Hi class, today I chose to analyze POOF- AI Short Film.

I enjoyed watching this film but couldn’t shake the suspicion that it was 100% AI-generated. Turns out I was right. I watched the behind-the-scenes video and it turns out the creator used a substantial amount of traditional digital effects to stitch together the AI footage. Fun watch, linked here:

Anyway, the film is about these monsters that work in an office space but they’re all bored out of their minds and end up spontaneously combusting. It’s clear that these clips are AI-generated. you can spot it from a mile away. The animation of the character is odd and the background tends to morph. Having said that, it’s still pretty dam good for AI.

In the case of POOF, the creator mimics a real-life aesthetic and interestingly tries to recreate the special effect of traditional puppetry with the special effect of AI. In this way, the creator is replacing old special effect techniques rather than altering or enhancing them.

It’s hard to say what AI is going to do for the world of cinema. Will it simply replace existing special effects techniques, or create new ones? How will it interact with the world of traditional cinema? Will it add/aid it, or replace it entirely? In my opinion, AI is just another tool, like all the other procedural forms of digital enhancement. Of course, the underlying tech is completely different, however, my point is only that it will afford us the same opportunities to shift the production of cinema to a more digital and individual base environment – one that inherently prioritizes quantity or quality. (If you watch the breakdown video, which I highly recommend you do, you’ll notice he repeats something along the lines of “This isn’t perfect but it works for my purposes”… quantity over quality).

If we track the rise in digital technology in the cinematic field (digital cameras, digital video processing/editing, 3D digital rendering) it’s clear that with it comes an increased rate of content production and at a decreased cost. The creator of POOF was able to cheaply and by himself make something that 10 years ago would have cost at least tens of thousands and required a substantial crew with specialized skills. We also see a shift in the content itself to a more sustainable form of media that can handle a high rate of production (short-form internet content). Will AI take over traditional film production? Maybe for a short time but not in the long run. What’s more likely is that it will harbor a style of media production that is entirely different from the fundamentals of traditional cinema. Long-form narrative content will slowly die out and with it, cinema will go too. Makes me sad to say.

The first scene in this sequence takes place in the hallway that connects the rooms of the siblings (I’m assuming their siblings). Shooting this scene in such a tight narrow space can be difficult bust also afford unique opportunities. Much like a road, a hallow is very linear, people travel up and down it. For the most part, we can get a sense of the space by following the main character with the camera as she goes back and forth to the different rooms. Here the director is using the movement of the subject to guide the camera and give the audience a sense of the space.

Then we have a nice transition to the breakfast scene. We cut from the action of the previous scene to an empty frame of the fridge door before a hand comes into the frame. Here the director tells us that we are already in the kitchen before cutting to the wide shot. Then we get a series of intercutting between wide shots and chaotic closeups. However, this works because the director has already done his job and made us familiar with the space and the placement of each character in that space. The audience is already well acquainted with the space, so the director can afford to jump back and forth across the 180-degree line without confusing the audience.

I think this sequence is a good example of how showing the audience a space early in a scene can afford you more freedom later on.

-Luca Wagner

The reason I chose to talk about this scene over the others is only because of the very first shot. Spielberg starts this scene with a fantastic single shot that is, in my opinion, the most important shot of the scene. In one single shot, Speilberg sets up the space of the scene and lets us know who the bad guy is. It begins on one side of the 180-degree line, behind our protagonist. The camera moves down the line to show us the truck is in front of the red car, and then it settles on the opposite side of the 180-degree line at a low angle to the truck, making it look large and menacing. Spielberg also uses sound to his advantage. In one shot we went from a small quiet red car to a big dirty loud truck, a perfectly executed juxtaposition.

Then we cut back to the small car, its quiet… tension is starting to rise. We get a POV shot as the David passes the truck. (a clever POV shot because we see through his eyes and at them at the same time). As David passes he shoots a few glanes at the truck, giving Speilberg and excuse to put some motivated POV shot of the truck in their and show the passing in its entirety.

Then we get a reverse shot of David and a key moment of tension where the truck barely peaks into the frame as if it’s going to pass again (something the audience is probably already anticipating). Then we cut to a reverse of the previous shot – an over-the-shoulder of David and the truck is a little louder but we can’t see it. Then Speilberg puts in a couple of close-ups to break up the tension a little (calm before the storm) and resests to another angle. He shifts to a side shot of David where the window is conveniently framing a big empty space… and of course, the truck passes through it. What’s interesting is that the audience surely is expecting it but it’s still very effective. As the truck passes we get another over the shoulder of David to see it better. Then it’s back to the passenger POV from earlier (with his eyes reflected in the rearview). This shot is used as a motif to show David stuck behind the truck again. The fact that almost the entire rest of the scene is reusing these angles is a testament to their effectiveness.

Chungking Express, directed by Wong Kar-wai, is a fantastic movie with a fascinating visual style. Upon first watch, Kar-wai’s fast-paced, handheld cinematography may seem chaotic. But if we take a closer look we’ll see that there is a method to Kar-wai’s madness and that the chaos of certain sequences is clearly controlled.
Before reading any further, I highly suggest watching this scene several times. (And if you have the time, maybe even the whole movie). https://www.youtube.com/watch?v=2FrNkCk9niU&t=2s

We begin with a jarring shot that introduces our main character and sense of urgency. The camera (handheld) glides quickly past several faces before finding the woman in the glasses. How do we know this is the main character? Well one, she is the only woman in the scene so far and is conspicuously dressed. But more on the topic of framing the camera actually begins to follow her instead of just gliding past like it does the extras.

Next, we get a series of shots that indicate her POV as she navigates the crowd. Framing-wise, Kar-wai keeps the camera handheld and at eye level to tell us we are looking through someone else’s eyes. We also get characters looking directly into the camera and making eye contact with the audience, which helps cue the audience to understand that these are POV shots. These shots are also all medium shots, shot with focal lengths similar to the human eye (maybe a little longer). These POV shots are interjected with cutaways to our main character. Kar-wai chooses to keep the camera handheld and at eye level when tracking our main character. Why does he do this? I think, for a few reasons. He’s trying to create a claustrophobic and fast-paced scene and doesn’t want to break up the rhythm. More importantly, he wants to establish her character as nervous and always looking over her shoulder(which she literally does a few times). What better to put the audience on edge than to make them feel like someone is following our main character? By putting the camera at eye level Kar-wai cleverly achieves that effect.

Our character seems to get where she’s going the camera stops tracking her and is cut off but a curtain, ending the scene. Then we get a break from the action in the form of a title screen and some establishing shots. These simply get us acquainted with our setting. These establishing long shots also allow Kar-wai to transition into another sequence that is very similar to the last without confusing the audience. I would have liked to cover the second part of this intro but for the sake of the length of this blog post, I’ll end my ramblings here.

Thanks for reading!

-Luca Wagner

Medium eye-level shot following her

Medium shot of her getting in the elevator. Helps pass time and takes us to a new area.

Part of a series of Medium and Medium closeup shots of passersby making eye contact with us/camera/main character. Well call this shot A

Shot B

Shot C

Shot D

Shot E

Back to the tracking shots of our main character (Medium). Still at eye level and still handheld. This keeps the rhythm and tension of the scene going.

A Medium Closeup from behind.

She turn around… are we following her, or is someone else?

Medium of her walking away.

Medium of her entering a room. This time the camera stops tracking and that is how we know she’s reached her destination.

The curtain draws, the scene ends.

Titles.

Long Shots of the city establish the setting further and gives us a break from the action.

Testing 123