FAQ / Recording a clean course voiceover
How do I record a clean voiceover for an online course?
Course narration is long, technical, and almost always gets revised at least once. The winning strategy is to write a fully scripted lesson, record paragraph by paragraph so revisions cost minutes instead of hours, and export with captions on every video.
Script the whole lesson, do not improvise
Improvised course narration sounds friendly for about 30 seconds and then turns into "um" and "right" and "so what we wanted to do here." Write the whole thing. Read it out loud. Cut the parts that feel stilted. The first draft of your script is too long. It always is.
One paragraph per teaching beat
Break the script into one paragraph per teaching beat. "Here is the concept." "Here is an example." "Here is the gotcha." "Here is the recap." This maps directly to the way VoiceOverAndOver works: each paragraph becomes a row you can record, re-record, crop, and reorder independently.
Use your normal teaching voice
A common amateur trap is to put on a fake "professional" voice. Students learn better from instructors who sound like a real person explaining something they actually understand. Sit forward. Smile when something is interesting. Slow down on the tricky line. Your real voice is enough.
If a paragraph keeps coming out flat, rewrite it. The line is probably too long, too jargon-heavy, or trying to do two things at once. Recording will not save a paragraph the script broke.
Equalize loudness for the lesson
Course lessons are listened to back-to-back. If lesson 3 is noticeably louder than lesson 4, students will adjust the volume, miss the next intro, and rate the course lower. Pick a loudness target (the merge step's -16 LUFS default is a sensible starting point) and use the same target for every lesson in the course.
Always export captions
Accessibility is non-negotiable for paid courses. Tick SRT or VTT on export. Most course platforms (Teachable, Thinkific, Kajabi, Podia, Vimeo) accept either format directly. The captions are timed off your paragraph durations, so they sync without manual nudging.