Subtitle Translation Services

Professional subtitling protects meaning and readability with tight timing and layout rules. Typical targets are 1–2 lines per subtitle, about 37–42 characters per line, and roughly 15–20 characters per second so viewers are not rushed. Files are time-coded to frames (for example 23.976, 24, 25, or 29.97 fps) and delivered in standard formats like SRT or VTT. Good workflows include at least two checks, one linguistic and one timing, which reduces rework later. Share your final video and a short 60–90 second sample to lock style and speed before full production. Confirm if transcription is included and whether on-screen text will be localized. Tomedes assigns native subtitle specialists and a project manager, follows these reading-speed and line-length norms, and supports over 270 languages with timed files ready for upload.

Quality comes from native linguists plus structured checks. A common baseline is a two-pass review, one translation edit and one timing/QC pass, covering CPL limits (about 37–42), CPS (about 15–20), and shot-change alignment within a small tolerance window, typically ±2–3 frames. Style guides cap lines at two and keep minimum cue duration near 1.0 second to remain readable. Ask for a 60–90 second pilot to confirm tone, key terms, and pacing, then apply the same rules at scale. Provide a glossary and preferred spellings upfront to reduce revisions by an entire cycle. Tomedes uses a documented checklist for line length, reading speed, and spotting, then validates the final files against target platforms before delivery.

Yes, with parallel teams, shared glossaries, and batch delivery. A common plan is to pilot 1–2 minutes, then process the full runtime in daily or twice-weekly batches, which keeps feedback loops under 24–48 hours. Multi-market launches often span 5–30 languages, so consistent CPL/CPS targets and a single termbase prevent drift between versions. For hour-long videos, per-language turnaround typically spans a few days when full QC and retiming are required, faster if transcription is not needed. State your go-live dates, rush needs, and priority languages so staffing can scale early. Tomedes coordinates linguists across time zones, locks style at the pilot, and phases handoffs so each market ships on schedule.

Most providers charge per minute of video because timing and reading-speed work sit on top of translation. A practical planning range for common languages is about $6–$16 per video minute, with rare languages or SDH subtitling adding a premium. Rush delivery can add roughly 20–50 percent. For example, a 60-minute video budgets to about $360–$960 at the base range, excluding rush or complex retiming. Ask what the price includes, such as transcription, a second QC pass, and revisions, and request a tiered table by language so you can allocate spend per market. Tomedes provides clear per-minute quotes, bundles common tasks, and offers volume and multi-language options for predictable costs.

Standard outputs are SRT and WebVTT, with TTML/IMSC available for streaming pipelines that need XML and richer styling. Files are delivered in UTF-8, aligned to your master frame rate (for example 23.976, 24, 25, or 29.97 fps), and kept to typical limits like 1–2 lines and about 37–42 characters per line. If you publish to multiple platforms, ask for a simple format matrix so the team exports exactly what each channel needs. For social, you may also want a burn-in version for quick posting. Tomedes supplies SRT, VTT, and TTML/IMSC, checks encoding and frame rate, and validates samples on your target platforms before final delivery.

AI can create a fast first draft, but publishing-ready subtitles still need human timing and style control. Platforms expect 1–2 lines, about 37–42 characters per line, and roughly 15–20 characters per second, none of which AI applies reliably without rules. Human spotting also fixes shot-change overlaps and minimum cue duration, which is commonly set near 1.0 second. A safe workflow is AI draft, then human edit and timing QC, then a short 60–90 second review sample before scaling. Share key terms and tone so edits are consistent across languages. Tomedes blends AI assistance with human subtitle specialists who enforce CPL, CPS, and platform rules and provide final QC.

Yes. SDH adds non-speech cues and speaker IDs to improve access for deaf or hard-of-hearing viewers. Readability targets stay similar, for example 1–2 lines, about 37–42 characters per line, and a minimum cue duration near 1.0 second, with careful placement to avoid UI or lower-third graphics. If you publish to the web or apps, aligning with modern accessibility guidelines improves consistency for users and regulators. Provide a list of recurring sounds and character names so cues are labeled the same way in every episode. Tomedes produces SDH tracks, applies consistent cueing rules, and delivers files ready for review and upload.

Yes. Rare languages require native linguists and early sampling to get pacing and style right. A good approach is a 60–90 second paid sample to confirm tone, then the full file in batches, each checked for CPL and CPS so reading remains comfortable. Expect the same delivery formats, usually SRT or VTT in UTF-8, matched to your frame rate. Share audience region, preferred terms, and on-screen text so adaptation choices feel natural. If the script is dense, agree on tighter line targets up front. Tomedes maintains native subtitle talent in Sinhala and many other less common languages and delivers timed files that meet platform norms.