Some clothes

At least, since I junked GTA4 10 minutes after I finished installing it, I had more time to work on tcod portraits… Here is the complete good happy family with clothes. Those versions are bigger than the in-game portraits. The guy still has those weird plastic hair. I’ll improve it later.

Since it’s rather difficult to find pictures of people in medieval clothes, I’m using several sources of inspiration. There are first a few on-line medieval clothes shops :
here and here

I’m also using a few very good comics with medieval settings :
Lanfeust of Troy
and François Bourgeon’s awesome Companions of the dusk

It’s very cool to take a little refreshing break, drawing a few new layers from time to time. The next steps are the emotion engine and more eyes/nose/mouth variations. I have first to figure out what set of eyebrows/eyes/mouth I need to represent the main emotions. I improved the engine so that it works directly with the full size layers (220×220). Once the final portrait has been created, the game uses a high quality supersampling technics to scale it down to the appropriate size (either 64×64 or 80×80). This feature has been recently added in libtcod svn version. It allows not only to work with the original pics without having to scale them by hand, but it makes it also possible to do a few new tricks, like randomly translating vertically the eyes / nose / mouth position just by a few pixels before the scaling operation to increase the portrait diversity. I’m also thinking about splitting the eyes into 3 layers : eyeball, iris+pupil, eyelid. This would allow to move the pupil inside the eye which would be great for the emotion engine.

4 comments so far

Add Your Comment
  1. Good looking bunch! Don’t forget the wretched hag and the demented prisoner 🙂 But for 90% of the characters in full-blown adventure I’m sure that portraits based on these will be enough!

    This gives me an idea: give each image a series of tags, and when creating a random portrait, you can request for a particular set of tags to be present. This can be used for emotions but also for broader categories like good/evil, rich/poor, pretty/ugly, armed/civilian… (I used dichotomies but there are actually no restrictions in form, if you get what I mean.)

  2. And now it gets complicated: different sentient races deserve to have portraits as well! Yes, I’m evil 🙂

  3. Mingos, currently, I’m not planing to have anything but humans and non-sentient creatures, so this won’t be an issue 🙂

    But the really touchy part is indeed how to filter all the random possibilities to get a portrait suiting some specific needs without completely killing the randomness. Tags might work, but some nose will look lovely when associated with a specific mouth, but ugly with another one, so I think it’s more complicated than this…

  4. [with Scottish accent]
    No dwarves, aye, laddie? Ye think us dwarves don’t deserve to enter ya game, doonya? ‘Tis bloody racism! Meebe a whack in the head with me axe here help ye changer yer mind?