Úr Linguistic Team International Wiki
Stökkva á: flakk, leita

LTI WikiLTI GuidelinesOur Methodology | English Guidelines | Quick Reference Guidelines | Glossary(edit)

This page aims to simplify and update our guidelines to make them as user-friendly as possible. Should anyone wish to address the issue of the guidelines, please raise it during one of our bi-weekly meetings . You can also raise it in the LTI forum.


  • It goes without saying that members should be aware of their own fluency of the English Language and should possess a decent command of the language if they want to contribute to the English Team.
  • Members must realize that collaboration and communication is crucial if the Linguistic Team is to achieve its aim of being the sole transcription/ proofreading and translation hub for all TZM / TVP-related multimedia. Please join and contribute within your Language Team's primary communication medium, which can be found in the Language Team Information page.
  • Transcribing / Proofreading may become harder as one takes on longer projects and it's important for members to be able to finish what they have started. As a result, these are usually broken up into smaller chunks to be handled by more than one person. Please contact your project coordinator if you wish to help with them.
  • It is equally important that the more active members understand the absolute need to take occasional mental breaks, both within longer sessions, as well as overall. Some members have mentioned that this kind of effort can be highly addictive, which makes mental burn-out a very real concern.
  • Before opening a video project to start or continue a transcription or timestamp-shifting task, it is important to clear the cache memory from your browser and reboot your computer, so that the video timings presented by your browser are trustful. For information on how to clear the cache memory of your browser:
    • On Firefox: Preferences => Advanced tab => Network tab => Click the 'Clear now' button
    • On Explorer: Tools => Internet Options => On Temporary Internet Files, click on 'Delete Files'
    • On Opera: Menu => Settings => Preferences => Go to Advanced tab => "Click on History" from the menus on the left => Click the button 'Empty Now'.
    • On Google Chrome: Preferences => 'Under the Hood' tab => "Clear browsing data..." button. => Check 'Empty the cache' => Click 'Clear browsing data' button

Guidelines for video content

Key actors

This is an overview of the main roles to take part in the process of transcribing a video:

  • Transcribers: Transcribers write down the captions for the video. They have a good command of the English language, being able to distinguish every single word that is uttered in the video so that the transcription is as close to the material as possible.
  • Timestamp Proofreaders, or "Timeshifters": They make sure that captions are perfectly synchronized with the video, and that the subtitle strings structure (the way the text is cut in different strings) maintains a harmonious reading flow, both in reading speed and in sentence structure. This activity does not require a perfect knowledge of English, because its main focus is on the presentation of captions themselves, and not their content.
  • English Proofreaders: These members are usually English native speakers with a strong understanding of the language and, therefore, have a sufficient command of the English language to make the final corrections in the subtitles ( grammar, punctuation).

General Guidelines

Every video project passes through 5 stages before it is ready for the translation teams:

  • 1- Transcription
  • 2- First Round of proofreading
  • 3- Timestamp-proofreading (timeshifting)
  • 4- Second Round of proofreading
  • 5- Final Review.

Transcribers and translators have been included in the process so as to agree upon our methodology, which has led us to the following guidelines :

The blue indicates that the guideline is a major concern for the transcriber of the video. In spite of the fact that the time-shifter will proofread it later, an increase in text strings that doesn’t follow the guidelines can result in the time-shifters taking ~10x more time to complete the task.

The brown indicates that the guideline is a major concern for the time-shifter when he/she is proofreading the transcription. The transcriber doesn't need to spend too much time trying to respect all of these guidelines.

  • Consider the strings structure through 2 main parameters: a) the grammatical structure of the sentences

when determining the transition point from one string to the next one, and b) the character per second ratio in the subtitles strings display time, in order to achieve the best reading comfort possible.

  • Minimum display time for a subtitle: 1.5 seconds

Consider leaving at least 1.5 seconds for 1 line of text or less (42 characters), and at least 2 seconds for more than one line (more than 42 characters), if possible.

  • String length consideration for a subtitle string, ~70 characters (including spaces),

Since the maximum in a TV screen is 84 chars (two lines, 42 characters per line), this reference of 70 chars is taking other languages into consideration (those that use more letters than English to say the same thing). Some languages need up to an extra 25% of string space, but be aware that splitting the speech in smaller strings can also compromise the reading comfort.

  • Start subtitles exactly at the same time as the speaker’s voice, but without compromising the reading comfort in a way that might cause the viewer to miss any content.
  • If there is silence after a subtitle, extend subtitle time up to 2 seconds,

to allow more reading time when possible; this measure is subjective to the length of the string, of course. (Example: add 25% of silence after the speech pauses).

  • Leave between 100 and 140 milliseconds between strings.

This provides a subtle visual cue to the reader that a new subtitle has replaced the previous one.

  • Maximum display time for a subtitle : 6 seconds

Anything longer might encourage readers to read the same string multiple times, which is unnecessarily distracting. Such long strings are easily dividable into two separate 3-second strings.

General Advice

The Four Basic Principles of Captioning (‘CARE’)

Please remember that four principles should be kept in mind when transcribing :

A uniform presentation of text strings facilitates viewer understanding.

An accurate transcription is always the goal.

The subtitles are displayed with enough time to be read completely, and synchronized with the audio.

Equal access means that the full contextual meaning of the materiel is completely preserved.

The Process

While transcribing or time-shifting a transcription, don't focus only on the synchronization of the subtitles with the speakers' voices. Also consider the HARMONY OF THE FLOW OF THE SUBTITLES WITH THE SPEECH AND SENTENCE STRUCTURES. Listen to the first 10 to 15 seconds, so you can form a mental preview of how the speech might be most appropriately split into strings and evaluate how well any existing timings are in accordance with any of the guidelines, and if it could be done better.

In more detail...

- 1st consideration, the length of the strings. If a string is longer than 1.5 lines (~70 characters), it is not leaving 25% extra space for other languages that will need it. Split the strings that are too big or transfer a few words from 1 string to the next, in order to correct it. If the strings are too small, the transcription can be a strain to the eyes and increase the effort of the reader (changing continuously, not allowing the necessary rests so that the viewer can observe the image as one reads). Other guidelines are implicit on this point:

  • a) a minimum of 1.5 seconds of display time for each string - In this case, if the extension of the timing is not possible, a merging of strings or transference of words from one string to the next one are 2 solutions possible;
  • b) if there is still space (silence) after the speech until the next string starts, extend the display time of the previous string up to 2 seconds, subjectively, according to its size/display time (25% more display time).

- 2nd, but equally as important as the first point, as both are basic conditions for a video transcription/translation to be comfortably absorbed, the synchronization of the beginning of speech with the display of subtitles should be exact. If these two are not synchronized, the ability to focus on the message that's conveyed in the video will be compromised.

- 3rd, adding 100 to 140 milliseconds time between strings is one detail that represents a measurable impact on the viewer. Subtitle strings that don't take this break into account stress the eyes of the reader, who will not be able to absorb the message comfortably. A minimum of 140 milliseconds is necessary for the desired effect. Sometimes, 100 milliseconds is enough when an image transition on the video helps create the awareness of the subtitles' transition.

- 4th, The harmoniousness of the text should be considered when we split strings, considering the grammatical structure of a phrase. Cutting a string between a noun and a verb, or between a verb and a predicate, can have a very different result from cutting it according to the sentence segments.

Style Guidelines for Original Writers (Newsletter Contributions)


  • Natural languages such as English have unspoken rules regarding eloquence. We must heed these rules if we are to have our material easily read by the masses we wish to reach.


  • The use of the same word over and over again, even if it fits the meaning, is not easy to read. Thesaurus.com should be used liberally.


  • This should go without saying, but all information should be written without heavily accented dialogue. Levels of eloquence, however, must be maintained as a standard.

Continuous Review

  • Once you feel you have finished your work, leave it and go do something else. Then return later to take another look at it, further editing as you feel necessary. For difficult passages, you might want to repeat this a number of times before submission. Writing in natural language is not an exact science, so rechecks are very likely to uncover places that can be improved beyond your first attempt. Before you distribute it to the RBE Community, however, please pass it on to another proofreader as the 'peer-review' process is what allows us to produce the highest quality translations possible from within an all-volunteer system.

Logical Flow

  • Stay focused on one subject per paragraph. If you notice the original material going off on a tangent, remove the sidetracked information and group it appropriately into another paragraph.

Additional Resources


  • In order to help translators, members are encouraged to paraphrase complex English terms to facilitate translations. This does not apply to Proper nouns and titles such as Resource-Based Economy.
  • While this should be automated in time, translation teams are advised to create their own glossary of terms for use in future works. These language glossaries should be used to store collaboratively chosen translations for difficult, but frequently handled terms and phrases.

Linguistic Team Forum

Further Reading

Canadian Association of Broadcasters, "Closed Captioning Standards and Protocol for Canadian English Language Television Programming Services" [1]

Described and Captioned Media Program, "Captioning key : Guidelines and preferred techniques" [2]

The Annenberg Public Policy Center of The University Of Pennsylvania, "The state of close captioning services in the United States : an assessment of quality, availability, and use"[3]

Fotios Karamitroglou, "A Proposed Set of Subtitling Standards in Europe" [4]

LTI WikiLanguage Team Contacts | FirstTimers-General Guidelines | LTI Guidelines | Transcription Team | Proofreading Team | Translation Teams | Translation Proofreading Teams | Linguistic Tech Team | Media Center | Project Teams | LingTeam Development | Meetings (edit)