If your library draws viewers from southeastern Nigeria, or from Igbo-speaking communities in the United States, the United Kingdom, and Canada, a request to translate video to Igbo will arrive before you have a process for it. The request sounds like a single step: take the English master, produce an Igbo version, publish. In practice it is closer to five steps, and four of them are decisions rather than tasks.

Igbo rewards care in places where most localization workflows have no machinery. Tone carries meaning at the word level, so a string of letters can be several words depending on how it is pitched. Dialect variation is wide enough that a supposedly neutral choice still sounds like a particular place to a particular listener, and the codified standard, Igbo izugbe, is not the variety most people grew up speaking at home. The orthography uses dotted vowels that a surprising amount of tooling strips or misrenders.

If you do not speak Igbo, you cannot listen to the output and know whether it is right. The useful work is not in training your ear but in building a process with named checkpoints, a written brief, and a reviewer who answers narrow questions. This guide covers the audience, the linguistic features that break automated output, register and vocabulary decisions, orthography in subtitles, timing overflows, voice selection, and the checks a non-speaker can run.

Who the Igbo audience is, and why it changes the release plan

Two audiences with different viewing habits

The largest concentration of Igbo speakers is in southeastern Nigeria, across Abia, Anambra, Ebonyi, Enugu, and Imo states, with related communities in parts of Delta and Rivers states. A second, substantial audience sits in the diaspora, chiefly in the United States, the United Kingdom, and Canada. The two groups do not consume video the same way, and the difference should drive what you publish first.

  • Viewers in Nigeria are frequently mobile-first and often watch with sound off, which makes subtitles non-negotiable and dubbing an upgrade rather than a baseline.
  • Diaspora viewers, particularly second generation, often understand spoken Igbo better than they read it, or the reverse, so audio carries more weight than text.
  • A release with Igbo audio and English-only subtitles loses the sound-off group entirely.

The standard variety is not everyone's home variety

Igbo izugbe is the codified written and broadcast standard, drawing on several dialect clusters. Most speakers acquire a local variety at home first, which means "Igbo" in a project brief is underspecified. Naming your variety in the brief and in the video description gives the audience a frame; not naming it invites the comment section to supply one. The choice also determines the voice you cast, the pronunciation notes in your glossary, and a large share of the reviewer's comments, so make it first and record it.

The linguistic features that break automated Igbo output

Tone, at the word level and in grammar

Tone in Igbo is not decoration or emphasis. It distinguishes words and carries grammatical information: the letters akwa can carry several unrelated meanings, and the tone pattern separates them. Everyday Igbo writing usually omits tone marks, so a text-only pipeline receives no tone information and has to infer it.

That inference fails first on proper nouns, the strings a model has never seen in a tonal context. Two mitigations help: supply tone-bearing input for glossary names where an authoritative spelling exists, and have the reviewer listen to those segments at reduced speed rather than the whole track at normal pace.

Dotted vowels and the subscript n

Igbo uses ọ, ị, and ụ as distinct letters, plus ṅ. They are not stylistic variants of o, i, u, and n; stripping the dot changes the word. Two failure modes follow.

On the input side, optical character recognition, manual transcription, or a copy-paste through a system that normalizes to ASCII can drop the dots, and the engine then processes the wrong word. On the output side, fonts, players, and export routines that handle combining marks badly render a dotted vowel as a box, a floating dot, or a fallback glyph. Both failures are invisible in a plain-text editor and visible in the player.

Use precomposed Unicode forms rather than combining sequences: U+1ECB for ị, U+1ECD for ọ, U+1EE5 for ụ, and U+1E45 for ṅ. Normalize to NFC, export as UTF-8, then test rendering in every player your audience uses, including the mobile web player and one smart TV app.

Dialect spread and the limits of a single voice

Igbo dialect variation is not a matter of accent alone; vocabulary, verb morphology, and idioms differ meaningfully across regions. A single voice cannot be neutral for everyone, so the goal is not neutrality but a stated choice, applied consistently, that matches where your audience sits. Diaspora-heavy viewership makes a voice from the standard cluster the safer bet; an audience concentrated in one state responds better to a closer regional voice and draws fewer corrections.

Running the first Igbo translation

Lock the source before the run starts

Every downstream artifact depends on the English master. If you re-cut or re-record the source after translation begins, the cue timings and the dub timing both break, and you pay for the translation twice. Freeze picture, freeze the script, then start.

Build the glossary before the first translation

The glossary is the highest-leverage artifact in the project, and it must exist before translation rather than after. Record for each term the English source, the Igbo handling, a pronunciation note, and whether it appears in speech, on screen, or both. Put term decisions in the brief so the translator is not making a hundred unrecorded calls alone.

Map speakers to voices

For multi-speaker video, speaker diarization splits the audio into turns, and each speaker needs a distinct target voice that stays consistent for the whole piece. Errors here are audible even to a non-speaker: two characters who should sound different come out identical, or one character's voice shifts mid-scene. The video dubbing workflow handles the turn-taking and mixing, while the underlying video translation pipeline produces the translated script and timing the dub is built from. Verify the speaker map before the full render, not after.

Subtitles: orthography, fonts, line length, and speed

The character set problem comes first

Before judging any translation, confirm the dots survived. Search the exported SRT for each dotted character, then open the file in the player and confirm it renders. If the file is correct and the player is wrong, the font or the container is at fault, and burning in the subtitles fixes it. If the file itself is wrong, the pipeline stripped the marks, and no amount of reviewing will help.

Line length, and what long Igbo words do to it

Common subtitling practice allows roughly 42 characters per line and no more than two lines at once. Igbo aggravates that budget, because the language builds meaning with prefixes, suffixes, and vowel harmony, so a single word can consume a third of a line. That is where a translator working alone tends to truncate, and truncation is the most common way a subtitle ends up wrong in a way that looks fine to a non-speaker.

Reading speed and text expansion

A widely used target for adult viewers sits around 15 to 17 characters per second, with a minimum cue duration of about one second. Translated Igbo lines run both longer and shorter than the English source, and neither direction is predictable per line. Where the text expands, the cue needs more time or the sentence needs restructuring. Subtitle translation against an existing English timing file is the right starting point when a caption track exists; when the English track is poor or missing, automatic subtitle generation from the audio gives a cleaner timing base.

When a line no longer fits the timing

Work through these in order, stopping as soon as one resolves the line without losing meaning:

  1. Extend the cue into the surrounding pause or gap, up to the point where the gap closes.
  2. Rewrite the line in fewer characters while keeping the full meaning, which is usually possible and is the step most often skipped.
  3. Split the line across two cues if the grammar allows a clean break and both cues remain readable.
  4. Move the information to a place with more room, such as a neighboring cue or an on-screen label.
  5. As a last resort, drop a clause that carries no plot, safety, or compliance information, and record what was dropped in the project log.

The same problem appears differently in audio. A dub has to fit the scene, and if lip synchronization is enabled, it also has to fit the visible mouth movement. With lip sync off, the constraint relaxes to natural-sounding pacing, which usually produces a better result than forcing a tight match. Decide which constraint applies before the translator starts, because the two produce different scripts.

Register and formality decisions to specify up front

Address and pronouns

Igbo distinguishes a singular and a plural second person, and the plural or a title can carry respect. Decide whether the presenter addresses the audience as one person or as many, and whether the register is informal or formal instructional. Consistency matters more than the specific choice: a track that switches address form halfway through sounds like two different presenters.

Loanwords, borrowings, and coined terms

For each technical term there are three options, each with a cost. Keeping the English word sounds contemporary and matches how many speakers actually talk about technology. An established Igbo borrowing sounds natural but may not exist for your category. A coined term is the most principled and the most likely to sound like a textbook. Record the choice per term and never let two runs make it differently.

Code-switching is a legitimate style, not a defect

Real Igbo speech mixes in English frequently, especially for administrative, technical, and brand vocabulary, and a dub written in pure Igbo can sound more formal than the English source. Decide the target mix as a rule: keep English for product names and technical nouns, translate everything else.

Put all of it in a one-page style sheet

Audience, variety, register, address form, glossary, code-switching rule, and pronunciation notes. Give the same page to every translator, voice director, and reviewer. Without it, two releases in the same series will differ in ways your audience notices and your team does not.

Names, brands, numbers, dates, and units

Build these into the glossary before the first run, because each category hides a decision.

  • Personal and place names: authoritative spelling, tone marking where known, and whether the name is spoken or displayed.
  • Brand and product names: usually kept in English, but record the pronunciation the voice should use.
  • Numbers: decide whether digits appear in the subtitle and how they are spoken in the dub. Igbo has both a traditional counting system and a decimal-influenced one, and large numbers are frequently spoken in English. Specify the rule rather than leaving it to the voice.
  • Currency and units: localize or retain the original, and be explicit about which.
  • Dates: day-month versus month-day ordering, and whether months are spoken in Igbo or English.
  • Acronyms: whether to spell them out in speech, read them as letters, or keep them written only.

Choosing an Igbo voice

Cast against the variety decision, not against a general impression of pleasantness. Then test the voice on your actual text: does it handle dotted vowels correctly, and does it pronounce English loanwords the way a speaker would, rather than spelling them out or stressing the wrong syllable? Those two tests eliminate weak candidates quickly. The speech generation engine is where voice choice, pacing, and pronunciation behavior are set, so test before committing to a full render.

Cloning a recurring presenter's voice

Voice cloning is a working feature, and it is the right answer when a series has a recognizable host whose voice is part of the format. A speaker's voice can be preserved across languages with authorization. Authorization should be written, scoped, and specific: which languages, which projects, and for how long. If the presenter is under an employment or union agreement, check whether synthetic voice is covered before generating anything. Disclosure is a separate obligation from consent: label synthetic dubbing where the platform expects it, and always for advertising and political content, where several jurisdictions require it.

What you can catch without understanding Igbo

You can verify structure without understanding meaning: no segment missing or duplicated, each speaker holding one voice, and no clicks, abrupt cuts, or loudness jumps between the original and the dub. A source segment that ran twenty seconds but produced twelve seconds of audio is a flag worth raising before you publish.

The quality review a non-speaker can run

Run these checks before a native reviewer sees the file, so their time goes to language rather than mechanics.

  • Segment count: compare source and target counts, and investigate any gap.
  • Timing: no cue under about one second, none over about seven, none with zero duration, no overlaps.
  • Encoding: search the exported file for each dotted character and confirm it survives.
  • Glossary terms: confirm each agreed term appears in the agreed form.
  • Numbers and dates: spot-check ten instances against the source.
  • Voice map: confirm the speaker-to-voice mapping has not drifted.
  • Loudness: compare the dub against the original and aim for a close match.

Briefing a native reviewer

Give the reviewer the style sheet, the glossary, and the source alongside the target. Ask for two passes: an accuracy pass on the script, then a listen-through at reduced speed for tone and pronunciation. Ask them to mark problems rather than fix them, using four categories: wrong meaning, unclear, right words but wrong tone, and sounds like the wrong region. Ask for a short written summary naming the three worst issues. A reviewer who reports categories gives you something you can act on and compare across releases; one who reports a general impression does not.

A realistic first project

Which videos to start with

Choose three to five pieces between two and six minutes each. Prefer a single presenter, a clean background, minimal on-screen text, and no wordplay; humor depends on language-specific cues and is the hardest thing for a non-speaker to evaluate. Avoid anything with legal or medical claims, licensed music with lyrics, or cultural references that need explaining. Start with the video that already draws the most Nigerian or diaspora traffic, which gives you a baseline to compare against.

What to measure

Track the Igbo version's completion and watch-through rate against the same video's English performance over a comparable period. Read the comments for corrections about dialect or tone specifically, the most direct signal available, and look at retention in the first thirty seconds, where a wrong voice shows up immediately. Record your reviewer's issue categories per release to see whether the glossary and style sheet are reducing problems, and track cost per published minute to know whether scaling is justified. Current plans are on the pricing page, and the API surface for automating any of this is documented in the developer documentation.

Frequently asked questions

Is Igbo the same as Ibo?

Yes. "Ibo" is an older spelling that persisted in colonial-era texts; "Igbo" is the standard modern form for briefs, metadata, and descriptions.

Do I need a native speaker if the tool does the translation?

For anything you publish, yes, at least for review. The tooling produces the draft and the timing; a speaker catches tone, register, and dialect problems no non-speaker can detect.

Can I publish Igbo subtitles without dubbing?

Yes, and it carries less risk as a first test. Subtitles are cheaper to revise, easier to review line by line, and they serve the sound-off viewing pattern common among mobile audiences.

Will dotted vowels survive an SRT export?

Often, but not reliably. Export as UTF-8, normalize to precomposed Unicode, then open the file in your actual player, not just a text editor.

Should I use standard Igbo or a regional variety?

Depends on where your audience is concentrated. Standard Igbo is the safer default for a mixed or diaspora audience; a regional variety lands better locally. Either way, state the choice and keep it consistent.

Yes. Get written authorization from the speaker, scoped to specific languages, projects, and durations, and disclose synthetic dubbing where the platform or the jurisdiction requires it.

How long does a first Igbo release take?

The review loop sets the schedule, not the machine steps. Translation and rendering are fast; the two-pass review, the glossary decisions, and the resulting fixes take the calendar.

Conclusion

Start with one video, not a library. Pick the piece with the most existing Nigerian or diaspora traffic, build the glossary and the one-page style sheet before translation begins, and publish subtitles first. That sequence gives you a reviewable artifact at low cost and tells you whether your audience wants the content badly enough to justify dubbing.

Then decide on the voice. If your format has a recurring presenter, cloning that voice keeps the series recognizable, provided you have written authorization and a disclosure plan. If it does not, cast a voice against the variety decision you made in the first week and test it on your real text before committing to a full render. Quality is determined not by the engine but by whether you wrote down the audience, the variety, the register, and the vocabulary before anyone started translating, and whether a speaker checked the result against those decisions.