The Next Anime Industrial Revolution is Near

Animation is Going through a Massive Revolution Soon

Probably a lot of you are wondering what exactly am I talking about regarding the title of this article, back in the beginning when we saw the first animated films and TV shows these were all basically hand-drawn, and no computers were involved in the process of the animation. Fast forwarding to the late 1960’s we started to see early experiments with computer graphics and gradually over the next decades more software would come out that would make the whole animation process easier and faster than it did in the past.

AI Is Improving at a Very Fast Speed

AI Tools like Pika also have very impressive prompting capabilities on their latest version as you can see in the image above. Now people are still going to say “Hey! That looks uncanny or deformed” but if you did pay attention to the quality of Midjourney or early versions of Stable Diffusion, in the course of ~6 months or less massive leaps were made in technology and when it comes to video prompting we are expecting an almost similar level of leaps in a similar timeframe.
With Pika too you can pick existing footage and alter its style. So we can expect like old 60’s or even recent films totally turned into animated versions basically, this sounds very exciting component of this technology because there are a lot of old films that have outdated special effects, but after restyling them, they can look very fresh entries essentially.
Since things are built from scratch the post-editing is going to be harder unless the AI tools can simply adjust certain regions’ movements or remove objects from the background (Which we are already seeing with some image tools) so we can expect a bit different post-editing process, especially if the character designs are not staying consistent over all prompts.

Stylizing Existing Footage with Domo AI

Another very cool entry is Domo AI which allows you to stylize pre-existing footage into anime style or other presets like Pixel Art, Comic Book, Playstation 2, and many others.  I’m very impressed by this piece of technology because it allows you to turn any existing retro film into an anime or comic book movie. There were similar technologies introduced late last year which used a lot of the same techniques but required Google Collab and a Stable diffusion fork that was doing masking – these required a lot more work from the user and didn’t reach as same quality as Domo.

Currently, Domo AI is behind a paywall, but you are able to join their discord server where you can see different techniques and styles used by people, I think its only a matter of time until we see more industrial versions of this tool that alter heavily pre-existing footage.

Right now you don’t have much control over how heavy the style is being applied to pre-existing footage, but I clearly can tell that clips with slower movement are converting a lot better and there are still issues with things like hands, hats, and some other objects.

Tooncrafter is Animating Your Frames

Other tools like Tooncrafter allow you to input the first & last frame of an image and the tool will basically generate everything in between, which saves an insane amount of time for a lot of animators, granted it’s still a proof of concept prototype, but seeing how far they have come so far with this current version of the tool, i can see lot of upgrades in the future (Given that the team and universities are still going to work on this open source software).

The best part about Tooncrafter obviously being its Open source and anyone can use it for free, many people have done video reviews about the Demo and showcased its potential. It’s hard to say what type of AI technology will be levied most at these animation studios in the future, will it be prompting from scratch or will it be ones like tooncrafter that will add those extra frames to pre-existing footage? Most realistically I assume multitude of different tools will be used whether it’s upscalers or things that create automatic subtitles.

ArtCels are Blocking Progression

Just like when the first cars were introduced there were a large number of Horse owners and cart riders who were against this new piece of technology citing things like “Cars are very dangerous” and not to say they didn’t have valid points it was just much more convenient to have cars due to the horse having speed and endurance limits over long distance travel. Within the Art, Anime & Gaming industry there is already a blockade formed by Artcels who are highly critical of products that are using any form of AI – these opinions are mostly formed around the idea that AI is stealing existing art and that artists will be unemployed.

While it is true that AI is trained on a set of data, but the same people who cry about this factor do not care that basically that 95% of the current Web2 Internet is collecting user’s data with/without their consent and selling it onward to governments and interested companies. Data ownership isn’t a sexy position to have so that’s why you don’t hear it really outside Web3 space usually, but because artists are a celebrated class of people in these fields there is a grassroots/astroturfed set of people fighting for their positions in the workplace.

Ultimately when you look at the current production costs of modern movies, games, or anime – the prices are crazy high and there ought to be ways to cut down the costs on those insane production costs. Especially in gaming usually around half of the Triple-A games budget may go to “Graphics” which obviously entails a lot of things including UI, 3D Models, and all 2D Concept art that the game may use. CGI is another field that is also very expensive to play around with right now and AI will make things easier and affordable for Indie studios to access that cinematic flavor.

Artcels can only stop the progression for a while, but eventually, the sheer cost proposition is too good to pass by and it doesn’t always automatically mean that artists are out of a job, they are likely going to be trained to work a lot more with AI and probably might get paid less.

Next Generation Filmmakers will Change the Game

Anyone who has been on the internet for the last 10-15 years must have heard some drama and banter around modern Hollywood movies and their narratives, this has leaked over a great deal to the gaming space, and it is starting to tangle its roots in Anime industry too, which has stayed relatively isolated due to Japan’s geographical situation and obvious language barrier.  
Whether you think modern Hollywood is better or worse is irrelevant, the fact is that there is a growing number of people domestically and outside of America that have a less favorable view of the Hollywood industry and people running it compared to what it was 20-30 years ago. Now this is creating a perfect storm for change – where hit movies are going to be produced by indie studios that are completely going to do the films from scratch or massively alter their shot footage…
The days of needing 100M to do a blockbuster movie are over soon, and the new class of filmmakers will utilize AI to do crazy-style content that can be artistically sustainable (Because costs are relatively low, we don’t have to worry about cancellations as much). As people are already getting tired of Hollywood’s content there will be a large audience migrating into watching this AI-created content that will be visually close to par to the real thing, plus it won’t have the same political baggage and virtual signaling.

As I said earlier, I don’t think Japan will be jumping into this technology first, but I’m expecting a huge wave of editorial content coming from Southeast Asia and the West in general. A big puzzle piece will be distribution still, will there be truly decentralized movie or video platforms surfacing in this new era, that won’t be working on devious algorithms and censorship like the current wave of offerings? The idea of going into cinemas might become a massive niche in the future too, as ticket costs have gone up.

To be fair though… That current talent for filmmaking has to exist out there for this revolution to happen. I’m quite optimistic that there are bunch of people who are talented and creative, but just don’t really have the tools or the capital to produce these types of films right now without AI. Copyright issues will be issues too and perhaps the way these companies make money in the future will just revolve around selling licenses to companies and individuals to use certain names and characters, it’s too early to tell.