New Planet Tech is KILLER of Exploration (all terrain is tiling/repeating/not procedural/random)

TBF:
  • some planets are distinctively "pattern-broken" to the level it make me wince
  • there are (probably) not that common, otherwise this would be a thing instantly after release
  • explorers have right to be indignant at (possible) abandoning of procedurally generating planets, because more pattern usage means less (or none at all) diversity in exploring experience
I think that's it, the rest is usual "GAME IS RUINED/DOOMED" vs "THERE IS NO PROBLEM" polarization.
Steam reviews show less of an exact polarization more like a 70 / 30 kind of a thing.
 
Honestly I'm more bothered by the LOD problems and that some planets look like they joined the "You cannot AA us" club. There's some really gnarly looking planetary textures at times, and awesome ones at other times.
Odyssey is broken in many ways, but I will restrain myself from the "drama queen mode" and "this is the end, nothing will get better, the new tech has DOOMED hardcoded"
 
Odyssey is broken in many ways, but I will restrain myself from the "drama queen mode" and "this is the end, nothing will get better, the new tech has DOOMED hardcoded"
I just keep telling myself "Guess the Kickstarter Premium Beta paid for this and didn't have t obuy it now" So in that regard, it's more sad and disappointing, then "DooooOOOm". I was super excited and came back because of EDO, after the dumpster fire that was multicrew. I have no one else but myself to blame 😂
 
I just keep telling myself "Guess the Kickstarter Premium Beta paid for this and didn't have t obuy it now" So in that regard, it's more sad and disappointing, then "DooooOOOm". I was super excited and came back because of EDO, after the dumpster fire that was multicrew. I have no one else but myself to blame 😂
I had a choice. Backing ED or SC. Since I absolutely loved Freelancer, I went for SC. I always have wondered if I did make the right choice. But now, I'm wondering if there was a good choice to make.
 
Long time explorer and first time poster here. Sorry everyone. I'm possibly going to ruin Horizons' procgen for you as well. It's one of those "once you see it you can't unsee it" things... ;)

So I just lifted this screenshot at 1:50 from the video about terrain quality in Horizons vs Odyssey that someone linked earlier in this thread and marked some of the pre-generated terrain assets used by the Horizons procgen algorithm. See the top two tiles for Horizons terrain and notice the similar patterns baked into the Horizons terrain? A few of them are obviously somewhat distorted (see Horizons night vision screenshot top right). Perhaps it's due to being rescaled/squished in the y axis by the Horizons procgen or maybe due to the fish-eye effect in the original screenshots, but they're clearly drawing from the same assets. There are more repeating patterns in the screenshot of course but I only marked the, at a glance, most egregious ones.

View attachment 230708

This is as far as I know the only way to get procedural generation to work with anything resembling reasonable performance on today's hardware. Having every little piece of the landscape entirely procedurally generated on the fly, down to every rock and bump, for an entire planet at 1:1 scale is simply unworkable given today's hardware restraints (i.e. local storage, memory and computing power). In particular if the procedural generation is being done in real time on the user's own CPU/GPU. Also, it's extremely resource intensive and complex (if even at all feasible) to simulate and procedurally generate more complex geological formations, such as those formed by wind, liquid erosion or tectonic stresses, using real-time mathematics and achieve anything resembling 60+ fps at the same time. Perhaps when we all are running quantum computers sometime in the future...? This is 100% why Frontier uses these large pre-generated ground assets and try to blend them together as best as they can when Odyssey is generating planet surfaces.

The trick is to have enough variety of pre-generated assets and arrange these assets in random and subtle enough patterns to make it less obvious to the human eye that they're actually based on the same meshes and textures. (Even that won't work all of the time as one can see in Star Citizen and No Man's Sky procgen, or in the screenshots above...)

IMO the issue with Odyssey is not so much that it uses pre-generated assets on planet surfaces. As you can see Horizons does that too and for performance reasons it needs to do that! The issue is that Odyssey's lighting, planetary colour variety at a distance along with a (quite likely) lower pool of larger, more detailed and comprehensive mesh assets compared to Horizons. All of this makes the baked in patterns in Odyssey stick out like a sore thumb compared to Horizons more subtle approach to surfaces.

Could Odyssey do better as far as planetary surfaces go for us explorers? Absolutely! I fully agree that it needs more work and a more subtle approach to asset re-use. Is it unfixable? Most likely not. Horizons uses the same approach Odyssey does well enough that most won't ever notice any repeating surface assets in Horizons, as this thread is evidence of.

The devs need to tweak the sharp Odyssey surface lighting (yes please!), increase the variety and subtleness of ground mesh/texture assets and improve surface colours (i.e. make them better blended and with more subtle colour transitions between different spots on the surface, in particular when looking at a distance) and if possible also try to limit the adjacent re-use of pre-generated assets enough in Odyssey to make its surface assets blend as subtly as Horizons procgen does it. If the devs can pull that off then Odyssey procgen will fool the human eye at least most of the time, just as Horizons procgen algorithms manage to do.

Cheers!

PS. Again I'm sorry if I've ruined the Horizons procgen for anyone. Once you see it you can't unsee it... ;)
Yes, but my main point has always been Planetary Tech 2.0, if you like, should be a step forward, and not have asset re-use be MORE obvious. I'm pretty sure those copied zones are much bigger than anything in Horizons-- and the ones in the latter look more like small-scale hills or depressions. I mean those examples in Odyssey are entire geological regions.

Craters I think I can forgive a lot more, because well craters all look pretty much the same, and you don't need such a huge variety of types (oblique impacts and such) to make them not look like a kid has just discovered the clone tool in Paint.
 
Yes, but my main point has always been Planetary Tech 2.0, if you like, should be a step forward, and not have asset re-use be MORE obvious. I'm pretty sure those copied zones are much bigger than anything in Horizons-- and the ones in the latter look more like small-scale hills or depressions. I mean those examples in Odyssey are entire geological regions.

Craters I think I can forgive a lot more, because well craters all look pretty much the same, and you don't need such a huge variety of types (oblique impacts and such) to make them not look like a kid has just discovered the clone tool in Paint.
I think those craters always were repetitive, but the terrain did a good job hiding it. Like those optical illusion, the brain can be tricked. But now that the terrain is very bland and limited, they are suddenly very easy to see.
 
Terrain quality. You literally checked them all but that one. Above the rest.

View attachment 230752But it's for CAPTURE, it's in the name, not for gameplay, and shouldn't be used as a base for balancing and tweaking.

nope, still not an option

ultr.jpg
 
Yes, but my main point has always been Planetary Tech 2.0, if you like, should be a step forward, and not have asset re-use be MORE obvious. I'm pretty sure those copied zones are much bigger than anything in Horizons-- and the ones in the latter look more like small-scale hills or depressions. I mean those examples in Odyssey are entire geological regions.

Craters I think I can forgive a lot more, because well craters all look pretty much the same, and you don't need such a huge variety of types (oblique impacts and such) to make them not look like a kid has just discovered the clone tool in Paint.
Yes. That's exactly what I'm saying:
IMO the issue with Odyssey is not so much that it uses pre-generated assets on planet surfaces. As you can see Horizons does that too and for performance reasons it needs to do that! The issue is that Odyssey's lighting, planetary colour variety at a distance along with a (quite likely) lower pool of larger, more detailed and comprehensive mesh assets compared to Horizons. All of this makes the baked in patterns in Odyssey stick out like a sore thumb compared to Horizons more subtle approach to surfaces.

Could Odyssey do better as far as planetary surfaces go for us explorers? Absolutely! I fully agree that it needs more work and a more subtle approach to asset re-use. Is it unfixable? Most likely not. Horizons uses the same approach Odyssey does well enough that most won't ever notice any repeating surface assets in Horizons, as this thread is evidence of.
The devs need to tweak the sharp Odyssey surface lighting (yes please!), increase the variety and subtleness of ground mesh/texture assets and improve surface colours (i.e. make them better blended and with more subtle colour transitions between different spots on the surface, in particular when looking at a distance) and if possible also try to limit the adjacent re-use of pre-generated assets enough in Odyssey to make its surface assets blend as subtly as Horizons procgen does it. If the devs can pull that off then Odyssey procgen will fool the human eye at least most of the time, just as Horizons procgen algorithms manage to do.
With more subtle Odyssey lighting, better blended (and more varied) surface colours and better variety of pre-generated assets the patterns won't be as obvious.

Cheers!
 
I had a choice. Backing ED or SC. Since I absolutely loved Freelancer, I went for SC. I always have wondered if I did make the right choice. But now, I'm wondering if there was a good choice to make.
My £30 or so on Star Citizen is just money burnt as far as I see it.

Despite my criticisms ED is still a good game, and there are beautiful vistas to be found on the new planets. It's the inconsistency that bothers me, and I don't think repeating regions so large on a planet is acceptable. Also I've yet to see any 'magic' zones like Mars' Valles Marineris. Maybe you need an active world to get such features? Well, I see no reason why some of these planets can't be extinct ones that once had plate tectonics, some kind of vulcanism or maybe even a water cycle.

Anyway we have plenty of real moons with some pretty impressive features to look at. Miranda for instance.
 
Hm maybe it's time for a sequel.

Maybe Cobra engine / stellar forge / the system that links players to each other has reached the limit of what it can do?

It's like when a high mileage car starts going wrong all the time and needs constant fixing 😁 😔 😢
 
Ultra for capture, as the name imply, is intended for screenshots and gifs. Not for actual gameplay. Whether you can run it or not is irrelevant, since it's not supposed to be the base for your average gameplay.
never said it is. just throwing some screens into the mix for reference, thats all.
 
never said it is. just throwing some screens into the mix for reference, thats all.
No harm done. It just seems that the setting was made for promotional pictures, and as such are a special case. In fact, it's likely the setting was left by mistake.
 
The first image where it's very obvious and the DSS pattern both look like dragons (the first one like a wyvern breathing fire, looking at it from the side, the second like an eastern dragon)
I'd like to call it the dragon landscape bug.
 
Back
Top Bottom