Nvidia Broadcast Eye Contact Setup
557 words · Reviewed for accuracy

NVIDIA Broadcast has an Eye Contact feature that subtly adjusts your gaze on video so you appear to look at the camera even when you're actually reading faces on your screen. For video interviews — where natural eye contact with the lens feels deeply unnatural — it's a genuinely useful piece of kit. Here's how to set it up properly and use it wisely.
The key idea: eye contact tech is a polish layer, not a performance layer. It fixes the camera-gaze problem so your attention can go where it belongs — on the conversation. Get your core video setup right first; this is the finishing touch.
What you need:
- A compatible NVIDIA GPU (RTX-series; the app checks on launch)
- The free NVIDIA Broadcast app installed (or the feature's newer home in the NVIDIA app, depending on your driver generation)
- A decent webcam — the feature works with most standard USB and laptop cameras
- Enough system headroom to run it alongside your video call without the fan drowning your audio
Setup, step by step
- Install and open the app. Let it update its drivers and effects models on first run.
- Select your camera in the app's camera tab, then enable the Eye Contact effect. You'll see the preview update immediately.
- Tune it while looking naturally at your screen. Read a paragraph of text off your monitor and watch the preview — the corrected gaze should look steady but not stare-locked.
- Route it into your meeting app. In Zoom, Meet, or Teams, choose "NVIDIA Broadcast" (or the virtual camera it exposes) as your camera source — not your raw webcam.
- Test with a friend or a recording. Do a full mock call and review the footage. Check edge cases: looking far off-screen, glasses glare, and quick head movements.
Using it wisely in interviews
Keep the effect on the subtle side — the goal is "engaged person," not "unblinking news anchor," and overly aggressive correction can drift into uncanny territory, especially at steep gaze angles. Position your interviewer's video window directly under your webcam anyway; the smaller the correction the software has to make, the more natural you look. And always have a fallback: if the app crashes mid-call, knowing how to switch back to your raw camera in five seconds is worth more than the effect itself. This pairs naturally with the delivery fundamentals in the body language guide — tech handles the gaze; you handle everything else.
What to avoid
- Testing it for the first time on interview day. Burn it in across several calls first.
- Relying on it while reading notes off-screen. The correction has limits, and reading patterns are visible no matter what your eyes do.
- Stacking it with heavy virtual backgrounds and beauty filters. One invisible enhancement reads as polish; three read as a render.
FAQ
Is using eye-contact correction dishonest? It's a presentation tool in the same family as good lighting and a clean background — it doesn't change what you say or know. Interviewers care about your answers; this just keeps the camera from misrepresenting your attention.
What if I don't have an RTX GPU? The low-tech solution works remarkably well: shrink the call window, drag it directly beneath your webcam, and glance at the lens when making key points. Practice it in a few mock sessions and it becomes second nature.
Put this into practice
Continue with the Aissence workflow this guide supports.