Mayo's 3D modelling workshop

[Mayo Master]

I'm getting a move on in the past few days, after a lot of time stalling because of my Survival playthrough of Fallout 4 *ahem*
Anyways, I've more or less finished the fence, and I got started on the materials. Namely, I've made good progress on the the body of the bus, which had been facilitated by my previous works

Spoiler: for Team Avalanche

.

Spoiler: So, here's where I'm at

 

[Mayo Master]

Alright, finally getting an update on this project. I've spent some time working on the snow (the original scene takes place in snowy Hokkaido winter). The difficulty here is to create a displacement map (actually a lot of tedious work with Gimp) which actually "sculpt" the snow, so that I could create footprints, sled tracks, and and all the bits of snow pushed out of the way. I'm pretty happy with what I managed so far.

Spoiler: Here is the most recent update

It's still not done yet. First I don't know if I should revise my snow shader (the "glitter" effects of the snow are not really visible, maybe a scaling problem). Second, I need to add some "burnt" winter grass around the patches that are not covered with snow. Then, maybe I'd have to add a few features in the landscape behind the bus, because it looks a bit bare. But, I'm slowly getting there.

 

[Mayo Master]

New update!
Reworked the scale of the snow shader for getting the "glitter" effect, added some stuff in the background (actually in the original setting, the bus is in the backyard of an elementary school, which here is casting shadow over the scene), added some blades of withered grass and some trees. I still need to make a couple of tweaks on the population of trees (so that they look less like "repeats of instances") and that of grass (it's too sparse right now), but it's getting there.

Spoiler: Update of doom

 

[Mayo Master]

Ok, update on that bus project. I modified a few things, it may not be so apparent at first glance. Mostly, I tweaked the distribution of the grass (which barely shows through the snow) and created more diversity among the trees.

Spoiler: Here's the image

(View image in another tab for full resolution)

 

[Mayo Master]

Lately I've returned to my Sci-Fi project, as I've got a bunch of ideas for expanding it and flesh it out. Maybe I'll open a separate thread about the whole setting - in short it's "Speculative Science" (ie, "hard" Sci-Fi) about a society established in floating cities and airships in the Venusian atmosphere - the Airship project on the top of this page is part of that. These past days I've been exploring designs for floating cities,

Spoiler: here is an example

(the "transparent tubes" are filled with air - which is a lifting gas in the Venusian atmosphere - and the "habitable space" is made by the thin strips at the base of the tubes).

 

[Shademp]

Fuck yeah, Venusian floating cities!

 

[Mayo Master]

Fuck yeah, Venusian floating cities!

That's an excellent video that makes a good summary of most of the advantages of creating a colony on Venus. Maybe I should start that thread on this Sci-Fi project - I'm not opposed to the idea of collaborative work
Among other things, the main elements needed for this colony is an energy efficient way of maintaining altitude, and breathable air on Venus has 2/3 of the lifting power that Helium has on Earth (because the Venusian atmosphere is made of CO2 for the most part).
To make up the nitrogen of the synthetic air, it's very easy, since the Venusian atmosphere has plenty of N2. Even though it's only 5% of the Venusian atmosphere, it's actually almost as much N2 as we have on Earth. We can also have 2 sources of O2, not only from CO2, but also from the decomposition of SO3 into SO2, SO3 resulting from a chemical reaction which produces water from sulfuric acid.
The other things is about how you make structures resilient to the atmospheric sulfuric acid, and we can get that covered by teflon and FEP (Fluorinated Ethylene-Propylene), the later having the advantage of being transparent to sunlight.
Anyway, that's enough for a quick overview, I should open a new threads if you guys are interested in the concept.

 

[Mayo Master]

Spoiler: First attempt at modeling Venus...

I tried to create a "realistic" atmosphere and render the Rayleigh scattering (which causes the blue color of the sky on Earth, and as the atmosphere of Venus is a lot more dense, it would yield an orange sky on the Venusian surface). You can see the effect as the light fades out in the atmosphere. It's pretty cool since I managed to do get that orange color with a grey planetary surface (contrarily to most images you see about a cloudless venusian surface, the surface should be mostly grey because it's essentially basalt). However, my spherical gradient texture seemed to be acting up, the radial "fading-out" didn't work as well as I would have liked.
For now I only modelled it with a layer for the atmospheric haze (between 10 and 45 km above the surface) and just with a single cloud deck based on photographs. Ideally, it would be awesome to make that with multiple procedurals (which would enable me to make the clouds move and change with time in an animation), but one difficulty is to recreate the Venusian cloud pattern, especially the crazy polar vortices.
Anyway, I hope you like this one.

 

[Mayo Master]

Spoiler: Making some progress on modelling Venus even if it doesn't look like it

Okay, I've done a fair amount of work in the past few days on modelling Venus, and below is my latest update.

Now, you might wonder "how is that supposed to be better than the very first attempt from the previous post?"
The answer is: "Procedurals", my friends, Procedurals.

So the next question is going to be "well what's the big idea with these Procedurals anyway?"
The big deal with Procedurals is the generation of images, textures, mapping, etc. using mathematical functions, often involving a fractal component. You build some more or less complicated functions, have a few input parameters, and voila, you get an image. In this particular image, the Venusian clouds are not created using an existing image or a photograph (whereas that was the case of the first attempt from the previous post), they are all generated by mathematical functions.

Then, I would guess the next question to be "what's interesting in using an image generated by mathematical functions instead of a photograph?"
If you generate an image via mathematical functions, it is possible to compose these mathematical functions with the inclusion of variables depending on time. The big idea behind it is to set time-dependent parameters so that, when generating a series of images for an animation sequence, the generated image will change for each frame. In other words, I can make the clouds change shape during an animation sequence (besides the basic animation moves such as overall rotation). Eventually, my goal is to render a 10 seconds animation of the planet, using 250 frames, with an estimated time scale of 1s in video = 1 hr real-life (knowing that Venusian clouds take around 4 days to revolve around the planet, along the crazy phenomenon of "Super-rotation").

So, that's where I'm at. Before I get to work on the animation itself, I still need to fix some issue on my atmospheric volumetric shader which is aimed at recreating Rayleigh scattering (the physical effect that causes the sky to be blue on Earth). Thankfully I have some pointers about why it was acting up, but I still need to fix that. Among the pointers, I should end up with a result where the shaded part of the clouds should be yellowish, not pinkish, so I'll tune the setting to achieve this type of result. I find it quite inconvenient that real-color images of Venus are actually very scarce, so it's very hard to get good references.

 

[Mayo Master]

Spoiler: Fixed the atmospheric shader

(by the way, if you right click to "View image", you have it in full HD resolution)

Okay, this is pretty much the result I wanted. It might look a bit less "glamorous" than previous images, but actually this one is fairly realistic. I went to some pain to remodel the atmospheric shader, using a radial function of volumetric scattering proportional to the atmosphere pressure profile with altitude.
There's one big downside, though: it took me 8 hours to render this image (the complex interaction between multiple volumetric scattering shaders makes rendering pretty complicated). I made a test with fewer rendering requirements which decreased the render time under 2 hours, but the result was a very noisy and I didn't like it.
So, if I need to render 250 frames for my animation, well... I'll have to check out some render farms!

 

[Mayo Master]

Animated clouds!

 

[Mayo Master]

By the way, for the anecdote: since I want to make an animation of 250 frames, and that each frame takes 8 hours to render on my computer, I looked for a free render farm to get some help.
I managed to get my animation render launched at BURP (the Big and Ugly Rendering Project), for which I'm grateful. At the current pace, the animation should be rendered in about 18 days (that's still 4-5 times faster than what it would have taken on my own!). Looking forward to it!

 

[Mayo Master]

Finally!