A fundamental problem is that audio/video are quantizable in fairly cheap ways for recording (microphone/ccd) and playback (speaker/display), in very low dimensionality (typically 1-2 for audio, 3 for video)
Smell/taste exist in much higher dimensions, and it's unclear what both the recording as well as the playback "canonical devices" would be like. I've tried academic demos that give zaps to your tongue in order to recreate taste, and sure they sometimes triggered some faint taste for me but it still mostly feels like something zapping your tongue.
For smell, there doesn't seem to be much of a way other than having some sort of device with lots of tanks of various compounds for various flavor profiles that get puffed in your face, which you then have to refill as they get used up. A personal version of such hardware would likely be very cumbersome and expensive for limited upside? Some movie theaters do have such systems, but it remains a rare novelty.
Touch has a similar problem - all haptics feel like things buzzing under your fingers/in your hands, we don't really have direct paths to recreating the richness of tactility in the world. The one difference is that while the use cases for smell/taste seem pretty limited, touch is of huge relevance right now for ar/vr interactions.
When we're at the point where we'll be directly interfacing with brains though, things will likely be different. But that's not for a while.
I didn't understand the nature of the challenges, disappointing to hear consciderering how smell could add to the depth of storytelling. I wonder in the future if you could get a toungue transplant to recreationally change your taste buds.
