Every explainer video starts with math. Take the total duration of your motion-graphic video in seconds and multiply it by 2.5. That number is your maximum word count for the narration. A 60-second video gets 150 words. A 90-second video gets 225. This keeps the script tight and the audio from racing ahead of the visuals. Going over the budget is the most common mistake. The voice ends up competing with the screen instead of guiding it.

Writing a Script That Works With the Visuals

The script should complement the visuals, not narrate them. If the screen already shows a checklist appearing item by item, the voice does not need to read each item out loud. Instead, it speaks to the meaning behind what is shown. It answers the question the viewer is silently asking: why does this matter to me? Short sentences work best. One idea per line. When you write the script this way, the final video feels like a conversation, not a lecture.

Once the script is ready, the avatar video is generated through HeyGen. The avatar is set to cover fit inside a vertical or square crop so it fills its frame cleanly without awkward borders. Balay ni Bruno & Co. generates the avatar clip as a standalone video file first, separate from the motion graphic. This makes the compositing step much easier to control.

The bookend approach is often cleaner than a full overlay. Let the avatar open the video, hand it off to the motion graphic, then close with a direct call to action. The viewer gets a human face at the moments it matters most, and the explainer content gets its own space to breathe.

Two Ways to Composite the Final Video

The first approach is full Picture-in-Picture. The motion graphic runs at full size and the avatar sits in a corner, visible throughout. FFmpeg handles this by layering the two video files, positioning the avatar window at a fixed coordinate, and rendering a single output file. The timing is matched so the voice and visuals stay in sync from start to finish. This works well for technical walkthroughs where the audience benefits from a visible presenter the whole way through.

The bookend approach is the cleaner option for most content. The avatar opens the video with a 10 to 15 second introduction, the full motion graphic plays uninterrupted in the middle, and the avatar returns at the end with a direct call to action. FFmpeg concatenates the three segments into one file. The result is a video that feels personal at the beginning and end without the distraction of a floating face over detailed graphics. Balay ni Bruno & Co. uses this format as the default for client explainer series because it respects both the human element and the visual content equally.

Key Takeaways

  • Word budget: duration in seconds x 2.5 = maximum narration word count.
  • Avatar as bookend intro and outro is often cleaner than a full narration overlay.
  • cover fit on HeyGen avatar generation fills the frame and prevents letter-boxing.
  • FFmpeg composites both video layers without re-rendering either source.