Greetings everyone, i am new around here. I came here to see if anyone could help me with making sprites, a full sprite sheet, etc.
As the title says, i want to be able to make 8-bit sprites from existing sprites, but are either 16 or 32-bit. Take for example these two images: https://imgur.com/a/zfutT
How do i "convert" or re-create them in 8-bit?
The issue I see with your question is...what defines 16/32-bit as far as sprites go? I'm guessing you are referring to the limitations some systems had, for example, the PS1 sprites are sometimes considered 32bit, SNES/Genesis as 16bit, and then NES as 8bit. If that is the case, it would be good to specify.
The NES sprites had limits in both the palette they could choose from, and the amount of colors that could actually be used for each sprite. If I remember correctly, the palette had around 56 colors(and was lacking in certain colors, like browns), and you could only have 3 colors in use for a single sprite(with the "4th" color being transparency). If you want to lower "bits" of more modern sprites, a good place to start is simply using less colors per sprite. You could also find or create a small palette to choose from for all the sprites in a game(or art piece or whatever). The other thing to consider is lowering the resolution, and a good resolution for "8bit" sprites would be 16x16 in my opinion.
The final catch with this...I don't think there is going to be a good automatic way to do this at all. You technically could lower resolution in an art program, but details will be lost and will need to be rebuilt(if the new resolution has space). And you could do some kind of color amount reduction, but I've never seen good results come from that either, at least not with the amount of colors under discussion.
The colors aren't much of an issue for the moment, i know how to handle that. what i do want to know is how to make the sprites look "smaller" in simple terms, so they can "fit" in a NES game. For example: using a variation of the 32 bits image i gave you earlier, take a look at this one: https://imgur.com/a/5WRTv
The image from the left is made for the game boy color (or the normal game boy? i don't know) but it has obviously less details than the image of the right, and is also smaller.
Sorry if those things i mentioned doesn't define what a 8/16/32 bits sprite is, as i don't know much on the matter.
EDIT: Also, credits goes for devianart user MegaRed225 for the image used on the left.
I still am much confused what problem you are trying to solve here. To eliminate the confusion, please only talk about the image, and not about systems that may display it.
An image has a size in pixels (in 2 directions), and each pixel has a colour, usually in 3 or 4 channels (normally RGB and an A channel for opaqness). Each channel has a bit depth (number of bits available to express the fraction of eg Red). Further limitations may be that only a limited number of unique RGB combinations may be used (eg gif has 256 colours available from the 24bit RGB space), or that the set of available RGB combinations is fixed (eg the system has already fixed the palette).
EDIT: For valid sizes, you may want to use cropped image sizes, eg not the large white area that you have in some of the examples, as that is non-relevant to the problem.
Since you mostly excluded reduction in colours from the question, I assume you are talking about a reduction in size. You can map several pixels from the source image onto one pixel in the destination, by blending their colours.
As an example, converting an image from 10x1 to 5x1 can be done by blending the colours of the first two pixels into the first pixel of the destination, the colours of the 3rd and 4th pixel into the 2nd pixel at the destination, and so on.
The reduction doesn't have to be a power of 2, if you go from 10 pixels to 6 pixels, each pixel in the destination gets 10/6th pixels from the source, using a weighted average, for example. Enlarging is the same process, except you'll very little colour information available.
None of these automated techniques will give you good quality images. The computer doesn't understand that one line has more importance than a filled area near to it, for example. In any case, you'll need to improve, or even redraw the images, if you want good quality.
7 hours ago, Alberth said:As an example, converting an image from 10x1 to 5x1 can be done by blending the colours of the first two pixels into the first pixel of the destination, the colours of the 3rd and 4th pixel into the 2nd pixel at the destination, and so on.
The reduction doesn't have to be a power of 2, if you go from 10 pixels to 6 pixels, each pixel in the destination gets 10/6th pixels from the source, using a weighted average, for example. Enlarging is the same process, except you'll very little colour information available.
First of all, sorry for all the confusion. I understand very little about this, so most if not all my explanations lacks the "definition" of what sprite making is...
As for your answer, Yes, i want to create them myself, as stated in my first post. If an automatic process gives a bad quality image, then how can i do all of that manually?
22 minutes ago, daniel hernandez said:As for your answer, Yes, i want to create them myself, as stated in my first post. If an automatic process gives a bad quality image, then how can i do all of that manually?
You'll have to draw them manually, using an image editor (e.g. paint.net -- there might be better suited pixel graphic editors I'm not familiar with).
Instead of starting each image from scratch, I would suggest loading up the original image and resizing it to the intended size, and saving that as your starting point. Using the resized image as your base and the original image as a reference, make changes until you're happy with the result.
For more general advice on creating any sprites, I would recommend searching online for sprite-making resources.
There are a few dedicated tools designed to make such a thing easier. For example, DavitMasia sells a nice little utility called PixaTool explicitly for this.
I see... Sorry for confusing matters.
So, i take it that this forum doesn't deal with art questions, then? because i have one more... But i'm beginning to think that it doesn't belong here.
There is a visual art forum, you may want to try it there.
While beginners forum is for anything, almost all questions here are related to programming, and most readers are programmers. Together with a confusing question about bith depths of images, we (in hindsight incorrectly) considered your question to be a programming question about image formats rather than "make it look like a retro image".
