0.8 Dev Update 15

Hey, folks!

Maid Lyriel is back!

If you missed her last year, don't miss her now! ➡️https://rift.ink/4y6f6aS 

➡️v0.8.14 is now in Beta

And coming to EA later this week. We have a new story event for Blair, and, of course, Lyriel's First Anal Sex Scene!

We're going to add two cheeky little Alchemy recipes before EA, but that's basically it for now.

I know, not the biggest of the updates, so let's dive into it.

Current Situation

We have another dev up where I'm going to be mostly talking about Our New Pipeline™️. In short, Blender, or rather Eevee, is being a bitch. I don't remember if I ever went into detail about Our New Pipeline™️ before, so why not now.

First, a bit of a background. We decided to start moving away from our old Unity-based workflow after doing a few scenes on Naomi's ship and then in the big city in MC's world. NPC-filled scenes, scenes with wide open spaces, all those were hell to work with. Everything was extremely slow even on top-tier hardware, we've been encountering visual bugs, lighting issues, and just plain old crashes. And after all that hassle the end result was very underwhelming.

Oh, and the animations. We just straight up hit the ceiling of what was realistically possible to do with the tools we had.

So we decided to start moving to Blender. Horny Herring - our sister animation studio - already had plenty of experience with it, we've made multiple Specials in Blender, and you folks seemed to like them. No-brainer, right?

There was one little thing. We decided to go with Eevee instead of Cycles. I assume most of you are not that familiar with Blender, so I'll try to give you a rough overview. There are two basic approaches in turning 3d scenes into 2d images: path(ray)-tracing and rasterization. Path-tracing (Cycles) is the "realistic" option. It tries to simulate the way the light actually moves through the scene, while rasterization (Eevee) fakes basically everything. You probably have seen the RTX reflection comparison videos, right? It's basically the same thing, only not really.

Anyway, here's an illustration.

Now, Eevee looks kinda bad here, but it's usually not that bad. We picked Eevee because in theory it should have been much closer to our Unity-based visuals. Aaaand it was probably a mistake. We've encountered a whole boatload of issues and we keep encountering more even now. And we couldn't even really emulate the look, at least not to the extent I wanted us to. The lighting systems are just too different. I won't go into too much detail (because I don't really understand it that well myself), but Blender lighting system is much more realistic, while our regular light setup was very much dependent on Unity quirks.

To add to that, the performance issues are demolishing our productivity. For example, the instancing craps the bed in the forest scenes from Blair 19. We had to rent RTX PRO 6000 with 96gb of VRAM just to render those. 

Yeah, just these scenes. Yeah, they are not, like, ridiculously complex or anything. Hard to imagine modern Blender being worse at rasterizing foliage than Unity 5 - a 10 year old game engine - but that's how it is.

Still, we've already made the decision. A couple of months back we had to choose how to prepare new environments, NPCs and props, and we decided to go with Blender. We've already made some, we're making more, and it's like 99% certain we'll keep going with Blender.

What that means is that we're going to drag our feet for a few more updates. The amount of finalized content we can produce in a month is not huge at the moment. We're still working on our tools, porting/remaking characters and NPCs, updating shaders, fixing issues, etc, etc.

I'd say the work is going fine overall. We've made our own plugins to make working with large quantities of static scenes easier, we've mostly finished remaking the main girls, the House and some other environments, characters and props.

That said, we definitely underestimated the scope of the migration. We also have a few things we need to sort out before I'm happy with the quality.

And we lost some QOL features, too. We had a primitive, but mostly working dynamic physics for working with hair and clothing in static scenes. Can't do that in Blender, need to run a separate simulation, which is not always possible, so we're mostly making do with manual work, which is not exactly awesome.

We will persevere nonetheless.

Let's go over some good things while we're at it. You've probably noticed already, but I just can't shut up about the animation improvements. They are so, so much awesomer now!

Look at that plapping. JUST LOOK AT IT.

What's more, we can make actual, detailed, dynamic cum animations now! You don't know how I loathed the fact that we couldn't make a good ending animation in the old pipeline. Hated it! No sex scene is complete without a good cum animation!

And that's not the limit! We'll keep going until our in-game animations are as good as our Specials!

Right.

We're going to run a poll next week, after the EA version is out. I want to know what you folks think about the events and animations made in Blender. We'll keep working on making them look better in any case, but we might adjust the course somewhat. Maybe we'll even move to the path-tracing renderer - Cycles. That will definitely take care of some issues, but could very well introduce some new ones.

Btw, if you have anything you want to say about the graphics right now, you don't have to wait until the poll! Always appreciate your comments, folks!

v0.8.15 and Beyond

Let's end with a small preview for the next update. We'll be keeping our focus on Blair, adding at least one story event for her...

And a not-so-story scene!

When we're done with her current arc, we'll move to bringing Yona up to speed and to advancing the main story. I really hope we'll solve most of our Blender issues before the end of this year and get up to having 4-5 new events with each update.

And then we'll probably be so locked in, our next project could very well become an interactive movie.

See you soon!