eLearning Narration with AI: Pace, Pronunciation, Terminology
Updated September 3, 2026
TL;DR: teach at 140–150 words a minute, signpost every screen, define acronyms once and then add a pronunciation rule, keep one narrator across the whole curriculum, and produce in a Studio project so a changed slide costs one paragraph. Captions come from the narration timeline, not from ASR.
Pace
Explainers on YouTube run 150–160 words a minute; learners replaying a lesson prefer 140–150. On an AI voice, set it once — director's note "steady, patient, clear" on Ultra, or choose an educational voice like Carl — and keep it for the whole course. Vary energy between modules with a tag at the start of a section, not by changing the voice.
Terminology
- Define an acronym in full the first time ("S-a-a-S, software as a service"), then add a rule so "SaaS" is read your way everywhere.
- Alias rules (respell the word) work on every voice tier; Leap Turbo voices also take IPA and Leap Natural voices take pinyin for exact phonetics.
- Product names, people, places: one rule each, entered before you generate module one.
- Numbers and units: write them the way they should be said — "ten megabytes", not "10MB" — unless the voice already reads them correctly on your sample.
Structure
- One Studio project per course, one chapter per module, one paragraph per screen.
- Signpost aloud: "On this screen…", "Next…", "To recap…" — the voice tells learners where they are.
- Keep screen narration under 60 seconds; split long screens.
- Put assessments and answer keys in their own paragraphs for easy updates.
Updates
Courses change. When a screen changes, regenerate that paragraph; version history keeps the old take. When a term changes, edit the rule and regenerate the paragraphs that use it. Nothing else moves.
Accessibility
Export SRT/VTT with each module — built from the narration text, so they're exact — and keep the script available as a transcript. Learners on mute, learners with hearing loss and search inside the LMS all benefit.
Voices that teach well
Frequently asked questions
Which voice tier for courses?
Standard tiers are usually right: clear, consistent, economical. Use Ultra when you need multilingual delivery or emotional range in scenario-based learning.
How do I make an acronym read as letters?
Add an alias rule: pattern "API", replacement "A P I" (with spaces). The rule applies across the project.
Can I localise the course?
On Creator and above, translate the project into eight languages and re-narrate with the same setup; review each edition with a native speaker.
Does LeapFun produce SCORM?
No. Export mastered audio and subtitles per module and import them into your authoring tool or LMS.