Reading lips is only possible with a lot of context (at least for me). The same is true for any audio with no captions: if I know what the topic is, that limits the words you could potentially be using, so I can tell that you said "cool" and not "yule." But if you say something outside a conversation with a clear topic, I am lost in thousands of possibilities.
This has the really sad side-effect that jokes are way more hard to understand, because they usually break out of the context. So if we are talking about something and you suddenly reference something else, jokingly, I have to sift through my entire vocabulary to put together what you said, instead of the couple hundred words we were using before.
I think this is part of why I like jokes that you have to put together in your head-- that's how I have to do all jokes, so it feels like a "true joke." A lot of times I will make a joke and no one will laugh, and then someone else will make literally the same joke but put together so it is more obvious, and then everyone will laugh.
More often, people will make a joke and I will have to ask them to repeat it, or I will just miss it. Sometimes I laugh politely at what I thought was a joke, and then I worry that they said something shitty and I just accidentally expressed approval of it. So while I like jokes, I often find them stressful if the situation doesn't allow me to ask them to repeat themselves, or if they won't repeat themselves exactly.