The Wasteland - A test of the first working version of the custom pipeline


A couple of weeks past reaching Save Point pipeline 1.0, and I thought I'd post a collection of clips from the first set of final output templates being developed. I talked a lot about snowballing speeds once we completed the pipeline. This is the very beginning of that snowball. As the number and refinement of shot templates grow, it will become possible to do things like assemble a half hour animated show in a few days for 1% of studio cost.

I suspect some will question whether this visual style is an improvement on just using the single engine method. There's a few major reasons why an illustration look is stronger for this project.

1. The Uncanny Valley

People have trouble connecting emotionally to 3d characters. It's much more difficult to get people invested in a storyline if they don't care about the characters, so, it's kind of a big deal.

2. Speed

This stage is where the immense speed advantage of the Save Point system actually takes effect. 15 hours to create and render a cell becomes 15 minutes. This is where it becomes possible to do interactive fiction as film in a streaming economy.

3. Flexibility

Unreal Engine looks great, but ultimately, it looks one way. The twin engine route means that visuals can be in any style we need, without rebuilds, including just UE5 style. The big deal though is that the visuals can be constantly improved over years, and it's built so improvement is "built in" and requires very minimal rebuilding of the pipeline. Layman's terms - we can build this into photorealism in a few years and we don't have to keep rebuilding the engine.

Tech research never stops, and we're always looking for ways to improve, but this is essentially the point in time where Save Point stops working on tech, and starts producing films for public consumption. Tons of work left, but this month was the turning point. Just a few months to get the hang of using the new system, and I can probably launch the smaller "Labyrinth" project (the cat project), and fire up production on Save point episodes.
 
Upvote 3
1. I would like all tests to be shown in real-time motion, as slow motion can mask realism.
2. The cat still looks off, like he is waking with stiff wooden legs on his tippy toes.
3. Looks good though, these are just observations. Hope this helps?
 
looking really cool, are you still thinking about doing a Flood short film ?
Yeah, I liked that film idea, and had it developed a bit. I think it's still a good candidate for a non interactive side project.

I probably will make it at some point soon, right now though I need to just finish Labyrinth and get some income rolling in. As the snowball grows, I'll reach the speed to do both sometime this summer. This is kind of what the whole speed throughline is about. Do you want to walk to California today? No. Ok, here's a bicycle, you want to ride it to California? No. Ok here's a gulfstream, you can go to California and back in 6 hours, can you do that? Yes. Ok, let's start going to California for lunch on Fridays. Different speed, different mentality.

I want to make all kinds of films, it's just too slow and expensive. So basically instead of working on films, I've been working on making them less slow and expensive.

Flood is planned as a feature, because it would take a lot of work, and short films don't ever earn money. Since most people have never heard me talk about it, here's the only concept development reel that was made for it so far. I won't talk about Flood a lot here, but basically it's like a combination of Fallout and Bambi, with a lot of water SFX in a city full of mysterious perils. So "Watership Town", lol. The major reason it was shelved for now was because I was too close to the engine switch, and I don't really think a purely UE5 version would sell.

Warning to people who haven't seen this, it's not a film, it's just a video where I collected a bunch of loose concepts as I was designing this idea. It does have some interesting moments here and there.


I've thought about this project here and there, and expanded it a lot from the beginning when this primitive demo reel was made, but it just got temporarily shelved during the pipeline changeover, which has been a couple hundred of hours of work already just by itself. This first 1.0 demo film in the OP took maybe 40 hours, but that's just growing pains. If I did the same thing again this month, that 40 hours could be 5 hours, and soon enough I'll be able to put out a reel like that in an hour. When I got that fast with the UE5 pipe, I basically just quit posting videos, because it was starting to overwhelm the feed.
 
Last edited:
I still think you could build a little 8-minute interactive short film out of it, and then market it to netflix to develop the remaining episodes and release a series


That still feels viable to me

 
Last edited:
1. I would like all tests to be shown in real-time motion, as slow motion can mask realism.
2. The cat still looks off, like he is waking with stiff wooden legs on his tippy toes.
3. Looks good though, these are just observations. Hope this helps?

I do tend to overuse slow motion, for various reasons. There are many shots in there that are not slow motion though. The longer it takes me to make a video, the greater tendancy to economize in various ways, such as milking some shot. As the speed increases, this will give way to faster action and story replacing cinematography as the primary pace motivator. If it takes me 5 hours to get a tree blowing in the wind, I'm going to want to show that off for 10 seconds, once it's fully automated, focus will shift to the characters, and you'll see something much more like a normal product.

I've already solved the animation issues, but the new method wasn't ready to deploy in this video. There's probably another 1-200 hours of prep work needed to get the cat to be realistic in every scene. I did show a proof of that aspect working a few weeks ago though.

In case you missed it.


Before long most of the animation will move like this. There are some places where I have to use the artificial cat, because for some reason, I can't get my actual cat to walk in a straight line for the camera.
 
I still think you could build a little 8-minute interactive short film out of it, and then market it to netflix to develop the remaining episodes and release a series


That still feels viable to me

I've thought a bit about doing some small kids version to get the ball rolling. I'm not super enthusiastic though, for a variety of reasons. I'll think about it more. If Netflix would buy it instantly, that would be a major deciding factor. The big deal is that my main goal was to change the way people perceive interactive film. People have been making the cheapest, quickest thing that would make money for a long time. It's kind of had the "VR effect", meaning that the people developing the new thing were so underfunded that when VR launched 7 years ago, people started making these 1 hour games, kids stuff, fruit ninja, mobile phone grade stuff, because it was cheap and easy. Unfortunately this became the default for the industry for half a decade, and at the critical point where a lot of people were deciding whether or not they cared about VR, it was just this endless parade of low effort cash grabs. I need funding, no question, but I want to introduce this concept of full length episodes, because that's what's actually new about what I'm doing. Of course, you can always have more than one brand under more than one company name.........
 
Back
Top