Free tools Windows power users keep installed
One-click scans. No signup required.
iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Subtitle reading speed, measured in characters per second (CPS), is the number of characters displayed divided by the seconds they remain on screen. There is no single CPS limit that works for every language, platform, audience, or program: Netflix’s published guidance, for example, sets different caps in its general templates and its US English guide. And meeting a cap does not guarantee a subtitle is readable. Transcription, translation, text breaks, and timing all have to work together.
What CPS means in subtitles
CPS is calculated by dividing the number of characters in a subtitle event by the time it is displayed. A 40-character subtitle shown for two seconds runs at 20 CPS. It is a practical way to flag text that may be difficult to read in the available time, not a universal measure of what every viewer can comfortably read.
| # | Preview | Product | Price | |
|---|---|---|---|---|
| 1 |
|
Blackmagic Design USB Davinci Resolve Editor Keyboard | $669.00 | Buy on Amazon |
Counting rules matter. Netflix says its software counts spaces and punctuation toward both character count and reading speed, and calculates CPS for each subtitle event as well as for the file average. A manual count that excludes spaces or punctuation may therefore differ from a platform validator’s result. Netflix Partner Help Center explains its reading-speed measurement.
What is a good subtitle reading speed?
Use the specification for the platform, language, audience, and program you are preparing; do not treat one service’s cap as a general human reading threshold. Netflix’s current guidance illustrates why the exact guide matters:
#1 Best Overall
- Designed for professional editors who need to work faster and turn over quickly
- Designed for DaVinci Resolve 16
- Integrated search wheel integrated directly into the keyboard
| Netflix guidance | Adult programs | Children’s programs |
|---|---|---|
| General subtitle templates | Up to 17 CPS | Up to 15 CPS |
| English (USA) Timed Text Style Guide | Up to 20 CPS | Up to 17 CPS |
These are production limits in two Netflix guides, not competing estimates of a universal reading speed. Check the applicable guide and its current requirements for the specific delivery. A different way of expressing reading pace is words per minute, but converting WPM to CPS is unreliable without accounting for language and the actual text: words vary in length.
Why an acceptable CPS can still be hard to read
A subtitle can fall below a CPS cap and still be difficult because the number says nothing by itself about phrasing, synchronization, or what viewers must watch. Dense language, an unnatural break between lines, a late subtitle, or text held across an unsuitable shot change can all make reading harder.
Timing is an editorial decision, not just a matter of leaving text on screen longer. Netflix’s timing help notes that shot changes can take precedence over extending a subtitle, and that text may need to be edited to meet reading-speed requirements. For overlapping dialogue and on-screen text, Netflix advises prioritizing the most plot-pertinent message rather than severely truncating text or slowing it excessively to include everything; dialogue-heavy material may require timing adjustments alongside truncation. Netflix’s timing guidance describes these trade-offs.
Why AI subtitle workflows can miss the mark
“AI subtitles” can mean several different operations: automatic speech recognition (ASR) transcribes speech, machine translation (MT) renders it in another language, and automated spotting or segmentation assigns display times and line breaks. A plausible transcript does not prove that translation is accurate, and accurate words do not prove that the result is well segmented or timed.
Transcription and timing are separate problems
A 2024 study tested two ASR tools on English and Italian material, including a British talk show and a US film dubbed into Italian. The authors reported that both tools fell short of the industry expectation assessed, particularly in Italian, and found segmentation and timing relatively poor in both languages. They concluded that substantial customization and human post-editing were needed for broadcast-ready subtitles. The findings apply to the tools and material studied, not every current speech-recognition system. Read the 2024 ASR study.
Translation can sound plausible and still lose meaning
A 2024 case study examined Google Neural Machine Translation for Chinese-to-English movie subtitles. It reported errors in approximately a quarter of the subtitles in its collected film sample, with functional-equivalence errors most common, followed by acceptability and readability errors. The authors identified issues including semantics, idiomaticity, grammar, line length, proper nouns, culturally specific terms, and incomplete sentences. That sample-specific result is not an industry-wide error rate for AI subtitles. Read the Chinese-to-English subtitle case study.
Post-editing still involves structural decisions
A 2022 comparison of Swedish subtitle samples found that the post-edited subtitles in its 2020s sample were faster, more oral, less cohesive and complete, and lower in the assessed quality areas than the 2010s sample. This documents an outcome in a particular sample and study design; it does not show that machine translation invariably reduces quality. Read the Swedish subtitle comparison.
A 2020 process study involving 12 professional subtitlers also identified segmentation and timing as important challenges in MT post-editing. Editors are not merely correcting words: they must make subtitle units readable and synchronize them with speech and picture. Read the 2020 post-editing study.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How to review an AI-generated subtitle file
Assess the whole subtitle task, not only the average CPS. These checks combine error areas identified in the cited studies and platform guidance; they are not a claim that every study used the same checklist.
- Check transcription against the audio. Listen for omissions, substitutions, speaker changes, and errors in names or terminology.
- Check translation in context. Confirm that meaning, idioms, cultural references, and proper nouns fit the scene, not merely that each sentence reads grammatically.
- Apply the intended reading-speed rule. Use the relevant platform and locale guide, including its character-count convention, and inspect individual subtitle events as well as the file average.
- Review segmentation and line breaks. Break text at natural linguistic points so that a subtitle reads as a coherent unit rather than a string of short fragments.
- Check synchronization with dialogue and picture. Make sure the subtitle appears and clears appropriately, and consider speech pace and shot changes.
- Review completeness and consistency across the scene. Check that important content is not missing and that names and terminology remain consistent.
When comparing two subtitle versions, record the governing CPS limit and counting method, language and audience, transcription and translation accuracy, segmentation, timing, and completeness. An average CPS alone cannot reveal whether a few individual events are too fast or whether the subtitles are awkwardly divided.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →
