Maths notation in captions: Unicode, LaTeX or plain text?
Caption files are plain text. There is no MathML in an SRT, so a formula has to be written in characters. Three conventions work, and the right one depends on where the captions are going.
Unicode: for caption players
Superscript and subscript digits (x², xᵢ), Greek letters (σ, π, θ), the integral and sum signs (∫, Σ) and the arrows and relations (→, ≤, ≈) are all single characters that every modern player renders. This is the default and what most courses should use. Its limit: not every letter exists as a superscript, so “x to the q” falls back to x^q.
LaTeX: for notes and MathJax
If the transcript is going into lecture notes, a MathJax-enabled LMS or Overleaf, LaTeX keeps every expression exact. Stemscribe wraps each expression in $…$ so it renders in place, and leaves prose untouched.
Plain: for screen readers and strict systems
Some captioning pipelines strip or mangle non-ASCII characters, and some screen readers read “x²” as “x”. Plain notation (x^2, sqrt(x), pi) is the safest for those, and it is what many accessibility guidelines ask for.
| Spoken | Unicode | LaTeX | Plain |
|---|---|---|---|
| x squared over two sigma squared | x²/2σ² | $\frac{x^{2}}{2\sigma^{2}}$ | x^2/2sigma^2 |
| the square root of two | √2 | $\sqrt{2}$ | sqrt(2) |
| the integral from zero to pi of sine x d x | ∫₀^π sin x dx | $\int_{0}^{\pi}$ $\sin x$ dx | integral[0,pi] sin x dx |
| the sum from i equals one to n of x sub i | Σᵢ₌₁ⁿ xᵢ | $\sum_{i=1}^{n}$ $x_{i}$ | sum[i=1..n] x_i |
| big o of n log n | O(n log n) | $O(n log n)$ | O(n log n) |
| h two o | H₂O | H_{2}O | H2O |
| nine point eight meters per second squared | 9.8 m/s² | $9.8\,\mathrm{m/s^{2}}$ | 9.8 m/s^2 |
| alpha is less than or equal to beta | α ≤ β | $\alpha$ $\leq$ $\beta$ | alpha <= beta |
What is not converted
Only expressions in a mathematical context are typeset. “Two plus two people came over”, “she squared the deal” and “pi day” stay English, because the operands around them are words, not variables. A Greek letter becomes a symbol only next to an operator, a number or another variable.