No replies to this topic

[Zonzai]

Real Girls will be leaving alpha soon™.  Real Girls 1.0 beta will be the first version that is actually good enough to be used throughout your whole game experience without any extra files.  It will include custom .msn and .sk files and vampire files as well as the necessary lip tint mask (courtesy of Bella).

But before we get there I need to modify the base skin texture some more.  Version 07 will be the last alpha version for this project.  There's a list of things that need work posted in the file description.  I'll keep at it in between doctor's visits and everything else.  I won't have an ETA but I can say that it will be sooner than I had originally expected.

In the meantime, I felt like maybe an explanation is in order.  I get a lot of questions and comments from knowledgeable and uneducated people alike. Most of the time I don't have the time to explain things and only post very brief replies.  And I think I come across as being short or uninterested in feedback. I hope I can explain why now.  This will be a huge wall of text, but I will try to keep it simple.

COMPLICATIONS OF PHOTO-COMPOSITE IMAGING
It's important to understand that this is a different sort of texture.  I am not painting it as one would a normal skin texture (or any texture).  Likewise, I am not taking a 2d image of a mostly flat object (like a brick wall or a dirt road) and applying it to a mostly flat object.  

Painting a convincing human skin texture is extremely complicated.  Painting a good one is an enormous task.  Stitching and layering images of human skin together to create a single convincing human skin texture is also extremely complicated.  But in addition to the challenges of painting a texture, it has it's own set of challenges.  

The most time-consuming part of creating a texture this way is finding the right image.  This isn't always as easy as you might think.  An image might look right at first glance but be cropped wrong, be too dark or washed out or maybe just out of focus.  I've often had to go back to images I didn't think were going to work out because of their smaller pixel size to find that they worked better because they retained more detail due to one of these issues. I also often find a better, more detailed texture for an area that I've already completed and have to go back in and add the new image on top of it. This is a constant process.  I think they call it iteration in the gaming business.

There are other issues specific to this type of texture.  Blending textures takes a lot of time and iteration if I want to retain clarity and not just have it look like a jumbled mess. Lightening and darkening areas where part of the image was lighter or darker than the rest (because it is a 3D object) is a constant battle.  The little painting that I do have to do has to be extremely high quality or it won't match the rest of the texture.  And then there is the issue of detail retention...

Brightness, contrast and sharpening are the three major adjustments I have to constantly toy with in order to translate the minutia of detail of skin in to a finished product.  A painted texture doesn't have the full range of detail in dark, mid and light tones that actual human skin does.  It is much, much more forgiving.  But for this texture, too light or too dark can be just as bad as being out of focus. I use curves to alleviate this somewhat.  But every piece of hobbled-together skin texture has to have it's own carefully adjusted curve, also tone matched to the existing texture.  Get it wrong and some areas look too blurry and others look too detailed (yes that can happen too).  

Take a look at the hips on v06.  You will see just to the inside of each hip an area that looks a bit plain and out of focus.  That is because the base image was slightly out of focus there and I didn't adjust the curves properly for that area.  Look at the wrinkles on the rib cage.  They look way too dark and almost dirty.  Again, that was because I didn't get the curve just right.  I needed to pick up on the bottom end of it and brighten the overall tone, then drop the middle of the curve back to where it was originally so that it retained the high-end detail but wasn't as dark or fine in the dark areas.  But because I've already blended that area I can't fix it without reworking the entire thing now.  Fortunately, I can lasso copy the flattened image in that area, feather it out and drop transparency to 80% or so and apply my sharpening filter only to the merged layer.  That nets me an overall decrease in the appearance of dirt (compared to what it would be after sharpening) while retaining a portion of the sharpening, bringing the areas that aren't too sharp in-line with the areas that are.  Or at least, that's the theory.  

You may not have understood all of that but that's not important.  What matters is that, that explanation is actually oversimplified.  There is a lot of complexity to adjusting a curve or adjusting the final filtering.  And I have a long history with Photoshop.  I know how to get the most out the filters the tools that I use.  For example, when I run the unsharpmask filter I blow my images to about 10x their actual size first because the softening enlargement filter breaks down each pixel into 10 new pixels blended to better match their surrounding pixels.  That way when I run the sharpening filter I get more natural-looking sharpening and less jagged edges.  But then that takes even more time.

Despite my long history with Photoshop, I am also a modding noob.  Skyrim is my first modding experience, AOM was my first mod.  So I still have a lot to learn outside of my limited area of expertise.  (Which has been fun but also challenging.)  And that has cost me some time that I would have not had to spend had I known what I was doing to begin with.  It has even led to some pretty large mistakes on my part.  I don't know how much time I spent trying to perfect the seams with a version of Mudbox that didn't support 8192 texture files.  That cost v06 a lot of time.  But my biggest mistake was trying to do everything in Photoshop to begin with because I was too intimidated to try using Mudbox.  So I guess I shouldn't complain.  

I've also had to unlearn a lot of Photoshop habits that don't work for textures.  But this is getting a bit long so I'll skip that explanation.  And then there is the issue of my car accident.  Even though it gave me time to work on Real Girls, it is going to be a time-sink for me for the next few months too.

TLDR
If there is such a thing for a post like this, it would be time.  It takes a lot of time to make a texture this way.  Which is probably why nobody else does it (or at least does it as detailed).  That's also why I don't reply with very detailed responses and ignore a lot of PM's.  I'm sure this texture seems much more simple when you are looking at it in game.  But please just trust that I read your PM's and posts and - because I want RG to be the best that it can be - do take them into consideration... even if only for a moment.  

Anyway, you guys are great.  If I come off as short-tempered it is probably just because I don't want to (or more likely can't reasonably) explain in detail the complications that have lead me to do something as I have.  Please be patient with me waiting for v07 and 1.0.  I won't stop working on this texture (again) until I am reasonably satisfied with it.  I'm also having personal issues so that is going to set me back a bit.  But I do know what I am doing (at least in Photoshop) and I do have enough time to dedicate to this project to get it done.  

I do still need your feedback though.  Please don't stop giving me feedback on the texture; even bad feedback.  Just understand that I probably won't reply with a detailed explanation because there are a lot of things I have to consider every time I make a change or add a new layer.  And if it seems like I'm brushing you off or ignoring you, I'm not really.  It's just difficult to explain everything and it takes free time I should spend working on the texture instead.

Thanks for reading, expect v07 soon™.