Taking in a Stray - Dev Diary

[NandabaCanti]

My system is pretty much on it's death bed now. Rebooted on me just about 50 times at this point over the past 4 days. But luckily I've got a bit of help from a friend of my family who offered to loan me a bit of money to get a few essential upgrades to replace part of this damn system.

Nothing too special, but way better than I've got now. Decided to go with an AMD Ryzen based system to save a bit. Went with one of MSI's Tomahawk motherboards (B450 chipset), the apu is just a Ryzen 3 (3.5 GHz quad-core), and then got 16 GB or RAM to go with it (2x8 GB 2400). Going to have to stick with my old GTX 750ti for now though, as well as the old case, PSU and HDDs, but they should all be fine for now.

Probably won't show up until around (or after) new years if I had to guess, so there isn't much I can do for now but wait and spend some time vegging out with the tv and doing some traditional drawing on paper to pass the time until I get the parts in to rebuild. Hopefully though everything works out with the rebuild and I can finally get back on track with creating some damn art, lol.

I've still managed to make a lot of headway with the puzzle game, got the basics of a lot of the major mechanics worked out in test scenes in Godot. But I think at this point I will wait until I get things back up and running with the new parts, it's just too much hassle trying to get anything done with all the sudden reboots. I'm not sure where this leaves the other games, at this point I've started about 3 of them, scaling back each in their technical complexity to match the ever worsening state of this system. I'll have to give it some thought while I wait and decide my priorities for when I can get back to work.

 

[NandabaCanti]

After some thought, I feel I want to focus mainly on my first passion, which is the game this thread was started for - Taking in a Stray. But I don't want to abandon any of the other projects either. The picture puzzle game will be the easiest and fastest to finish by far. So I will probably make that a priority as soon as I get this system back on its feet. But my creative desires pull me in a lot of different directions. So when I'm not working on that puzzle game, my 3D work will focus here on my main project, TiaS. Though when I'm in more of a painterly mood, I will work on the background matte paintings for my text adventure game, The Ever-Setting Sun.

If I find myself wanting to mess around with low poly stuff for a bit, I will work on the town for the most recent project I was previously talking about in the

You must be registered to see the links

post. I think I also want to mess around with Blender 2.8's new grease pencil features to do the sprite work for that game as well. Though that game, being the latest addition and the least developed will get the least work for quite a while to come. Though I've got to say I had a lot of fun mapping out the town for it, I basically already know where everything belongs in that one and created a rough map for it in Blender shortly before this system started pushing the reboots into overdrive.

So the jist of all this is that I want to get the puzzle game done as quickly as I can, while most of my art efforts will go into TiaS and on the side when the mood strikes I will switch gears to do a bit of work on the other 2 games. I technically also have at least 2 other games I can work on as well - a 3D island community sim where you play the part of a minor diety who can possess the locals and a short 2D VN about a chance encounter with a monster girl in the wild - but those will stay firmly on the back burner until I get at least 2 or 3 of these projects done first, lol.

 

[NandabaCanti]

Got my new parts in on Friday and I've finally got a properly working system again. Other than the RAM, it is only a minor upgrade spec wise (went from a 3.4Ghz quad-core to a 3.5Ghz quad core CPU), but the AM4 boards seem to have a fair bit higher throughput than my old AM3 stuff. RAM wise though it is a nice upgrade, even if I am still using kind of lower end offerings for this gen of hardware, but 16 GB of 2400 DDR4 is a big step up from 8 GB of 1600 DDR3.

The important part is that I can work on stuff without reboots now. A lot of stuff happens a lot quicker too, like building software, packing/unpacking archives, etc all are 3-4 times faster for me now. I'm probably going to lose a fair bit of time catching up on a lot of gaming I haven't been able to do over the past year or two, heh. But I make it a point to spend at least a few hours a day working on my projects. Got another of the core mechanics for that puzzle game worked out today, the animated tile swapping. Previously I've worked out how to import images and slice them up for the tiles, now that I can move them around the most important bits are all out of the way. Now I just need to tie them together with some GUI bits and do a ton of little things, like the puzzle editor that will let you choose the tile counts and what area of the imported image you want to use for the puzzle and a lot of the customizable cosmetic options and such.

The big issue now though is just getting back into the right mindset to focus on TiaS after this long break and getting comfortable with the changes in Blender 2.8. I've got to pick up the pieces and start figuring out a proper priority list for all the work I need to do for that. Somewhat unfortunately, during the last few days of my waiting for the parts to arrive, I found myself going through some more Godot tutorial series to pass the time and ended up putting together a good chunk of a basic match-3 puzzle game, so now I've got yet another project on the stack as well. >.<

I had a fun idea for it though, basically the different types of gems for that game correspond to different sexual targets (vaginal, anal, oral, etc). When you make a match it adds points towards that type of interaction, So as you gain points you unlock controls for engaging in those various sex acts with the character taking up the other half of the screen opposite of the puzzle and you have to unlock enough to reach orgasm in at least one of them before you run out of puzzle moves or you will leave your partner dissatisfied and lose the game, heh. Some times the gems will also have toys in them, when you make a match with those you'll unlock new sex toy options to utilize as well.

It isn't a game I plan on investing much time in right now, just a fun little side project to mess around with from time to time, much like the low poly town exploration game. And in fact it could be a good warm up project for that, since I haven't done any 2D character animation before. So something like the match-3 game could be a good place to experiment and get some practice in with that side of things, which will prepare me for doing the animations for the low poly game as well, since I still plan to use 2D sprites for the characters in that one.

I'll probably be able to get the puzzle side of that game done fairly quickly, since it is about half way there already. But the animation work is a huge bottleneck, so I doubt that game will be in any sort of suitable release state for a long time to come, especially since it will just be a side project and lucky if it sees a few hours of development a week until some of my main projects get done.

If all goes well, and I don't lose too much time to gaming, I hope to finally have some art to share by around mid-to-late January. I already know for sure that I want to get the base model for the homeless girl out of the way first in terms of TiaS progress. I'd also like to get a good chunk of the material/texture work for the models I've made for the park scene done and maybe the exterior of the house model to share some shots of that too.

 

[NandabaCanti]

2019's First Update
I did a terrible job of avoiding the temptation of getting lost in gaming after I got my new hardware. I spent a huge chunk of the first two weeks spending most of my days gaming, catching up on stuff I've wanted to get around to the past few years while I've been struggling with hardware issues. I'm starting to get it a bit more under control, but most of this month has been a write off so far, heh. I haven't got around to finishing and polishing that puzzle game; I haven't really done much of anything with Godot this month yet, but I've started to get the urge to get back into it again over the past day or two.

Spoiler: Taking in a Stray Progress

You don't have permission to view the spoiler content. Log in or register now.

I originally wanted to wait until I had her a lot further along to start sharing art, but with all my setbacks and recent time wasting with binge gaming, I'm getting tired of holding back on sharing at least a bit of something, even if none of it's even close to a finished state yet. I'm sure any of you actually interested in all this don't want to wait any longer to start seeing some images either, so I've decided I will start posting the early WIP images and update more regularly with new images to show the progress being made. So with that in mind, please be aware this stuff is really early work that will change quite a bit. Almost nothing is textured at this time and most of it is straight up viewport snapshots in Blender's default solid shade mode. No rendering or any of that going on, I'm not even at a point where I've started to move things to 2.8 to start working on their materials in Eevee yet.

Spoiler: Screenshots

You don't have permission to view the spoiler content. Log in or register now.

 

[215303j]

Here is the rough sculpt of the homeless girl

Looks good, but she looks pretty well fed (not fat but not starving either). Not sure if that's the image you'd want to portray. Given the topic, you could make here skinnier, to the point of becoming unattractive and then maybe gain some weight after the story progresses.

 

[NandabaCanti]

I can see what you mean, but she's only been on the streets a few months and has done a fairly good job with figuring out reliable locations for dumpster diving and the like. While her mother was still alive she wasn't exactly cared for, but she never left her completely neglected. She at least made sure her daughter had a little money to buy food and necessities. So she learned a fair bit of self-sufficiency and picked up a frugal nature pretty early on in life. Her problems came more from social and emotional neglect rather than things like malnutrition or other physical abuses.

I'll be spending today retopoing her though and I do plan to tweak some things a bit when that is done. Also that screenshot was done in an orthographic view, so the appearance of some of her features may be a bit misleading due to the lack of perspective in the image. When dealing with high poly sculpt models like that you get better performance staying with orthographic views, that way Blender (or any other sculpt program for that matter) doesn't have to do the perspective calculations for all those faces on top of everything else. I was periodically switching to perspective to check on things though. But the other reason I went with ortho for that shot was that in a T pose like that, it's hard to center her without the closer arm getting cut off by the edge of the image while having the camera so close on her and the perspective was a bit too much at that range, distorting things more than I'd like and I didn't feel like fiddling with draw distance values and camera tricks to fix it when the orthographic view was a key press away, heh.

It's also worth noting that down the line there will be other models to choose from for both body and face. It may be post release, but I do want to make that character quite customizable. But for now, my vision for the first version of her is that in many ways she would seem rather plain or average at first glance, wearing baggy/shapeless clothes that hide her figure, but kinda cute when you do actually take notice of her and with very expressive eyes that pull you in. The eyes (and face in general) still need a fair bit of work because I wasn't yet working at a good resolution to define them properly, rough sculpting with dyntopo is good for quickly blocking in larger shapes, but it falls apart when trying to get very fine and precise detail without pumping up the geometry to nearly unworkable levels. Even with the low poly version I'll be making today I will have a lot more control over that sort of stuff and then it will only get better when I do the detail sculpting on top of that later this week.

 

[215303j]

In particular, I was thinking about her belly and upper thigh area.
Of course, IRL, it doesn't mean that all homeless people are starving, but on the other hand, it's also about the visual image that you want to portray.

 

[NandabaCanti]

Ya, those areas in particular will seem bigger than they actually are with the lack of perspective in that image, since they are further from the camera and your mind is expecting there to be some foreshortening going on. Though part of the problem is that I didn't do a very good job of defining the different areas of the stomach after combining the shapes into one and smoothing things out. As I was getting into the retopo in that area just a bit ago I was realizing the area between the belly and the sides needs to be better defined with a deeper separation. But that's something I can easily fix by moving a few verts in the low poly when the retopo is done and I can apply the shrinkwrap and hide the rough sculpt and focus on the final refinements of the base shape.

 

[NandabaCanti]

So today was mostly a bust. I did the retopo most of the way through twice, but wasn't happy with either of them. So I decided to clean up the sculpts a bit more before going to bed and do a really quick and simple sculpt for a placeholder hair and eye brows, just to make the model look a bit less weird until I can get a real hairdo made. Then I took a few new screenshots, one with her in perspective this time and another with a quad view so you can get a better idea of how she looks.

 

[215303j]

Looking good, but this time I think that the head is a bit on the large size compared to the body.
What's her height?

 

[NandabaCanti]

I went with a stylized ~6.5 head tall basis for her proportions. I'm not shooting for a very realistic art style with this game, but I'm also not going for straight up anime style either. I want something in between with the environments done in a style very similar to anime backgrounds, though perhaps slightly more detailed. Characters with more realistic faces, but with stylized body proportions similar to what is common in many anime. In terms of materials I'd like to go for something a bit like the Illusion games, but with a better material setup for full PBR support. I always liked the way their characters have realistic elements to their appearance, but are still quite clearly stylized works of art that you wouldn't confuse with reality. Though I'm probably going to tend to go even more towards the stylized end of the spectrum than they have, for instance I'm planning to do the hair with solid shapes instead of particles or semi-transparent cards (which is the route I think Illusion and many other game makers take).

I may still adjust her head proportions though, but I want to wait until after I get her default hairdo made and see how things look then. Oh and here is

You must be registered to see the links

of what a 6.5 heads setup tends to look like in a typical 2D anime style as a point of reference. This may just be personal preference, but I always prefer art to be easily distinguishable from reality, since I don't much care for reality in most cases. I want the game worlds I go into to offer me something I can't find in reality, so I always drift towards more stylized work. What I'm after is a bit like the way the

You must be registered to see the links

series does their characters, but higher poly, no cell shading lines, and pbr materials. I like that hand painted style they go with for the skins, but something a bit cleaner since most of their characters tend to be rather grungy. They also use the solid geometry hair models and stylized proportions I like (but they go even more so in the stylized direction than I plan to with mine).

So with all that in mind I guess I am going for a style somewhere between Honey Select/Play Home and Borderlands, with a better material/lighting set up than any of those.

 

[215303j]

Oh, OK, I thought you were going for a realistic depiction. That makes sense, thanks!

 

[NandabaCanti]

So I finally got a retopo I'm fairly happy with, might rework some parts down the line, but at this point I'm satisfied enough to move forward with the model. There are still a few things I have to do with it, mainly clean up the fingers, toes and ears. As there was only so much I could do with them while still using the shrinkwrap modifier. I may just end up making a vertex group to exclude those bits from the shrinkwrap so I can clean up the geometry placement in those spots.

I may even go back into the rough sculpts for a little touch up. The forearms are a bit wonky, the bottom of the rib cage sticks out a bit too far in the front, the corners of her lips and lower cheeks can be pulled back a little further and I feel like the thighs can use a bit more definition. Right now I went with a pretty standard straight grid setup that most game models use, but thighs are a lot more complex in reality, with a lot of crisscrossing muscles that cause very defined edges cutting across the front and back of the thighs in a variety of poses, even with people who aren't very muscular. But that bit will have to wait until some other time once I've given enough thought to the matter and done a few anatomy study sculpts for that region to better solidify my understanding and come up with a better topology approach for it.

I also still have to model her genitals and the inside of her mouth. Plus the eyes and eyebrows also need to be replaced with their final versions as well (those I will probably be making today if I don't waste too much time on the retopo clean-up). The genitals I'm intentionally putting off because I need more geometry down there still and the rough sculpt just got a simple indentation to demarcate the area. The way I worked out the retopo was that I basically did it at a quarter of the face density I desired for it. That way I could rely on the shrinkwrap modifier and 1 level on a subdivision surface (subsurf) modifier to fill in the gaps, which effectively cut the number of verts I had to position by hand to about a quarter of what it would be otherwise. In the new screenshot below, imagine there is a vertex in the middle of every face you see displayed on that model with edges that connect them in a grid pattern as well as new verts at the spots where that new grid of edges cut across the existing one, basically turning every quad there into 4 new quads and every triangle into 3 new tris.

Once I'm happy with the state of things with the current setup, I will apply that subsurf modifier to bring in all those extra verts I just mentioned to fine tune some areas, at which point I will have enough geometry to connect in a model for her crotch and put the final touches on her fiddly bits like the ears and the tips of her fingers and toes. The inside of the mouth will already be kinda dense with detail, so I will likely model that separately and add it in after the subsurf has been applied to the body/head.

For those curious about geometry stats, as it stands the retopo of the body and head right now is coming in at about 20k tris with the single level of the subsurf accounted for. I have a feeling that by the time I finish her nether region, mouth, eyes and hair, she'll be a lot close to 40k tris.

 

[Vicktor Karloff]

WOW. Good work on that model. Will that "rendering" (I don't know the correct term) be able to be animated? Is that type of modeling doable in Blender? That is what I'm using to teach myself how to model. I can't wait to see how your game turns out.

 

[NandabaCanti]

@Vicktor Karloff Thanks, that is possible in Blender, considering that was what I made it in, heh. I've been using Blender as my primary 3D design tool for over 12 years now (used Maya for 6-7 years previous to that). It isn't rigged yet, but it will likely be at some point in the next month or two, depending on how motivated and focused I can keep myself. My main project will be in full 3D with animated models. (Though you probably won't be able to explore the environment much, beyond moving the camera during sexual encounters. It will be using maps to navigate around the world and interactions will mostly be handled VN style. If you've tried anything like

You must be registered to see the links

, it will be a bit like that in regards to how you get around and interact, just with the static renders replaced with 3D environments and character models.)

I strongly encourage you to stick with Blender though, if you are serious about getting into 3D design and indie game development. It is a very robust piece of software that can do most everything you are likely to need to do in 3D and then some. I initially moved away from Maya, because I wanted to clean up my system, get rid of all the pirated productivity software and go legit, so if the time ever came that I could start selling my art, I wouldn't have to worry about companies like Autodesk or Adobe potentially coming after me some day. For a similar reason I started to dual boot Linux around the same time, and within a year or two had switched to using that full time as well. But within a month or so of starting with Blender I was more productive with that then I was after years with Maya and from then on I was hooked.

Also, if you are just starting out, don't let yourself get discouraged. 3D design is a daunting field with a steep learning curve that you are going to have to fight through at your own pace without giving up if you hope to eventually start making works of art you can be proud of. It helps if you have some sort of background with physical sculpting and or 2D art (drawing/painting/etc), but not at all necessary. They just tend to speed up the time it takes to find your legs with 3D art. It never hurts to get into a daily habit with something like drawing though, even if it's only 10-20 minutes a day at first working on stuff like gesture drawing, it will make a huge difference in your ability to see things properly and work confidently.

If you have any areas in particular you'd like to get to know better in Blender, you can let me know and I'll try to point you towards some tutorials and channels that may be pretty helpful for you. I'd done a little of everything with Blender at this point and have watched countless tutorials and timelapses over the years, so I can usually recommend something good for just about any topic, heh.

 

[215303j]

Looking good mate!
Keep it up!

 

[NandabaCanti]

I kinda took it easy over the weekend, so haven't got a lot done recently. But I have got the inside of the mouth sculpted out into 4 separate parts atm, got to retopo them and join it all together and connect it to the head, Yesterday and today I started to play around with Eevee, Blender 2.8's new real-time PBR viewport render engine. It's rather nice, I've mostly just been messing around in the LookDev view mode, which gives you access to some built in HDRI maps and simplifies some stuff so you can have a nice and responsive view that will give you a good idea how your materials are looking.

I made a quick model and UV unwrap of a single breast and painted a really basic albedo (color) map for it and then started to mess around with the principled shader settings until I got something reasonably good looking out of it. I could definitely improve things by adding some additional lamps or painting maps for other channels as well, but using the slider values are good enough for rough approximations when just looking to evaluate things quickly. I'm pretty happy with the results it offers and from what I've heard, Godot 3.x does a fairly good job of reproducing results achieved with Eevee when it gets fed the same values, which should be the case for all PBR setups really. That was the whole purpose behind why that material model was developed in the first place, to help standardize materials across all DCC and rendering packages that implement it (though some times you may need to invert a channel in cases where they may be using something like a smoothness value instead of the more popular roughness, like with Marmoset).

Anyhow, here was the result of that test, figured I'd post it for shits and giggles.

 

[Vicktor Karloff]

figured I'd post it for shits and giggles.

I think you meant to say shits and wiggles.

 

[Cirro84]

...

Well. I see a tit. One only. Now I *have* to see the chick that belongs to it. Or have you gone all Jack the Ripper now?? Know you do deliver in part but still?

 

[Vicktor Karloff]

Well. I see a tit. One only.

Don't they usually come in pairs?