Sales FAQ

• What cameras do you integrate with?
• Do you have plans to integrate other PTZ cameras in the future?
• Can the system follow more than 1 person at a time?
• Can the system do ‘group framing’?
• Is Seervision compatible with live sports productions?
• How many Seervision servers do I need?
• Is a detected person permanently remembered even if they walk out of frame?
• Can the system run on its own without an operator?
• Can the system discern between different people wearing similar clothes?
• Can the system track the subject when a part of a person’s body is obscured?
• Can the system track subjects other than people?
• Can the Seervision Suite be used jointly with a green screen / chroma key?
• Can the Seervision Suite detect a person who is speaking, and track them?
• Can Seervision be used for augmented reality or virtual reality?

Technical FAQ

• What components is the computer unit (Seervision Server) made of?
• Does the Seervision Suite integrate with venue control infrastructures?
• Does the Seervision Suite integrate with external panels / PTZ controllers?
• How do you operate the Seervision Suite remotely?
• Does Seervision add latency on the video footage?
• How does Seervision work together with Shure’s MXA920 ceiling array microphones?

---

What cameras do you integrate with?

As of January 2023, the following camera models are compatible with Seervision. For optimal performance, FreeD capable cameras are recommended.

Panasonic PTZ Cameras:

• [NEW] Panasonic AW-UE160 (FreeD) – untested
• Panasonic AW-UE150 (FreeD)
• Panasonic AW-UE100 (FreeD)
• Panasonic AW-UE80 (FreeD)
• Panasonic AW-UE50
• Panasonic AW-HE40
• Panasonic AW-HE70
• Panasonic AW-HE130
• Panasonic AW-HE145
• Panasonic AW-UE20 – untested

Visca over IP PTZ Cameras:

• Sony
  • BRC-X1000 (FreeD)
  • BRC-H800 (FreeD)
  • BRC-X400 (FreeD)
  • Sony SRG X120

• Canon
  • [NEW] CR-N700 (FreeD) – untested
  • CR-N500 (FreeD)
  • CR-N300 (FreeD)

• BirdDog
  • P100 (FreeD)
  • P200 (FreeD)
  • P400

• Datavideo
  • PTC-280 (very limited support)

• Vaddio
  • Roboshot 30 NDI (limited support)
  • Roboshot 40 (limited support)

• QSC
  • Q-SYS NC-20X60 (limited support)

• PTZ Optics
  • PT20X-SDI (very limited support)

Robotic Heads:

• Ross Furio VR100
• Ross Furio VR600
• Ross CamBot XY
• Seervision Pauli Robotic Head

Do you have plans to integrate other PTZ cameras in the future?

At the moment Seervision supports the whole (flagship) Panasonic line AND most VISCA-over-IP cameras, which need to be tuned individually. Other camera control protocols may be integrated in the future.

Can the system follow more than 1 person at a time?

Currently each camera running on Seervision can track one person at any given time. The more cameras that are automated, the more people can be tracked individually (1 per camera). This doesn’t mean that the system doesn’t recognize multiple people at the same time. Up to eight people in the frame can be recognized simultaneously (before the system starts to slow down), and the operator can then select who to track. The subjects are available to be tracked as soon as they enter the camera view.

Can the system do ‘group framing’?

No, currently the Seervision Suite only follows one person at a time and does not account for other people in the field of view when framing a shot. We can only do group framing based on audio information, such as when using Shure’s MXA920 ceiling microphones.

Is Seervision compatible with live sports productions?

We don’t currently support any sports use-cases beyond a prototypical level. In the past Seervision has only been used for tennis productions. We are, however, constantly developing our solution in a way that will enable us to overcome the technological barriers brought by computer vision, for example, with external UWB tags.

How many computer units (Seervision Servers) do I need?

With SVs1-2U you can automate 1 camera per server (2U Rack Mount)

With the SVs4-2U you can automate up to 2 cameras per server (also a 2U Rack Mount).

We are working on increasing the amount of cameras you can automate per server. This will come over time via software updates and some hardware upgrades.

Is a detected person permanently remembered even if they walk out of frame?

By default, each person currently visible in the frame is assigned a unique ID. Seervision however does not “remember” talents for longer than 8 seconds if they walk out of the frame. If they remain in the frame but are fully obscured, Seervision will continue to search for the talent for 5 seconds before “forgetting” the talent’s identity.

If a person walks out of stage and comes back after a while, they will have a new ID and will have to be tracked again as the system won’t remember them.

It is possible to enable Person Specific Automation features that persistently store the identity of a person, for example by uploading a picture. For more information, see our Person Specific Automation manual.

Can the Seervision Suite run on its own without an operator?

Yes. Our main focus at Seervision is to enable more autonomous workflows for educational, stage and corporate use-cases by using features such as ‘trigger zones’ and the ‘exclusion zones’ and to reduce (or completely eliminate) the need for an operator during live productions through workflow automation.

Our system is designed to adapt to the needs of the user and use-case, whether this is used with an operator to supervise, or as a fully autonomous lecture-capturing system.

Can the system discern between different people wearing similar clothes?

Yes. The Seervision Suite is tuned to function even when the subjects are wearing similar clothes. There might be some edge cases where tracking might fail, for example, in poor lighting conditions and/or when they wear exactly the same clothing and they are moving very fast (team sports), but for the majority of use cases, the system should perform very robustly in these cases.

Can the system track the subject when a part of a person’s body is obscured?

The system is able to keep track of the person in the majority of use cases, even when lower portions of the person’s body become hidden. However, Seervision needs to see the subject at least from their shoulders and above.

However, the system’s performance depends on the zoom level to an extent. The system would not be able to track consistently if the person’s image is too small in the video frame (less than 30% of the frame height-wise), as there wouldn’t be enough pixels of the subject to create an ID and track them reliably.

Can the system track subjects other than people?

Not currently. The current algorithm is designed and optimized specifically to track people. Animals, vehicles, or any other subject will not be tracked by the system.

Can the Seervision Suite be used jointly with a green screen / chroma key?

Yes. It performs even better because the background is plain.

Can the Seervision Suite detect a person who is speaking, and track them? 

This is possible with certain beam-forming microphones. We recommend the Shure MXA920 as it gives us very precise coordinates about the speaker’s location. This allows us to translate the information of where the sound is coming from and point the camera in that precise direction.

Can Seervision be used for augmented reality or virtual reality?

The Seervision Suite cannot provide information to AR or VR solutions.

Nevertheless, if the camera or robotic head that is used can send this information (i.e. output FreeD data) and it does not interfere with the position feedback that the Seervision Suite needs, then Seervision can work alongside those solutions.

 

Technical FAQ

What components is the computer unit (Seervision Server) made of?

Weight: 11 kg / 24 lbs
Operating System: Ubuntu Linux 18.04 LTS Bionic Beaver
Access: Via LAN, needs to be on the same network as the Seervision Server
Control: Mouse and Keyboard, or Skaarhoj PTZ Extreme, or API
Supported Video Protocols: SDI, NDI, and NDI|HX

Details	SVs1-2U	SVs4-2U
Form Factor	2U rack mount	2U rack mount
CPU	i5-12500 (12-gen)	i7-12700 (12-gen)
GPU	Nvidia RTX 3060 Ti	Nvidia RTX 3080
RAM	16GB DDR4 RAM	32GB DDR4 RAM
SSD	250GB SSD	500GB SSD
SDI Grabber Card	Blackmagic DeckLink Mini Recorder HD	Blackmagic DeckLink Duo 2 Mini
Power Supply	450W Power Supply	600W Power Supply

 

Does the Seervision Suite integrate with venue control infrastructures?

The Seervision Suite can be integrated with venue control infrastructure like Crestron and AMX via Seervision’s API. We typically use Node-RED, which allows you to make logic-based flow diagrams.

Does the Seervision Suite integrate with external panels / PTZ controllers?

The Seervision Suite currently has a direct integration with the Skaarhoj PTZ Extreme panel, this means you can use the joystick simultaneously whilst tracking. Furthermore, Seervision can work in parallel with other hardware controllers, but not simultaneously. By clicking the ‘Release’ button within the Seervision User Interface, you can take over manual control with a joystick, and then click ‘Regain’ once you want Seervision to take over control again. The Seervision Suite and the joystick can’t both send commands to the camera at the same time.

How do you operate the Seervision Suite remotely? 

To operate the Seervision Suite remotely, follow the instructions here.

Does Seervision add latency on the video footage? 

Seervision adds no discernible delay to the video feed when tracking is used. Seervision also does not affect the image quality. The system only sends commands to the camera on how it has to move.

In regards to the latency of manual control, there is some delay depending on the LAN latency and the latency introduced by the different protocols used. This latency is minimal.

How does Seervision work together with Shure’s MXA920 ceiling array microphones?

It works by processing the raw 3D positional information from the active speaking participant that the microphone provides. This data allows you to build dynamic single or multi-camera workflows that  adaptively point towards participants in the room – without having to depend on presets. You can read more details here.