|
loren schmidt
@
lorenschmidt
|
|
images, words, games, and other digital art. she / they. please support my work on patreon: patreon.com/vacuumflowers
|
|
|
57.966
Tweetovi
|
2.250
Pratim
|
19.945
Osobe koje vas prate
|
| Tweetovi |
|
loren schmidt
@lorenschmidt
|
7 h |
|
i think people forget that realism and depth of simulation are often at right angles to internal harmony
|
||
|
|
||
|
loren schmidt
@lorenschmidt
|
7 h |
|
eternal looks so colorful and tactically solid. it seems like it's even more up my alley than 2016 was. the increased emphasis on motion + embracing the game's artificiality are absolutely the right direction.
|
||
|
|
||
|
loren schmidt
@lorenschmidt
|
7 h |
|
trying to finish a nightmare run of doom 2016 by the time eternal comes out. it's so good. i really respect the risks they took with the design.
|
||
|
|
||
|
loren schmidt
@lorenschmidt
|
7 h |
|
i wouldn't be surprised, though i don't recall finding any images offhand
|
||
|
|
||
|
loren schmidt
@lorenschmidt
|
16 h |
|
being human is like horse racing except you bet on all of them at once and also you are all the horses
|
||
|
|
||
|
loren schmidt
@lorenschmidt
|
16 h |
|
you're right, i gotta do that and side clipping anyway. might as well let them help me out.
|
||
|
|
||
|
loren schmidt
@lorenschmidt
|
20 h |
|
i just came across a reference to this! i am trying to implement it now. that's really neat.
|
||
|
|
||
| loren schmidt proslijedio/la je tweet | ||
|
Sarafina Nance
@starstrickenSF
|
1. velj |
|
me, an astrophysicist who has published 2 (nearly 3!) papers on Betelgeuse and specializes in supernovae: *tweets about Betelgeuse*
men in my mentions: “well actually it may have already exploded do you know how light works here’s the Wikipedia link also calm down”
|
||
|
|
||
|
loren schmidt
@lorenschmidt
|
20 h |
|
i'm writing a software renderer on top of SFML
|
||
|
|
||
|
loren schmidt
@lorenschmidt
|
21 h |
|
thank you! i think there are probably different matrices for this which all work (though maybe i have the wrong one for my other components?) trying yours seems to give me the same symptoms, however. squashing along z axis, weird near plane blowout.
|
||
|
|
||
|
loren schmidt
@lorenschmidt
|
21 h |
|
yeah i don't have near plane clipping yet. that is my suspicion as well.
i also suspect there may be some kind of weird z-axis related thing happening. it seems like everything is squashed down to z = 0 (or very close) by applying the perspective matrix.
|
||
|
|
||
|
loren schmidt
@lorenschmidt
|
21 h |
|
i thought you couldn't do shift from (x, y, z) to (x + dx, y + dy, z + dz) using a matrix multiply, and you needed to do an addition in between applying the object scale, rotation matrices and the camera matrices
|
||
|
|
||
|
loren schmidt
@lorenschmidt
|
21 h |
|
wait how do you combine your model matrix into the same set of operations? as far as i can tell you need to apply that, then add to get your model transform to work, then apply the camera matrix to the result.
|
||
|
|
||
|
loren schmidt
@lorenschmidt
|
21 h |
|
yeah it should be backward, right? i keep swapping things around to superstitiously try and get it to work (that's why the z element was positive) but so far no dice.
|
||
|
|
||
|
loren schmidt
@lorenschmidt
|
21 h |
|
i am also suspicious that i may have this part wrong. if i put the perspective first, it's really off but looks like it is in perspective (see video i just posted) . if i have it at the end of the set of operations, it seems to be in ortho but scaled down uniformly with depth. pic.twitter.com/cwEChdjaqk
|
||
|
|
||
|
loren schmidt
@lorenschmidt
|
21 h |
|
swapping the columns and rows? i tried that and the weird thing is it doesn't do anything visible, as far as i can tell.
|
||
|
|
||
|
loren schmidt
@lorenschmidt
|
21 h |
|
perspective matrix bug, with cube array pic.twitter.com/StS0ongPKS
|
||
|
|
||
|
loren schmidt
@lorenschmidt
|
21 h |
|
i think maybe my problem is my matrix itself. i thought i had this right, but i don't get the math well enough to actually spot errors. pic.twitter.com/IQyFB1Lhnp
|
||
|
|
||
|
loren schmidt
@lorenschmidt
|
21 h |
|
yeah i am fuzzy on exactly what's happening in 4-space there, but i do have that step correct!
|
||
|
|
||
|
loren schmidt
@lorenschmidt
|
21 h |
|
ok, progress: i have isolated the problem. if i turn off my perspective matrix completely, i have a perfectly respectable ortho projection. so the problem is somewhere in the perspective operations. pic.twitter.com/nkMEyDCQGZ
|
||
|
|
||