Subtitling & voice-over services for global audiences.

A media agency comparing suppliers should request a six-language, asset-by-asset adaptation plan, not a single translation quote. Separate what is spoken, what is displayed on screen and what must be approved for each market. The creative team still owns final editorial decisions, platform specifications and advertising claims.

• Map every variant and intended audience. Provide the approved master edit, script with speaker IDs, shot/timecode list, target languages and countries, campaign claims, product names and a matrix of 6-, 15- or 30-second cuts where relevant. List aspect ratios such as 9:16, 1:1 or 16:9 only where the planned channels require them; confirm all actual platform safe areas, limits and export specifications with the publishing team before production.
• Brief timed script adaptation, not literal length matching. Ask how suppliers will keep the intended call to action and tone while adapting idiom, voice-over pace, pronunciation and pauses to the exact available seconds. Maintain timecoded source/target scripts and flag lines that cannot fit without changing the message, extending a shot or re-editing. Brand, legal and claims approvals are required before recording or publication; marketing transcreation remains the route for a separately scoped creative campaign rewrite.
• Include graphics and end cards in the localisation map. Inventory burned-in titles, supers, captions, price overlays, offers, disclaimers, buttons and end-card calls to action, recording their on-screen time and editable source-layer ownership. Check translated text expansion, typography, contrast, subtitle-overlap and vertical-video crop. Agree who creates/replaces motion-graphic layers and who approves any revised pricing, conditions or market-specific claim.
• Define subtitle and speaker QA. Request timed captions or subtitles aligned with dialogue and shot changes, readable segmentation, correct product terminology, and speaker identification where two voices or off-screen narration would otherwise be ambiguous. Compare each subtitle track with the source audio, language-specific reviewer feedback and a full watch-through. The W3C WebVTT specification defines timed cue data, while the W3C captioning guidance explains the distinction between dialogue subtitles and accessibility-oriented captions, including significant sound information; platform-specific limits still need testing.
• Set delivery and acceptance by channel. Request an agreed matrix of language/version IDs, reviewed source and adapted scripts, a cue sheet, SRT or VTT when applicable, optional recorded VO stems, editable graphics if in scope, approved burned-in and clean masters, and export files that the client's editor can open. Test each crop with subtitles enabled and muted playback, review font/diacritics and line breaks on a phone, and log fixes and approvals. For UK-directed paid advertising, the ASA/CAP codes are a reference for advertising claims; local-market legal review remains the advertiser's responsibility.

For comparable quotations: send the six target languages and varieties, territories, edit durations, video/graphics source files, voice and speaker notes, brand glossary, on-screen text list, planned channels, caption/VO choices, required formats, review contacts, deadlines and version count. Ask each supplier to separate costs and responsibilities for translation, creative adaptation, linguistic and in-video review, caption spotting, voice talent/recording, graphics replacement, final encoding and revisions. Confirm actual production capacity and platform compatibility rather than assuming they are included. Request a scoped multilingual video quote. Plain spoken-word episode transcripts belong to the podcast transcription service, not this video-ad workflow.

For recorded lectures or lessons, compare suppliers on the caption track, readable transcript and in-player learning experience as separate acceptance items. A subtitle file that translates speech alone may miss a lecturer's diagram explanation, a question from a second speaker, or an important audible demonstration. Define the learner needs and each target language before choosing deliverables.

• Scope every recording and learning asset. Supply video duration, source language, target languages/varieties, class level, specialist glossary, lecturer and student speaker count, existing notes, slide deck, whiteboard images, relevant diagrams and any permissions for learner voices. Mark confidential or personally identifiable classroom content, and agree anonymisation, access and retention needs before file transfer.
• Require accurately synchronised, useful captions. Ask for timed speech, correct names and subject terms, punctuation and readable cue breaks checked against the audio. Include speaker identification when a change of voice would otherwise be unclear and relevant non-speech information, such as a demonstration sound or audience reaction that changes the lesson's meaning. The W3C's prerecorded-caption criterion (WCAG 1.2.2) and captioning guidance explain why dialogue-only subtitles are not always an adequate accessible caption track.
• Specify the transcript separately. Ask for a selectable, searchable, correctly ordered transcript with meaningful headings, speaker labels, section or time markers and agreed technical spellings. Decide whether a basic speech transcript is sufficient or a descriptive transcript should also communicate slide titles, formulas, drawing steps and important screen actions needed to follow the class. W3C transcript guidance distinguishes these formats; caption timing files alone are not automatically a complete accessible transcript.
• Assess visuals and audio description. Check whether key information appears only in charts, slides, lab demonstrations, gestures or on-screen text. Ask the course owner whether narration already communicates it, whether a descriptive text alternative is needed and whether audio description must be separately produced. Under WCAG 2.2, 1.2.3 (Level A) allows audio description or a media alternative in applicable cases, whereas 1.2.5 (Level AA) calls for audio description where relevant visual information is otherwise unavailable through the audio. Translation alone does not establish accessibility conformance.
• Plan two layers of human review. Request a bilingual educational/subject-terminology review for the target captions and transcript, followed by an in-player check against the final edit. Test caption start and end times, speaker ambiguity, quiz references, mathematical symbols, proper names, subtitle overlap with slides, readability on mobile, language track labelling and keyboard access to player controls. Ask the teaching owner to resolve content ambiguity and approve corrections and publishing rights.
• Agree formats and handover before production. Identify the actual video player or learning-management system, whether it accepts separate language tracks, and whether the publisher needs SRT, WebVTT, editable DOCX or accessible HTML transcripts, timed cue sheets, an audio-description track or a corrected master. Request a sample import/test on the target platform and named responsibility for upload, track switching, accessibility testing and later lesson revisions. These integrations are not automatically included in translation.

For a comparable quotation: provide a representative recording, durations, languages, speaker counts, course-level glossary, media and slide files, requested transcript detail, confidentiality constraints, LMS/player details, target accessibility specification, reviewer contacts, deadlines and correction rounds. Ask for separate prices and acceptance criteria for transcription, translation, captions/spotting, descriptive transcript preparation, bilingual review, audio description if needed and final platform QA. UK public-sector accessibility requirements may apply according to the provider and publishing context, with exemptions and exceptions; the education provider should establish its own obligations rather than assume universal WCAG compliance. Request a recorded-class project quote. Audio-only discussion or podcast transcripts can be scoped via podcast transcription.

For a financial-results webcast, shortlist media-language suppliers on a confidentiality-approved production plan, not just a per-minute subtitle price. Distinguish English same-language captions (including speaker changes and relevant non-speech sounds) from Arabic translated subtitles. Specify whether each must appear during the live event, in an edited recording after authorised public release, or both. Live, bilingual and post-production services have different latency, reviewer and platform requirements; confirm which are actually available for the event before committing.

1. Lock the disclosure and access timetable: the issuer's investor-relations/legal team identifies any inside information, approves who can receive unreleased scripts, figures, draft slides, rehearsal recordings and Q&A materials, and sets the release/embargo time. For UK MAR-covered information, FCA guidance on controlling inside information says selective disclosure must be necessary in the normal exercise of employment, profession or duties. An NDA is a safeguard, not automatic permission to disclose. Ask about named authorised recipients, relevant insider-list or access-record duties where applicable, conflict controls and a leak/escalation contact.
2. Prepare English caption inputs: provide only the material authorised for the agreed stage: confirmed speaker roster and titles, approved financial terms, company/product names, reporting periods, slide terminology and permitted number formatting. Ask for speaker-change attribution, relevant sound cues, live-caption latency expectations, treatment of overlapping speech and a process to correct a misheard figure or name without altering the issuer's approved disclosure.
3. Define Arabic output and review: specify the target Arabic variety/register, right-to-left display in the chosen player, whether the audio remains English, and whether subtitles will be generated simultaneously or translated from a reviewed English caption/transcript file afterwards. Verify numerical amounts, currencies, percentages, year-on-year comparisons, dates, speaker attributions and financial terminology against the issuer-approved English source. Agree a named Arabic linguistic reviewer and an investor-relations sign-off; do not imply simultaneous Arabic subtitles can be guaranteed without an actual live workflow and rehearsal.
4. Test the stream, recording and accessibility: record the webinar/webcast platform, live caption interface, player subtitle-track controls, timecode/feed requirements, audio delay, frame rate and fallback plan. WCAG 2.2 criterion 1.2.4 addresses captions for live synchronised media at Level AA, whereas criterion 1.2.2 addresses prerecorded captions at Level A. Translated Arabic subtitles alone do not establish accessible English caption coverage. Run an authorised dry run, checking timing, line wraps, speaker labels, Arabic direction and whether users can select the correct track.
5. Specify controlled deliverables and retention: ask for English caption files, Arabic subtitle files (for example SRT or VTT only where supported), a reviewed bilingual glossary, timed cue sheets, speaker labels, correction log and publication-approved recording/transcript as required. Agree encrypted or equivalently appropriate transfer, minimum authorised access, approved subcontractors, access revocation, deletion/retention time, draft watermarking where practical and an embargo-hold procedure. If personal data is included, ICO data-security guidance informs proportionate access and handling controls.
6. Request a comparable quote: submit the planned event time zone, webcast duration, expected speakers and Q&A, English/Arabic delivery windows, pre-release versus post-release access policy, audio/slide sample cleared for suppliers, platform/recording specifications, reviewer availability and release time. Separate pricing for live English captions, live Arabic translation if available, recorded Arabic subtitling, bilingual QA, rehearsal, platform integration, urgent corrections and cancellation/change terms.

For an approved spoken-word transcript or post-release summary, use the podcast/spoken-content transcription route as a separately scoped supporting service. Request a confidential project-scope assessment initially with a redacted example, then arrange protected file access only once the issuer authorises disclosure and the supplier confirms specialist capacity. No legal clearance, regulated disclosure, specific live-caption platform capability or real-time Arabic availability is promised here.

When selecting a supplier for documentary interviews, request a sample of the entire source-audio-to-English-subtitle workflow, not only a rate per minute. The edited film needs an English rendering that preserves what the interviewees meant while remaining readable within the actual shots. Agree whether a source-language transcript, an English reference transcript and independently timed English subtitles are all required; they are distinct deliverables.

• Identify the precise source audio and speakers. Share the approved interview clips or edit, audio/video durations, frame rate, starting timecode and a speaker map. State Spanish and Portuguese language varieties where known, such as European or Brazilian Portuguese and relevant regional Spanish speech, plus dialect shifts, code-switching, technical names, sensitive subjects and any consent or confidentiality restrictions. Ask how the supplier will distinguish an interviewer from a respondent and handle off-camera, overlapping or unclear speech.
• Specify transcription and ambiguity handling. Ask for a time-referenced Spanish/Portuguese source transcript with consistent speaker IDs, faithful quotations and a stated convention for interruptions, pauses, unintelligible words and overlaps. Decide whether the source record must be verbatim or edited for readability. Do not allow inaudible speech to be silently guessed or an interviewee's certainty, negation or meaning to be changed; flag contested passages with source timecodes for producer review.
• Require translation from the correct source language. Define whether English subtitles will be translated directly from the Spanish/Portuguese recording using the audio plus transcript, or from an approved bilingual intermediate document. Ask for review by a language-appropriate human linguist who can check idioms, factual names, cultural context, pronouns and terminology against the original speech. Keep an English reference transcript when the editor needs searchable quotations, and note where subtitle condensation differs from a fuller translation.
• Test spotting against both dialogue and picture. Ask who sets and reviews subtitle in/out points, cue segmentation, reading speed, line breaks, shot boundaries and transitions between speakers. Verify the timebase and account for locked-picture changes or revised interview cuts. A WebVTT cue carries timed text and can identify speakers; Netflix's published guidance illustrates the practical value of audio and edit synchronisation, but its specific frame rules should not be imposed unless the actual distributor requires them. Test speech fidelity and readability on the final picture.
• Separate dialogue subtitles from accessibility captions. English translation subtitles generally convey speech. If the production also requires an English accessibility caption track, specify relevant non-speech sounds, speaker identity and the target viewer requirements; W3C captioning guidance describes the distinction. Confirm whether original-language captions, forced translations of important on-screen signs or archive-text overlays must be supplied as additional assets, and prevent overlap with lower thirds.
• Define review and deliverables for the editor. Request a source transcript and English reference transcript if needed, a timecoded subtitle file (for example SRT or VTT where compatible), a cue/issue list keyed to timecode, and any agreed burned-in review video or distribution format. Agree separate bilingual linguistic QA, a final in-picture watch-through, producer approval of disputed quotations, and a correction pass after picture lock. Do not assume platform-specific exports, specialist subtitling software or guaranteed frame accuracy without a sample-file acceptance test.

For comparable quotations: supply representative sample footage, total minutes, language/region and speaker inventory, source audio quality, rough-cut or picture-locked status, required transcript fidelity, subtitle style guide, timecode/frame-rate details, review roles, privacy and release restrictions, delivery file types and deadline. Ask suppliers to itemise source transcription, audio-to-English translation, subtitling adaptation, spotting, linguistic review, in-video QC, format conversion and revision rounds. For a scoped proposal send an approved sample and brief. Audio-only episode transcription is a separate use case covered by podcast transcription services.

For a product demo aimed at French- and Italian-speaking users, decide whether viewers should hear the demonstration in their language, read translated speech while retaining the original audio, or receive both. The right approach depends on how quickly the interface changes, how much text is already on screen, whether users need to watch the cursor, and how often the product UI will be updated. Translation of the narration script, recording and video engineering should be separately specified; supplier capabilities and availability must be confirmed for the particular files.

1. Choose subtitles where a viewer can read and follow the product screen together. Keep the original presenter and use timed French or Italian dialogue subtitles (SRT or VTT where the player supports them), checking line length, screen action overlap, cut timing and on-screen calls to action. If rapid cursor moves, menus and dense UI text already demand constant visual attention, subtitle-only localisation may overload the screen; test a representative 30–60-second sequence with users or the client's reviewers.
2. Choose voice-over where listening frees attention for the demonstration. Ask for timed, recordable French (France, or other specified variety) and Italian scripts, not just a literal translation. Preserve approved feature names, settings, numbers, menu labels and claims; plan for natural sentence length, pronunciation guide, voice style, pauses for clicks and demonstrations, and retakes. Specify whether the original voice is replaced, briefly audible underneath, or kept as a separate track. Voice-over is not necessarily lip-synchronised dubbing; quote lip-sync and specialist recording separately if required.
3. Consider a combined or alternate version. Offer translated subtitles alongside an optional French/Italian audio track where the host platform can switch them, or make separate video masters when it cannot. Maintain same-language captions for the final French/Italian soundtrack when accessibility requires them. The W3C captions and subtitles guidance distinguishes translated dialogue subtitles from captions that also represent important sounds and speaker information; one does not automatically satisfy the other's purpose.
4. Coordinate narration with the visible interface. Supply the approved master video, time-coded English source transcript, screenshots, UI string list, feature glossary, script/scene identifiers and editable source graphics. Decide which menu names remain in English and which are displayed in a localised product. Check the spoken instruction matches the actual French/Italian UI build rather than translating a button the viewer cannot find; the software/product owner approves technical terminology and on-screen demonstration accuracy.
5. Set a clear review sequence and deliverable list. Request bilingual review of script meaning and pronunciation, voice audition approval before full recording (if voice-over is commissioned), audio-level and timing checks, subtitle spotting and in-player legibility review, and final video play-through for both languages. Specify editable scripts, time-coded SRT/VTT tracks, WAV audio stems where requested, pronunciation glossary, mastered MP4/MOV outputs as compatible, correction rounds and platform/aspect-ratio versions. The W3C guidance on description of visual information is a separate consideration if important screen actions are not expressed in the audio or another accessible alternative.
6. Ask suppliers to quote by production pathway: provide video minutes, speaker count, target locale (French for France versus other varieties; Italian for Italy), target platforms, frame rate, source/project-file access, on-screen text count, timing lock, recording specifications, existing English narration/stems, accessibility needs, language/technical approvers and deadline. Compare costs for translation, script adaptation, voice casting/recording, subtitle spotting, mixing, graphic/UI updates, review and revisions; do not assume a single flat per-minute price includes all work.

For audio-only interviews or episodes, use the podcast transcription route rather than a video-production quote. Send a redacted demo clip and approved terminology for a scoped French/Italian quote; exact voice, recording capacity, production schedule and file compatibility will be confirmed for the project rather than guaranteed here.

A corporate training producer commissioning multilingual safety video localisation should ask suppliers to scope the translated spoken script, time-aligned captions or subtitles, on-screen warnings and technical review as separate but coordinated deliverables. Accuracy matters beyond fluent speech: workers must be able to identify hazards, understand the correct protective measures and follow emergency actions shown in the film. A translation agency can assess the language-production brief, while the employer's qualified health-and-safety team retains responsibility for the underlying instructions, workplace suitability and confirmation that workers understand the training.

1. Inventory every safety-critical asset: supply the approved master video, timecoded source transcript, narrator and speaker IDs, target languages/varieties, audience roles and literacy needs, site/country scope, approved equipment or PPE names, warning signs, emergency routes, diagrams and text embedded in video frames. Provide editable graphics and narration stems if available. Keep scene IDs and a version register so changed procedures are updated in each language.
2. Preserve instructions rather than inventing safer-sounding wording: use employer-approved terminology for lockout/isolation, hazards, controls, machine settings, prohibited actions, PPE and emergency procedures. Check negatives, conditional phrases, quantities, units, signal words and demonstrator actions against the source. A bilingual linguistic reviewer should flag ambiguity; the client's competent safety lead must approve the actual technique, local legal requirements and equipment-specific details, including differences between overseas sites.
3. Plan voice-over and subtitle timing together: specify whether the original speech remains, is replaced by recorded local-language narration, or is supplied with translated subtitles only. Agree script pacing around step demonstrations, pauses for warnings and screen actions, how overlapping speech and multiple speakers are labelled, and whether speaker names or roles are shown. Test translated cue timing, line lengths and placement so captions never mask the hazard, control or critical instruction being demonstrated.
4. Distinguish translation from accessibility deliverables: foreign-language dialogue subtitles are not necessarily full captions. Under WCAG 2.2 criterion 1.2.2 (prerecorded captions, Level A), captions convey speech, speaker changes where necessary and meaningful sound cues. Where essential safety information is shown only visually, consider separately scoped audio description or an adequate media alternative, as described by criterion 1.2.3. The client should determine the accessibility obligations for its delivery channel and jurisdiction.
5. Require comprehension and platform checks: arrange linguistic and safety-specialist sign-off of a sample scene, then proof the full video with each subtitle/audio track in its actual LMS, video player or offline training format. Inspect warning visibility, text-direction and glyphs, captions and important noises, narration levels, speaker attribution, links to supporting procedures and completion/quiz prompts if included. UK HSE employer guidance advises plain language and checking that workers understand risks, precautions and emergency arrangements; a translated film alone does not prove effective training.
6. Quote the deliverables and owners explicitly: provide duration, language and country pairs, narrators, source files, script and on-screen word counts, video/frame rate, approved glossary, confidentiality constraints, reviewers, platform and deadlines. Request separate costs and acceptance criteria for source transcription, technical script translation, independent review, voice talent/recording, timed SRT or VTT where supported, audio stems, graphics localisation, mastered video files, accessibility work, revision rounds and future safety-procedure updates. Confirm which production tasks and file formats the supplier can actually deliver.

For accompanying written manuals or risk-control procedures, use technical training document translation. Audio-only speech transcription can be scoped via podcast and spoken-content transcription when suitable. Send a representative approved safety clip and terminology list for a scoped quotation. No translation, recording, platform integration, regulatory certification or worker-understanding guarantee is implied without a defined, approved project.