Odyssey vs. Horizons Performance - Some comparisons

Background

CPU - i5-4590 (3.3ghz)
GPU - GTX 960
RAM - 12gb
Hard Drive - SSD

I've played ED on the same machine since about 4 months after release of the base game with a fairly long break before the Odyssey Alpha. Ive generally played it on either High or Ultra settings with various settings tweaked (often a mix of: shadows, texture quality ambient occlusion, terrain quality, bloom) to get the best performance vs frame rate. I dont mind dropping to 30fps on planets as long as it looks pretty and it's still relatively smooth to play. In stations it would often drop textures on buildings close to PADs and then occasionally load them back in if I stayed there for some time. It would often drop the textures on inside of the SRV ship hangar as it lowered and brought me up through it, or my legs and hands if I looked down in the cockpit. All in all though it was playable on these higher settings.

Since Odyssey Alpha and the full release this has become really difficult. Textures being a muddy mess (most of the time) and polygons displayed in stations has made it a real chore to play and I didnt do any of the Alpha once the exploration part started. With the minimum spec changing my card fell just outside of this, mainly because it's only 2gb

Testing

So this afternoon I ran a few tests with the updated version of Horizons vs the Odyssey release. Originally I was only testing in and outside a station as I didnt see the point in comparing planets as they would be so different and rather pointless to compare. However, I did and I noticed something very surprising, but in a good way. But more of that later.....All vanilla settings, with Vsync off, in open with the frame limit turned off and in 1080p. I've not included GPU useage as they were all 100%. I also deleted my graphics file as per FD suggestion.

Armstrong Dock (Landing pad with external camera)

1621605043163.png

Horizons - Ultra mode

1621608998386.png

Odyssey - Ultra Mode


1621605087147.png


I then flew outside the station and parked 4kms away

Outside Armstrong Dock @ 4km distance

1621608143739.png


Flying into Armstrong Dock
1621605505082.png


I was a little surprised that the FPS in Ultra was so high without anything turned down, but it's been a while since I played so I thought I'd make a trip to a planet in Horizons and look at it there. I went to Aritimi 5b and took a trip out in the SRV. I drove about and was surprised to see that the frame rate didn't drop below 60 fps. On Ultra.....Here's two static cam shots, showing

Aritimi 5b - SRV (Horizons)

1621608750888.png

Horizons - Ultra Mode

I then increased supersampling to 1.25, which I know to be a real killer for my card. To my surprise it still ran upwards of 40fps..

1621608800747.png

Horizons Ultra Mode (1.25 Super Sampling)

I'd include a photo of the same spot in Odyssey for a 'comparison', but for some reason it's just in the dark side of the planet, whereas Horizons isn't!....I'd point out though that the Odyssey version is hovering about 20fps on Ultra.

Conclusion

Im not really sure what to think. Has there been some optimisation of Horizons long before Odyssey was released that Im unaware of? Has it been optimised in advance of Odyssey? Has Horizons been downgraded? Has deleting my graphics config made any difference? Either way Im definately getting a much better performance out of Horizons than I did previously, and obviously the gap between it and Odyssey when comparing 'like for like' situations is very significant. Anyone else with a 960 able to do a test in Horizons and check if they are getting different results? @Agony_Aunt I think has a 960....
 

Attachments

  • 1621605184292.png
    1621605184292.png
    11 KB · Views: 144
  • 1621607623644.png
    1621607623644.png
    791.1 KB · Views: 167
Last edited:
Wow. As someone running on a GTX 1070, many of your frame rates are eerily similar to mine. Makes me wonder how much optimization is needed to actually get us both to a similar level of performance.

My entire rig is newer and more powerful than yours but to be scoring such similar levels of performance is...interesting.
 
So I just turned the supersampling up to 1.5 on Ultra and Im still getting between 30 and 35fps. That's on a GTX 960 with 2gb RAM. It looks glorious...

What's going on? o_O

I'll try and record some video later......
 
Note: oops, accidentally cut a bit off the title, it should read: "Some comparisons....."

Background



I've played ED on the same machine since about 4 months after release of the base game with a fairly long break before the Odyssey Alpha. Ive generally played it on either High or Ultra settings with various settings tweaked (often a mix of: shadows, texture quality ambient occlusion, terrain quality, bloom) to get the best performance vs frame rate. I dont mind dropping to 30fps on planets as long as it looks pretty and it's still relatively smooth to play. In stations it would often drop textures on buildings close to PADs and then occasionally load them back in if I stayed there for some time. It would often drop the textures on inside of the SRV ship hangar as it lowered and brought me up through it, or my legs and hands if I looked down in the cockpit. All in all though it was playable on these higher settings.

Since Odyssey Alpha and the full release this has become really difficult. Textures being a muddy mess (most of the time) and polygons displayed in stations has made it a real chore to play and I didnt do any of the Alpha once the exploration part started. With the minimum spec changing my card fell just outside of this, mainly because it's only 2gb

Testing

So this afternoon I ran a few tests with the updated version of Horizons vs the Odyssey release. Originally I was only testing in and outside a station as I didnt see the point in comparing planets as they would be so different and rather pointless to compare. However, I did and I noticed something very surprising, but in a good way. But more of that later.....All vanilla settings, with Vsync off, in open with the frame limit turned off and in 1080p. I've not included GPU useage as they were all 100%. I also deleted my graphics file as per FD suggestion.

Armstrong Dock (Landing pad with external camera)

View attachment 228552
Horizons - Ultra mode

View attachment 228593
Odyssey - Ultra Mode


View attachment 228553

I then flew outside the station and parked 4kms

Outside Armstrong Dock @ 4km distance

View attachment 228586

Flying into Armstrong Dock
View attachment 228561

I was a little surprised that the FPS in Ultra was so high without anything turned down, but it's been a while since I played so I thought I'd make a trip to a planet in Horizons and look at it there. I went to Aritimi 5b and took a trip out in the SRV. I drove about and was surprised to see that the frame rate didn't drop below 60 fps. On Ultra.....Here's two static cam shots, showing

Aritimi 5b - SRV (Horizons)

View attachment 228590
Horizons - Ultra Mode

I then increased supersampling to 1.25, which I know to be a real killer for my card. To my surprise it still ran upwards of 40fps..

View attachment 228591
Horizons Ultra Mode (1.25 Super Sampling)

I'd include a photo of the same spot in Odyssey for a 'comparison', but for some reason it's just in the dark side of the planet, whereas Horizons isn't!....I'd point out though that the Odyssey version is hovering about 20fps on Ultra.

Conclusion

Im not really sure what to think. Has there been some optimisation of Horizons long before Odyssey was released that Im unaware of? Has it been optimised in advance of Odyssey? Has Horizons been downgraded? Has deleting my graphics config made any difference? Either way Im definately getting a much better performance out of Horizons than I did previously, and obviously the gap between it and Odyssey when comparing 'like for like' situations is very significant. Anyone else with a 960 able to do a test in Horizons and check if they are getting different results? @Agony_Aunt I think has a 960....

Yeah, i've got a 960 but no plans to get into Horizons or Odyssey in the coming days at least.
 
So I just turned the supersampling up to 1.5 on Ultra and Im still getting between 30 and 35fps. That's on a GTX 960 with 2gb RAM. It looks glorious...
I have the same frame rate (GTX 970) on legacy planets in Odyssey at low settings. At high, I lose a maximum of 5 frames per second. In the same places in the "horizons" - 60+ frames per second (Vsync enabled)

How does it even work? How could this ever be released?
 
I have the same frame rate (GTX 970) on legacy planets in Odyssey at low settings. At high, I lose a maximum of 5 frames per second. In the same places in the "horizons" - 60+ frames per second (Vsync enabled)

How does it even work? How could this ever be released?

I'm not sure what it means... I'm wondering if it shows that any engine update now has Horizons is running much better, but shows Odyssey is still lacking optimisation? It seems unlikely that clearing the graphics config is making that much of a difference. Previously the idea of running my 960 on Ultra at 1.5 supersampling and getting 30+fps was a pipedream...

Maybe @Kay Ross can tell me I'm imagining it all and to lay off the onion head. 🤯
 
Last edited:
For me personally, clearing the graphical configuration increased the unstable FPS in the stations. It used to be 35-40, and now 35-55.
But it's not playable anyway.
but shows Odyssey is still lacking optimisation?
There is no optimization here, since most often (in normal companies) optimization is done at the end of the development cycle.
All those who say that the game has become incredibly beautiful compared to Horizons feel the placebo effect.
There are no perceptibly qualitative changes in the game.

The most annoying thing is that instead of admitting that performance problems are one of the main critical mistakes, Frontier mentions this in a small paragraph and propose a solution from the guys at Reddit.
As if this problem is observed in a small number of users and is not as important as the correct display of positions in the frontier store.
In this case, I have a question about the competence of the employees. Why does Odyssey use Horizons config files?
 
For me personally, clearing the graphical configuration increased the unstable FPS in the stations. It used to be 35-40, and now 35-55.
But it's not playable anyway.

There is no optimization here, since most often (in normal companies) optimization is done at the end of the development cycle.
All those who say that the game has become incredibly beautiful compared to Horizons feel the placebo effect.
There are no perceptibly qualitative changes in the game.

The most annoying thing is that instead of admitting that performance problems are one of the main critical mistakes, Frontier mentions this in a small paragraph and propose a solution from the guys at Reddit.
As if this problem is observed in a small number of users and is not as important as the correct display of positions in the frontier store.
In this case, I have a question about the competence of the employees. Why does Odyssey use Horizons config files?

Agreed, optimisation often comes last of all. I'm not sure there hasnt been any, but clearly there's either a considerable amount to do or there's some other issue.

Edit: David Braben has just commented here.
 
Makes me wonder how much optimization is needed to actually get us both to a similar level of performance.
There is no optimization here, since most often (in normal companies) optimization is done at the end of the development cycle.
Agreed, optimisation often comes last of all. I'm not sure there hasnt been any, but clearly there's either a considerable amount to do or there's some other issue.

I have zero faith 'optimization' is going to close this wide a performance gap. I can't recalling any developer releasing an 'optimization' that doubled frame rates, and if anyone can I don't believe FD is capable of doing it.
 
I have zero faith 'optimization' is going to close this wide a performance gap. I can't recalling any developer releasing an 'optimization' that doubled frame rates, and if anyone can I don't believe FD is capable of doing it.
As I suggested, I suspect that there's more to it than just that, and that it possibly/probably isn't the main/only issue here.

My post might be complete coincidence and/or rubbish, but im at a loss as to why Horizons is suddenly running at top specs on my machine all of a sudden, when for the last 6 years it was a no-no. If anyone else has a similar set-up and can disprove the performance increase please do!
 
Wow. As someone running on a GTX 1070, many of your frame rates are eerily similar to mine. Makes me wonder how much optimization is needed to actually get us both to a similar level of performance.

My entire rig is newer and more powerful than yours but to be scoring such similar levels of performance is...interesting.
Sounds like the game must have a bottleneck or multiple bottlenecks outside of CPU and GPU performance capabilities that are hampering performance in general.
 
Back
Top Bottom