Sunday School - Sermon 1 - An exploration of Collision Checking in GameMaker

[@PixelatedPope]

This video was commissioned by a client for their purposes. If you found it useful, yay! If you didn't find it useful, I don't really care since you didn't pay to have it made. Feel free to leave a comment below telling me what you didn't like about it though, it will help with my engagement numbers.

[@Stefan-bn4wh]

great thx for sharing!

[@DarkKnightDad]

Honesty is part of the commandments 😅👍🏾

[@soreyJr]

Glad to see you back!

[@Mattstov]

The first time I’ve actively loved Sunday school!

[@NanoGamingGamer]

Wow welcome back! It's been a long time.

[@bradleebowers8203]

Stoked to see something new from you!

[@martinperon4576]

38:00 man, that is exactly what I needed! Thank God for the PixelatedPope!

[@scotsparaman]

Man… you’re back… 🙌🏼

[@sunguh]

nice to see you are back

[@HlootooThunderhammer]

And on the third year year did Pixelated Pope return! Rejoice!

[@xotmatrix]

Very interesting. I learned some new techniques.

[@VamosFener]

Great !! better then school

[@Evanski]

finally, collision code explained

[@NickyNoo]

Omg the legend is back

[@CoffeeAI201]

Awesome!

[@Poetax]

Just a question. On pixel perfect collisions, if a donut shape was used against a small enough pixel shape that would fit inside it without touching the donut, would it count as a collision or not. Is the collision just the outside shape?

[@PixelatedPope]

You mean if a pixel was "in the donut hole"? That would not be a collision if precise collisions are enabled for the donut sprite.

[@TheRealZen00]

The new collision system is fundamentally broken. If I have a 16 pixel square, and I set its collision mask to be pixel precise, then check for a point collision at x = 15.4, a hit will register, but if I test for a collision at x = 15.6, no collision will register. So it's pretty much unusable for subpixel world movement (using the tile/object combo system you devised)

[@PixelatedPope]

My latest top down movement and collision system uses a whole number locked coordinate system, and then draws the object "lerped" to create sub pixel motion. I'm never going back to non-whole number coordinates. You can check out the demo (both precise tiles and object based walls) over on itch: pixelatedpope.itch.io/tdmc

[@TheRealZen00]

@@PixelatedPope I broke the demo somehow, got it to tank the FPS when you move straight/diagonally into a flat surface. Otherwise it looks pretty smooth now.

[@TheRealZen00]

@@PixelatedPope Though yeah, that was my eventual decision, non-pixel perfect positions are just a hassle. I did leave the subpixel code in my game, but since I set the pixel accuracy to 1, it essentially turns it off.

[@JaguarPanda]

nice, ty, do you have plans to explain tilemap collsions too?

[@PixelatedPope]

No current plan. I do have a article on the Yoyo Tech blog that covers it pretty well. gamemaker.io/en/blog/precise-tile-collisions-by-pixelated-pope

[@JaguarPanda]

@@PixelatedPope Thanks for the link, i had a look at your blog post, it works pretty good for detecting collisions, but what i'm missing and cannot find anywhere, is how to resolve them (it's pretty easy for full tiles but it's a nightmare when it's about slopes and what not).

[@PixelatedPope]

@@JaguarPanda Yeah, what you are looking for at that point is any tutorial that allows for slopes. Then you can replace their "place meeting" with your "tile meeting" and it should work

[@alessandromangiapia7082]

If it’s from pixelated pipe it will be interesting guarantee. Hit like before even seeing it

[@ronbriscoe888]

Video was good. I liked it. but if you didn't want to use a list to do the collision to detroy all the bullets you could use a while loop and just keep executing the while loop until the collision function returned noone. (just an observation).

[@PixelatedPope]

While true, it is usually safe to assume that if there is a built in way to do something, it is going to be significantly faster than doing it in GML yourself. So when possible, I defer to built in functions. That being said, I didn't like benchmark it or anything, so there may be little to no difference, and that would be a perfectly acceptable method

[@flaviopsilva_]

Nice!

[@hakuzosionnach]

3 years in the making.

[@timthaman5778]

Have you returned to us?

[@PixelatedPope]

Seems so for now 😅

[@overidon]

FIRST!