Practical hardware guide

Camera Hub background works in preview but disappears in another app

Check the processed virtual-camera route, supported RTX requirements and application selection when a Camera Hub effect is missing.

START HERE

If the effect works inside Camera Hub, select Elgato Virtual Camera in the receiving application. Selecting the physical Facecam or Cam Link can bypass that processed route. If the effect fails inside Camera Hub too, check the supported hardware and matching NVIDIA Video Effects SDK first.

A verified processed image in the actual destination app, plus a simple fallback camera choice.

In this guide
Before you start
  • A camera, NVIDIA RTX GPU and software combination supported by the current Camera Hub effects documentation.
  • A receiving application with a camera-device selector.

Before you change the setup

There are two useful checkpoints: Camera Hub's own processed preview and the image shown by the receiving application. Testing them separately avoids reinstalling effects software when the second app is simply using a different camera source. Start with one receiving app and close other camera consumers during the comparison.

Elgato's documented AI-background workflow requires a compatible RTX GPU and the NVIDIA Video Effects SDK for the correct GPU series. Those requirements are specific to this effect path. Do not assume that having Camera Hub installed, or seeing a normal webcam picture, proves the AI component is available.

Run a controlled check

  1. Confirm the upstream effect

    Open Camera Hub and select a simple background effect. Move a hand across your face and check whether the background visibly changes there. If it does not, note the exact error or unavailable control.

    Look for: You know whether the problem exists before the video reaches the destination app.

  2. Resolve requirements only if needed

    Compare your GPU model with the current official requirements and SDK download choices. Install the matching supported package through the official links and restart where instructed. Keep a note of the versions used.

    Look for: The software prerequisites match the documented hardware, rather than an unrelated package selected by a similar name.

  3. Select the processed device

    Open the destination app's camera settings and choose Elgato Virtual Camera. Compare it briefly with the physical camera entry so you can identify which image each provides. Keep Camera Hub running as required by its workflow.

    Look for: The destination shows the same intended background treatment, rather than the original physical-camera feed.

  4. Check the complete composition

    Make a local recording or private test in the actual app. Include hair edges, hands, headphones and anything you hold up. Check the picture after resizing it into your usual layout.

    Look for: The effect is acceptable in context; seeing a background change is only the first functional check.

  5. Rehearse under normal workload

    Repeat while the game and normal stream tools are running. Compare with the effect disabled if video becomes uneven. Keep a quick way to select the physical camera or a simple scene.

    Look for: You have evidence about this computer's workload and a usable fallback without assuming a universal performance cost.

Read the result and keep a working setup

If Camera Hub looks correct but the destination remains unchanged, inspect the selected device again and restart that receiving app after changing the route. If the destination shows a frozen image, check whether another app has claimed the physical camera. If the entire computer struggles only with the effect active, reduce optional processing and compare a local recording before changing stream settings.

Background replacement is an image treatment, not a privacy guarantee. Review the output yourself whenever private surroundings would matter, particularly during quick movement or when leaving the chair. The suggested rehearsal is editorial guidance rather than a Foxdip benchmark. No frame-rate, latency or segmentation-accuracy result is claimed for an untested computer.

Keep a short route note: physical camera into Camera Hub, processed virtual camera into the destination. That note becomes especially useful when several virtual-camera tools are installed. Choosing one processing path deliberately makes future troubleshooting simpler than stacking background tools in several applications.

Keep these distinctions clear

Keep these distinctions clear
CheckpointInterpretation
Effect absent inside Camera HubCheck supported requirements and SDK.
Effect visible only inside Camera HubCheck destination camera selection.
Effect visible but motion degradesCompare processing load under a local rehearsal.

Try it on your own setup

Start small: try the free files or continue with a guide for the part you’re working on.

OBS Virtual Camera on Discord: How to Set It Up ↗

Sources & references

Documentation review checked 4 November 2026 using the primary sources listed below. This is not a hands-on product test: the diagnostic sequence and rehearsal suggestions are Foxdip editorial guidance. No measured performance, search volume, price or universal compatibility result is claimed.

  1. Elgato Camera Hub — How to use AI Background Effects ↗Checked 2026-11-04

Spotted something out of date? Let Alex know.

How we put our guides together ↗
READ NEXT

A few places to go from here.

All stories

OBS Virtual Camera on Discord: How to Set It Up

Start OBS Virtual Camera, pick it as your camera in Discord, and choose what Discord sees: the whole program, one scene or one source. Set the mic apart.

Webcam Freezes in OBS When Another App Opens

Check camera ownership and reconnect the existing source before rebuilding the scene.