Architectural Prompting: Midjourney now creates videos, and works surprisingly well

|

Emma Potter

Dear readers, well back to our usual appointment! In the latest articles we have often had the opportunity to talk about new features, because Midjourney continues to update with an impressive rhythm. This week there is also a novelty, and perhaps it is the more important until now: Midjourney has finally introduced the possibility of generating videos.

Starting from a generated image, we can now transform our scenes into small animationsopening new creative and narrative possibilities. I have not lost time to try it, and I confess that – to be a version 1, and to have just come out for a week – It works surprisingly well.

>> If you want to receive news like this directly on your smartphone subscribe to our new Telegram channel!

What is Midjourney Video

Midjourney Video (V1) allows you to animate an image generated or loaded, creating one Video sequence of about 5 seconds. The goal? Add Movement, atmosphere and visual storytelling to our work.
There are two ways to choose from:

  • Auto Mode: does everything alone. Midjourney imagines a small movement and generates the video.
  • Manual Mode: Here you can write what happens, inserting a personalized motion prompt.

In both cases, you can decide if the scene will be Low motion (firm room, small movements) or High motion (intense and dynamic movements, but more subject to “oddities”).

How the flow works

  1. Genres an imageas you have always done.
  2. Click on “Animated Image” And choose whether to use car or manual.
  3. After a few seconds, Midjourney shows you 4 Videos of 5 seconds.
  4. If you like the result, you can use the option Extend To lengthen the video of another 4 seconds at a time, up to a maximum of 20 seconds.
  5. You can too save The video optimized for social or editing.

All this works both with images generated on Midjourney, and with images loaded from outside.

Quality, costs and durability

  • The videos are in resolution 480p (not too highbut the perceived quality is good).
  • Each “Job” video generates 4 clips e It costs about 8 times an imagewe can say that it is not too high by calculating the work of the machine and the number of frames to be generated to allow an adequate fluidity!
  • Currently the system uses the fast hours of your subscription (but are planning the usual “relaxation” mode also for videos).

A first example

As always, I wanted to test everything with an ad hoc built image. What you see below is the result of the first generation: an atmosphere night scene, inspired by vintage surveillance cameras. The quality of the image is already remarkable – but moreover we know well that Midjourney is now very solid in the generation of static images.

Architectural Prompting: Midjourney now creates videos, and it works surprisingly well 00 digitalalchemysitudioimg tutorial

But it is With the video generation that happens the real magic. As you can see from the clip below, the fluidity of the chamberthe way in which The scene comes alive And the subject comes to life – everything is surprisingly coherent. The movement is credible, natural, and succeeds in transform the image into a small narrative fragment.

Second example: an animated interior

The second example I want to share is closer to our design world: it is a interior seen from abovegenerated, as usual, totally inside Midjourney:

Architectural Prompting: Midjourney now creates videos, and it works surprisingly well 01 Digitalalchemysitudioimg tutorial

Again, the result is surprising: The Movement of the Chamberwhich descends with a fluid pan towards the sofa, is natural and cinematographic. But what really strikes is the Consistency of the result: the geometriesthe lightthe shadowthe arrangement of the furnishings – everything is reconstructed in the video in line with the original photo.

Because it is interesting

Now imagine how powerful such a tool can be in our work: not only can we show hyper -realistic images of our spaces and architecture, but we can also sap them, explore themadd a narrative and spatial depth which was previously accessible only with long modeling and rendering phases.
A tool that really opens a new world of possibilities. The function is perfect for:

  • create more engaging social content;
  • develop animated storyboard;
  • testing on moving scenes for environments, characters or product;
  • Imagine architecture or objects in action, with a cinematographic touch.

Limits (for now)

Is one version 1therefore it still has some limit:

  • low resolution;
  • No audio (again);
  • You can “prompt” only the initial frame, not the final one;
  • The complex movements (hands, bodies that jump) are sometimes not perfect.

But so the potential is already enormous. And Midjourney announced the arrival of:

  • Video in high resolution;
  • audio synchronized;
  • Upscaler Video;
  • prompt initial + final.

In summary, Midjourney Video is a New powerful and accessible toolwhich allows us to bring life, rhythm and narration to our images. It is not just an extra effect: it is a new way of imagine, design and communicate visually.

See you soon, with new tests! In the coming weeks I will continue to experiment with this and new functions, trying to understand until we can push ourselves. As always, I will share examples, tests and new ways to integrate Midjourney into our daily work.

The weekly section “Architectural Prompting” is curated by the experts Luciana Mastrolia, Giovanna Panucci and Andrea Tinazzo
>> If you are interested in these issues, also subscribe to the Free LinkedIn Newsletter Ai & Design for Technicianslet’s talk about it here!