YouTube has added a pile of improvements to the closed captioning service it launched in 2006. Automatic captions use speech recognition to build captions from the audio track, and they're now available in Japanese and Korean in addition to English. To test the feature, we let YouTube listen to a Japanese video on sign language (below), and while we thought we were hearing a discussion about Kawaji Sensei's teaching methods, the subtitles we got had something to do with low air pressure, a 7500 yen debt, and a Japan-Korea war. That said, the player reminds us that the feature is in beta, and we expect it to get better with time. Transcript syncing syncs the words in a text file to the video's audio in order to get the captions onscreen at the right time, and Google's making the service available in all three languages, too. The new changes also include a ", cc" search token that you can add whenever you want to limit your results to videos with captions, as well as new color and font preferences, and support for a bunch of new formats. The changes are live on the site, and with 1.6 million captioned videos and 135 million videos that support automatic captioning, the world is your oyster.