Because both of those are physical phenomena (the emission of photons, and waves of pressure).
Smells are chemical. They require the release/adsorption of molecules.
Expanding on this: all light is made up of photons, and a TV screen or whatever else can easily create “matching” photons to the original image. All sounds is made up of vibrations in the air at certain frequencies, and a speaker can recreate those frequencies to give a copy of the original sound.
But smell needs specific chemicals. If something smells minty, that’s because there is menthol (and other stuff) drifting through the air into your nose. In order to reproduce that smell, a device would need to release the right concentrations of menthol and other chemicals into the air. Since it can’t just “create” those chemicals, it would have to have a stockpile of them. And there are many, many chemicals that create smells. I’m gonna guess millions. Any device that reproduced smells like a TV or a speaker can reproduce light and sound would have to have an enormous stockpile of chemicals in it. It would be prohibitively expensive.
To expand on this, our eyes work with 3 dimensions of perception – red, green, and blue – where we have about 400 different types of smell receptors. Capturing light takes one type of receptor. Reproducing light takes three similar emitters tuned to slightly different wavelengths. A smell recorder would have to perform 400 chemical reactions which, by their nature, would probably consume some sort of element that reacted with the chemical and need to be replaced periodically. Reproducing the recording would take emitting this pattern of 400 chemicals in a way that they would arrive at the intended audience in the specific desired concentrations.
There are smell generators. They’re just disposable cartridges. Used in perfumes and military training and household cleaners etc.
You can generate arbritrary images from pixels of primary light colors. Sounds from vibrating speaker cones. We don’t get have the technology to generate random atoms of scent from scratch (nuclear fusion or at least chemical reactions of arbitrary types) so we just have to store molecules of scent and release them when wanted. We can record and recreate smells easily only when we have some of it lying around.
TLDR light is a mixture of basic colors. Sound is vibration. Smell requires actual molecules of a substance, or some close proxy.