Input example
approved script + final voiceover audio for this publishing workflow
Subtitle workflow guide
By Yana Li
Validators catch broken timestamps, overlapping subtitles, and malformed syntax. They catch none of the things that actually make viewers give up on captions: line breaks that fight the sentence, lines that outrun reading speed, and captions that collide with on-screen text. This guide is a ten-minute review for that layer, built from publishing the same scripted show in five languages.
Input example
approved script + final voiceover audio for this publishing workflow
Output asset example
SRT/VTT subtitle assets plus quality notes for downstream upload or editor handoff
Common quality signal
Late narration edits shift subtitle timing against the approved script.
Decision points
Where a subtitle line breaks is where the viewer's eye pauses, and a break in the wrong place forces the brain to re-parse. 'He said he would never / sell the company' splits the phrase at its point of emphasis; 'He said / he would never sell the company' breaks at a natural speech pause. Same words, same timing, same valid file — only one of them reads cleanly. Most auto-generated captions break lines purely by character count, so this check almost always finds something.
Every captioning house style sets a reading-speed limit, usually in characters per second. The exact number is a style decision, not a legal one: WCAG requires captions to be synchronized and equivalent, and it does not set a characters-per-second figure. Whatever threshold you pick, the failure mode is the same — a subtitle flashes by faster than a person can read it. The fastest lines cluster where narration accelerates or where two sentences were merged into one subtitle.
Captions sit on top of a picture that often contains its own text: titles, lower thirds, on-screen quotes, UI recordings. When the caption and the picture demand the same eyeballs at the same moment, the viewer reads neither. This failure lives entirely outside the caption file, which is why no validator can see it.
Nobody owns it. The editor treats captions as an export setting, the captioning tool never sees the final picture, and the reviewer checks only that captions exist and match the words. Readability falls between those chairs because it requires looking at the words, the clock, and the picture at the same time. If you publish in more than one language, the failures move around: reading-speed and line-break rules calibrated for Latin scripts behave very differently on Japanese and Chinese.
Practical workflow
Scrub through the video and read only the line breaks. Flag any break that lands inside a tight phrase: verb plus object, negation plus verb, a name split from its title.
Find your three fastest subtitles and play them at normal speed with the audio muted. Muting matters: when you can hear the narration you already know what the caption says, and you will believe it is readable when it is not.
Watch the video once paying attention only to the bottom third of the frame, and note every timestamp where a caption overlaps burned-in text, a chart label, or a face during an important beat.
Fix what each pass found: rebreak lines at phrase boundaries, split or retime the fastest subtitles, and move or retime the ones that collide with the picture.
If you publish in multiple languages, repeat the review per language rather than trusting one global threshold.
Product boundary
TimedSubs creates subtitle files and quality notes for downstream video workflows.
FAQ
No. WCAG requires captions to be synchronized and equivalent to the audio. Reading-speed thresholds like characters per second are house-style decisions that make captions comfortable, not legal requirements.
No. Validators check syntax, timing validity, and overlaps. A file can pass every technical check and still break phrases badly, outrun reading speed, or collide with on-screen text.
Yes. CJK scripts carry more information per character and have no spaces, so line-break and reading-speed rules calibrated for Latin scripts do not transfer. Review each language on its own terms.
TimedSubs checks reading speed and line length while timing subtitles from your approved script and audio, and the free subtitle checker reports long lines in existing SRT/VTT files. The picture-collision pass is a human review by design.
Related guides
View all guidesWorkflow guide
Prepare YouTube SRT or VTT subtitle files from an approved script and final voiceover without relying on auto-caption wording.
Text to SRT guide
Create an SRT file from text, TXT, or a finished script by pairing it with matching audio, preserving the wording, and checking subtitle timing before export.
Workflow guide
Prepare SRT/VTT subtitle assets before importing into CapCut or Jianying short-video workflows.