Thought this was maybe worth a separate thread / discussion (merge it with main 3D dev thread if not).
Basically, at the moment we've got a lot of "bits" in different places. If you look at Graf's trackers, there's a lot that's been done so far, but there's some other bits that need doing to connect everything together. i.e. we could make another 100 additional armours, but if they're never rigged or implemented by code, they'll never do anything except look awesome on a few screenshots. Another option is to go through all the existing stuff and improve it - this may however involve re-doing work we've already done - but in theory, that'd probably need doing eventually anyway.
A while back, I made a
sort of list for someone asking about "when will it be ready". There's been good progress on a lot of that stuff, but I wonder if we should re-focus it a little.
What I was considering is that we could move the "fully testable" bit forwards by putting some of the "extra stuff" later - so ignore the suit, long coats etc for the time being, and focus on making leather, metal and combat armours as good as is currently possible. Basically to make the 3D versions of everything that currently exists in the game first, then add the extra bits afterwards.
So first, we'd try and get this stuff sorted :
- A final agreed shadow / light / specular setting (I'm happy with the current one, but if it changes in future, we'd likely have to re-change the textures)
- White skinned Male bald, male short hair, male long hair and female models
- All animations for these
- Various death animations
- Various weapon effect animations
- Bluesuit, leather jackets, leather armours, metal armours, combat armours
- Generic pistol, generic rifle, generic SMG etc (i.e. like the original sprites, we initially use one model for all weapons of that type
- floor models for those armours and weapons
- something about the timings of animations - currently the enemy is "hit" before you've moved your gun or fist
- Models and textures linked up and implemented to appear in hand / on body via game mechanics, rather than "param codes"
The most problematic ones are weapon effect animations, some of the death animations and the animation timing. I think the other stuff is all much more straightforward. It's mostly adjusting and tweaking. I know Karpov is working on those animations at the moment, but the timing thing might be related to the original engine, rather than something we can change. I'm not sure.
These are current model-to-sprite comparisons. Now we've got the specular stuff working, we should probably redo these textures, or tweak the models a bit. Try to get them just right. Reshaping the models a little until they match the shape, messing with the textures so they're exactly the right shade, tweaking the specular bits until it matches the way the light works on the original sprites, adjusting the skin texture until it matches etc
Once those are done, it's testable, maybe usable in the main game. After that's done, we can add the extra skin tones, extra texture variations for armours, extra body types, start doing the same "tweaks" to try and perfect our additional armours, start adding the extra weapons and hand / floor versions of objects in etc.
Obviously this is only a possible suggestion, and there are many other ways we could focus things (i.e. rig and implement all models and objects we've already made, then do the tweaking and refining later). One problem is that doing things such as the death animations may take a long, long time - and we'd have to pause other things until that's done. There's probably some sort of compromise where we can finish a lot of stuff AND add new stuff, but it needs thinking about.
Essentially, I've recently been experimenting, "researching", tweaking and testing lots of different things - but finishing nothing
If we agree on a focus or "road map", or at least a priority order, then I (and everyone else helping out) could focus on one thing until it was complete, then move to the next on the list. Much in the way that we have the 3D development trackers, we could maybe do with a "tracker-tracker" - outlining the bullet points of all the bits that need doing.
So, any thoughts on this or other possible ways of focussing the work?
