As a scientific animator, I am always on the lookout for new tools and technologies that can improve the quality and efficiency of my work. Recently, I had the opportunity to test the new version of Midjourney AI, which just got a version update to version 5. To say that I am impressed would be an understatement. The results of this AI tool are simply astonishing.
To test the new version, I asked it to regenerate some of the things I had asked for in the previous version with the same prompt. The difference in the results between the two versions was night and day. In the blog post, I will include several images that compare the previous work to the results of the current version. My favorite prompts, such as Nikola Tesla on a space station, in a restaurant, or doing various things in the laboratory, were particularly impressive.
One of the most significant benefits of generative AI for my studio is that it saves a lot of time while also improving the artistic qualities of the final results. For instance, every time I start a new animation project, I ask Midjourney for direction and inspiration for lighting and composition. With the new version, the results are even more inspiring, and we can get clearer visions of how to best visually represent the concepts that I'm trying to turn into an animation video explainer.
Moreover, I use AI-generated images for blog posts and press releases, where it's very important that the visual content will be inclusive to all members of the scientific community. Fortunately, the new version has improved its representation of diverse ethnicities. This comes to show that, at least to some extent, the people behind these A.I's are listening to the complaints of the users and acting to address moral issues.
The most interesting perspective on this situation is actually comes from comparing the results from version 5 to version 3. We are comparing results that are less than a year apart but have an immense difference in quality. The result I got from version 3 in the attached example was somewhat helpful with lighting and materials inspiration, but the result I got for the same prompt from version 5 can almost be used as a cover for a scientific journal, as it is stunning and looks like expert artwork.
It is remarkable how fast AI has improved, and I am excited about what I will be able to do for my clients with the AI tools of the future. The capabilities of text-to-image AI for scientific animation are astonishing, and I am looking forward to incorporating this technology into my workflow.
Comments