What I would like is to see a game where the player can recreate a reasonable approximation of their own face in game. To achieve this the game should provide 250 base head models for the humans.
That is the number they need to provide sufficient diversity and appeal so that all bases and both sexes are covered. Once achieved, the different fantasy races should key off those human face models with the various fantasy racial abstractions morphs for whatever features tend to get played up for that fantasy race. So basically you select a human head, and then when you change to dwarf or elf it doesn't present you with a totally new head, but instead the Dwarf or Elf version of the face you have currently selected.
I don't think any game exists right now, certainly not in D&D crpgs, that has achieved this. It's pretty straightforward though. Basically they just need to cast a wide net and scan until you hit the 250 in variety. Its a small enough set that the player can still navigate (especially if organized in columns and rows, rather than cycling 1 at a time.) Essentially they need Mombi's cabinet of many heads. They could totally do it, and that's the sort of content they should push out as soon as it's available.
Unlike environments, custom character content is a bit different. The more variety we have during EA the more likely it is that each playthrough feels unique, even if the actual gameplay content is still fairly limited. It feels new, when the PC feels new. I'd put new faces and voices and choice of clothing color at the top of the list for any new content drops. I think its more important honestly than new areas to explore. If EA is mainly about systems testing and gameplay content is necessarily limited by their repeated statement that we'll only get act 1, then having lot of char customization on offer would at least incentivize people to keep playing with new characters.
250 is a lowball number too, but it covers most of what we'd need as proof of concept, the stronger number pushes about twice that at 500, but 250 is a solid spread. Basically the core morphology of the skull divided into rows and columns, not surface cosmetic stuff like skin tone or freckles, but the major markers. General shape of the face round, square, oblong, kite etc which is basically jawline relative to cheekbones and forehead plane, shape of the eye orbits (square to round), height of the nose bridge, width of the nasal cavity etc. There are quick references one can scan of famous film actors that show the kind of diversity that exists, but with that number you can cover most of everything that we'd tend to describe as general diversity of morphology, beyond just shorthands for racial diversity. The helpful thing is that most of the major markers are related to each other and have to do with skull shape. Simple example would be stuff like eye and browshape relating to the height of the nose bridge and shape of the orbits, stuff like that. Just hire like an anthropologist or a bone doc, or someone with a background in casting that likes to deep dive on it, but I think 250 is the magic number. Less than that and there's just not enough, and huge swaths will get overlooked. Once you hit 250 and up things drift into the "you look just like 'so and so'" territory and its serviceable that way. I think they got the right idea by scanning, but they're doing it live and random whereas I think they should try to approach it a bit more focused in trying to cover the broader spread. I don't have the time tonight, but I'd like see every current head in Bg3 displayed in a single image at once. Then it would be easier to point at where there are gaps that need filling. But the total number right now is pretty low, even the larger mods I've seen top out at like 80. We really need 250
Last edited by Black_Elk; 29/10/21 05:59 AM.
