Planetary Tech Q&A - What's new in Odyssey
Planetary tech is a very important part of Elite since we'll be flying back onto the planet surface once again, visiting them closer than ever before! To answer more player questions, Frontier hosted a dedicated stream last week covering a myriad of planetary things...
These questions were answered by Dr Kay - a Lead Render Programmer at Frontier and also one of the minds behind the Stellar Forge.
If you missed it, you can watch the entire stream below as well as read through the Q&A summary.
Q: What is the Stellar Forge?
The Stellar Forge is the framework in which we input data and get out the space stuff that we want! We build this framework with what raw data we have on our Milky Way to hand to feed it, and knowing what we need out of it in the end - it then procedurally and deterministically offers a physically plausible series of star systems which represent our Milky Way! Examples of the Milky Way data included are: layout, mass distribution, age regions and lots more!
Q: Can you describe something funny that's happened with the Stellar Forge?
For sure! We've had to deal with something called the 'Hairy ball theorem' - if you have a sphere (or an object) and it's got hair pointing out all around it - you can't. There's going to be a seam somewhere, that's mathematically the case in an object like that. However, in generating the surfaces of the planets through Horizons, Beyond and Odyssey, things can be sensitive when inputting data! If something gets inputted wrong, what we have ended up with is essentially a very hairy planet to show the very frustrating Hairy ball theorem! There are tiny spikes everywhere and it looks really fuzzy - it's a very expensive way of rendering fur and we don't recommend it!
Q: So you've built the galaxy, how do you go about the finer details?
So, the last time we did a talk on this, I went into how planets start not as spheres but as cubes, with square patches that make up the cubes. Everything then becomes a sphere and an offset is generated to represent terrain. For Horizons, that terrain is generated entirely mathematically. There was a lot of effort put into representing the kind of shapes a hill or canyon makes, just using maths.
For Odyssey, there's an entirely different approach for making these offsets. It takes the view of starting big and refining small. A planet has all sorts of Stellar Forge data associated with it. Things like the gravitational stress its crust goes under, the proportion of its depth which is crust or magma or core, the stresses its under, how cratered it would be, if it's tidally locked, and more. From those, and the general knowledge of what the planet is made up of, we decide its top layer information. That is describing general shapes and regions of interest across the entire surface and the general types of formations which will be there. Things might make the terrain more chaotic, make more flat plains or basins and things like that. From stochastically sampling that information, the next layer of information goes down - we're talking up 100km worth of terrain for example, which are now generated offline into terrain shapes that we know are formed.
We have resources that represent the shapes and the types of features that we'd want to generate in the first place. Then we've got some stochastic sampling and blending to put those together, not only in the right places across a planet, but now also rotated correctly so they flow into each other in the right direction - which makes quite a difference in the visual quality of a vista.
Q: Would you say your approach for Odyssey is more accurate?
Yes, we've got the layers going from the top level down. We're trying to get the detail in all the way in, and represent it from a distance, so you can see the detail from further out which makes a lot of difference. The system we made for Horizons was good for Horizons, I'm very proud of what the team did and the results you can get from it. I still from this day enjoy looking at screenshots on Twitter and lurking in streams! But to look forward to whatever comes next for Odyssey, I wanted a system that was robust enough to handle things without having to do another remake of any of the tech, that's why it's changed and we have all that detail.
Q: What type of planet is this new tech going to be applied to?
Every planet you could land on before, and the new ones opening up, will be using this approach. The old surfaces can't be represented in this new approach and you're going to get a larger variety using this new tech.
Q: Are non-landable planets seeing changes too?
No, the planetary tech is focused on the planets you can land on to see the detail up close.
Q: Will console users see the changes at the same time as PC users?
No, console players who purchase the expansion will experience the planetary tech changes upon the release of Odyssey on their platform.
Q: Does it make you sad that players might not see the perfect alignments of planets and stars before moving on, missing out on some stunning views?
There's a range of feeling regarding that! There's the "can we test every single possible combination in the galaxy?", "what does the lighting here look like in 2 years time?" - we don't get to see that, but we do get to see your screenshots on Reddit, Twitter etc. I do spend a lot of time lurking, not to stroke one's ego but to see the joy of players when discovering sights in the galaxy.
Q: How comparable are the changes you are making to Odyssey to those you made for Horizons?
It's the same but different, we've talked a lot about the planets, but the second half of what my team does is the lighting, which is another thing that's changed a lot for Odyssey. Horizons was a fantastic and heavy undertaking to introduce full scale planets, which people don't tend to do because planets are hard! When it comes to Odyssey, we've got a framework there of how to make a planet and surface, but the process in which the shape of the surface is decided, how the resources are streamed in and out, and combined, that's also a large undertaking! It hasn't just been a weekend or two's work!
Q: Were you excited by the challenge?
We felt it was a great opportunity to make those changes I was thinking of before, to make this expandable to more and more afterwards. I was excited to see what could happen with the code and art side working together to create these assets which are blended for planets.
Q: What can players look forward to in regards to lighting in Odyssey?
I want to give good credit to the render people that I've worked with over the years to provide the visuals going forward. Not only that, but provide a consistency in how lighting behaviour works. We've made quite a few changes on the rendering side, to list a few examples: We now have some per-pixel lit particles, we have more shadowed spotlights working together, and we also finally have physically based materials that have information about how rough they are and how they should respond to light. If we then feed in these realistic lighting values that we have now, things work right, together, and consistently. That's part of the reason why the new planets are looking very nice now, it's a combination of lighting from the atmosphere, the star, and any lights around your body.
The lighting on a world with an atmosphere is different to that of a world without. You don't have the atmospheric contribution as much, it's a slightly starker look - and I love that we're able to see that now in our engine. Not only are we generating the Milky Way as accurately as we can, we're also getting the looks down.
Q: Are some of the planet types getting some more love?
The planet classes are the things you see in the galaxy and system map, how it categorises them. There are somewhat arbitrary dividers between what fits into each 'box'. In Horizons, the landable classifications were rocky, metal-rich, high-metal content, rocky-ice, and icy. However, there were two main noise graphs (collection of nodes that we put together using that old maths to generate the shape of the terrain) used and rocky-ice ended up using the icy graph. Due to that, the rocky-ice worlds were somewhat not distinct from icy worlds, however in this new system where we are deciding which resources for generating terrain go to which planets depending on their parameters, the rocky-ice worlds are going to have a much more distinct look. They'll have specific terrains and materials on them, they'll be firmly represented, which warms my heart a little!
Q: Will large worlds feature tall mountains or geological features?
This is an interesting question! There is a reason why the features are shallower on larger planets. With the increased gravitational strains you can't maintain as tall a natural feature with the strength the material is made out of. so you'll end up with shallower features. I'm afraid it's just how the maths drops out for those, there's a good range of planets but the tiny ones tend to be able to support the more extreme terrain because the large things aren't being destroyed by the gravity pulling it down.
Q: Does the planetary tech upgrade allow for new land features such as caves in the future?
Caves aren't currently part of the plan for Odyssey, the focus instead is on planet-wide improvement.
Q: How may rock scatter be different in Odyssey and how is it determined?
Scatter has changed quite a bit! With Horizons and Beyond, small rocks are placed around the player as they go along - there's not too much logic driving it, it was a CPU based placement. For Odyssey, we have what we call the 'scatter system'. This is sending data through another noise graph, and we have essentially different collections of things that are expected on different areas of different types of planets - so you get a wider range of things being put down in different patterns and densities, depending on what should be there. This is going to make the surface more visually interesting to drive and walk around!
Q: Will atmospheric worlds feature weather effects such as wind and dust?
Yes, there'll be some effects visible on the surface.
Q: Will the changes include a multi-source lighting system to reflect the presence of multiple stars in systems that have them?
As we're aiming for similar specs to base game for Odyssey, we won't be including a multi-source lighting system, for performance and art reasons.
Q: How will the new planetary tech affect volcanism and the integration of volcanic features into the terrain?
We will be using the scatter system to ensure that the volcanic regions are placed in the correct areas of volcanic worlds.
Q: How does the Stellar Forge place planetary bases and installations?
There are various ways that we place settlements on planet surfaces. The outposts and the ports use a simple radial flattening of an area where we know a port or outpost is going to be - or a little bit of a deviation if they're a crater port. For things like the Guardian sites, we use 'stamps'. We have an authored resource of what the terrain should look like underneath this point, and we can tailor what goes where and decide how it blends in. With Odyssey, we've expanded this a little bit, so that across a 'stamp' there can be variations of the height it's calibrated to, which leads to more interesting levels.
Q: Will atmospheric planets have visible poles with different biomes?
After internal debate, we've ended up using the term 'geomes'! There are geomes across the planet, there are polar caps where there should be, plains, mountain forming regions, and all of the nice stuff!
Q: Will aurora be visible on some atmospheric planets?
Whilst planets with atmosphere will look different, there are no plans for visible aurora at the moment.
Q: More of a personal question, do you feel more fear about space the more you learn about it?
What I've found at university, is that things start becoming larger and larger, and you start using scales which don't make sense to humans anymore, in particle physics for example. One can start dissociating the universe in your mind in these models, and it's somehow so far removed from ourselves and our tactile interactions with it. It's less scary, and more 'indescribable'.