Build a Termbase So Your Dubs Stay Consistent
When you're dubbing your second, fifth, or fiftieth video into the same language, you'll notice a problem: a term that meant one thing in video one means something slightly different in video two. Your brand name, a product feature, a recurring joke, or a technical term drifts across videos because there's no single source of truth for what it should sound like and mean.
A termbase solves this by centralizing the decisions you've already made about how to translate and deliver specific terms. It's not a dictionary, it's your team's memory of what you've chosen, and why, so no future video goes off the rails.
What goes in a termbase
A termbase is simply a list. Each entry has a few fields:
- English term (source) - the exact word or phrase from your original script
- Translated term (target) - how it should sound in the dubbed language
- Context - where and when you use this term
- Notes - pronunciation hints, tone, or the reason you chose it
- Category - product terms, brand names, technical jargon, etc.
Example for Spanish dubbing:
| English | Spanish | Context | Notes | Category |
|---|---|---|---|---|
| "optimize" | "optimizar" | technical how-to videos | verb form for voice, not noun | feature |
| "dashboard" | "panel de control" | SaaS demo | consistent with UI labels | UI term |
| "folks" | "chicos y chicas" | casual vlogs | friendly tone, not formal "gente" | tone |
| brand name | (no change) | any | never translate proper name | brand |
Start simple. You need English, translation, and context. Pronunciation hints and category tags are helpful but secondary. The goal is speed and clarity, not perfection.
Who should own the termbase
Assign one person to be the termbase owner. This is typically a translator, localization specialist, or producer who reviews every dub. They decide what goes in, make judgment calls on edge cases, and update it as your brand or product evolves.
Everyone else on the team should be able to read it and must follow it. If a voice actor, subtitle creator, or fellow translator needs to add a term, they flag it in a comment or a shared document, and the owner reviews and adds it formally.
A decision framework helps when choosing your owner. Ask yourself these questions in order:
- Do you have a dedicated translator or localization manager on staff? They should own it. They understand both languages and will use it daily.
- Is your producer or project lead the most familiar with your brand voice and product? They can own it if the translator reports to them on new terms each week.
- Are you a solo creator with an occasional contractor? Assign it to yourself but build time each week to review and update it together.
Without a clear owner, termbases grow stale and contradictory. With one, they become the guardrails every future video leans on.
How a termbase prevents drift
Drift happens when each video is dubbed in isolation. Video one translates "dashboard" as "panel de mando". Video two calls it "tablero". Video three uses "interfaz". Your audience gets confused because the same UI looks different when described in your dubs.
A termbase creates one source of truth. The translator for video five opens the termbase, sees that "dashboard" is always "panel de control", and uses it, even if they would have chosen differently. Consistency across fifty videos feels professional and builds trust.
More importantly, a termbase makes it fast. Your translator doesn't spend time re-deciding the same terms. They spend time on the hard parts: pacing, tone, and making the script sound natural.
A worked example: Building a termbase for a YouTube channel
Let's say you run a YouTube channel about productivity software, and you want to dub your first five videos into German. Your first video is a tutorial on task management.
You finish the dub and review it with your translator. You stop at three moments where terminology felt off. Your translator says "scheduler" felt clunky as "Zeitplaner", but she translated it as "Aufgabenplaner" because it matched your UI. You approve this choice. You also debated whether "workflow" should be "Arbeitsablauf" (literal) or "Workflow" (borrowed English). You chose the German term to feel more native. You also have your brand name, which never translates.
That's three entries. You add five more from the script: "project", "filter", "collaborate", "deadline", "automation". With notes on why each choice matters.
After your second video, your translator reports two new terms: "tags" versus "Kategorien", and whether "nest" (for organizing projects) should be "verschachteln" or left as "nesting". You review both, approve both, and add them. Now you have ten entries.
By video five, you'll have thirty to forty terms. Your new translator for video four and five opened the termbase, never had to ask what "dashboard" means, and the dubs feel like the same voice across all five videos. Your audience notices the consistency. Comments say "these feel natural" instead of "this sounds weirdly different from the other one".
That termbase started with one hour of work after your first dub. It saved hours of re-work and questions across videos two through five.
Where to keep your termbase
A termbase can live in:
- A shared spreadsheet (Google Sheets, Excel) - easiest, searchable, can comment on rows
- A structured text file (CSV, TSV, JSON) - version-controllable, integrable with tools
- Translation memory software (SDL Trados, Wordfast, or free tools like Omegat) - built for this, but overkill for small teams
- A simple wiki or Notion table - if your team loves documentation
Most solo creators and small teams start with a spreadsheet. It's free, visible to everyone at once, and needs no special software.
Name it clearly: "Spanish_Termbase_2026" or "Dubs_German_Terms". Add a "Last Updated" row. Include notes on how to request changes. If you're using an AI dubbing tool, some platforms support importing a termbase or reference glossary directly, which feeds term preferences into the voice synthesis. Check your tool's documentation.
Common mistakes when building a termbase
Mistake 1: Too many columns too early. New team members open a spreadsheet with fifteen columns and abandon it. Start with five: English, Target Language, Context, Notes, Category. Add more only if you use them.
Mistake 2: No context. A translator sees "run" and has no idea whether it means "execute code", "manage operations", or "operate a business". Always write the context. One sentence is enough: "SaaS demo, referring to starting the software".
Mistake 3: Letting it sit. You build a termbase after video one, then skip video two and three because you're busy. By video four, the termbase is months old and you're not sure if it's still accurate. Review and update after every dub. Fifteen minutes per video, every time.
Mistake 4: Owners who disappear. Your translator left the team and took the termbase knowledge with them. Document why you chose each term, not just what it is. "Panel de control because it matches the Spanish UI language" is more useful than "dashboard: panel de control".
Mistake 5: Glossary, not termbase. A glossary is one-way: English to Spanish, that's it. A termbase is two-way and includes tone, context, and pronunciation. It's a decision log. Make sure your version actually captures why you chose what you chose.
Growing your termbase over time
Start with brand names and product terms. These are unambiguous and reappear often. After your first dub, you'll have ten to twenty entries. After five dubs, you'll have fifty to a hundred.
Review the termbase every quarter. Archive terms that rarely appear. Break it into categories if it gets large: technical terms, brand/UI terms, tone markers, recurring phrases.
As you add languages, some terms might work across multiple languages (e.g., brand names, technical acronyms). Keep a small "shared" section, then language-specific sections below.
The termbase is a living document. It changes when your product changes, your audience changes, or you learn that a previous choice didn't land as well as you hoped.
When consistency matters most
You don't need a termbase for a single dub. You need one when you're committing to ongoing multilingual content. A YouTube channel with plans for ten Spanish videos should build its termbase after the first video. A SaaS with demos in five languages should build one before recording demo two.
If you're dabbling in one language for two videos, skip it. If you're serious about growing in that language, build it early. The cost of building one is an hour of structured thinking. The cost of not having one is hours of re-work when you realize video three contradicted video two.
Termbase and AI dubbing tools
If you're using an AI dubbing platform, a termbase becomes even more valuable. When you generate dubs, you can sometimes reference or import your termbase as a custom glossary. The system uses these terms during voice synthesis, which means your brand names, product terms, and tone preferences bake directly into the generated audio. This gives you one more layer of control and consistency without needing manual re-recording.
Some platforms let you tag terms as "never translate" (for brand names) or "prefer this form" (for technical terms). If your tool supports this, use it. It saves the translator or AI from guessing.
Troubleshooting: What goes wrong without a termbase
Problem: Your translator says "Let me check the last video for how you said this." This is a sign your termbase doesn't exist or isn't being used. Your translator is wasting time searching through previous videos instead of consulting a single reference. Build the termbase and link it in your project brief.
Problem: The Spanish dub sounds off, but you can't pinpoint why. Often it's because key terms were translated differently than in previous videos. Your audience notices the inconsistency even if they can't name it. A termbase prevents this by locking in how recurring terms sound.
Problem: A term works in one context but not another. Your product manual uses "activate" formally, but your casual tutorial vlog should use "turn on". A termbase handles this with context notes. Same English word, different translations based on where it appears. Your translator sees the context and knows which version to use.
Problem: You hired a new translator for video four and they took completely different choices. They didn't know your previous decisions. A termbase is your onboarding document. Your new translator reads three minutes of context and knows exactly how your channel talks about concepts.
Problem: You made decisions three months ago and forgot why. You set "customer" to "cliente" instead of "consumidor", but you can't remember if it was a branding choice or just a guess. Good termbase notes say why. They're insurance against re-deciding the same term when you expand to another language.
Frequently asked questions
Q: Should I include phrases or just single words? Both. Include single terms like "optimize" or "dashboard", but also common phrases like "best practice" or "technical roadmap" if they recur across videos. Phrases often have tone decisions baked into them, which makes them worth documenting.
Q: What if I don't know the term until I'm dubbing a video? Flag it with a question mark or "TBD" in the termbase. Your translator notes it, the owner reviews it after the dub is done, and you make the call then. Future videos use that approved term. This is better than letting each translator guess.
Q: How do I maintain the termbase if I work with many different translators? Assign one owner who does final review on every dub. Translators propose new terms in the spreadsheet comment column. The owner decides and adds it formally. Translators read the termbase before starting work. Simple flow: read, work, propose, owner approves, move on.
Q: What if a term evolves? Do I keep the old one or replace it? Keep both. Add a new row with "v2" or note the date it changed. Example: "filter (v1): Filtro, archive date: Sep 2024 / filter (v2): Selector de filtros, current as of Dec 2024, reason: UI updated". Your old videos used v1, new ones use v2. It's a historical record, not a burden.
Q: What if I have ten languages? Start with one language and prove the termbase works. One language, five videos, solid termbase. Then expand to a second language. You can reuse some terms (brand names, technical acronyms), but each language needs its own context and tone notes. A master spreadsheet with tabs per language is clean. Or separate sheets with a shared "brand names" tab.
What to do next
Create a spreadsheet with five columns: English, [Language], Context, Notes, Category. Add twenty to thirty terms from your first dub. Start with brand names, product terms, and any phrase where you made a deliberate choice about tone or phrasing.
Share it with everyone who will touch that language. Go through your translation and dub once more, adding any term you debated. Mark it "Owner Approved" or add today's date so you know when it's current.
In your project brief or channel notes, link to the termbase and set one clear rule: translators consult it before starting work, and they flag new terms for owner approval before the dub is finalized. This one-sentence rule prevents surprises.
Review the termbase after your second and third videos. Add new terms, update notes if something didn't land the way you expected, and archive terms you haven't used. By video five, you'll have a reference that saves time and keeps your dubs sounding like one professional voice, even across dozens of videos. Your audience will notice the consistency, and your translators will thank you for reducing the re-work. Start today with just one language and one dub. The investment is small. The payoff compounds with every video you add.
🚀 Start Dubbing Your Videos Today
DubLab uses AI to translate your videos into 92+ languages in minutes.