Super long response; not strictly for the OP, but for the broader subject of sensory exposition in general.
Lots of good ideas here, but we have to be careful of strong one to one adjectives, similes and metaphors that can make everything sound like a noir detective thriller, especially as a repetitive style. For me, a taxi doesn't drive by, it splashes water from a puddle; a bell doesn't ring, it echoes down the narrow street. So, thing + action + emotion whenever possible.
I try to immerse the reader within the current emotional experience of the scene, or narrative as a whole, depending on the situation. In the cavernous room example above, why is the size of the room important? Why should the reader care? Are we inferring something about the owner, such as wealth, or are we drawing attention to the character experience? Do they feel small by comparison, lonely, afraid, or do they admire the construction, are they envious? The size, smells, sound/echoes, even the feel of the wood or uneven stone floor beneath their feet should reinforce where they are in their character arc, the same goes for establishing the experience as a metaphor.
How characters react to similar experiences throughout the story can be a good indication of where they are relative to their overall arcs. For example, how they perceive an experience in the beginning of the story may be completely different from the end, such as dressing a deer for the first time as opposed to once they've become a seasoned hunter. And then you can take that down another layer and use the deer as a metaphor for other arc experiences. In a thriller, the metaphor might reflect killing a person.
This also extends to how they perceive the world. In the beginning of a post-apoc story, a character may notice the buildings and streets of the city, survivors and how frightened everyone is, whereas by the end, he may only see concealment, enemies and targets - threats. Perspective, and the changing of that perspective relative to sensory exposition, is a tremendous filter for communicating character arcs, mood and even tension.
Descriptions, including the use of the senses, should always reinforce character or mood or tension or whatever the story needs. But, if we take it one step further and reference the emotion behind the senses, as opposed to a standalone call-out, we can hit the reader deeper, because it allows them to fill in the blanks with their own experience, which engenders more engaging empathy. This strategy allows us to fill in the scene over time, reinforcing mood, rather than relying on an info-dump or single line reference, which then allows us to weave the other senses in as we go. For example, when we enter a space, we don't take it all in at once, we discover it slowly as we experience it. In a horror scenario, we may get a sense of dread as we enter, and then that mood is reinforced as we move through the scene, the apples don’t smell of cinnamon, but rot; the once-polished floors now blackened and molding between the boards.
On the flip side, in a chase or confrontation scene, or one of extreme immediacy, I'd dispense with all of that and just be super blunt. For example, by using more immersive descriptions leading up to a serious action beat, then the sudden absence of adjectives and starkness of the prose will make the action more emotive because it will stand out and grab the reader’s attention. This is an example of rhythm, where the action beat is the punch line. In this case, the action is hearing or seeing or smelling or feeling (experiencing) the 'thing'.
We can apply the same reasoning to the stepping/squishing example above. What are we trying to convey? Which sense are we trying to stimulate? For a lighter beat, we might go with the audible sound, because squishing is usually funny. But for a more serious beat (using horror again) we might just have the character feel the thing give beneath the weight of their foot, if barefooted, then perhaps they feel the gush of warm liquid...like stepping on a disembodied organ.
If the 'sound' (event) is heard rather than felt, then I think it should be set up through the pacing of the scene as noted above. If we’re going for tension, then maybe lead up to the sound by describing the tomblike silence, hearing/feeling their own heartbeat…the sound of their breathing…they swear they can hear it echoing in the space…every footfall sounding like a performance of Stomp. Noting these observations in a succession of ever decreasing sentence lengths will naturally build tension. And then, end the sequence with a one word fragment that reflects the sound’s real-life origins: snap (stick or branch)...crunch (broken glass underfoot or gravel)...snick (a sidearm hammer being pulled back).
Another option is to just say whatever it is plainly: He stopped. Someone just c*cked a pistol.
Another option is the callback metaphor. Let's say the character suffered a trauma at a vacation house as a kid, and the strongest memory is the crunchy sound of hard-sole shoes on sandy floors. When he hears it, he thinks of the location of the beach house where the trauma occurred - Plum Cove. Once that is set up (way back in the story), now, when the character hears that sound, the prose defaults to the shorthand fragment: Plum Cove; and the reader, instantly, not only knows the actual sound, but also that it indicates something bad - something dangerous, and finally, that that bad thing is close. When appropriate, this setup can draw the reader in much deeper than simply noting a sound - the old show vs tell thing.
We can also bypass the actual sound beat, and put the emphasis on the character's reaction, like Starling does in Gumm's basement.
How one handles these situations totally depends upon the desired emotional result of each instance, and then, of course, the overall style of the prose. So, the good news is there's no right way to do it, sadly, the bad news is that there's no right way to do it.