It might be but I wouldn’t describe that as narration or a scene
Well, it is but I do understand where you are coming from. Compared to my typical example of play, your scene setting or frame setting is very sparse. You are expecting that the player will fill in any necessary details that he doesn't have with questions, and thus you are speeding play because you aren't giving the players details they don't want.
In practice though, your artificial example feels really artificial. It works only presumably because the PC's already know quite a bit about the setting and there is almost nothing about the setting that they need to know. I suspect something like the following would happen in my game:
ME: You get inside the city of Tung-On and the street you are on is filled with merchants and shop stalls
PLAYER A: Can we see the Lucky Mountain Gambling Hall nearby?
ME: Yes, it is just down the street, across from a bunch of inns.
PLAYER B: I lead the way to the gambling hall
ME: You get to the door there are men in blue robes at the entrance.
PLAYER B: I step inside and look around for this Iron God Meng guy.
ME: The door is in fact closed. When you reach to pull it open, one of the men in blue robes grabs you and says, "Where in the nine hells are you going? This is a private club."
PLAYER A: Wait? He grabs me? I'm going to use my Tiger Kung Fu to resist as soon as he tries to touch me.
ME: OK, roll for initiative.
PLAYER B: This is great. I thought the plan was to call Iron God Meng out, not end up in a fight with the entire Turtle Clan.
PLAYER A: You know, I assumed since you didn't say otherwise that the door was open. I wouldn't have even tried to go inside if I'd known the door was shut.
ME: Err... yeah, it's a set of large stout doors, painted red like most of the building, with decorative iron plates covering most of the door.
PLAYER A: Ok, so I jingle my coin string and say to the guards, "I've heard this was a fine tea house. I'd like to obtain a membership."
ME: Ok, make a diplomacy check.
PLAYER A: *fails* Crap.
ME: The biggest and ugliest of the guards says. "This is private club. Get lost or you will regret it." You can tell he is spoiling for a fight and doesn't like you very much.
PLAYER B: Crap, nice roll. Looks like we made them hostile. Ok I take a look over the building to see if there is another way in.
ME: The building is highly ornate with florid wood carvings, repeating the theme of a Garuda. The metal plates on the doors feature a Gurada tossing coins in the air, and the corners of the roof...
PLAYER B: Wait.. the sign of the gambling Garuda is on the doors?
ME: Yes.
PLAYER B: Why didn't you say so in the first place? I go up to the guards and say... Wait, damn it. I want a retcon. If you'd described to door in the first place, I would have never let Player A make the diplomacy check. There is no way I would have missed the sign of the Gambling Garuda if it was right there on the door.
ME: *Sigh*, ok...
Ect. Etc. Etc.
Now maybe you could train the players to ask questions about everything, including things that might not be obvious and requires leaps of intuition, before they make an actual proposition, but that would tend to slow play as well. I find that there is a fine line between being too wordy and not wordy enough. And in particular, while I've gotten a few hints about what your setting might be like, there is so much detail missing that I can't be certain that I'm imagining what you see in your head. I can remember numerous cases where I discovered in the middle of play that one or more players at the table had been taking actions based on a different understanding of the scene than I thought I had framed. For example, I can remember describing a room that was bisected by a chasm, and some players envisioned this the same way I did, with a chasm cutting the room in two in such a way that it blocked progress toward the end of the room. However, one player had envisioned the room in a way that the chasm separated the room in to two parts 'left' and 'right', so that he could advance all the way down the left side, but not reach the right side of the room. Several rounds of actions transpired before I figured out that his plan was based on seeing something in his head entirely different than the rest of us.
That issue is big enough that regardless of how you handle it, having a process for getting everyone on the same page so that they all see the same thing (part of how I define "cinematic") is something I consider really important to tabletop play. That's why I lean toward "both".