Transcription Guidelines for Math content

CaptionSync allows you to submit Math content for Closed Captioning. This article details the transcription guidelines for such content.

Math content is particularly difficult to transcribe due to the usage of special symbols, many of them non-Latin characters. Because determining when and where to use symbols rather than words can be difficult, we have set several transcription guidelines, which we use for Math content. These guidelines are motivated by the following: 

• Special symbols in the text can lead to uncertainty in exactly what was said.

• Many special symbols are not included in the standard character set for some caption formats.

• Our goal in captioning is to try to capture what exactly was spoken, without imposing our own interpretation of the content.

Guidelines:

• Numbers zero through nine should be written out, while 10 and above can be represented as numerals.

• Use words for numbers of things (e.g., three axes; two points; the one exception), but use numerals for numbers as numbers (e.g., the number 5).

• When a number is the first word in a sentence, write it out, except if it's part of an equation. Examples:

Four customers walked into the store.
4 is the square root of 16 

• Use numbers for degree and radian amounts (e.g., 270 degrees; 3 pi radians; 45-degree angle).

• Use commas in numbers with 5 or more digits: 1000 but 10,000. For numbers 1 million and above, use "34 million" or "245 billion", rather than writing the numeral with all the zeros.

• Use first, second, third, tenth, but 11th, 12th, 13th, and so on.

• Use the word "zero" if it is the only number that appears in a sentence; use the digit "0" if other numbers appear in the sentence.

• If it is an even dollar amount, do not use the decimal unless cents are spoken (e.g., use $15, not $15.00).

• Write out operators as words: plus, minus, equals, is less than, is greater than, times, etc; unless as a part of a formula or equation.

• In general, do not spell out numbers or operators in formulas or equations. However, the operators "times" (x) or "divided by" (/) can cause confusion; "x" can be confused with a variable name, and "/" can be confused with fraction notation. If there is any room for confusion, revert to spelling out these two operators.

• In broadcast outputs, the following characters cannot be used: *  \  ^  _  `  {  |  }  ~ . For other outputs, those characters are allowed, but some of them need to be escaped: \^ , \* and \\ . Examples:

7.3 10\^4     (exponentiation)
3\*8     (multiplication)

• When spelling out the operator, use the term that the speaker used in the audio. For example, if the speaker says "x over 2" then use "over"; if the speaker says "x divided by 2", then use "divided by".

• When reading formulas or equations, speakers can omit the "times" operator and say something like "3x" (meaning 3 times x); in these cases, transcribe as spoken: "3 x". Put a space between the operands (the 3 and the x, in this case).

• If a formula or equation contains additional information, such as an interval of validity, apply the equation rules to the entire phrase. Examples:

y = 4 – x squared on the interval negative 2 to 2
s(n) = the sum i from 1 to n of  6 - (2 i over n)

• Write out the words "negative" and "positive" preceding a numeral (e.g., "negative 5").

• Write out superscripts and subscripts (e.g., "x⁶" as "x to the sixth"; "the point x₁, x₂" as "the point (x1, x2)").

• Write out the words "squared" and "cubed" (e.g., "4 x squared" for 4x²; "4 b squared c" for 4b²c).

• Write out fractions (e.g., one-fifth).

• Write out the following words: infinity, factorial, percent, prime, degrees, pi, theta (or any Greek letter).

• Write out all trigonometric functions: sine, cosine, tangent, cotangent, cosecant, secant.