While the AI image market is now pretty crowded, Midjourney was one of the first to do it, turning words into images years ago. But since then, the company has fallen behind, being outpaced by a variety of competitors.
This was most evident in the development of AI video generation. The technology has blown up in recent months, seeing huge improvements from companies like Google and OpenAI. While Midjourney has been quiet on this front for a long time, it finally launched its first video generator.
However, while others push boundaries, competing to offer the most advanced package, Midjourney’s first attempt at a video generator was surprisingly limited. There is no ability to prompt for the video, only to add images — either your own or ones you’ve made with Midjourney. While audio isn’t common in video generators yet, it's missing from Midjourney's tool
However, despite these limitations, in my time using Midjourney’s video generator so far, it has quickly become one of my favorites.
What makes Midjourney a good video generator?
Midjourney has always stood out in the world of AI generation for one good reason. Where the likes of ChatGPT and Gemini are designed to create lifelike images and videos, Midjourney is hyper-focused on creativity.
Before you even use it, you need to rank hundreds of images, giving the model an idea of your style preferences. These preferences can be turned on or off, but with the personalization on, image and video results are clearly pushed to styles that fit me.
While you can’t directly prompt for a video, the process is isn’t much more complicated. After creating an image on Midjourney, you’ll have the option to “animate” the image. This can be done automatically, allowing the AI to choose what happens in the video, either with low or high motion. Or you can manually choose what happens.
This turns Midjourney into a similar version of an AI video generator as the big competitors… just with a few extra steps to get there. You can also upload your own images, turning those into videos.
Despite some big concerns around copyright right now, Midjourney has also put a strong emphasis on avoiding deepfakes. It won’t edit images of real people and is surprisingly unwilling to create something that might resemble celebrity figures.
The video generation from Midjourney is clearly designed for the same group as the image generator. It’s built for people wanting to make creative projects or design things that are clearly separate from real life. Scroll through Midjourney’s explore page and you’ll be greeted by moving comic book strips, anime fights and stylised car chases.
Midjourney also seems to have put a lot of work into its prompt understanding. Previously, the model would struggle to create good results without incredibly specific details. Now, it works in a similar way to the likes of ChatGPT, able to create images and videos from short descriptions.
From the short time I’ve used it so far, I’ve got high hopes for Midjourney’s video generator. They have warned that prices could change as they test the model so now is a great time to give it a go.
More from Tom's Guide
- OpenAI court filings just revealed new details on AI device with Jony Ive
- Google Gemini just got a great video feature ChatGPT can't match — here's how to use it
- I've been using this one ChatGPT prompt for years — and it works in absolutely any situation
You must confirm your public display name before commenting
Please logout and then login again, you will then be prompted to enter your display name.
