My first impressions after playing 4.0 Horizons client

[Morbad]

Can you suggest a better comparison?

Direct comparisons are difficult as few scenes will show areas where Odyssey is actually destroying information or resulting in a scene that cannot be salvaged with appropriate settings. Best ones I can think of are ships and other assets that have the same geometry between 3.8 and 4.0, in lighting conditions that are supposed to be similar, with neutral colored paints, where possible. Cockpits themselves, especially Imperial ships with white interiors, aren't bad.

The skybox is one of the most customizable parts of the view and I just tune it to look more or less what I think I would see from the vacuum of space, with minimal embellishment, on either client. It is darker in Odyssey, and there is a deliberate muting of the more fanciful elements, but it's also not outright missing information like other parts of the scene can be.

Minor JPEG compression artifacts aside, this is the Milky Way I see from within the bubble (specifically LTT 5455, 32k LS from anything), with my usual settings:

But I can also make it look a considerably brighter by changing a handful of variables:

Swap out another two for a five and stick a random seven in there it gets really absurd (I logged in and thought Praxis had exploded):

Some people did prefer Old Horizons though, and it certainly did look spectacular...but I really doubt the milkway looks like that to any human eyes from anywhere:

Keep in mind I have my display calibrated so that I can see everything that's in the first image. If turning up brightness reveals more, chances are one's display has a fair degree of black crush. Black crush is also what Odyssey itself does to some dark scenes, but not so much the skybox.

 

[LordLocksley]

Dr. Kay talked about that during one of her last appearances. One of the design goals was future-proofing the planetary generation system because (as I understood it) they painted themselves into a corner with Horizons procgen which couldn't do atmospheric planets the way they wanted to. That then motivated the decision to change the rendering to PBR so the lighting works correctly for all types of environment. And to her (and the teams) credit, when it works the lighting is absolutely gorgeous - dusk and dawn settings especially are among the best I've seen over the years. Of course there are still plenty of areas left where the quality simply isn't good enough and the whole experience breaks down, but as much as I miss Mt. Neverest and other mountains, Horizons never sold me on a planetary landing as much as Odyssey can.

...

The problem here is that Elite Dangerous is first and foremost a Space Sim (at least, it was, and it's the one reason many players have paid for it, even now). You cannot sacrifice the visual quality of the main target of your game for getting instead gorgeous dusks and dawns when wandering on planet surfaces, which is for a non trivial amount of players just an "accesory".

This is the fundamental flaw that I've found so far with Odyssey.
I'm perfectly ok with FDev wanting to include space legs in the game. I don't like that feature, but I don't have to, I still have the Space Sim part. Other people might get attracted to the game for that. The problem is the amount of sacrifices and compromises that the space legs feature have forced on those who liked the original game in the first place. Noone can justify that "to her (and the teams) credit, when it works the lighting is absolutely gorgeous" when Space Sim players of Elite Dangerous spend 90% of their time in areas where it "does not work" (or it works worse now than the product we had before).

 

[Old Duck]

The problem here is that Elite Dangerous is first and foremost a Space Sim (at least, it was, and it's the one reason many players have paid for it, even now). You cannot sacrifice the visual quality of the main target of your game for getting instead gorgeous dusks and dawns when wandering on planet surfaces, which is for a non trivial amount of players just an "accesory".

Yep, this has often been my thought when looking at Odyssey screenshots. There are some amazing sunsets and sunrises, but I already have plenty of games with amazing sunsets and sunrises - Elder Scrolls Online, Red Dead Online / RDR2, Microsoft Flight Simulator 2020, and many more. These games also have amazing weather, animals, trees, snow, fishing, and other things to enjoy during those sunsets. That's not to say that I don't want pretty sunsets in Elite, but I'm not buying a DLC just for pretty sunsets, because I have loads of "pretty sunset simulators" in my library already.

 

[Riverside]

The problem here is that Elite Dangerous is first and foremost a Space Sim (at least, it was, and it's the one reason many players have paid for it, even now). You cannot sacrifice the visual quality of the main target of your game for getting instead gorgeous dusks and dawns when wandering on planet surfaces, which is for a non trivial amount of players just an "accesory".

This is the fundamental flaw that I've found so far with Odyssey.
I'm perfectly ok with FDev wanting to include space legs in the game. I don't like that feature, but I don't have to, I still have the Space Sim part. Other people might get attracted to the game for that. The problem is the amount of sacrifices and compromises that the space legs feature have forced on those who liked the original game in the first place. Noone can justify that "to her (and the teams) credit, when it works the lighting is absolutely gorgeous" when Space Sim players of Elite Dangerous spend 90% of their time in areas where it "does not work" (or it works worse now than the product we had before).

I think you make a good point but pretty sunsets on distant worlds is solid sci-fi content imo. Like you I have little interest in on-foot combat but I care more about the lighting on planets than I do in space.

 

[Rat Catcher]

Fortunately we don't need to guess what it looks like.

Odyssey doesn't look anything like that. Client 3.8 gets closer to it.

I don't see much orange...

 

[Red Rocket]

There is also a big difference in planet tech between Horizons and Odyssey.

Try for your self, drop into a planet in Horizons and you should start seeing large mountains and crevasses as you get closer,
As you get closer pay attention to the rendering and how the rendering will be mostly done before landing.
Look at the verity of mountains and deep valleys that Horizons offers also the rendering is mostly done.

Now do the same in Odyssey on the same planet, notice that the mountains and crevasses are gone and the surface is mostly flat.
As you approach the render never slows down even when landing along with the possible view of patterns.
Now driving around the rendering is always ongoing and the lighting is flickering, rocks,plants popping in and out.

At this point I find Horizons planet tech more interesting and enjoyable because of the faster rendering to a finished state and verity.

 

[Treya]

It's kind of weird that they had to change the surface geometry to accommodate atmospheric shaders. By the way, No Man's Sky has all kinds of atmospheric effects and is still 100% procgen (explaining its tiny install size of only 13 GB).

It does but NMS has a big advantage in that it doesn't need to look realistic, just interesting. And they take a lot of liberty at all stages of their procgen, like nothing that resembles a Stellar Forge or using any kind of orbital mechanics. Doesn't make it a bad game, its just a different approach. By the way, NMS had at least one update that changed their planet generation to accommodate more variety but wiped a lot of the existing ones in the process - pretty much what happened with Odyssey.

I thought FDev switched from procgen to prefabs because that made it easier for them to align thousands of new settlements with the underlying terrain.

It's been a while since I last seen most of the dev interviews but if memory serves me right that was at least part of the reason, yes. But the prefabs are only one stage of the process which at its core is still procedural. They also increased they amount of data that Stellar Forge is feeding in. The result is a bit of a mixed bag as discussed in this and countless other threads since release. Some parts are really nice, like having polar caps or areas that look like they're the result of tectonic shifts. Others not so much (the tiling/banding issues or those weird spiky protrusions come to mind).
One big problem in these discussions is that FDev has been so damn tight lipped over the last couple years. They probably tried a few things internally to get atmospherics working in Horizons before deciding to do The Big Rewrite. It would've been interesting to see what it looked like earlier during development.

The problem here is that Elite Dangerous is first and foremost a Space Sim (at least, it was, and it's the one reason many players have paid for it, even now). You cannot sacrifice the visual quality of the main target of your game for getting instead gorgeous dusks and dawns when wandering on planet surfaces, which is for a non trivial amount of players just an "accesory".

This is the fundamental flaw that I've found so far with Odyssey.
I'm perfectly ok with FDev wanting to include space legs in the game. I don't like that feature, but I don't have to, I still have the Space Sim part. Other people might get attracted to the game for that. The problem is the amount of sacrifices and compromises that the space legs feature have forced on those who liked the original game in the first place. Noone can justify that "to her (and the teams) credit, when it works the lighting is absolutely gorgeous" when Space Sim players of Elite Dangerous spend 90% of their time in areas where it "does not work" (or it works worse now than the product we had before).

The reason I came to ED was Stellar Forge and the suspension of disbelief it creates for me. The reason I stayed all these years is the complexity of the various systems. So the Space Sim is right at the top for me as well. Planetary landings however are also a part of that for me. And just like with the galaxy it needs to be realistic enough for me to suspend my disbelief. Odyssey does that better for me in some parts but also fails in a few others.
I don't have an opinion on how the galaxy background should look like, any of the versions Morbad posted above would be fine with me because the important part for me is knowing that I could (theoretically) travel to all those stars. But I can appreciate that it is important for others.

Do these changes justify having broken features? Of course not. Do the working parts of Odyssey look better than Horizons? I would say yes, but that is mostly based on personal preference. And before I stop babbling: Was it worth it? Right now its still a bit of a mess but I think its a salvageable mess that is moving in the right direction. Going back to what Dr. Kay said about future-proofing in that video I linked to: Assuming we will see another DLC, that is when this question will ultimately find an answer.

 

[The Old Coder]

Warning: Copious amounts of armchair development below this point
One thing to keep in mind is that Odyssey was approx. two years in full-time development prior to release. Even without a pandemic that is at the low end for a project of this scope - creating what is essentially a new game (FPS), (re)writing a render engine and procgen system, and integrating everything with all the legacy systems and established content. Just compared to other games it seems ludicrous to me to expect anyone to deliver on that in that timeframe. Yet somehow someone made the decision to release it when they did, presumably motivated by having to release regular shareholder reports. I'm obviously just speculating here, but I can't imagine that this wasn't questioned internally beforehand.

Warning: Speculation from an actual developer with nearly four decades of experience.
I think very little work on Odyssey was done inhouse. IMO, they outsourced almost the entire thing to some low-bid, barely competent, overseas firm. It would explain a lot. Usually when doing this you have people assigned to monitor what's being done. You should be getting regular reports, working code you can test yourself and the like. "Talk is cheap, show me the code" should be your motto.

As I see it, none of that was done. The management in charge just took the happy talk provided by the outsourcers, passed it on to senior management who in turn fed it to the shareholders and players. ED just wasn't important enough to bother doing more. Then came the day when a dumpster fire was rolled up to FDEV's door and they realized that after all the glowing reports about how it "was shaping up nicely" they were stuck with releasing it. No way could they admit that a) they'd outsourced it and b) they'd screwed that up. So out the door it went and the negative reaction was blamed on "connectivity issues".

As I said, this is speculation but it's based on having been in a similar situation including the reports about how great it was all going to be. What we got from the outsourcing firm barely ran. Database queries were taking five minutes or longer, the UI was "unique", and the servers were begging for mercy when running it. The documentation looked like it had been through a machine translator backwards. My team and I spent eight long months of overtime pounding it into shape but in end there only so much we could do. The code was simply too bad. It worked but nothing like should have. It was the first module of a planned series but all the rest were canceled as there was nothing to build on. I did have the satisfaction of watching the management team responsible be summarily fired and kicked to the curb. I doubt FDEV did the same.

 

[thistle]

I think you make a good point but pretty sunsets on distant worlds is solid sci-fi content imo. Like you I have little interest in on-foot combat but I care more about the lighting on planets than I do in space.

 

[thistle]

Warning: Speculation from an actual developer with nearly four decades of experience.
I think very little work on Odyssey was done inhouse. IMO, they outsourced almost the entire thing to some low-bid, barely competent, overseas firm. It would explain a lot. Usually when doing this you have people assigned to monitor what's being done.

I've read this a number of times, mostly from the angry ex boyfriends of the forum, and never seen a shred of evidence for it. Is there any evidence for it?

You appear to have a similar trajectory as me to this point, development but not games. I've seen utter cluster****s from highly paid contractors and outsourced "experts" in specific tools that could and would have been far better coded (and properly designed) internally. But I'm not aware of any suggestions the cobra engine work for EDO was outsourced. I'm not sure it makes sense and it seems to me someone threw the idea out there once and others have run with it.

The vast majority of your post, from the 2nd line of the first para through until the end, are predicated on this idea it was outsourced.

I'm not buying it.

 

[The Old Coder]

I've read this a number of times, mostly from the angry ex boyfriends of the forum, and never seen a shred of evidence for it. Is there any evidence for it?

I did say it was speculation, didn't I? But I don't read the forums that often so if others brought it up I wasn't aware of it. I based my post entirely on my own experiences.

There's no solid evidence for or against it unless somebody at FDEV spills the beans. But I think FDEV committed the bulk of their inhouse developers to other, more potentially profitable games; especially ones where the IP was licensed. ED is a niche game and if push comes to shove it's the one getting shoved.

If FDEV knew they had a disaster on their hands why were putting out the happy talk about it? It would have made much more sense to talk about how they were working as hard as they can but the lockdowns were really making things rough on development. Then finally announce that they had to delay it because of the pandemic. That's what I would have done. Sure, there would have been the usual moaning and groaning but it would have bought them some time and been far better than the utter debacle they ended up with. That cost them an enormous amount of customer goodwill that will never return. Many of the 70 people I used to play with deleted their Frontier accounts swearing never to have anything to with them again. The rest just walked away.

It also all but killed the game FDEV built their reputation on. When you're promoting yourself as the company that people should license the IP to for competent game development... well... it's not a good look. I can't see Braben walking into this if the game had been developed inhouse.

Anyway, that's how I see it. Your milage may vary.

 

[BL1P]

I think one of the biggest problems is 4.0 looks better in screenshots than 3.8.
Yet for myself and others it seems, 3.8 looks much better and runs far smoother than 4.0 when being played.
I would say top priority should be - Performance, Antialias, Shadow flicker, Too Dark.

 

[zimms]

[...]
If FDEV knew they had a disaster on their hands why were putting out the happy talk about it? It would have made much more sense to talk about how they were working as hard as they can but the lockdowns were really making things rough on development. Then finally announce that they had to delay it because of the pandemic. That's what I would have done. [...]

All of that did happen. But apparently/allegedly at some point somebody said "No more delays, we'll get this out by the end of this fiscal year!"

 

[TheTechnobear]

as a developer (also) with nearly four decades of coding experience...
Id say there is zero evidence of it being outsourced, either overseas or otherwise.

frankly, the most likely scenario, is the normal issue of staff moving on... getting new staff, and the 'old code' being less understood.
basically, whilst code does not 'age' (degrade)... everything else changes around it, so it needs constant maintenance/improvement.
if you don't, then you just end up building up technical deficit year on year.

also, generally with code bases, you also try to 'move forward'.
the planet tech may have some issues, but as a developer, you'd want to try to resolve them ...
it would take some very dramatic to actually go back to the old code base.
(that won't happen.. fdev will 'can' the project well before that happens )

then we have more project manager/resouces/money issues ... developers (and their time) are not cheap!
if ED is not bringing in significant revenue, the resources will be moved on to other projects.
similarly you have to reduce costs as much as possible (reduce number of platforms, code bases etc - familiar?)

so unfortunately, even if you prefer 3.8, you have to hope 4.0 will succeed.
if not, then fdev, will (slowly?) pull the plug on the whole project, as it just won't pay its way.

 

[GJ51]

ED 5.0 should be on a new engine with a ground up re-write

Remember MS-DOS 5.0 was the best OS MS made...

 

[Rat Catcher]

I would say top priority should be - Performance, Antialias, Shadow flicker, Too Dark.

I agree wholeheartedly with the first 3, I like the distinct lack of orange everywhere, the darkness suits my play.

 

[LordLocksley]

Well, as the OP of this thread, and after about a month of playing with the new client, I've definitely decided to get back to Horizons 3.8 for my regular gameplay. I really tried hard to give the new 4.0 version a chance and use it by default, in preparation for any feature that might be exclusive to 4.0 in the future. Unfortunately, so far there isn't any compelling reason to keep playing in the new version.

At this point, the only advantage I've found on the 4.0 Horizons client over the older one is the inventory management and the outfitting screen. Everything else is (for my taste) either slower, more cumbersome, uglier, or just plainly worse than in the 3.8 client. From the slower interfaces and menus, the awful new galaxy map design, the slower performance overall, the less readable cockpit interface, and above all the lightning issues (darkness issues more appropiately). The fact that you have to turn on the Night Vision feature just to be able to properly dogfight almost anywhere on any system is just proof that the "new" lightning technology is not properly designed and/or implemented (except maybe if you are almost next to the system star). I don't care if it is "realistic" or not. It's just a bad gameplay experience regardless.

I really hope that in the future things change and the new 4.0 client branch catches up with 3.8. But in the meanwhile, for Horizons-only owners 4.0 is just an inferior version of the same game as of today.

 

[Ryox]

Well, as the OP of this thread, and after about a month of playing with the new client, I've definitely decided to get back to Horizons 3.8 for my regular gameplay. I really tried hard to give the new 4.0 version a chance and use it by default, in preparation for any feature that might be exclusive to 4.0 in the future. Unfortunately, so far there isn't any compelling reason to keep playing in the new version.

At this point, the only advantage I've found on the 4.0 Horizons client over the older one is the inventory management and the outfitting screen. Everything else is (for my taste) either slower, more cumbersome, uglier, or just plainly worse than in the 3.8 client. From the slower interfaces and menus, the awful new galaxy map design, the slower performance overall, the less readable cockpit interface, and above all the lightning issues (darkness issues more appropiately). The fact that you have to turn on the Night Vision feature just to be able to properly dogfight almost anywhere on any system is just proof that the "new" lightning technology is not properly designed and/or implemented (except maybe if you are almost next to the system star). I don't care if it is "realistic" or not. It's just a bad gameplay experience regardless.

I really hope that in the future things change and the new 4.0 client branch catches up with 3.8. But in the meanwhile, for Horizons-only owners 4.0 is just an inferior version of the same game as of today.

Yes, EDH4.0 is just if you want to get the narrative updates and new content. EDH3.8 is excluded from those (end of life).