The parent post even used google translate to generate some of those sentences. You can't do that with video/audio.
You could enter the sentence into google translate, have it attempt to synthesize a pronunciation, and then you verbally mimic it to the best of your abilities. But now, not only are you compounding the inaccuracies of google translate, plus the inaccuracies of the speech synthesizer, plus the inaccuracies of your own pronunciation, which means the viewer has no reliable way to return to the original source material. But the entire process is still reliant on text at one point.
I happen to understand that, and if I haven't I would just google translate it.
Now, if you made a video saying that and I couldn't understand you - no such luck.
Text is the only medium for which we have reliable operations working on semantic level.