While in 3D it is relatively easy to create millions of characters that all look different. I was wandering if there was a relatively easy way to do similar in 2D.
My first idea is one I've seen discussed before that is creating images that can be connected together to replace various parts of a combined image.
That actually takes a hell of a lot of work. You would also be forced to draw any clothing to then fit all the various body parts and so on.
However, I did come up with a some what easier solution last night. That is SVG and transforming the image. You can create SVG images and the parts are all separate and coordinates can be modified on the fly. That means you could also simply adapt the clothing to the same coordinates and make it fit rather easily.
The SVG code looks like this. As you can see the coordinates are easy to distinguish.
Programs like Krita and others make creating SVG files quite simple.
Rendering an SVG isn't that difficult.
All elements are stored in the file in the order of the painters algorithm. That means what is in back is rendered first. Each new element after moves closer to the viewer.
Example:
If you had a female character you would have her basic skin tone and then shading and lighting on top of that.
You can also fill with patterns / images to add further detail.
By doing this I could then create one set of images that I can then modify to create many different looking characters.
There are already tools that can convert a 2D image to SVG however the results aren't exactly what we need for this type of project.
I was thinking about how hard it would be to create software that could break the image into sections that would be grouped or labeled for different body parts.
I could then identify areas as eyes, nears, nose mouth, hair, breasts, chest, stomach, butt.... and so on. After doing that I would have a set of data I could simply modify for all the different images.
While this is still a good bit more work that using a 3D model and giving it modifiers it's doable vs trying to create the crap load of images if they were all hand drawn.
It would allow for a library for generating characters that could be used in all sorts of games or stories.
The reason I thought of this was the games I played that are hand drawn or 2D images generally don't have lots of different positions in them. They tend to get reused over and over again just on top of a different back ground. In most cases if you played 1/3 of the game you have seen most of the various types of sex scenes you will see. All that really changes is the background after that point. There are some exceptions to this but by far that seems to be the way most 2D games are.
Having a library like this the artist could further customize the work to meet what his desired style. It would give them a potential base to work from if they chose to use it.
It could also be built into a simple software where someone tells it what type of character they want via a single image they can manipulate which then uses that info to assemble all the other images for that character.
That all said if someone wanted to do this themselves they could by either creating their own software or using something like Krita and create the individual parts for the image separately so they can be assembled later. It would be a good bit of work though
My first idea is one I've seen discussed before that is creating images that can be connected together to replace various parts of a combined image.
That actually takes a hell of a lot of work. You would also be forced to draw any clothing to then fit all the various body parts and so on.
However, I did come up with a some what easier solution last night. That is SVG and transforming the image. You can create SVG images and the parts are all separate and coordinates can be modified on the fly. That means you could also simply adapt the clothing to the same coordinates and make it fit rather easily.
The SVG code looks like this. As you can see the coordinates are easy to distinguish.
SVG:
<svg xmlns="http://www.w3.org/2000/svg" width="500" height="300">
<path d="M480 150 C 300 200, 200 150, 20 200 C 15 150, 15 150, 20 100 C 200 100, 300 150, 480 150 Z" stroke="black" fill="red"></path>
</svg>Rendering an SVG isn't that difficult.
All elements are stored in the file in the order of the painters algorithm. That means what is in back is rendered first. Each new element after moves closer to the viewer.
Example:
If you had a female character you would have her basic skin tone and then shading and lighting on top of that.
You can also fill with patterns / images to add further detail.
By doing this I could then create one set of images that I can then modify to create many different looking characters.
There are already tools that can convert a 2D image to SVG however the results aren't exactly what we need for this type of project.
I was thinking about how hard it would be to create software that could break the image into sections that would be grouped or labeled for different body parts.
I could then identify areas as eyes, nears, nose mouth, hair, breasts, chest, stomach, butt.... and so on. After doing that I would have a set of data I could simply modify for all the different images.
While this is still a good bit more work that using a 3D model and giving it modifiers it's doable vs trying to create the crap load of images if they were all hand drawn.
It would allow for a library for generating characters that could be used in all sorts of games or stories.
The reason I thought of this was the games I played that are hand drawn or 2D images generally don't have lots of different positions in them. They tend to get reused over and over again just on top of a different back ground. In most cases if you played 1/3 of the game you have seen most of the various types of sex scenes you will see. All that really changes is the background after that point. There are some exceptions to this but by far that seems to be the way most 2D games are.
Having a library like this the artist could further customize the work to meet what his desired style. It would give them a potential base to work from if they chose to use it.
It could also be built into a simple software where someone tells it what type of character they want via a single image they can manipulate which then uses that info to assemble all the other images for that character.
That all said if someone wanted to do this themselves they could by either creating their own software or using something like Krita and create the individual parts for the image separately so they can be assembled later. It would be a good bit of work though