Raycaster | : feelsgood : A simple raycasting experiment

I tried to get the index of the particle I crossed with my mouse cursor while using raycaster. however, When I log intersects, it returns a millions of object even if I only came across one point.

I could access to the exact match only when I go with index zero of the intersects array.

And I have to pass the params.points.threshold to get the intersects.For test, I put 0.1 as a experiment. But I don't know the meaning of it quite clearly.

Does anyone know the reason?

My code is as below.

1. Setting raycaster and pass in threshold parameter.

   ...

I am working on my Raycaster engine for some time, that I am runing on slower machines. The most challenging problem I occures was/is the efficient floor and ceiling casting.

My question is: what other faster approached can I use? (I am not sure how Doom floors and ceilings are rendered)

So far I tried two typical solutions:

• verical and horizontal - casting as described in well know lodev tutorial: https://lodev.org/cgtutor/raycasting2.html

The horizontal approach is of course much faster, but I additionally optimized it with fixed point variables.

Unfortunately even that approach is a performance killer - quite big fps drop even on faster cpus, and an slower cpus its a bottle neck.

My other ideas:

• I figure out an algorithm that was converting visible floor/ceiling map tiles to quads that I splitted to two triangles - and rasterized them as in regular scanline rasterizers. It was much faster - also I could sorted tiles by texture id to be more cache friendly. Unfortunately I got into "perspective correction texture mapping" in that case - to fix this I must add some divisions, that will lower the performacnce.. but also there are some optimalizations that can be done..

• using horizontal casting with every 2 ray (in column, row or both) - i will fill the blank spaces with averaged texture coords

• I could also try to combine my algorithm from 1 point with horizontal casting - I could sort the textures by ID then for example, I think that there would be no texture distortions

• mode 7 ?

my progres so far: https://www.youtube.com/watch?v=u3zA2Wh0NB4

EDIT (1):

The Floor and Ceiling rednering code (based od lodev tutorial the horizontal approach) but optimized with fixed point. Ceil calculations are mirrored to floor.

https://lodev.org/cgtutor/raycasting2.html

This approach is faster than the vertical approach, butlots of calculations is inner loop and random accesing to texture pixels hits the performance..

I have created a three.js element to show a number of points on the screen and I have been tasked with clicking on two and calculating the distance between them. I am doing this in a Angular (8) app and have all the points visible and mouse events (pointerup/down) set up correctly. My idea is to ray trace from the mouse point when clicked and highlight a vertex (I do only have points no lines or faces). So I have attempted to set up Three.js RayTRacing on my scene but every time I call setFromCamera the camera is undefined even though I srill have the points visible on the screen at all times.

So I have an event signature like this:

My binary gltf file(modelled in blender and animated using mixamo) is not detecting on raycast. I read bunch of tutorials and questions about it to try to fix it but it does not work what so ever:(

I created a class that helps me to show a 3D model using A-Frame. In this class, there are some spheres created at runtime ad inserted into the scene. I'm trying to add an event listener (I have to show a message when those spheres are clicked)

Here is the code:

I have this piece of script that I made and I get this error and I don't understand why

I made a voxel raycaster in Unity using a compute shader and a texture. But at 1080p, it is limited to a view distance of only 100 at 30 fps. With no light bounces yet or anything, I am quite disappointed with this performance.

I tried learning Vulkan and the best tutorials are based on rasterization, and I guess all I really want to do is compute pixels in parallel on the GPU. I am familiar with CUDA and I've read that is sometimes used for rendering? Or is there a simple way of just computing pixels in parallel in Vulcan? I've already got a template Vulkan project that opens a blank window. I don't need to get any data back from the GPU just render straight to the screen after giving it data.

And with the code below would it be significantly faster in Vulkan as opposed to a Unity compute shader? It has A LOT of if/else statements in it which I have read is bad for GPUs but I can't think of any other way of writing it.

EDIT: I optimized it as much as I could but it's still pretty slow, like 30 fps at 1080p.

Here is the compute shader:

I have had problems with the OnDrop method, it is that it is not called, I was reading and maybe it has something to do with a Raycaster component, but I'm not sure, I don't even have knowledge of it, if someone could explain this to me I would greatly appreciate it.

Here is my code in c # plus an image of my hierarchy in Unity2D: