In the regular flow, you should not need to create a Transmission manually. Enable live viewing on the source, and let Toposync reconcile the generated artifacts.

RTSP and ONVIF​

Toposync supports both RTSP and ONVIF, but they solve different parts of the camera workflow.

TechnologyWhat Toposync uses it for
RTSPReading video streams from cameras, NVRs, Frigate, go2rtc, or other local restream services.
ONVIFDiscovering cameras on the local network, reading camera metadata, capturing snapshots, finding stream profiles, receiving events, and controlling PTZ when supported.

Many cameras use both: ONVIF helps Toposync find and describe the camera, while RTSP carries the actual video stream. Some cameras expose RTSP without useful ONVIF support, and some networks block ONVIF discovery even when manual camera access works.

Use discovery when it works. If ONVIF discovery does not find a camera, add the camera manually with its RTSP or ONVIF address and credentials.

Source roles​

Use roles to describe how a source should be used:

RoleTypical use
subGrid tiles, thumbnails, low-cost monitoring
mainFullscreen viewing and inspection
zoomPTZ, autotrack, or low-latency interaction
customUser-defined rendered output or a special camera feed

The viewer chooses by context. A grid can prefer the sub source, while fullscreen can switch to main. This is different from choosing a technical quality profile.

Live viewing​

When a source is marked for live viewing, Toposync creates what is needed for that source to appear in the dashboard:

  1. a publication intent;
  2. a live-view variant;
  3. a generated transmission;
  4. a continuous implicit pipeline.

See Live viewing for transport behavior, playback states, and troubleshooting.

Pipelines​

Use automatic camera publication for ordinary live viewing.

Use a manual pipeline when you want to transform the video before publishing it, for example:

  • crop one corner of a camera;
  • overlay or filter an image;
  • publish a detection-focused view;
  • create a custom main, sub, zoom, or custom variant.

In manual pipelines, add the Transmit video operator and publish that output as a visible variant. You still should not need to create the technical transmission by hand.

Cinematic transmission is another optional pipeline helper: it links an existing Streaming transmission to a generated pipeline and cuts between cameras from notifications. That is a specialized workflow, not the normal camera live-view setup.

Mapping​

When the camera is placed in a composition, Toposync can use it for spatial context:

  • where the camera is installed;
  • which areas the camera sees;
  • how detections map into the world;
  • how live video can be projected in 360 View render modes.

See Camera mapping for calibration, PTZ poses, and spatial video projection.

What to validate​

After adding a camera, validate this in order:

  1. Snapshot works.
  2. At least one video source works.
  3. The source has the correct role.
  4. The source is published for live viewing if you want it on the dashboard.
  5. The dashboard shows the camera as live, warming up, or with a clear problem.
  6. PTZ works only if the camera supports PTZ and credentials allow it.
  7. Mapping is calibrated only after the image and live source are stable.

Avoid tuning advanced transport settings until one camera is reliable.