I really do think that the 'making it for the audience' part is secondary to WHY people do fan projects.
Its just the natural end point - but its not WHY stuff like this gets made.
Ive done lots of fan stuff - 90% of it just for me - some for other people - and some as stealth 'pitch work'.
Fan projects have a lot of advantages.
People who are fans will generally contribute ‘more for less’. Most creative people will never actually work on Star Wars or The Last Of Us - so often working on a fan project is a sort of wish fulfilment.
When it comes to production, working within a known IP is just, honestly, much easier. The rules are set - a lot of the groundwork has been done. You aren’t having to build from nothing - you get to stand on the shoulders of giants. It lets you focus on delivering the best slice of this ‘piece’, without having to focus your limited energy and resources on that foundation.
Built in audience - certainly helps to know that at least there will be some people who will enjoy your work!
But mostly - sometimes when you love something you want to be a part of it.
For every single creative person I know, the final product is just the icing. It’s the desert - not the mains.
The reason we do it is because the actual act of doing it is what brings joy. The late nights, the sleepless times of worry and work - the dreaming and planning and seeing it all slowly be stitched together.
If you’re gonna do all of that - why not do it for something that inspires you?
I do hope that this branches off into other projects - perhaps their own IP, or other collaborative works.
But sometimes it’s just about making something you want to make.
I probably comes more down to the performance and filming side. Sometimes if its gonna take longer to do a costume swap, or reset makeup, or even just small things like having him not scrunch his face while he removes it - honestly in a lot of cases the whole 'we will fix it in post' does make more sense. :D
Is this done with a complete CGI takeover, or a collapsing 'shell' in the mask, and then blending out the seams at the neck?
The last 5 or 10 frames says 'CG takover' to me, or complete CG mask from the start - but I can see how it could be done practically.
His hand and finger movements are spot on (which I imagine would be incredibly difficult to 'mime' - and doing a full hand/arm replacement seems like overkill).
I think that they are always wanting to keep the silhouette and 'face shape'. The Spiderman mask itself is usually made of a hard plastic thats just painted to look like its a soft fabric.
Specifically his hair would likely deform the mask quite a bit if it wasn't rigid.
I know they also have different suits with different mask attachments - so one where the mask and the suit are 'one' - so there is no seam at the neck line and it reads as one single outfit - and then they do one where he can remove the mask - so it's split at the neck.
Here it looks like they are transitioning from the 'complete' suit (no neck separation) to the suit where you can remove the mask.
Usually what they do here is they would just clean up the line on the neck so that it looks like its one piece - and then blend it into the 2 piece suit.
But the mask itself doesn't appear to deform when he takes it off - specifically the eyes.
His hair would probably also have more of a 'hat hair' feel to it - and possibly even break continuity between takes if it was different mussed with each take.
That hair is perfectly sculpted to look messy.. :D
My assumption is that this was shot without the mask at all - that he is miming removing it - and they CG'd in the mask.
OR they did a plate with the mask - and then one without - and are doing a blend between the two... which would probably be the easier of the two.
Or its a mix of both, with a projection mapped mask from a take with it, onto a CG mask to do the deformning.
This tool is very performance friendly - BUT what it does do is let you be a little more.... relaxed with your asset use. So while the plugin itself is very light for what its doing - it can lead to bad practices.
But I've found that with most of the really big recent advancements. If they are paired with good practices, you can get INSANE results.
I haven’t done any comparisons - but doing it in a shader has some very cool advantages - specifically with things like using masks to reveal the damaged areas as mesh bends and buckles, etc, all tied in with a single scalar.
You can still do that with blend shapes - but having it all in once spot just keeps things nice and easy to work with.
This is why I love unreal - there are 50 ways to do anything - and none are right or wrong. They just ‘are.’
Not sure if this is a stupid question - but is there a reason why NANITE and traditional LOD's cant be run concurrently?
Ie - have nanite streaming the geometry - but use traditional LOD's to load/unload the geometry as needed, with nanite handling the blending of the LOD's?
@LincolnMargison@WinGameDev The Max script lets you set it up with a morph target - but the info is written to the UVs.
https://t.co/SRYeJIkcB9
Its very cool - I've used it for things like pushing foliage out of the way, or even deforming/bending metal for damage states.
I think that WPO's are pretty cheap in general - but if you're doing some sort of casting to get the shape you're wrapping to, I imagine that could be pricey... but yeah - you could bake it easily enough.
There is a MAX script that lets you set up a morph target that you can control with a blend shape - so prebaking the vert positions with that, and then using a distance shader to actually warp it probably WOULD work.
It could be a fun experiment to try - sort of a pre-baked-LOD-Vertex-Wrap.
It sounds like the kind of thing I could get lost in, that would have no practical application beyond being a cool thing to try.. :D
That would be a pretty cool experiment to run. :D Use a WPO to reproject the geometry down onto each lower LOD - warping it based on distance to avoid all poping.
But yeah - I imagine the overhead would be... well.. 'just use nanite' levels.
Honestly this is more of a though experiment than anything practical. Sort of like trying to get as many textures and models as possible in a single draw call. You CAN do it - but at a certain point the overhead and complexity just doesn't really make sense.
Honestly it was more just me wondering outloud.
I know for nanite you have to hold the whole mesh in memory - whereas with LOD's you effectively hotswap em out.
But in retrospect - with LOD - you still gotta hold em in memory. :D I sort of assumed that you would keep the 2 on either side at a time - so lower and higher - and juggle them. But with nanite you only hold the highest - and then it decimates them lower - but the highest is always being loaded in.
I still gotta do a proper deep dive into the tech behind nanite. I understand it on a super surface level - but wanna go further!
@WinGameDev Ah yeah I see what you mean. You'd have to have some sort of parity of point to point in the chain where its almost 'morphing' the vertexes into place to avoid popping, which would also destroy and UV Coordinates and other data per model.
If you can’t use masks - you’ve gotta adjust the geometry.
Could be done in 2 ways dynamically.
1 with a WPO - attach it to a scalar, and ‘suck’
It in when not needed - and push it out when needed.
You’re already moving the vertexes around so the hit would be negligible, and you can control
It quite easily.
2 - you could try to use a pixel depth offset with a mask. It can be heavy with lots on the screen - so profile where you can - but I’ve used it successfully in the past to ‘nudge’ stuff away a little.