How to Fix Eye Contact in Camtasia

Option 1: correct it live with NVIDIA Broadcast (Windows, RTX only)

If you record on Windows with an NVIDIA RTX graphics card, NVIDIA Broadcast is free and includes an Eye Contact effect that corrects your webcam feed in real time. It appears as a virtual camera, so you select it as the camera source in the Camtasia Recorder and the corrected image is what gets recorded. TechSmith's own compatibility page lists NVIDIA Broadcast 2.0 and later as tested with Camtasia on Windows, with Eye Contact among the tested features.

For people who have the hardware, it's the best option for future recordings: no extra step and no cost. It has real limits, though:

• It needs an RTX card on Windows. Camtasia on Mac, and Windows laptops without an RTX GPU, can't use it.
• The correction is baked into the recording. If it looks wrong on a take, there's no uncorrected version to fall back on.
• It can't fix anything you've already recorded — your existing lesson library stays as it is.
• It runs on your GPU while Camtasia is also recording, which can matter on a laptop that's already working hard.

Option 2: correct only the webcam, after recording

If live correction isn't available to you, or the footage already exists, correct it afterwards — but correct the webcam footage, not your finished export. A finished Camtasia video is mostly screen recording, which has no eyes in it, and in a picture-in-picture layout your face may be a small bubble in the corner. Correcting the whole export means paying for every minute of screen demo, and giving the correction a face only a few dozen pixels across, which gives poor results.

Exporting the camera stream at its own full resolution fixes both problems: the correction sees your face at full size, and you pay only for the minutes the camera was rolling. File-based tools such as GazeFix work this way. Captions and Descript also correct eye contact, but inside their own editors, which means importing the webcam clip there, correcting it, and exporting it back.

Workflow: correct the webcam before you edit

The cleanest routine is to fix the camera stream straight after recording, before you make any cuts. Then the edit is built on corrected footage and nothing needs re-syncing later.

• Separate the streams. In current versions of Camtasia a recording arrives as Rev media; right-click it and choose Unpack Rev Media to put screen, camera and audio on their own tracks. Note that unpacking is one-way — you can't send it back to Rev afterwards. In older versions, a .trec recording with a camera already places the camera on its own track.
• Export the camera at full size. Create a new project whose canvas matches your webcam's resolution — 1920×1080 or 1280×720 for most webcams — add the camera clip and your microphone audio, make the camera fill the canvas, and export an MP4. Keep the audio in: it makes it easy to check sync later.
• Correct the file. Upload it, and download the corrected MP4. GazeFix keeps the clip's duration and audio.
• Swap it into your project. Import the corrected clip into the Media Bin and drag it onto the original camera clip on the timeline, choosing to replace it. TechSmith's replace feature keeps the clip's properties, transitions, effects and animations, so your picture-in-picture size and position carry over.
• Check sync, then edit as normal. Play the last few seconds and confirm your lips still match your microphone audio.

Already edited? Correct the camera track from the finished timeline

Once a lesson is cut, the camera is in many pieces, and replacing each one by hand is tedious. A faster route is to export the edited camera track as one file, correct that, and lay it back over the top.

Select all the camera clips on their track and group them, then temporarily scale the group up to fill the canvas. Disable the other tracks so only the camera shows, and export the full timeline with your audio. After correcting, import the result, place it on a new track above the original camera track starting at the very beginning of the timeline, and scale and position it to match your picture-in-picture layout. Then disable the original camera track and undo the temporary scaling. Because the corrected file covers the whole timeline from zero, every cut lines up.

Only correct the parts where your face is large

Most Camtasia training videos show the webcam full-screen for the introduction and wrap-up, and as a small picture-in-picture during the demo. Gaze drift is obvious at full size and barely noticeable in a small corner bubble, especially when the viewer is watching your cursor.

So for many lessons, the efficient approach is to correct only the full-screen camera segments. Export just those stretches, correct them, and replace them; leave the picture-in-picture sections alone. A ten-minute lesson might have two minutes of full-screen camera, which is two credits instead of ten. Every upload costs at least one credit and is rounded up to the next minute, so put several lessons' intros back to back in one export and correct them together.

Export settings that keep the correction clean

• MP4 at your webcam's native resolution, 1080p or smaller. Larger files are scaled to 1080p anyway.
• Include your microphone audio. It's how you'll confirm sync after correction.
• Keep it under 500 MB, the upload limit. Webcam footage compresses well, and around 5 Mbps is plenty for a talking head — that fits roughly 13 minutes into 500 MB. Split longer recordings.
• Your face reasonably large in the webcam frame. If you sit far back from the camera, your eyes are too few pixels to correct convincingly.
• Light your face from the front. A bright window behind your monitor leaves your eyes in shadow, and the correction needs to see your pupils.

Sync, frame rate and resolution

Webcam streams are often recorded at a slightly variable frame rate. GazeFix converts variable-frame-rate video to a constant rate (at most 60fps) while keeping the clip's duration, so the corrected camera should line up with your screen and microphone tracks. Check the end of long recordings anyway, since that's where any drift shows first.

Output is watermark-free at up to 1080p. That's the native resolution of most webcams, so in practice nothing is lost; a 4K webcam's footage comes back at 1080p, which is still larger than a picture-in-picture bubble ever needs.

What it costs for a training library

One credit corrects one minute of video. A training team publishing ten lessons a month, each with three minutes of full-screen camera, uses 30 credits a month or 360 a year. The Studio pack is 600 credits for $199, about $0.33 a minute, and the Pro pack is 200 credits for $79, about $0.40. Credits never expire, so a quiet month between course launches costs nothing, and there's no subscription to keep paying for.

If you already pay for Descript or Captions and edit there, their built-in eye contact may cost you nothing extra. The case for a file-based tool is that you keep editing in Camtasia and don't add a second editor to the workflow.

Record so there's less to fix

Every option works better when your gaze was close to the lens in the first place. Put your notes in a narrow window directly under the webcam rather than on a second monitor, use bullet points instead of full sentences, and look into the lens at the start and end of each section. Our guide to reading a script on camera covers the setup in detail.

Then test on a lesson you know is bad. GazeFix gives you one free credit when you sign up, with no card required — enough for a full minute. Export one minute of full-screen webcam from your worst intro, correct it, and swap it in.