Anatomy of a Subtitle Job: Spotting, Reading Speed and SDH

What arrives with the job: video, transcript, style guide and target platform

Every subtitle project begins with a delivery from the client. For a standard twenty two minute TV episode, most translators receive the full video file, a dialogue transcript, a style guide, and instructions for the target platform. The video is usually in MP4 or MOV format, with timecoded reference audio. The transcript may be a plain text file or a Word document, sometimes with embedded timecodes, but often it's just the dialogue as spoken.

The style guide sets out the rules of the road. It covers terminology, spelling conventions, treatment of numbers, and how to handle profanity. The client may specify character or line limits, speaker identification, and preferences for on-screen text translation. The target platform, streaming service, broadcaster, or DVD, dictates the technical requirements, especially the file format and any accessibility needs.

Some jobs arrive with spotting lists that already mark suggested in and out times for each subtitle, but many do not. In those cases, the translator must start from scratch, which adds hours to the timeline and affects the quoting process.

Keep reading: Where Post Editing Work Is Heading and How to Price It

Spotting the timecodes and why subtitle boundaries follow shot changes

Spotting is the process of defining the exact in and out time for every subtitle. This step sets the backbone for the whole file. Subtitles need to appear and disappear in sync with what is said and what is shown. The goal is to guide the viewer without covering key visual information or causing distractions.

Subtitles usually start a few frames before the relevant speech begins and end a few frames after the line is delivered. Timing must allow for natural reading, but also respect the editing of the original program. In most cases, subtitle boundaries are aligned with shot changes. This helps the viewer follow the story without jarring interruptions. When a scene cuts to a new location or camera angle, a subtitle that lingers across the cut can confuse or distract. Therefore, spotting often means adjusting in and out times to land within a shot, rather than running across edits.

Some scenes bring extra challenges. Rapid-fire dialogue, overlapping speech, and visual gags may require more precise timing or faster subtitle transitions. Good spotting balances viewer comprehension with fidelity to the source.

Characters per line, line count and reading speed limits

Most platforms and clients set strict rules for subtitle length. The maximum is often forty two characters per line, with a two-line limit per subtitle. This keeps text clear and prevents it from crowding the screen. Font choice, screen size, and aspect ratio all play a part in these decisions, but the line and character caps are non-negotiable on many jobs.

Reading speed is the next constraint. The industry standard often caps reading speed at around seventeen characters per second, though some streaming platforms prefer lower. If a subtitle exceeds that rate, viewers may not finish reading before the next one appears, especially if the content is dense or complex.

Subtitle files are checked for reading speed violations. If a subtitle breaks the limit, the translator must either condense the text or split it across two subtitles, which may not be possible if the dialogue is rapid and the timing is tight. This is a key reason that not every spoken word makes it to the screen.

Keep reading: A Week Inside a Two Person Language Services Company

Condensing dialogue without losing the meaning a viewer needs

Condensing is an art in itself. The translator's job is to keep the essential meaning, emotion, and tone, even when there is no space to fit every word. Subtitles must be shorter than the spoken dialogue, but they cannot omit critical information or nuance.

Direct speech is often trimmed to the core idea. Fillers, repetitions, and false starts go first. For example, "Well, you know, I was thinking maybe we could go to the park or something, I guess," might become, "Maybe we could go to the park." The core intent remains, but the subtitle is now readable within the allowed time.

Idioms and slang present another challenge. A literal translation may take too many characters or confuse viewers. The translator must find an equivalent phrase in the target language that fits within the caps and makes sense in context. Humor and wordplay are especially tricky, often requiring creative rewrites that keep the spirit, if not the letter, of the original.

The goal is clarity and natural flow. If the viewer has to pause to reread or puzzle through a subtitle, it has failed. The best subtitle files are invisible: the viewer forgets they are reading at all.

SDH and closed captions: speaker labels, sound effects and who they serve

Subtitles for the Deaf and Hard of Hearing (SDH), and closed captions, are not the same as standard subtitles. SDH files include speaker identification, sound effects, and other non-speech audio cues. These additions give viewers who cannot hear the audio a fuller understanding of what is happening on screen.

Speaker labels appear when the dialogue is off-screen or when multiple characters speak in quick succession. For example, a subtitle may read "JOHN: I'll get it," to show who is speaking, especially if several voices overlap. This is essential in scenes with group conversations or voice-overs.

Sound effects and important music cues are bracketed and added as separate lines. Examples include "[door slams]," "[laughter]," or "[intense music]." The goal is to signal moments that matter to the plot or mood. Not every background sound is noted, only those that add meaning or context.

These extra elements add to the time and care needed in subtitle creation. The translator must decide what to include, based on the style guide and the platform's norms. SDH subtitles tend to be longer and more complex than standard translation, which affects the quoting and scheduling of these jobs.

See how WordRate handles this for translation and localization

File formats: SRT, WebVTT, TTML and what the platform will accept

The last step before delivery is formatting the work into a file the client's platform can read. The most common format is SRT (SubRip Subtitle), a plain text file with numbered subtitle blocks, start and end timecodes, and the lines themselves. SRT files are widely accepted by most streaming services and video players.

Some platforms require WebVTT (Web Video Text Tracks), which is similar to SRT but uses a different timecode format and supports extra metadata. TTML (Timed Text Markup Language) is required by some broadcasters and offers advanced styling and positioning options. There are also proprietary formats for specific hardware or software. Each format has its quirks, such as different rules for line breaks, italics, or special characters.

Translators must export the file in the requested format and check for compatibility. File conversion tools can help, but manual checks are still necessary to catch encoding errors or lost formatting.

Quality check against the picture, then the delivery package

Before sending the final files, every line must be checked against the picture. This is the stage where technical and linguistic errors are caught: mistimed subtitles, text that overlaps shot changes, lines that extend past the reading speed limit, and typos.

Some translators use subtitle editing software with a built-in video player, allowing them to preview the timing and placement in real time. Others export the draft file and check it in the platform's native player. Problems like overlapping subtitles, missing speaker labels, or incorrect sound effect placement are corrected at this stage.

The delivery package usually includes at least two files: the standard SRT and the SDH or captioned version. Clients may also request a copy of the glossary used, or a change log showing edits made during review. The package is zipped and sent along with a signed delivery form or email confirmation, depending on the client's process.

How the runtime minute price was built from the hours this job took

Quoting for subtitle translation is often done on a per runtime minute basis. This rate reflects the average time spent per minute of finished video, but the actual hours can vary depending on the complexity of the job. Here is a breakdown of how the work stacked up for this twenty two minute episode.

• Spotting: Setting in and out times for each subtitle took about one hour for every ten minutes of video, since the job did not include a spotting list.
• Translation and condensation: This stage involved careful editing to fit line and character limits, and took about two hours for the full episode, including research for idioms and jokes.
• SDH additions: Speaker labels and sound effects added roughly thirty minutes, as the episode had several scenes with off-screen voices and important background sounds.
• Formatting and export: Creating SRT and SDH files, checking compatibility, and converting to the requested formats required forty five minutes.
• Quality control: The final review against the picture took about half an hour, with small corrections to timing and text.

The total time invested was just over five hours. Dividing this by the twenty two minute runtime gives a little over thirteen minutes of work per minute of video. This ratio guides quoting for future jobs, though the exact figure will rise or fall depending on whether the client supplies a spotting list, how dense the dialogue is, and whether SDH or only standard subtitles are required.

For language service providers and freelance translators, tracking these hours, managing rate cards by client, and storing project glossaries can save time and help ensure accurate, consistent quoting. Tools that combine these functions let operators focus on the work itself, rather than the paperwork that follows.