|
To be honest that sounds pretty realistic. Never take on a giant spider in a tavern brawl
|
#
?
Sep 7, 2014 23:08
|
|
|
|
| # ? Apr 25, 2024 09:47 |
|
|
Working on a problem set for our students. Welp. Guess the module is pretty specific about the BMP images it can work with.
|
|
#
?
Sep 12, 2014 17:51
|
|
|
Sebbe posted:Working on a problem set for our students.  Usually it's due to incorrect row padding or bpp.
|
|
#
?
Sep 12, 2014 18:21
|
|
|
Sebbe posted:Working on a problem set for our students. I recognize that glitch. Your stride calculation is wrong. It'll come out wrong like that if you picked the wrong pixel format for the decoder. For example, if you thought your pixel format was 32bpp ARGB but its actually 24bpp RGB, you'd overestimate the stride by 8 bits * image width. The error will propagate as you iterate the image top to bottom causing the slanted look.
|
|
#
?
Sep 12, 2014 18:40
|
|
|
I kind of like that bug because it reminds me of poor analog TV reception, or a cruddy VHS tape from the video store 
|
|
#
?
Sep 14, 2014 04:32
|
|
|
Sebbe posted:Working on a problem set for our students. 
|
|
#
?
Sep 14, 2014 15:27
|
|
|
 Playing around with physics in love2d and failing to attach the ball to the mouse movement correctly so every time I try to knock over the tower of blocks I miss. For comparison what I was actually trying to do: 
clockwork automaton fucked around with this message at 18:03 on Sep 21, 2014 |
|
#
?
Sep 21, 2014 17:59
|
|
|
Maybe not a glitch per se but I did some tinkering with an idea for "connect four IN SPACE" but it slowly dawned on me that it would be incredibly frustrating and fiddly.    Might work if I just did a bubble-bobble type thing instead of ~physics~.
|
|
#
?
Sep 21, 2014 18:10
|
|
|
My city generator might need some work 
|
|
#
?
Sep 28, 2014 17:19
|
|
|
ekuNNN, any explanation behind those giant floating eyes? 
|
|
#
?
Sep 30, 2014 17:45
|
|
|
I don't know, they just show up to silently judge me whenever I gently caress up my code
|
|
#
?
Oct 1, 2014 14:25
|
|
|
Don't worry about that, it's just the Observer Pattern.
|
|
#
?
Oct 1, 2014 14:36
|
|
|
LP0 ON FIRE posted:ekuNNN, any explanation behind those giant floating eyes? It took me a second to realize that that was the poster's name and not your reaction.
|
|
#
?
Oct 4, 2014 05:57
|
|
|
 chip8 rave party 
|
|
#
?
Oct 16, 2014 00:37
|
|
|
 I don't get iOS 8 Size Classes
|
|
#
?
Oct 26, 2014 17:51
|
|
|
Drastic Actions posted:
no no that's perfect
|
|
#
?
Oct 26, 2014 19:10
|
|
|
lossy compression
|
|
#
?
Nov 2, 2014 15:27
|
|
|
I just liked this little satanic glyph that would zoom all around the screen (supposed to be a circle)
|
|
#
?
Nov 13, 2014 18:43
|
|
|
That's really neat and you need to put that in the real game somehow. 
|
|
#
?
Nov 13, 2014 20:13
|
|
|
sarehu posted:lossy compression Intel's new architecture lookin good
|
|
#
?
Nov 30, 2014 13:03
|
|
|
 I...broke something...
|
|
#
?
Nov 30, 2014 17:41
|
|
|
Shoehead posted:
Looks to me like Skull Head Dude's position pointer is getting copied and used by some other aspect of the program. Either that or you have very specific memory corruption. Funky either way though.
|
|
#
?
Nov 30, 2014 20:05
|
|
|
TooMuchAbstraction posted:Looks to me like Skull Head Dude's position pointer is getting copied and used by some other aspect of the program. Either that or you have very specific memory corruption. Funky either way though. I *think* it's the flags for locking him in place unflagging too soon, or just plain not working. I wanna look at it but real life is calling atm. 
|
|
#
?
Nov 30, 2014 20:31
|
|
|
TooMuchAbstraction posted:Looks to me like Skull Head Dude's position pointer is getting copied and used by some other aspect of the program. Either that or you have very specific memory corruption. Funky either way though. This happened in a game I've been developing, and the seaweed starting behaving like an enemy. It was freaky.
|
|
#
?
Dec 1, 2014 16:45
|
|
|
LP0 ON FIRE posted:This happened in a game I've been developing, and the seaweed starting behaving like an enemy. It was freaky. I wish I had thought to record video of it at the time, but years ago I had a bit of code that loaded the editor logic inside the regular game (due to not cleaning up the regular game properly when switching to the editor), so you were able to control the edit cursor even though the game was running at the time. IIRC there was also a wild pointer or two that made it possible. I remember being totally impressed that it managed to line the heap up just right that it worked, because there were a lot of differences in the 'state' region of memory between the two modes. It wasn't quite perfect since it would aggressively crash when enemies noticed and tried to attack the edit cursor. Seat Safety Switch fucked around with this message at 05:14 on Dec 4, 2014 |
|
#
?
Dec 4, 2014 05:10
|
|
|
I decided to dive back into Blender and re-learn how to do animations. How do armatures work again? Probably not like this. 
|
|
#
?
Dec 7, 2014 02:15
|
|
|
TooMuchAbstraction posted:I decided to dive back into Blender and re-learn how to do animations. those weights  here's one of my own, i call it "the angry boulder test" 
|
|
#
?
Dec 7, 2014 05:32
|
|
|
Forgot to initialize some data on the GPU.
|
|
#
?
Dec 9, 2014 03:54
|
|
|
NorthByNorthwest posted:Forgot to initialize some data on the GPU.
|
|
#
?
Dec 9, 2014 04:03
|
|
|
Zereth posted:That's a cool effect, can you do it reliably and safely? Since it's just reading uninitialized data, the patterns are highly variable, and might not come out the way you'd like, If you need quick and dirty noise, it might work. I'd recommend sampling from a noisy texture, you'll have finer control over the base color and the style of glitches. If you're going for fake glitches, you'll probably get more fun by 'forgetting' to clear the buffer in between draws. A few other experiments: Normalizing an uninitialized vec3:  Multiplying the above by a sinusoidal time:  edit: Passing uninitialized data per vertex: 
NorthByNorthwest fucked around with this message at 05:06 on Dec 9, 2014 |
|
#
?
Dec 9, 2014 05:02
|
|
|
 From an assignment I had during uni. The goal was to create an algorithm for interpolation of pixels. Top-most picture is the source.
|
|
#
?
Dec 9, 2014 08:19
|
|
|
NorthByNorthwest posted:Since it's just reading uninitialized data, the patterns are highly variable, and might not come out the way you'd like, If you need quick and dirty noise, it might work. I'd recommend sampling from a noisy texture, you'll have finer control over the base color and the style of glitches. And just like that, Homestuck the Video game was created on accident.
|
|
#
?
Dec 9, 2014 14:20
|
|
|
Lowen SoDium posted:And just like that, Homestuck the Video game was created on accident.
|
|
#
?
Dec 9, 2014 14:58
|
|
|
Simple, but it got a big laugh out of me: https://vine.co/v/O66QLquDj0M  It's been a long day.
|
|
#
?
Dec 13, 2014 00:49
|
|
|
Oops
|
|
#
?
Dec 21, 2014 05:42
|
|
|
I was trying to bend traintracks      Edit: In progress fixing it to work without deformation in the last gif, but something is still Very Wrong  Edit2: Bah, still can't get it 
Jewel fucked around with this message at 01:06 on Dec 22, 2014 |
|
#
?
Dec 21, 2014 11:00
|
|
|
Shoehead posted:Oops I'm not sure I see a problem with this.
|
|
#
?
Dec 22, 2014 05:45
|
|
|
Jewel posted:
Did you manage to get this working? The first one looks good, but it feels like you'd want to just calculate the midline of the sleeper and then rotate the shape to position it Unless you're doing some interesting procedural calculation thing to generate these or something 
|
|
#
?
Dec 22, 2014 15:15
|
|
|
Going to bed so I'll paste what I wrote up on my blog and get back to any further questions later!quote:There’s some deformation in between transitions, but any other methods I try have far worse. This method was done using polar coordinates, where the angle was the x coord (0 to 1 space) converted to 0 -> 90 degrees, and the length was the y coord. From there the new x or y coords is just cos or sin(angle) * length. But yes, this is entirely just procedural on point data, not necessarily any special "tracks" or "sleepers". That's one of the reasons I wouldn't really use this in anything too, it's more efficient to just model it because you get cleaner non-tapered sleepers.
|
|
#
?
Dec 22, 2014 15:23
|
|
|
|
| # ? Apr 25, 2024 09:47 |
|
|
I helped Jewel a lot with this, so let me explain what the three gifs are. The first is a simple lerp between the fully bent and unbent shapes. Second was after I gave her some random suggestions and she somehow ended up with that. She didn't save the code for it. The last is the mathematically correct "rotate point N degrees around the center point", where N ranges from 0 to 90 based on the X of the input point. This sort of procedural bending modelling is clearly a solved problem, there are tools to do the modelling, but I haven't been able to find any papers or research on the subject, and I'm too dumb to think of a proper solution on my own.
|
|
#
?
Dec 22, 2014 19:36
|
|
