The biggest problem is that the vast majority of new joiners at the AU tend to vanish after some period of time, before joining a province and beginning to contribute. I don't know what is the best way to improve the retention rate, but one idea would be to bring them into the community as some kind of 'provisional' member, where they can hang out on discord servers and get a feel for the project overall until the point where their skills reach the level where they can effectively contribute. They might gain a small sense of belonging that encourages them to stick around. I would be against, at least for the 3D portion of the AU, giving trainees an actual claim. I think the fraction of output that would meet the expectation level of a province team would be low - let's be real here, if it's the first 3D asset someone makes, it's not going to be amazing - and this would be discouraging not only to the learner but to the teams hoping to benefit from the AU. A system like that could work better for other disciplines. Not throwing my hat in the ring for AU lead here, while I'd like to see it become a better engine of talent for BS I don't think I have the spare time to do the job well.
Added a couple pieces for a semi-modular minaret set. The first piece is the base piece, seen on the small house in the later pictures, and there is a spikey add-on piece with a shiny brass orb. I was just admiring the outline of the tower in image 2 when I noticed that it's ruined by the outline of the garbage floating in the sky at 0,0 =| we really need to remove that stuff.
I'm in favor of linking (if not combining) the AU with the AF for the main reason that the trainees will be more likely to join a province team and put their skills to use at BS, if they feel more of a connection with the project after working on an asset. Of course, it may be optimistic that someone's first asset will be high enough quality for our use, but it's worth a shot.
So are you saying there are surfaces between the planks that aren't really seen by the player, covered up by the other planks? You should connect the edges of the planks if so, so that you don't have big hidden surfaces.
I would suggest these islands to start with: one for each leg (4)one for each brace between the legs (2)one for the seat one for the seat back possibly one for the head rest, or combine it with the seat backFor each one you'll want to create an island similar to what is shown for the box in the video. Try to hide the seams in places that are less likely for the player to see, so on the inside of the chair leg for example. This isn't always possible but should be done when it can. By the way, how many polys do you have currently? If it's not too high, you might consider chamfering the edges in some places to give a smoother result in the end.
Alright, we can make a lot of improvements here. What you're doing is essentially breaking every flat surface into it's own UV 'island'. While that will guarantee no stretch of the textures, it doesn't necessarily give you the largest pixel density, and also makes it very hard to create a texture with a smooth appearance that lacks visible seams at the edges. Do you understand that last point? If you have two isolated UV islands placed on an image, and they are then rendered sharing an edge in 3D, the texture will appear to be torn between the two faces. There ARE ways to get around that to an extent, but it's better to leave faces with exposed edges together in an island if the UV distortion isn't too great. This video has some decent animations to help get the visual concept of what you want to be aiming for when you unwrap: https://www.youtube.com/watch?v=iIvTUDgaXik
Sure, if you have identical (or even similar) surfaces you can reuse the baked UV space. But I'm not sure the front and back of your chair are a great candidate for that since there appear to actually be some differences there. A place you could use symmetry would be the legs, since it looks like your right and left sides are exact copies. In the low poly model you use to bake, only use one side. Then, for your actual low poly copy the UV mapping to the other. And, I wouldn't be to worried about fitting everything for the chair into one UV map anyway, it doesn't have all that much surface area relatively.
For use in game I would say to aim for less than 1,000 triangles. To go from something like 600,000 to 1,000 with decimation master it can be helpful to do it in two or more rounds where you first decimate down to say 5-10% of the original and then go to your target geometry. It's likely though that you'll have to use a combination of decimation master and modeling the low poly by hand to get a good fit. The other option is to just model the whole thing by hand. Up to you, but for me at least starting with decimation master and cleaning it up is a time saver, and that's important because I find creating the low poly to be the most time consuming (and important!) step of the whole workflow. I put together a few examples of what you're trying to do. Here in the first one we're trying to make a good low poly mesh that captures the high poly mesh efficiently with the fewest vertices. This will mean sometimes being on the 'inside' of the high poly, sometimes not. After UV-unwrapping the low poly mesh and completing any changes you want to make to it, you can now create a 'cage' for the low poly. This model is used to calculate the angle and position where rays are simulated in the baking process and should be on the outside of the high poly at all positions. The easiest way to make this is by using the Push modifier, which moves a vertice along the normal vector. Some manual tweaking is necessary at times but Push does most of the work for you.
Now with the low poly and cage, I was able to take the painted high poly (left) and bake the polypaint to a texture (middle) and then use that texture on the low poly model within Max (right).
It's ok to add a little more geometry here. For an asset like this that will be seen up close it's fine to bevel the edges and conform the surface to the woodwork a little more. For instance, your high poly has inset squares in the back rest, you could add an inset square in the low which you might not expect to add much, but the ability to vertex paint those few additional vertices can really help the overall appearance. Do you have your high poly imported into max as well? If not, Go ahead and export your high poly as an obj and merge it into your max scene. From there, clean up and optimize your low poly, and finish by UV unwrapping it. Posting the uv layout here for feedback is probably a good idea. After that, we're going to bake our first maps in XNormal, so you need to make a 'cage' for baking the high poly information onto the low poly. The cage mesh essentially projects rays along the normal line, and needs to completely surround the high poly. This video provides a decent explanation of how to do it fairly easily: https://www.youtube.com/watch?v=yeczjVAqznM