To tell a story is to recount something that happened; to more fully convey it to the reader, one "shows" it, or expresses a narrative using language that allows the reader to experience it as one does in the real world. The cinema and the television almost always "tell," in this sense, by showing you exactly what happened--by reporting the facts within the narrative. Sleight-of-hand can trick you into missing something, but if the cut of the show's jig is honest portrayal, then you never have cause to question what is shown. The impressionist element of the moment is lost in a scene so well-rendered that it transcends the ability of a now-human to sensorily experience.
It is in that impressionist element that we actually live, where Degasian flashes compose most of what we see. Not only via mistakes in our predatory focus, but in the widescreen blur of our periphery, and the assumptions built upon sound and scent. These arguments are currently rank with the shitspray of gender relativism, but misappropriation does not destroy the underlying truth that, as computers, we are not sensorily perfect. Film and literature MFA apparatchiks talk about the benefits of "showing" like Cologne feminists talk about the drawbacks of rape, but their vulgar selective application of the terms should not be permitted to lead others to spurn art or morality.
It is acceptable for first-person narration to convey the message of melancholy, or of one's impression of a setting, which is in that case not truly telling, but showing, by virtue of showing the truth of the experience as it is passed through the imperfect lens of the reporting character, e.g. the man approaching the abode of Roderick Usher. The drug addict's blurred first-person narrative is reality, in the sense that it is his perception of it. The third-person omniscient narrative is the "tell" narrative, designed for those who cannot read or think well, and, more importantly, who cannot apply the same tools to their real life experiences.
How does a character's appearance make the reader feel? Tell him as an omniscient observer, and you have answered the question. Did the gaunt man wear a gray suit the first time you met him? Perhaps. Yet if you describe him as melancholy, you have become the television, answering all the questions and doing all the work. Critical thought is no longer necessary, because the voice of omniscience has spoken; it is, therefore, not required that the reader (in the story or in real-life) be intelligent enough to identify the combination of "gaunt" and "gray" as melancholy, foreboding, etc. If it is later revealed that the gaunt man was a clown about to hop out of his jacket and surprise the viewer, you've ruined it, begged the question, set up the joke unfairly, and told the audience how it must feel. If description alone makes the audience draw their own conclusions, they are fairly tricked, justly delighted, and encouraged in their capacity for rationally discriminating between impressions in the real world.
"Telling" can be done through "showing," in the sense of blatantly and unquestioningly portraying a truth, so the phrase, "show, don't tell," is disingenuous unless you're intelligent enough to make out the difference. Cultivating the ability to tell for yourself, and to do so based upon what is shown rather than on what you have been told you should see, is the difference between being able to dynamically react to an environment filled with both intrinsic and alien, deliberate and accidental, deceptions, and to make you a more intelligent person, as well as a better viewer, reader, and listener. Telling someone, "He was a short man, broad of shoulder, with the serious look of a judge," invites them to abey laborious thinking and merely summon up their templates of pre-loaded narrative imagery to fill in the slots. "His shoulder brushed the bar counter as he strode toward his gavel," accomplishes the same essential task. It places greater demands upon the audience's intelligence, leaving it arguable between parties how "short" short is--what kind of bar counter? The lower part of his shoulder, or the upper part? Well, he's still short. And is it his gavel, or someone else's? And even more importantly, does "stride" really sound more determined and serious than "walk" or "saunter"?
It takes more mental labor--and correspondingly greater mental gains--to drink deeply from the more complex goblet, so since Brawndo has what plants crave, why bother? The protein is pre-digested so that you don't have to chew; pretty soon, your jaw is slack from gulping down a simpleton's EZ-Eat promise, and you're halfway to the Road, unable to tell who is good and who is bad without a compelling voiceover. That's why Americans read reviews before going to Amazon: because without help from The Atlantic, they risk not knowing what it was they experienced once it's been completed, even after Hillenbrand and Brown have done their damnedest to lower the narrational mean to "rides for free" levels. Permitting a narrative to stand on its own feet, by suggesting character and setting in the way people actually experience them, betters creator and audience through the joint exercise of figuring out what's going on and how it possibly could go on. When you teach to the test, you increase future STEM majors while decreasing future inventions.
(For the spiritually inclined, Yaldabaoth tells, while Christ shows. It is the teller's embarrassed terror of his own blindness which places him in fear of those who claim to see more than automaterial propagation.)