Save Point - an interactive animated series

Save Point is a unique project that allows independent creatives to put their skills to work on a crowd-sourced project. Save Point is the name of both the organization, and the online series it produces. It's an animated series, which allows people all over the world to collaborate, producing one flagship product with a better chance of success in the market than would be possible with many tiny projects.

Here's how it works. You can learn about Save Point in one of three places.

1. The web site at https://www.savepointofficial.com
2. Here on Indietalk, where there are many posts discussing aspects of the project
3. By joining us on the Discord server, and asking us questions
you can join the Discord server using this link - https://discord.gg/bYY2GvzdJ7

Comp 5 (0-00-00-00)_1.jpg


In this blog, we will publish regular updates that show many different aspects of the development process. Here you can look at art, videos, scripts, technology, pipelines, and other aspects of the project. You can follow along as we build it, or if you're interested in collaborating, you can join the project and take on a more active role.
 
In this first section, we'll archive shot images from the work in progress. Our base resolution is 10k anamorphic, with a final output of 4k, so unfortunately we cannot post full resolution images here, but what you see here will be a pretty good representation of what it will look like for most YouTube viewers.

Comp 2 (0-00-00-00).jpg

Comp 1 (0;00;00;00)_54.jpg


Comp 1 (0;00;00;00)_52.jpg

Comp 1 (0;00;00;00)_47.jpg

Comp 1 (0;00;00;00)_45.jpg

Comp 1 (0;00;00;00)_46.jpg

Comp 1 (0;00;00;00)_42.jpg

Comp 1 (0;00;00;00)_39.jpg


Comp 1 (0;00;00;00)_37.jpg


Comp 1 (0;00;00;00)_34.jpg


Comp 1 (0;00;00;00)_31.jpg

Comp 1 (0;00;00;00)_30.jpg


Comp 1 (0;00;00;00)_29.jpg

Comp 1 (0;00;00;00)_28.jpg


35753.jpg


Comp 1 (0;00;00;00)_79.jpg

Comp 1 (0;00;00;00)_78.jpg

Comp 1 (0;00;00;00)_77.jpg

Comp 1 (0;00;00;00)_75.jpg


Comp 1 (0;00;00;00)_74.jpg

Comp 1 (0;00;00;00)_72.jpg

Comp 1 (0;00;00;00)_73.jpg

Comp 1 (0;00;00;00)_70.jpg


Comp 1 (0;00;00;00)_69.jpg


Comp 1 (0;00;00;00)_68.jpg


Comp 1 (0;00;00;00)_67.jpg


Comp 1 (0;00;00;00)_66.jpg


Comp 1 (0;00;00;00)_65.jpg

Comp 1 (0;00;00;00)_64.jpg


Comp 1 (0;00;00;00)_63.jpg


Comp 1 (0;00;00;00)_59.jpg


Comp 1 (0;00;00;00)_58.jpg
 

Attachments

  • Comp 1 (0;00;00;00)_51.jpg
    Comp 1 (0;00;00;00)_51.jpg
    49.8 KB · Views: 300
  • Comp 1 (0;00;00;00)_27.jpg
    Comp 1 (0;00;00;00)_27.jpg
    67.8 KB · Views: 303
Here is where we will display GIFs. These are tiny snapshots of the animation development process. The quality is very low on these and does not represent the final product, which is natively 4k 30 at this time.










 
Since the last time we published in this blog, there have been significant advances, including the switch to an entirely new production pipeline. This post is simply a collection of demos, dalies, etc, from the past month. By the end of the year, we will have tripled our output power, specifically when 2 major software platforms and 1 major hardware rendering platform arrive throughout the next month. It will take a while to put all this technology to work, and it's a difficult learning process. You can see a bit of it here though, including the long awaited first demonstration of our digital characters. Much more is forthcoming in the next 4 months, at which point we hope to begin production on the actual product. I expect many on the forum have seen the last few videos before, but I'm adding them here for people encountering this blog as the first point of contact with Save Point.

Here is the first look at the new gen animated characters. Animations are looped in the demo to save time.


Here is the current work in progress demo, which is about 60% complete. We'll publish the final in the videos section once it's complete.


Here is the biome development demo, this progress was achieved in a single week, week 29 of the project.


These last 2 videos were the first tests of the new engine, one for interior scenes, and one for exterior scenes.


 
Here's the dalies reel from this last week or so. This one shows some progress in a few key areas.

1. the first space scene, required construction of starfield, nebulas, planets, and flightpaths.

2. Fog and light shafts seen in jungle scenes, required development of volumetric fog and lighting workflows

3. Animated humans, required a lot

4. Most importantly, crowd movement, seen in the cyber city street scenes. required the most effort and time.

What's seen here is a first try implementation of these aspects, each one needs a lot of work and refinement. Cyber city scenes were rendered in low quality to save 10 hours of render time.

 
Last edited:
It's been a while since I updated this blog section. I mainly use it for screenshots now, since you have to scroll for 10 minutes to get to new posts, so I doubt anyone actually sees anything I put here. I would delete the top posts, but I want to keep them as a record of how this developed over time.

Maze stripe 3.mp4_snapshot_02.22_[2022.05.20_17.32.29].jpg



Maze stripe 3.mp4_snapshot_02.33_[2022.05.20_17.32.52].jpg



Maze stripe 3.mp4_snapshot_02.37_[2022.05.20_17.33.01].jpg


Maze stripe 3.mp4_snapshot_02.56_[2022.05.20_17.33.20].jpg


Maze stripe 3.mp4_snapshot_03.03_[2022.05.20_17.33.26].jpg


Maze stripe 3.mp4_snapshot_00.04_[2022.05.20_17.30.38].jpg


Maze stripe 3.mp4_snapshot_00.48_[2022.05.20_17.31.33].jpg


Maze stripe 3.mp4_snapshot_00.43_[2022.05.20_17.31.24].jpg


Jungle 7 and 8 stripe.mov_snapshot_01.08_[2022.05.20_17.12.08].jpg


Jungle 7 and 8 stripe.mov_snapshot_01.12_[2022.05.20_17.12.12].jpg


Jungle 7 and 8 stripe.mov_snapshot_00.13_[2022.05.20_17.11.48].jpg


Jungle water drone stripe 1.mp4_snapshot_01.09_[2022.05.20_17.11.00].jpg


Jungle water drone stripe 1.mp4_snapshot_00.13_[2022.05.20_17.08.44].jpg


Jungle water drone stripe 1.mp4_snapshot_01.16_[2022.05.20_17.11.17].jpg


Engine fog stripe.mov_snapshot_00.15_[2022.05.20_17.06.31].jpg


Engine fog stripe.mov_snapshot_00.22_[2022.05.20_17.06.54].jpg


Maerd - Maze 1.mp4_snapshot_00.26_[2022.05.20_17.05.53].jpg


Maerd - Maze 1.mp4_snapshot_00.21_[2022.05.20_17.05.48].jpg


Maerd - Maze 2 blue door.mp4_snapshot_00.30_[2022.05.20_17.05.11].jpg


Maerd - Maze 2 blue door.mp4_snapshot_00.10_[2022.05.20_17.05.01].jpg


go through the red portal.mp4_snapshot_00.28_[2022.05.20_17.03.52].jpg


Maerd - Spaceport stripe 2.mov_snapshot_00.05_[2022.05.20_17.02.54].jpg

spaceportr test clip 1.mov_snapshot_04.21_[2022.05.20_17.01.47].jpg

spaceportr test clip 1.mov_snapshot_01.58_[2022.05.20_17.00.56].jpg


Maze factory stripe 1.mov_snapshot_04.30_[2022.05.20_16.58.06].jpg


Maze factory stripe 1.mov_snapshot_05.03_[2022.05.20_16.58.21].jpg


spaceportr test clip 1.mov_snapshot_00.14_[2022.05.20_16.59.54].jpg


Flythrough.0032.jpeg
 

Attachments

  • Engine fog stripe.mov_snapshot_00.05_[2022.05.20_17.06.12].jpg
    Engine fog stripe.mov_snapshot_00.05_[2022.05.20_17.06.12].jpg
    52.4 KB · Views: 202
  • Jungle 7 and 8 stripe.mov_snapshot_00.17_[2022.05.20_17.11.51].jpg
    Jungle 7 and 8 stripe.mov_snapshot_00.17_[2022.05.20_17.11.51].jpg
    97.9 KB · Views: 220
That series of pictures is powerful. I could feel an emotion starting to rise. Maybe it's because I'm a cat person but I was truly drawn in by the cat, which appears to be on a journey of some sort. Bravo, Nate!
 
Thanks! I appreciate the feedback.

These frames are all from our final development demo, as we approach a point where we can begin construction and deployment of the real project. It's the story of a cat that finds itself trapped inside an evolving, surrealistic maze of worlds.

This is an interactive fiction film, so after a year of developing a visual pipeline that works, we are now getting some experience directing the choices and transitions.

Since you are a post guy, you might understand what I'm doing here a bit better than a layperson. I'm setting up for mid grade, high speed rendering, allowing a very high pace, maybe 8,000 4k frames of finished animation per day. Then we use a set of pre matched lighting and camera positions and movements, allowing modular combination of prerendered assets. For example, characters talking would always be shot from one of about 8 angles, and there is a pre rendered library of backgrounds matching those shots. As the library grows, it becomes a drag and drop construction workflow, with the advantages of ray traced rendering at 4k, but it's all real time inside resolve with a 3090 card. So sometimes I can make several 2 minute video sections a day. Each day the maze expands.

So basically it's like Vyond, except with 3ds max quality graphics. It's a platform built for speed above all else, because of the high volume needed for branching plotlines.

You can see a pretty decent sample of this via the "Maerd" post, the top link is the entry point to the maze. Unfortunately, I have been deep in development for a bit, so I haven't added a lot of the cells lately, and it's only got about 20 nodes right now, many of which are alpha or beta versions. I have about 15 more that I've made but haven't released yet. The goal is to average about 2 a day, and then multiply that with staff once initial investment arrives. Ultimately the idea is to roll out thousands of interconnected videos carrying ads for monetization. Since the format is chain viewing native, it should do well once it's large enough to be interesting.
 
Last edited:
I love the idea of interactive videos. I've toyed with the idea using Blue-rays programming capabilities for a movie that has gates chosen then compiled into the finished movie. The tone and even the outcome of the movie would depend on some simple questions you would have to answer before the program begins, either that or you would simply choose one of a half dozen options such as , happy, dark, tragic, off the wall, etc..

8,000 frames of 4K per day.. Hmm. That's about a frame every 10.5 seconds. So, since you're talking about using a library of pre-rendered image, I guess you're mostly going to be compositing the various layers together. Probably using a 3D camera in Nuke or Fusion. Wait, did you say you're doing this all in resolve? I just started working with Resolve but strictly for a typical picture/sound editing workflow. I don't know much about the Geforce 3090 card is but it must be pretty powerful. My system runs a Quadro P4000 . I admit I'm not savvy enough about wringing the maximum performance out of the hardware. I don't tinker with the default settings much. For the work I do, I use the CPU to do most if not all of the rendering. These days I usually don't even set up lights for a lot of my shots. I use Image Based Lighting. V-Ray renderer for fluid simulations and and Arnold for most everything else.
 
Last edited:
I love the idea of interactive videos. I've toyed with the idea using Blue-rays programming capabilities for a movie that has gates chosen then compiled into the finished movie. The tone and even the outcome of the movie would depend on some simple questions you would have to answer before the program begins, either that or you would simply choose one of a half dozen options such as , happy, dark, tragic, off the wall, etc..

8,000 frames of 4K per day.. Hmm. That's about a frame every 10.5 seconds. So, since you're talking about using a library of pre-rendered image, I guess you're mostly going to be compositing the various layers together. Probably using a 3D camera in Nuke or Fusion. Wait, did you say you're doing this all in resolve? I just started working with Resolve but strictly for a typical picture/sound editing workflow. I don't know much about the Geforce 3090 card is but it must be pretty powerful. My system runs a Quadro P4000 . I admit I'm not savvy enough about wringing the maximum performance out of the hardware. I don't tinker with the default settings much. For the work I do, I use the CPU to do most if not all of the rendering. These days I usually don't even set up lights for a lot of my shots. I use Image Based Lighting. V-Ray renderer for fluid simulations and and Arnold for most everything else.
It averages about a frame every few seconds. I just can't set up good shots as fast as it can render, so there's a lot of unused time each day. I guess I'm averaging about 6 hours a day rendering.

Davinchi has basically replaced CS now, combining AE and Premiere into something that works better than either of them. The trick to get super high speed cg renders was to build our pipeline around slowing down a real time engine, rather than trying to speed up blender, max, maya, etc. I used to work with all of them, and quality was a bit higher, but it's not so noticeable to most audiences, and the frame time is literally 100x higher than the modded UE5 engine when using something like vray.

3090's are the apex of current power, and some of nvidia's demos for them are pretty amazing. This is a real time demo of UE5, running on the equivalent of a 2080 card, the previous generation. Anyway, you can see from this demo that we're reaching the point where compelling fiction can be animated in near real time.

I do use HDRI maps at times, and it speeds up renders a lot, but there are obviously a lot of advantages to being able to use a lot of custom lights in a scene. For example, in dark areas, a weak invisible light follows the cat, a few feet above it, allowing it to be seen better. It's subtle, but without it the main character can be too dark. In that futuristic port city scene, every single one of those red lights is actually casting shadows, even the ones way back in the bokeh.

 
I remember using max for years, and really having to worry about adding 4 or 5 lights to the scene. Now I can use 90 point lights in a scene, and it's really nice. It's not that you need that many for a scene, but you can literally build an whole set and light it, and just forget about them, instead of having to constantly sweat optimization on that level. I recently lit an entire jungle manually with spotlights, so I could control the light shafts through the trees and fog, and now I can just focus on the camera and characters, which is great. The real time engine is really built from the ground up to auto optimize basically everything.
 
I've added about another 20 cells to the Maerd demo, and thought I'd post an update. Right now many paths are connected randomly just for testing purposes, but it's now possible to wander around in the maze for an hour or so before seeing repeating content.

Here's the v1 trailer, since most people missed it,


and here's the entry point of the Labyrinth


Pretty soon, once I expand the maze a bit further, I'll start hiding small amounts of money inside it, maybe 20 bucks, for testing purposes. When it's ready to release, I'll hide a grand somewhere inside. It will just be a code that anyone can redeem for a paypal transfer.
 
People probably wonder why I've disappeared for a few weeks.

Basically, I came to a point in the predictable curve in the technology meant it was time to just go ahead and make Save Point. This entire time I've been developing, quality, control, speed, etc. Mostly this has been done by working on the Labyrinth test/demo project. Now we're only a few small steps away from being able to retroactively transform footage to look X. This means that we can now focus more on the core work, branching plots, world building, cell editing and the like, and if it's not up to visual par, we can literally, for the first time ever, actually fix it in post.

So before I can launch the social media campaign, or trailer, or Kickstarter, I have to have a final trailer that explains the scope, mechanics, and story of Save Point.

That's what I've been working on for the last month or so. For most of that time, I've been composing footage, rendering, writing, scoring, etc. At this point, the film is mostly in the can aside from a few obligatory reshoots, and I'm trying to edit this trailer.

Here's why I may not be around a lot for a bit.

I've got it all on the cutting room floor here, and there are almost 2500 shots to edit down into maybe 6 minutes. For comparison, that's about the same amount of editing work as 2 average feature films.

It serves as a broad overview of the game, and concepts are presented visually at an incredibly rapid fire pace, showcasing a universe of unimaginable size.

This one trailer literally has as many vfx shots as Avatar.

Anyway, I'll post it when it's done. This could take a minute, by which I mean 160 hours.
 
Anyway, I'll post it when it's done. This could take a minute, by which I mean 160 hours.
I think it was Stephen King that I read, talking about working with a world-class New Yorker editor on a non-fiction thing he published there. He was surprised and appreciative of the amount of time the editor spent with him, and I remember the answer, simple but true: It takes as long as it takes.

Anyway, good luck. Looking forward to a release. :)
 
Last edited:
This one is going to be super crazy. I'm wondering if I'm actually screwing it up by going so big. It might actually be hard to remember individual imagery, simply because it's such a barrage. I could have cut this up and made 10 separate videos out of it, but I'm trying to convey the whole game in one video, and even trying to be incredibly concise, it's kind of a lot of information for a trailer. There's a whole section that's just about stellar cartography, and how you get prepared to venture into unexplored areas of the universe.
 
Back
Top