blender is proof that software can be good

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > YOSPOS > blender is proof that software can be good

«‹›11 »

Jenny Agutter: Mar 18, 2009

yesterday i asked myself what if i learned a little blender. today:
https://i.imgur.com/Dzun737.mp4

Jenny Agutter fucked around with this message at 00:57 on Aug 20, 2020

# ¿ Aug 20, 2020 00:54

Adbot: ADBOT LOVES YOU

# ¿ May 10, 2024 01:46

Jenny Agutter: Mar 18, 2009

echi's project

# ¿ Aug 20, 2020 03:05

Jenny Agutter: Mar 18, 2009

https://i.imgur.com/DD4enB7.mp4

# ¿ Aug 20, 2020 20:46

Jenny Agutter: Mar 18, 2009

the classic "donut tutorial"
https://i.imgur.com/r075rKw.mp4

# ¿ Aug 22, 2020 22:59

Jenny Agutter: Mar 18, 2009

experimenting with lux core render

verdict: iot's good but v slow

# ¿ Sep 5, 2020 03:59

Jenny Agutter: Mar 18, 2009

using the displacement node fed into a color ramp?

# ¿ Sep 5, 2020 15:59

Jenny Agutter: Mar 18, 2009

lux core render is just a plugin to blender but it does have its own shader system so you can't just use the blender nodes. this requires either redoing your scene or planning to use lux from the start. in return you do get bidirectional ray tracing, albeit only using the CPU which is why it's so slow (it also includes unidirectional path tracing with opencl or cuda support, nice an snappy). but you aren't getting caustics and dispersion from cycles

# ¿ Sep 5, 2020 19:44

Jenny Agutter: Mar 18, 2009

quick and dirty render playing around with shader nodes
https://i.imgur.com/Lmt17sZ.mp4

# ¿ Sep 7, 2020 02:04

Jenny Agutter: Mar 18, 2009

im not sure you can do something like that with cycles, it compiles the shaders and then sends out light probes so you can't really use the light source as an input. you also can't use the output of a BSDF as an input to a value socket, so you can't put a white diffuse material over the object and use its shading from the light as a key. what you can do is fake it by using the light source position as a vector input, like take the dot product of that and the object normals clamped between 0 and 1, feed that into the factor of your mix shader with an emission and other shader as input. here's a quick example I made

blue is emissive, yellow is diffuse
I made a python script to populate three value nodes in the material with the light source position since you can't use another object's properties as an input node

code:

import bpy

light_vec = bpy.data.objects['Light'].location

for node in bpy.data.materials['Material'].node_tree.nodes:
    if node.label == "light_x" and node.type == "VALUE":
        node.outputs["Value"].default_value = light_vec[0]
    if node.label == "light_y" and node.type == "VALUE":
        node.outputs["Value"].default_value = light_vec[1]
    if node.label == "light_z" and node.type == "VALUE":
        node.outputs["Value"].default_value = light_vec[2]

more accurate would be to get the difference between the object position and the light position. i don't know how to make this update automatically for such as animations either, you need to run it after every position update

# ¿ Sep 7, 2020 19:52

Jenny Agutter: Mar 18, 2009

if blender is proof software can be good how come it crashes when i haven't saved in an hour :negative:

# ¿ Sep 7, 2020 21:33

Jenny Agutter: Mar 18, 2009

looks great!

# ¿ Sep 7, 2020 23:27

Jenny Agutter: Mar 18, 2009

echinopsis posted:

and it hadn�t autosavrd? i think i cranked down my auto save interval. but i�ve honestly had sweet gently caress all blender crashes outside of some out of memories on the video card but that was when i used openclon my ati card, cuda works 👌

they said that in the other thread too but I didn�t see any auto saves. maybe I hosed up by loading an old save to quickly. also I think 2.9 is more crashy than 2.8x. I really need a new computer...

# ¿ Sep 8, 2020 00:31

Jenny Agutter: Mar 18, 2009

xpost

Jenny Agutter posted:

I think this is how bongs work no one correct me
https://i.imgur.com/rHlToVB.mp4

# ¿ Sep 8, 2020 03:47

Jenny Agutter: Mar 18, 2009

↑↑↑ could have baked that simulation much faster with more cores

# ¿ Sep 8, 2020 03:48

Jenny Agutter: Mar 18, 2009

i think voronoi textures is my favorite node

# ¿ Sep 8, 2020 21:18

Jenny Agutter: Mar 18, 2009

good poo poo op

# ¿ Sep 9, 2020 01:33

Jenny Agutter: Mar 18, 2009

wish I knew about drivers before I started writing python scripts

# ¿ Sep 9, 2020 01:34

Jenny Agutter: Mar 18, 2009

made a peach

https://i.imgur.com/i03PaIP.mp4

# ¿ Sep 10, 2020 14:37

Jenny Agutter: Mar 18, 2009

amberpos supremacy
https://i.imgur.com/6F8IyBU.mp4

# ¿ Sep 10, 2020 17:51

Jenny Agutter: Mar 18, 2009

echinopsis posted:

hell yeah how�d you do that

https://twitter.com/rodtronics/status/1304013733488861184?s=21

also wishing I could rip some adaptive subdivision threads coz that poo poo takes forever

make cube, make copy and hide -> subdiv surface (apply it) -> wireframe modifier w/ -1 offset -> emission shader

make plane -> solidify -> give transparent BSDF

unhide other cube -> boolean modifier with plane as object, do not apply

then just animate the plane through the solid cube

# ¿ Sep 10, 2020 21:04

Jenny Agutter: Mar 18, 2009

echinopsis posted:

occurred to me that the wireframe node triangulates everything so gives kind ugmo results

apply it and then add a bevel modifier, flattens it out

# ¿ Sep 10, 2020 22:44

Jenny Agutter: Mar 18, 2009

used a very different method to make this one, turns out boolean operations really don't like curved surfaces lol. it looks like crap because I wanted to make a transparent gif

# ¿ Sep 10, 2020 23:50

Jenny Agutter: Mar 18, 2009

echinopsis posted:

https://twitter.com/rodtronics/status/1216707370916564993?s=21

https://twitter.com/rodtronics/status/1235362941190324224?s=21

used the bevel modifier to create these grids from subdivided planes 😎

e: not the top one at all! that was done without blender wtf

both v cool

# ¿ Sep 11, 2020 04:51

Jenny Agutter: Mar 18, 2009

hell same except sub extremely stupid in for interesting

# ¿ Sep 11, 2020 05:32

Jenny Agutter: Mar 18, 2009

there's a much more disgusting version of the balls render

# ¿ Sep 11, 2020 05:33

Jenny Agutter: Mar 18, 2009

echinopsis posted:

https://twitter.com/rodtronics/status/1304532698489384963?s=20

ominous

# ¿ Sep 13, 2020 02:50

Jenny Agutter: Mar 18, 2009

# ¿ Sep 13, 2020 17:24

Jenny Agutter: Mar 18, 2009

echinopsis posted:

https://twitter.com/rodtronics/status/1305571407829504005

looks like a nudibranch

# ¿ Sep 14, 2020 22:06

Jenny Agutter: Mar 18, 2009

echinopsis posted:

what lol

different vibe with diff colours :
https://twitter.com/rodtronics/status/1305615531811069952?s=21

less repugnant imo

i guess closer to a fuschia flatworm

# ¿ Sep 14, 2020 22:49

Jenny Agutter: Mar 18, 2009

impressive work! did you do the chromatic aberration in blender as well?

# ¿ Sep 15, 2020 15:38

Jenny Agutter: Mar 18, 2009

Insanely cool dude. I always assumed you needed specialized equipment for photogram, can you use a normal camera and/or old phone?

# ¿ Sep 20, 2020 02:12

Jenny Agutter: Mar 18, 2009

and what conversion program is best?

# ¿ Sep 20, 2020 02:12

Jenny Agutter: Mar 18, 2009

oh yeah and blender guru said you end up with huge unoptimizrd models that are a pain to work with

# ¿ Sep 20, 2020 02:13

Jenny Agutter: Mar 18, 2009

Neon Noodle posted:

can�t you bake the extra detail to normal map?

I believe you can but it's one of those either it's free or its good things

# ¿ Sep 20, 2020 03:23

Jenny Agutter: Mar 18, 2009

cool, thanks for the info

# ¿ Sep 20, 2020 04:18

Jenny Agutter: Mar 18, 2009

yo echi i saw your post about blackbody radiation in the cc thread, radiance can be approximated to be related to temperature to the fourth power (T⁴). of course you need to get that down to an appropriate level for the emission strength input so multiply it by a very small value, if you want to be really physically accurate it's 5.67E-8 but you know, use artistic license

# ¿ Sep 20, 2020 21:40

Jenny Agutter: Mar 18, 2009

5.67x10^-8

# ¿ Sep 21, 2020 02:21

Jenny Agutter: Mar 18, 2009

echinopsis posted:

... maybe i�m just dense but what is that number.. what�s its purpose

its just a physical constant, specifically the stefan-boltzmann constant. it's the actual physical value you multiply T⁴ to get the total radiance of a blackbody. technically that will give you all of the radiance across the entire E-M spectrum, not just visible light, but assuming you're using typical lighting temperature values it's close enough. just adjust it if it's too bright.

# ¿ Sep 21, 2020 04:42

Jenny Agutter: Mar 18, 2009

ah i'm an idiot, that'll give you way too high of a value. disregard

# ¿ Sep 21, 2020 04:49

Adbot: ADBOT LOVES YOU

# ¿ May 10, 2024 01:46

Jenny Agutter: Mar 18, 2009

when you've forgotten most of your physics education

# ¿ Sep 21, 2020 05:20

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > YOSPOS > blender is proof that software can be good

«‹›11 »