General Guidelines for Transcription
When transcribing, you'll need to use one of the formats which will have
format-specific rules. However, please keep the following rules also in mind as they apply to every format.
NAMES & MENTIONS
-
Names are often censored for privacy reasons. If a name is censored, use the color the submitter used to
censor the name instead, i.e. **Red**, **Blue**, **Pink**, etc. If there isn't a color use User
1 or similar. If this is for messages, use colors or "Left" and "Right" or the initials provided in
the image. If those options aren't available, you can use Redacted or [Censored].
Please DO NOT use bold with brackets like [Censored].
-
Make sure that the community you're working with doesn't have any rules on sharing personal information.
For example, on Reddit, many subreddits do not allow any kind of personal information
(like names or profile pictures) to be shown. Ensure you're following the expectations of the community
as best you can.
-
Names of media (shows, artwork, etc.) should be in quotes, such as
Fred from "Scooby Doo"
or
Fred from 'Scooby Doo'
.
LINKS
-
Links should be described like an image instead of typed out to avoid spam filters. For example:
[Link to BBC News]
WHAT TO TRANSCRIBE
-
If the date does not provide any relevant information do not include a date and time.
-
Do not start your description off with "Image description: description" or "Image description -
description". Just start describing right away.
-
Do not include likes, loves, shares, etc. or the number of comments, unless it is relevant
to the
context and/or content of the post.
-
If the text to be transcribed includes emojis, please copy and paste the emoji into your transcription; we
suggest emojipedia.org.
-
If the text to be transcribed uses special Unicode characters (єχαмρℓє 1, ǝxɐɯdlǝ 2, ệẍäṃṗŀệ 3 etc.) or
script in a different alphabet/language (eg. Mandarin/Kanji) then please be aware that some screen readers
do not process them accurately. If they are included literally, they should also be described.
STYLING
-
For each subsequent reply in a reply chain, increase the number of >s on the side by one.
This will continue nesting the comments to indicate that each is a reply to the one above it. If the
comments are each replying to the post but not each other, this is unnecessary.
-
Do not include a horizontal rule between a post and its comments.
-
When text is on another line due to text wrapping (not enough space to show it on the same line due to the
image), keep typing it on the same line instead of making a new line.
-
Please ensure that any text in the picture is transcribed verbatim rather than just being paraphrased.
-
When transcribing humor, transcribe the picture in a way that the heart of the joke remains the same, so
that the transcription would be as funny to read as the image of the meme or comic.
-
When multiple paragraphs of a transcription are consecutive and indented with the same number of >, there
should also be a > on the empty line between the paragraphs.
-
When you are mixing in italics and bold in the same line, you MUST end each formatting
before you begin the next one or it won't format on old Reddit. This also may be an issue on other
platforms. This means that if you want to type out:
[This is a description that is bold]
You must type it out as:
[*This is a description* ***that is bold***]
-
In markdown, you must type two spaces at the end of a sentence to make a new line. If you want a paragraph
break, you will need to put an extra empty line between your sections. For example, you type:
text
text
...will render to text text
.
Using •
as our example space character, you would need to type:
text••
text
...to get:
text
text
-
When text is stylized in a way that your chosen platform cannot handle (i.e. highlight, underline,
hyperlinked), you can describe the beginning and end like:
[*Circled in red*] Text [*End circled in red*]
-
When different post types are embedded in one another such as a Facebook post with an image of a Twitter
post, you should use the format that matches the host site and describe the beginning and end of the embed
while indenting the embed by one > on every level. For
this post, the transcription should look like this image.