Technical Subtitling Services

What Our Technical Subtitling Services Cover

Companies come to us with video libraries that grew faster than anyone planned. A machine builder records a lockout-tagout demonstration on a phone. A service department films a gearbox teardown for new hires. Marketing hosts a webinar on a new inverter line. Two years later there are 60 videos, a plant opening in Monterrey, and nobody who can put accurate Spanish subtitles on any of it. That is the problem this service solves.

We handle technical transcription and subtitling as one continuous workflow: capture the spoken content, translate it, compress it into readable captions, and deliver files that drop straight into your LMS, YouTube channel, or video player. Typical source material includes maintenance and repair walkthroughs, operator training modules, commissioning and installation videos, safety inductions, product demos, and recorded webinars. The written companions to these videos have their own pages; see our guide to translating training material and the dedicated page on maintenance manual translation for the documentation side.

What sets a technical subtitling project apart from captioning a film or a corporate promo is density. A narrator in an entertainment piece speaks in loose sentences with pauses. A field engineer narrating a valve replacement packs part numbers, torque values, tool names, and warnings into every breath. The subtitler has to decide, line by line, what can be condensed and what must survive verbatim. That judgment call requires someone who understands the procedure, which is why we staff these projects from the same pool of specialists who work on our written engineering content.

Step One: Technical Transcription by People Who Know the Vocabulary

Every project starts with a transcript of the source audio, and this is where generalist vendors quietly fail. Automatic speech recognition hears "bearing preload" and writes "barren payload." It hears "megger test" and invents "mega test." We do run ASR as a first pass because it saves time, but a specialist then corrects the output against the audio, and against your drawings or manuals when you can share them. On noisy shop-floor recordings, that human pass routinely changes 15–20% of the words, and the words it changes are almost always the technical ones.

You choose between two transcript styles. A verbatim transcript keeps every false start, repetition, and "um," which matters for legal review, incident investigations, or compliance records of safety briefings. A cleaned transcript removes disfluencies and lightly repairs grammar while preserving meaning, which is what you want when the transcript feeds subtitles or becomes the script for a re-recorded voiceover. Most training content uses the cleaned style; most audit and investigation content uses verbatim. We confirm the choice at quote stage because it affects both price and turnaround.

The corrected transcript comes back to you as a timestamped document. Many clients keep it as a standalone deliverable: it makes video content searchable, feeds knowledge bases, and gives your other translation projects a reliable source text to work from.

From Transcript to Timed Subtitles: The Readability Rules We Apply

Subtitling is translation under physical constraint. A viewer can only read so fast, and a subtitle can only occupy so much screen before it hides the equipment being demonstrated. We apply a fixed set of readability parameters, adapted from broadcast norms and tightened for instructional content where viewers often pause and follow along with tools in hand.

Parameter	Our standard	Why it matters for technical video
Reading speed	Max 17 characters per second	Viewers split attention between text and the task shown on screen
Lines per subtitle	2 maximum	Three lines cover the lower third, often exactly where hands and parts are
Line length	42 characters per line	Keeps text legible on tablets used in the field
Display time	1 second minimum, 6 seconds maximum	Flash subtitles are never read; frozen ones get re-read and distract
Synchronization	In-cue within 2 frames of speech onset	A warning must appear when it is spoken, not after the action
Numbers and units	Never condensed or rounded	85 ft-lb must stay 85 ft-lb, even when the sentence around it shrinks

These constraints force choices. German runs roughly 30% longer than English, so a dense English sentence often cannot fit at 17 characters per second in German without cutting something. Our subtitlers condense connective language and keep the payload: the part name, the value, the warning. When the source speaker talks too fast for any faithful subtitle, we flag the segment and propose a wording with you rather than silently dropping content, because in a lockout procedure there is no such thing as an optional sentence.

Terminology Accuracy: Where Generalist Captioning Breaks Down

Media captioning companies are good at timing and formats. They are not staffed to know that a "traveling block" on a drilling rig is not a block that travels in the everyday sense, or that "energizing" a circuit has a precise meaning that "turning on" does not always carry. When those vendors translate video subtitling projects for industrial clients, the vocabulary damage is often invisible to the buyer, who does not read the target language, and painfully visible to the technicians who do.

We anchor every subtitling project to a termbase. If you already maintain a glossary, we load it and enforce it. If you do not, we extract key terms from the transcript, translate and confirm them with you before subtitling begins, and deliver the glossary alongside the finished files so your next project starts from the same foundation. Terminology checks run twice: once by the subtitler during translation, and once by an independent reviewer who watches the finished video against the source, a step our ISO 17100-compliant process requires on every deliverable, timed or not.

A recent project illustrates the stakes. A wind turbine operator sent us 31 nacelle maintenance videos, originally subtitled into Spanish by a media house, because Mexican technicians reported that the captions were "confusing." The old files rendered "pitch bearing" three different ways across the series, one of them the word used for a fruit pit. We rebuilt the subtitles against the operator's own glossary, using the vocabulary conventions documented on our Spanish technical translation page, and the follow-up survey of field staff came back clean. Companies in the renewable energy sector run some of the largest training video libraries we see, and consistency across a series matters as much as accuracy within one video.

Delivery Formats: Sidecar Files, Burned-In Video, or Both

The right output depends on where the video will live. We deliver any combination of the following, itemized separately on the quote:

• SRT files: the universal sidecar format, accepted by YouTube, Vimeo, most LMS platforms, and every desktop player. One file per language, easy to update later.
• WebVTT files: required by HTML5 players and many streaming platforms; supports positioning cues, useful when captions must avoid on-screen labels.
• Burned-in (open) subtitles: we re-render the video with subtitles fused into the image. Nothing to configure, nothing viewers can turn off, ideal for shop-floor kiosks, tablets in the field, and WhatsApp distribution to contractors.
• Timestamped transcripts: corrected, speaker-labeled documents in DOCX for records, search, or downstream translation.
• Multilingual masters: one MP4 per language, or a single file with selectable tracks where the platform supports it.

For burn-in work we match your brand's font and safe margins, and we return a preview for sign-off before rendering the full batch. Source video is never recompressed more than once: we render from your highest-quality master, so ask your videographer for the original export rather than the compressed upload.

When a Voiceover Serves You Better Than Subtitles

Subtitles are not always the right answer. If viewers must keep their hands and eyes on a task, listening beats reading. If the audience includes workers with limited literacy in any language, audio is the only channel that works. For those cases we produce translated voiceover: the cleaned transcript becomes a script, a professional voice records it with pronunciation of technical terms validated against the glossary, and our engineers lay the new track over your video with the original audio ducked or removed.

Cost is the honest trade-off. Subtitling typically runs 60–70% cheaper than voiceover for the same runtime, and updates are far simpler: changing one line in an SRT file takes minutes, while changing one sentence of narration means a studio session. Many clients split the difference by subtitling the long-tail library and commissioning voiceover only for high-visibility onboarding modules. If your videos sit inside interactive courses built in Storyline or Rise, the calculus changes again, and our elearning translation services page explains how we handle embedded media inside courseware, including synchronized quizzes and animations.

How We Price Transcription and Subtitling

Pricing is per minute of source runtime, with three components you can mix. Transcription is billed at one rate per audio minute, higher for verbatim style, poor audio, or multiple overlapping speakers. Subtitle translation and timing is billed per runtime minute per target language. Rendering, whether burn-in or multi-track packaging, is billed as a flat technical fee per video. Because the quote is built from runtime, you know the full cost before we start; there are no per-word surprises after the fact.

Repetition works in your favor. Training series reuse the same warnings, the same intro sequences, and the same closing instructions. Our tooling detects repeated segments across a batch, so a 40-video series costs meaningfully less per minute than 40 unrelated videos. Rush surcharges, when you need them, are quoted as an explicit percentage line, never buried in the base rate. For a batch quote, send us the videos or even just their runtimes and target languages through the quote form, and we will return an itemized figure within one business hour.

Turnaround Benchmarks and Who Uses This Service

One finished hour of technical video represents far more than one hour of work. As a planning rule, allow 4–6 business days for the first hour of runtime into one language, covering corrected transcription, translated and timed subtitles, independent review against the picture, and rendering. Each additional language adds 2–3 days, and languages run in parallel, so a five-language delivery does not take five times as long. Large batches follow a rolling schedule: you receive the first videos while later ones are still in production, which lets your training team start deploying immediately.

The heaviest users of this service are equipment manufacturers and site operators. Builders of excavators, cranes, and crushers subtitle operator and service videos for dealer networks; our heavy machinery page details how video fits into their wider documentation programs. Wind and solar operators subtitle turbine and inverter maintenance content for multilingual field crews. Manufacturers subtitle line-changeover and quality videos for plants in Mexico and Quebec. After fifteen years serving these industries from our Texas base, the pattern is consistent: video is where technical knowledge now accumulates fastest, and it is usually the last asset anyone remembers to translate. If you are unsure whether your library needs subtitles, captions in the source language, or full voiceover, ask us; as ATA and GALA members we will tell you plainly which option fits, including when the cheaper one is the better one.