First Hello! I looked for some kind of introductions forum for this part but didn't see anything but a general discussion. So hello everyone, been coming here for a while now but never had much need to post much. Some very amazing stuff on here and I can't wait to start putting some of my alterations of stuff up here as well. But for now I have an issue.
So I am trying to make an isometric tile set out of pre existing tiles / textures.
The source tile is 128 x 128. But I want it to be 64x32 for use in an isometric game.
I'm using Blender to do this, but I am getting an ungodly amount of resolution loss when rendering the image. So I am wondering if I am maybe going about doing it wrong. Up until now I had been using textures for these tiles that were large base sizes, and so it made a lot of sense to get so much quality loss.
My process is from what I can tell really standard. I set up a template project that has my blank square ready with it's base material on it. This allows me to simply change the image being used as the material to quickly render change the tile image, then render it at 64x32 (I actually quickly do a few different sizes so I have a range of tile sizes for using things in different ways.).
I'm using an camera set up for isometric capturing of the image.
Am I missing a step somewhere. For the most part I've been using texures that I had seperately made in Blender using bump maps and cycle rendering to get really good detail and color out of them. But this one in particular is a hand drawn tile that is set up for normal top down, and all I want to do is put it onto my tile template and make it isometric. I didn't think that there would be so much loss of detail when the image is already small to start.
Edit: I just noticed it's also making the image A LOT darker even if I turn the brightness way up on my lamp. (Using a Sun setting atm)
I think that's to be expected. If you just shrank something from 128x128 (16,384 pixels) to 64x32 (2,048 pixels) that'd be an 87.5% loss of information. But in an isometric projection it cuts the four corners of that grid off and has to fit everything in that little angled diamond shape that remains. At the end you're looking at something like only 6% of the data you started out with.
That is still applicable, but I was starting at 128 pixels, so while it's still a good rouch estimate, there isn't nearly as much information to start with so I didn't think that it would be as much loss, but I found a way around it anyway, it's a bit tedious, but I get a WAY better result with much more control of the tile.