human-gpu | Hello human, I'm sick to be your GPU!! | Graphics library

I've run into an issue while attempting to use SSBOs as follows:

I've come across this problem many times over the years and still live in hope that there is an easy way to do this that I have missed. I work with barcodes a lot. They are usually made of black dots or lines on a white background. Barcode readers generally work faster and more accurately when the edges are crisp and then size of the lines or dots are precise.

Most barcode generation algorithms will give you a compact barcode usually with the smallest element size being one pixel. A typical QR code could fit in a 21 x 21 grid. This would be too small to see if printed pixel to pixel on most printers and would typically be scaled up. The result of scaling it up depends on the method used and although sometimes you are given a choice, often you have no options that make the image suitable. Even printing directly will often give you expected gray artefacts or forms of dithering. The most consistent way I have found is to scale the images before they are use daily in other places such as Microsoft Word, lightburn and a few others I use that still give me a headache.

Below I will go through what I have tried and show the results. I am limiting this to bitmaps only because using vectors here is not something I need on my current project.

My current best resolution is not pretty, it is slow and although I could improve the speed by locking the bits in the bitmap, I am hoping someone has a really simple answer that I had totally missed on my search again this time.

Here is an image of a simple QR code blown up in GIMP.

The problem is, if it is scaled up, it'll often end up looking like this:

Below I created a small test program to go through all the different modes I know of and then generate a matrix of images which I have reproduced below. The version I currently use is Mode 99 which involves inspecting each pixel and drawing a square.

Does anyone have any better ideas?

I'm trying to wrap my head around calculating motion vectors (also called velocity buffer). I found this tutorial, but I'm not satisfied with explanations of how motion vector are calculated. Here is the code:

I have trouble to grasp how to use colors in CGA/EGA/VGA video graphics modes. The video modes I'm particularly interested in are 0Dh (EGA 320x200) and 12h (VGA 640x480). Both of these modes have 4 planes, thus 16 colors.

My (probably incorrect) understanding is that I should activate a set of planes by writing a bitmask to port 03C4h, then when I write to video memory, the data only gets written to the activated planes. Mostly I used this document to get my information, though I also encountered several other tutorials and discussions: http://www.techhelpmanual.com/89-video_memory_layouts.html

Now I'm trying to write pixels in all possible colors in the first word in the video memory (top left part of screen). I load 1 for the initial bitmask to AH and 1 bit to BX. Then in a loop, I increment AH and shift (SHL) the bit in BX to hit a different pixel next time. I OR BX to A000h:0000h to add each pixels by leaving the already existing pixels untouched.

What I'm expected to see is a line of pixels in all possible 16 EGA colors on the top left of the screen. What I actually see is 7 white and 1 bright yellow dots with black pixels in between them. What am I doing wrong?

Also, every tutorial says that I must write 0005h to port 03CEh before I start to use planes. What is the purpose of that? When I comment those lines out, I can still use planes (I mean, in other programs). Previously I had success using planes when I was writing to different words in video memory (so I didn't need different color pixels in one block of 16 pixels that's represented by a single word in video memory); and when I used BIOS functions (e.g. INT 10h/AH=0Ch) to write pixels, but still I want to understand how to use planar graphics without BIOS, as I believe the BIOS functions are slow.

Here is my code (indentation is optimized for 8-width tabs, so it kind of looks off here):

I am a beginner who is trying to implement simple graphics in VBE. I have written the following assembly code to boot, enter 32-bit protected mode, and enter VBE mode 0x4117. (I was told that the output of [mode] OR 0x4000 would produce a version of the mode with a linear frame buffer, so I assumed that 0x0117 OR 0x4000 = 0x4117 should have a linear frame buffer.

Here is a gradient color legend I created using rasterImage:

I have a model like this

I'm drawing a graph with two points of each point having a line with a weight.

for example graph: point "15" to point "16" line with the weight of 1.872 and point "16" to point "15" with the weight of 1.567.

take a look at my graph for now:

I want to draw a String with always parallel (adjacent) to the line.

I calculated the slope for the straight and the angel I did calculate is the arctan of this slope:

I had use this function to rotate the string:

How can I open a window and draw to it with SkiaSharp (without using winforms, wpf or anything like that)? I've tried using SFML.Net instead of SKIA but it lacks a lot of features (rounded rectangle, shadow, gradient).