If your audio has multiple speakers or speakers who are not known in advance, there is a workaround. To get enough accuracy for closed captioning, you need another person involved in the process, who is known as a shadow speaker, voice captioner, or re-speaker. The shadow speaker acts somewhat like an interpreter who translates from one language to another for a live event. Instead of listening in one language and speaking in another, they just have to listen to the words from different speakers and repeat the words out loud (because the speech recognition software only recognizes their voice).