Published 2026-03-20 · Updated 2026-10-22 · MachenVoice
Online courses need consistent narration across dozens of modules, but studio time does not scale when legal updates a compliance paragraph every quarter. Text to speech lets instructional designers ship MP3 narration from written scripts; MachenVoice adds block limits and export so you can regenerate only what changed. This guide walks through voice selection, LMS handoff, and revision loops you can document in your style guide.
When TTS fits e-learning (and when it does not)
TTS works well for policy overviews, software walkthrough voiceovers, onboarding sequences, and multilingual pilots. Learners expect a clear, neutral voice that matches slide titles.
It works poorly for emotional leadership messages, acting-heavy scenarios, or brands that invested in a signature human voice. In those cases, use TTS for drafts and hire talent for the published master.
Map slides to MachenVoice blocks
Treat one slide deck as two blocks minimum: block one states objectives, block two summarizes the exercise or safety reminder. If a module exceeds 2,000 characters spoken, add blocks at section breaks rather than speeding up the voice.
Name your blocks in an external doc (“Module 3 — objectives”, “Module 3 — lab reminder”) so you know which MP3 to replace after SME review.
Pick one voice per course catalog
Aria (English - United States) and Ryan (English - United Kingdom) are common starting points for corporate English. Document the voice ID in your style guide next to font and color rules.
For Brazilian Portuguese employee training, Francisca (Portuguese - Brazil) is a practical default; preview compliance scripts that contain legal terms before bulk generation.
Authoring tool handoff
Export MP3 from MachenVoice and import into Articulate 360, Rise, iSpring, or Camtasia. Most tools accept standard MP3; you handle SCORM packaging and LMS upload.
If your LMS caps file size, normalize loudness in the authoring tool rather than re-encoding repeatedly in multiple apps.
Revision loop with legal and SMEs
Share MP3 links internally before learners hear audio. When legal changes one sentence, edit that block only, regenerate, and replace the asset in the authoring tool.
Track version numbers on filenames (`module3-intro_v4.mp3`) so support knows which audio matches the published course date.
Accessibility pairing
Narration complements captions; it does not replace them. Provide synchronized captions for deaf and hard-of-hearing learners and allow learners to control playback speed in the LMS player.
For dyslexia-friendly programs, pairing TTS with plain-language on-screen text helps more than narration alone—see our accessibility guide for patterns that research supports.
SCORM and LMS packaging
MachenVoice delivers MP3 assets; your authoring tool still owns SCORM manifests, quiz banks, and completion tracking. Import narration per slide or per scene, then publish the package your LMS expects.
Test one module in a staging course before bulk upload. Confirm autoplay policies: some LMS players require a learner click before audio starts, which affects how you write on-screen instructions.
Quality assurance before learners hear audio
Run a three-pass review: instructional designer checks alignment to slides, SME checks facts, and a listener checks pacing on phone speakers. Log defects against block IDs in your script doc.
When translation teams localize text, regenerate audio per locale rather than reusing English MP3 under translated slides unless your policy explicitly allows it.
Voice matrix for global catalogs
Document a primary and backup voice per language in your style guide. Example: English US narration uses Aria with Guy as backup; UK corporate uses Ryan; Brazilian Portuguese compliance modules use Francisca.
Keep sample MP3s in an internal wiki so contractors do not invent new voice IDs per module. Consistency matters more than novelty for annual compliance refreshers.
Budgeting time for updates
Quarterly policy changes are cheaper with TTS when you only regenerate affected blocks. Estimate one hour per module for SME review plus ten minutes per changed block for MachenVoice export and LMS replace.
Batch legal edits: collect all string changes in one ticket, then regenerate in one session so voice settings stay identical across modules.
When to mix human and synthetic narration
Use TTS for knowledge checks, software demos, and policy refreshers. Record human talent for executive messages, empathy-heavy scenarios, and brand campaigns where timbre is part of the identity.
Hybrid courses can label sections (“Narrated overview” vs “Leadership message”) so learners know what to expect.
Structuring a forty-minute module
Open with objectives in block one, teach concepts in blocks two through five, and close with a recap block that repeats only the three actions learners must take on the job. Avoid narrating every bullet verbatim; summarize dense slides into spoken sentences.
Insert explicit pause cues before knowledge checks: “Pause the player and answer the question on screen.” That reduces the feeling that the voice is racing ahead of interactive elements.
Localization without voice drift
When you translate slide text, regenerate MachenVoice audio in the target locale with the documented voice ID for that language. Keep English master files archived so auditors can compare wording if a compliance question arises.
Do not assume one neural voice fits all dialects. Brazilian Portuguese and European Portuguese need different voice selections; preview legal terms in each locale before bulk export.
Analytics and completion rates
If completion drops on narrated slides, check whether audio autoplay is blocked, whether captions are missing, or whether narration simply repeats on-screen text word-for-word (learners skim and tune out).
A/B test shorter blocks: two three-minute narrations often outperform one six-minute wall of speech because learners can resume between blocks.
Vendor and procurement notes
Procurement teams sometimes ask whether TTS replaces voice talent contracts. Document that MachenVoice is a drafting and scale tool, not a talent agency, and that the upstream provider’s terms govern redistribution of neural output.
Include estimated regeneration time in SOWs when legal review cycles are monthly; TTS savings appear in update cost, not initial recording cost.
| Step | Owner | Done when |
|---|---|---|
| Script approved in doc | ID / SME | |
| MP3 exported from MachenVoice | ID | |
| Imported in authoring tool | ID | |
| Captions added or imported | ID | |
| SCORM published to staging | ID | |
| Production LMS upload | Ops |