Nostalgia flashback. Anyone else implemented a renderer based on fatmap2.txt[1]?
I came up with my own approach using bresenham and storing spans, but it was slow and sucked.
Then my buddy found fatmap2.txt on a BBS and gave it to me, as I didn't have a modem at the time. It was a revelation.
Programming in Turbo Pascal I was hampered by it being 16bit, but discovered I could prepend the assembly opcodes with 66h to turn them into 32bit instructions.
Later I tried converting the code to C, being only 300 lines and at that point well understood I figured it was a good task to get started with C. I was self-taught at programming but had read enough C to feel confident enough to try.
Converted it all line by line to C before compiling... and got literally over 3000 errors and warnings, and none of them made any sense.
Discouraged I left it, but a week later I was determined to get it working. After a few hours staring and bashing my head against the wall, I saw that the constant for the texture width which I had converted to a #define in C, had a semi-colon at the end...
So I removed it, and voila it complied flawlessly...
Only later did I realize my C compiler, djgpp, was actually a C++ compiler and that's probably why I got so many seemingly weird errors.
I had tough times with graphics under DOS. BorlandC C had a limited graphics library called Turtle. So I used DOS mode 13H for fancy 320x200 display with 256 colors.
Later, somebody gave me the source code for some graphics libraries together with some drivers and I could use 800x600 resolution with millions of colors.
After that I went to Windows and Linux and graphics APIs quit being an issue.
Yes I read fatmap. Writing software rasterizers was what got me into coding when I was a teenager.
Quite honestly, it was a great way to learn: algorithms, hardware, assembly, software architecture, all had to be taken care of to get a usable 3D engine.
I don't think I ever came across fatmap2.txt, but I do miss the good ol' times. I think my reference was 3DGPL, unfortunately can't find a live link to that.
Took a quick look at fatmap2 and surprisingly it doesn't do perspective-correct texture mapping (although it points to some other document in the introduction).
> I think my reference was 3DGPL, unfortunately can't find a live link to that.
This one[1], code being here[2]?
> surprisingly it doesn't do perspective-correct texture mapping
Yeah it was covered in some other document, can't recall which. I have some memory of fatmap3, but I can't find any references to it, and any files I had were on my IBM Deathstar[3]...
If you like this, there is also the 500 lines of C++ TinyRenderer, loading Obj files, texture mapping, clipping, and vertex/pixel shaders on CPU only: https://github.com/ssloy/tinyrenderer
I was super into this sort of thing until I hit triangle clipping concerns in homogeneous coordinate space.
Filling the triangle and spinning the cube are the perfect level of difficulty if you want to get your hands a bit dirty and achieve that sense of accomplishment. Anything beyond this starts to feel like a very risky time-reward tradeoff.
I strongly recommend this course “3D Computer Graphics Programming”[1] from Gustavo Pezzi. It walks you through the creation of CPU rasterizer from scratch in C.
I am working through it right now and I enjoy it a lot.
By the way, does anyone know how to modify the standard perspective projection to have a spherical cap instead of the flat rectangle as its far clipping plane? Having the flat rectangle means that you can see further from the corner of your eye than in front of you which is definitely not how the human vision works, and there are some games where this bizarre behaviour is very noticeable: you can barely see some thing poking out of the fog before you, then you turn the camera to the left, and now the thing is visible well and clearly near the screen border. Turn camera back, and it goes back into the fog.
2. clipping is a GPU fixed function done on x,y,z axes separately, so it will always be a planar cut
3. the effect you're talking about can be entirely removed by setting the frustum's near plane to fit within your collision hull, e.g. for a spherical collider
function near_plane_meters(collider_radius_meters, width_px, height_px, focal_length_px)
return collider_radius_meters * focal_length_px / sqrt(0.25*width_px*width_px + 0.25*height_px*height_px + focal_length_px*focal_length_px)
end
A spherical cuttoff limit is also not how the human eye sees. Light can travel from any distance to reach your eye. The view frustrum is an optimization, and there will always be cases where it fails.
> Light can travel from any distance to reach your eye.
Yes, but it won't necessarily register so unless we're talking about insanely bright distant object that should be visible through fog from any distance and which also is not a part of the skybox, this particular problem practically never arises. The flat far plane, on the other hand, is glaringly obvious in e.g. Minecraft, or any Unity game with first-person view and fog.
But it might more accurately reflect human visual perception. I can’t think of any case where your peripheral vision will perceive things that you won’t perceive by looking directly at them.
I love looking at stuff like this, working in the GPU space has only ever renewed my ambitions to work on similar projects. The hardest thing I always ran into with the more optimized fill algorithms was working around the single pixel holes that appear when doing everything with integers.
Small nitpick though, there seems to be an issue with the source code views, the file names and the first line of the code are on the same line with no spacing. looks like it might be a static site generation issue since there aren't any nodes to separate the name and the code in the raw html.
Edit: turns out the issue I'm seeing is somehow due to Firefox, seems to work correctly in edge.
It looks like for some reason Firefox isn't being served the same source as Safari or Chrome. In those browsers, the filename is wrapped in <center>, <a>, and <tt> elements, and followed by <hr> <br>.
But in the version of the HTML that Firefox receives, all that is missing and the filename is just some text that ends up immediately in front of the first line of code.
I once wrote a tiny cpu rasterizer in dlang. It is more efficient than what is shown here, since it uses only integer arithmetic (based on bresenham's algorithm). E.g. for a triangle it outputs the points in the triangle without ever considering a single pixel outside the triangle.
It's based on ranges (similiar to the ones added to C++) and the concept works very well imo. Could be a good exercise to translate it to C++23 (using coroutines with the range "generator").
Despite seeing code early in the series, didn't see any code that would actually put pixels on a screen, or even to write a static image into a file, at least not within the first 6 or 7 units. Somewhere around that point there is some reference to OpenGL, so presumably you could start to brighten patches of screen at that point?
Offering this as constructive criticism - the tutorials could be more engaging if the student could follow along and cause things to happen directly on the screen. It'd be awesome to organize the 1st unit as more than just confirming one has a compiler, to take it all the way to a kind of "hello, pixels".
Readit News
I came up with my own approach using bresenham and storing spans, but it was slow and sucked.
Then my buddy found fatmap2.txt on a BBS and gave it to me, as I didn't have a modem at the time. It was a revelation.
Programming in Turbo Pascal I was hampered by it being 16bit, but discovered I could prepend the assembly opcodes with 66h to turn them into 32bit instructions.
Later I tried converting the code to C, being only 300 lines and at that point well understood I figured it was a good task to get started with C. I was self-taught at programming but had read enough C to feel confident enough to try.
Converted it all line by line to C before compiling... and got literally over 3000 errors and warnings, and none of them made any sense.
Discouraged I left it, but a week later I was determined to get it working. After a few hours staring and bashing my head against the wall, I saw that the constant for the texture width which I had converted to a #define in C, had a semi-colon at the end...
So I removed it, and voila it complied flawlessly...
Only later did I realize my C compiler, djgpp, was actually a C++ compiler and that's probably why I got so many seemingly weird errors.
Anyway, good times...
[1]: https://github.com/rcoscali/ftke/blob/master/ogles/doc/fatma...
Later, somebody gave me the source code for some graphics libraries together with some drivers and I could use 800x600 resolution with millions of colors.
After that I went to Windows and Linux and graphics APIs quit being an issue.
Took a quick look at fatmap2 and surprisingly it doesn't do perspective-correct texture mapping (although it points to some other document in the introduction).
Edit: it was actually possible at the time.
This one[1], code being here[2]?
> surprisingly it doesn't do perspective-correct texture mapping
Yeah it was covered in some other document, can't recall which. I have some memory of fatmap3, but I can't find any references to it, and any files I had were on my IBM Deathstar[3]...
[1]: https://www.gamers.org/dEngine/rsc/
[2]: https://www.gamers.org/dEngine/rsc/3dgpl/code/
[3]: https://en.wikipedia.org/wiki/Deskstar#IBM_Deskstar_75GXP_fa...
[0] https://www.modeemi.fi/drdoom/3dica/3dica.htm
Filling the triangle and spinning the cube are the perfect level of difficulty if you want to get your hands a bit dirty and achieve that sense of accomplishment. Anything beyond this starts to feel like a very risky time-reward tradeoff.
[1] https://pikuma.com/courses/learn-3d-computer-graphics-progra...
2. clipping is a GPU fixed function done on x,y,z axes separately, so it will always be a planar cut
3. the effect you're talking about can be entirely removed by setting the frustum's near plane to fit within your collision hull, e.g. for a spherical collider
Yes, but it won't necessarily register so unless we're talking about insanely bright distant object that should be visible through fog from any distance and which also is not a part of the skybox, this particular problem practically never arises. The flat far plane, on the other hand, is glaringly obvious in e.g. Minecraft, or any Unity game with first-person view and fog.
Small nitpick though, there seems to be an issue with the source code views, the file names and the first line of the code are on the same line with no spacing. looks like it might be a static site generation issue since there aren't any nodes to separate the name and the code in the raw html.
Edit: turns out the issue I'm seeing is somehow due to Firefox, seems to work correctly in edge.
But in the version of the HTML that Firefox receives, all that is missing and the filename is just some text that ends up immediately in front of the first line of code.
It's based on ranges (similiar to the ones added to C++) and the concept works very well imo. Could be a good exercise to translate it to C++23 (using coroutines with the range "generator").
https://github.com/the5avage/shapes
Offering this as constructive criticism - the tutorials could be more engaging if the student could follow along and cause things to happen directly on the screen. It'd be awesome to organize the 1st unit as more than just confirming one has a compiler, to take it all the way to a kind of "hello, pixels".