[CivBase]

> It's faster to read text than to listen to someone speak, yet speech remains popular.

I suspect there are a few reason for that beyond the explanations you gave.

1) Speech is faster to produce than text. Professional captioners are among the few who can challenge that. For everyone else, it's usually faster to give a speech or say what's on your mind than to type it all out.

2) Speech frees up the eyes. My vision is usually in greater demand than my hearing, so I prefer to avoid committing it to a single activity for a long time if possible. Unfortunately, that's not an option for those without hearing.

3) Speech is easier to attribute to individuals. I can determine who said something with minimal effort just by listening to the sound of their voice.

4) Speech is more immersive for narrative content. Movies and TV shows feel more believable when characters are talking with each other face-to-face like they would in real life.

Unfortunately, a signer in the corner can't make up for any of those factors.

[bskrobisz]

I think these are reasonable. I should highlight, though, that 1, 3 and 4 are true for sign languages as well -- 3 and 4 doubly so (given how large a component individual expressions and gestures play in communication without sound).

[CivBase]

> 1, 3 and 4 are true for sign languages as well

That's true so long as the speaker is the one signing, which is probably true in the context of video calls but probably not in the context of TV.

[SamBam]

I believe 4 is. At least, in the sense that intent, rhythm and emotion can be conveyed by a signing translator better than text can.

[jcrawfordor]

I've specifically heard from a deaf person that they much prefer a sign interpreter over captions because the sign interpreter can convey nuance and mood more reliably than captioning.