The W3C has defined a markup language that provides for captions. SMIL (pronounced "smile") stands for Synchronized Multimedia Integration Language. SMIL became a technical recommendation of the W3C in 1998. At this writing, the current recommendation is SMIL2.1 and SMIL 3.0 is in last call.
It seems that you must use SMIL for RealPlayer and QuickTime, and SAMI for Microsoft Windows Media Player.
Here is a screenshot of a video example from the NCAM site with Captions.
These are closed captions, so you must let RealPlayer know that you want to see them, which is a task. If you can find the menus, use Tools > Preferences > Content and in the Accessibility section of the page, check both Use suppememtal text captioning if available and Use descriptive audio when available.
Closed captioning for TV was developed by the Corporation for Public Broadcasting television station WGBH in Boston. Its Caption Center, founded in 1972, played an instrumental role in the creation and passage of the Television Decoder Circuitry Act of 1990, the law that now requires built-in caption decoder circuitry in most new televisions. The National Center for Accessible Media, or NCAM, at WGBH, is working on the issues of captioning for the Web and is certainly the best resource in this area.
The NCAM site includes a number of examples captioned and video described science clips.
NCAM has developed MAGpie, the Media Access Generator, which is an authoring tool for making multimedia content accessible. With this freely downloadable, Windows-based authoring environment, developers can add captions in three formats: SAMI, QuickTime and RealText.
The home of MAGpie is the NCAM Rich Media Project. This is a valuable resource for tools and samples of captioning of multimedia on the Web.
In October 2007, AOL, Google, Microsoft and Yahoo! joined together to ask that NCAM establish and support ICF, the Internet Captioning Forum, to address the complex issues of captioning on the web.