How do voice and video change what informational quality means?

We had treated voice and video as output formats of the same content system. After several pieces we have changed our judgement: they are not formats, they are different models of quality.

A written report allows you to go back. If a sentence turns out to be ambiguous, the reader rereads it, compares it with the table and settles the doubt unaided. A narration offers no such way out: it is heard once, in order, and an ambiguity becomes a wrong conclusion that nobody corrects.

That shifts the weight of checking. In text, a good part of the load is carried by the reader; in audio, it has to be carried by the piece. We have had to write differently: less subordination, the figure before its qualification, an explicit subject where the text would have allowed a pronoun.

Video adds a problem that did not exist before. What is said and what is shown can stop matching: a figure updated in the script and not in the chart produces a piece whose two halves contradict each other. No text check catches that.

For the moment we have resolved it with human review of correspondence before publication. It works, it does not scale well, and it is a limit we prefer to declare rather than conceal.