#Tryng to find a check list to make something similar to john lin's graphics engine

23 messages · Page 1 of 1 (latest)

unborn elk
#

Nobody knows exactly how his tech works, he doesn't share the details

primal plank
#

that's realy sad

unborn elk
#

Meh it's his prerogative, if I invented something like that I would hold it pretty tight too I think

#

There is definitely some information about it around and if you search this discord you might see people discuss it but there's definitely not going to be a "checklist" of steps to make it

#

a checklist is not generally how you should be thinking about projects anyways

#

Tech like that is something you can build only if you have a very deep grasp of the fundamentals, it's not something you can cobble together from superficial understanding

#

Just start doing projects and experimenting and after a few years of experience you might be able to work your way in that direction

primal plank
#

the issue is that my math and physicis are not at the right level and won't be since math studies have ended for good in my batchelor

unborn elk
#

You don't need anything more than basic undergraduate linear algebra for all of graphics programming

#

Except for vector calculus for certain aspects of lighting (not really relevant to John Lin's project though)

#

You need strong computer science and engineering knowledge (frankly a prerequisite for all of graphics programming) and then an enormous amount of time and motivation to experiment and try (and fail) at doing ambitious projects

#

If you want to make really specialized or innovative tech you need to be way beyond the point of needing any outside guidance from tutorials or guides, that needs to be something you build up from your own knowledge

primal plank
#

starting whit DDA path tracing and learning the best graph algorithms sounds like a start

unborn elk
#

whatever interests you

#

ultimately that's all that matters

#

You'll never make yourself put in the thousands if not tens of thousands of hours it takes to get to that level if you don't enjoy it along the way

primal plank
#

I am interested in tiny voxels

#

for now I have thought of casting a ray from the camera to chunks of voxels on the world to see if wich should be visible and then cast it on the childs (voxels) of the visible chunks.

#

but I need to switch to vulkan

unborn elk
#

Why do you need Vulkan

primal plank
#

So I can control the pipeline to make it optimized for voxels

unborn elk
#

A voxel raytracing application is going to be bound by compute and memory bandwidth, not driver overhead

#

So there's not much that Vulkan will let you do that other APIs won't