Sunday, September 6, 2015

Shower and Models

Haven't been doing too much work on the demo for the past few days. Got preoccupied with work and reading "The Legendary Moonlit Sculptor". Not to say there's been no progress. I think found a model that I can use without violating some R-18, a loophole for using TDA in the demo, worked on normal maps for the avatar, added a few furnishings to the beach scene, and used the standard shader on a MMD model. At this moment in time, I'm working on the controls for the avatar more than anything as I want to try and get some of the rudimentary functions of my character ready for the RPG project that I want to work on after this. Hopefully the lessons I learn here will be useful down the road. For reference, I'll go into depth here.

The "loophole" I've found with the TDA models and many MMD models is the condition that they not be used for sexual materials. That however, does not posit that the models can't be used in systems that do. As such, I should theoretically be safer to use the models so long as I don't use them in the more sexually explicit scenarios. Under that, all I'd need to do is create two separate scenes, dividing the availability of the models to only the scenes that are tame in nature. In the end, most of the models are probably violating a copyright of some form in the end anyway and I'm probably overthinking things, especially considering the things people in the doujinshi scene are getting away with most of the time anyway so I'm not going to let this overly burden me for much longer. Worse case scenario, I'll just provide the demo's data itself and let you all reconstruct things on your end at your leisure. That's actually a smaller tidbit of a larger idea I have that can enable the circumvention of copyright under the current legislation governing matters in this area, but that's a matter for another post.

As for the normal maps, this is really just more of a learning experience for me since I'm not quite well versed in these matters and want to expand my knowledge set. I'd realized that the lower polygon version of my model could probably be improved by switching to using normal maps, so I decided to give it a shot and see what I got. I liked the results (would have liked them more if I actually like the model, it turned bum ugly) since it gave more definition, though I've yet to have a look at it from VR which makes that impression a bit tenuous as VR and normal maps are rather hit or miss. I've scene them work well, but I've also scene them fail miserably. I think it's a matter of using the shading to amplify already present details rather than in place of them to get the best results as it makes the trickery less noticeable. As an example, sticking a normal map of a brick pattern on a completely flat wall and expecting it to look good in VR right now is a joke. But, if you make the brick wall have the grooves inherently and add a normal map to flesh them out, they look far better and make things look good. Think of it as the difference between printing out a face on a piece of paper and sticking it on a ball versus actually sculpting the face then applying make up to make it look better.

I'm using Unity's pro water for the water and can say that it looks pretty good at a distance, but leaves A LOT to be desired up close (the whole normal mapping issue again), but the beach is looking more "beach" like with the enlarged waves. I added in a shower to get cleaned up and a boombox that will be able to cycle between songs at the user's preference (Yet another instance where I don't know about whether or not I should include the songs I have, learn how to let you use your own, or figure out a way to circumvent things). Apart from that, there's also the addition of some rudimentary soundeffects and audio. The ocean waves are pretty good and the shower water does actually sound like a beach shower.

From a shader perspective, I only really just changed all of the MMD shaders to different varieties of the standard shader, altering only whether it was translucent or opaque to see how it changed the aesthetics. I found it was rather nice since it works better with Unity's lighting system than the MMD shaders. I'll have to mess around with the smoothness and metallic properties later to even out the look, but all in all, I'm fairly content with the results. I'll probably go for a similar approach for all of the models as I make them to try and keep a unified aesthetic, though the production rate based on the amount of time it takes to render will likely leave things in the realm of 4-5 hours to get one done. That's a lot of time for me to burn, but I may be able to handle it if I leave things rendering while I'm at work, though that leaves the rate of production at one per day.

Outside of the models though, the avatar is what I'm focusing on right now. The rudimentary function I'm trying to work out the most right now is the left half of the controller, or the locomotion side as I've dubbed it. I've got it so that the left analogue stick operates on a tank control like system as the default state, with an alternate state called rotation mode that is meant to augment the core default state to enable a wide range of movement. They're both functioning as I intend right now from an axial perspective, but the rotation mode is still in need of the main item that combines these things together: a crouch. From the MMD demo side of thing, this is valuable to get a wide range of possible viewing angles on the models, which is a necessity for viewing panties and enjoying the ride poses. For the rest of the demos, it'll be key for the jumping and flight going forward, so I need to get this sorted out as quickly as I can.

The best solution I have in mind would be to make the player's avatar capable of ducking and that's what I'm working on right now. The medium term solution I have in mind is for the player's body to progress to a seated position then back based on the input of the left trigger on the Xbox controller. It's akin to making the Xbox trigger represent the hip joint flexion action. This is useful since the actions of ducking, and jumping are essentially using the same actions, merely with different applications of force. A jump is the by product of the force being used to raise the body back up being sufficient enough to push against the force of gravity and continue upwards in space a bit. Flight meanwhile would activate as a result of attempting to jump in air, a double jump if you will, and would change controls again to another system for flight. Specifically, in this state, letting the trigger go (0), would result in backwards movement, holding it down would be forward (1) and holding it in the middle would be needed to stay still. Rotational controls stay dominant in this state meaning controls would be something akin to a flight game in the vain of Star Fox, though with more control in the sense that you aren't trapped facing forward and without the ability to turn around fully at a single point. This is a bit more long term as both jumping and flight aren't really going to be useful in the H-demo (unless I want to add a skyhigh club scene, but I think that necessitates a plane not wings and the addition of Nintendo no immediately sending a cease and desist for letting players bone the princesses as Mario. I could do that by the way, the models and animations are there.

All in all, the demo is starting to look like something I would have downloaded myself had it been put up on the Oculus NSFW reddit. Just need to set up the ability to switch between some more animations and I'll upload the first version of the demo as soon as I can. Wish me luck. Maybe I'll have this darn thing done for joy on labor day :)

No comments:

Post a Comment