This success criteria, along with 1.2.2 and 1.2.3, all have a qualifier in the description: "except when the audio or video is a media alternative for text." Let's quickly explore what a "media alternative" means so we can remove it from over-complicating our understanding.
There may be circumstances when you have text that you want to also deliver as an audio or video alternate version. For example, you have an article that is long and may be difficult for some to get through. As an alternate way of delivering the content, you record a slow and clear audio version of the article. In this instance, the audio version itself is the media alternative to text, thus the audio does not need its own alternate version required by 1.2.1.
While we are simplifying definitions, let's look at "time-based media." All audio or video that is delivered over time (after starting to play) are considered "time-based." We can comfortably remove this from the description to simplify it further.
With that in mind, this criteria can be simplified to:
This criteria simply requires that no content is exclusively delivered as audio-only or as video-only. There must be an alternate way to access the same content.
If you have content that is being delivered by prerecorded audio-only or prerecorded video-only, you will want to include an alternate text version. This could be simply adding a transcript of text describing the audio or video. Be sure to include all dialogue, as well as descriptions of sounds and visuals that are part of the story.
If the text alternative is not hosted on the same URL as the original media, you are required to link to the text alternative immediately after or before the original media version, along with labeling the text as a text alternative to the original content.