Target upload costs for rigged meshes?

[AcidBingo]

So I've got my model, and have optimizeded my LOD meshes (seen here) and now that I've managed to claw my assets and costs down to something I consider semi-reasonable I have to ask... what is considered reasonable? I know LI and associated costs aren't necessarily indicative of actual render costs, but what is the average in this instance or your experience?

Currently my mesh and LOD's sit at about L$35 to upload, L$10 of that is for the model base price from what I can tell. I did a couple tests on stuff: replacing my HQ model with my LOD2, filling 1 & 0 with custom meshes and then generating the lowest. That only dropped my cost down about L$10, which to me is negligible, and I honestly want to keep my HQ model if possible (it's my first born, I don't wanna kill it if I don't have to). I'd say it's got a high tri count when compared to other meshes, but is that actually true? Having seen some walkthroughs and workflows from some creators on youtube and elsewhere, some clothing items in the wild have bigger polycounts for their hq models than all of my meshes combined! In this forum I've seen people state that they've made relatively small clothing items with counts as high as 20k, granted those same people got super aggro on their LODs, but still? 20k?! Others I've seen using full size materials with high poly meshes and normal maps for elements that can only be seen under a microscope. While I know that isn't supposed to be the norm, it kinda is anyway?

Given that, I guess I'm asking, is my model within acceptable ranges? The HQ is twice the polys that the base character is, but again, when I look at all the other stuff out there that kinda leaves me scratching my head. As for my materials, they aren't going to be that complicated. 2 textures (color and spec, maybe even just color) per material x 3 materials.

Thanks!

[OptimoMaximo]

Yeah, most creators don't give a crap about optimization, especially when it comes to organic models. Avatar pieces (body, head, hands and feet) and clothing have massive polycounts most of the times, because that's what "the market wants": the average user wants a render mesh, where if the most minute detail is geometrically detailed, the better the creator (🤮) and the more money they throw at them. Same goes for clothing. Then those same users cry because SL is laggy and LL should do something about it. 

If your high lod model sports up to 90K triangles ALL INCLUDED, I'd say it would be ok. Of course the fewer triangles the better, but considering what infernal poly-monsters go around the grid, that polycount should be fair enough.

Edited December 21, 2022 by OptimoMaximo

[AcidBingo]

Much obliged good dragon!

Truth be told, some of the poly counts I've seen thrown around here make my head hurt, my toes curl, and my hair turn white(er). It's taken all of my self control not to necro some old threads on this here forum and tell these damn kids what for! (That and to stay the hell off my lawn.)

I cut my teeth working on an MMO many, many years ago (sadly it died a horrible death prior to completion), and a lot of the principles I took from that still stick around with me to this day. Some are still useful in all contexts (work with the absolute minimum dammit!), some, in today's market of 40-90k hero characters, well, some aren't as useful from a bleeding edge portfolio standpoint. Meh, it's still fun, and working within constraints can actually be beneficial both from a creative side and the "we don't wants no lag" standpoint respectively.

As an aside and a little story from my past that no one asked for, my lead once came to me and asked for a 16 pixel texture for an item, which, ok, fine, but this was not a teensy weensy asset that nobody would see, but on one of our main weapon models! I'm not sure if they actually needed it or if they were just f***ing with me for the lols, but I did it anyway. It was pixelated as hell, but it worked dammit!

[Kyrah Abattoir]

A lot of us have been in favor of a hard cap for resources on models, an option quite a few people have been firmly against it: out of a fear of some kind of "loss of freedom" but also because it would push some creators out.