Support NeoGAF

[analog_future]

Gotta say, Directive 8020 is one of the first games to make me really think that maybe the UE5 haters aren't completely off-base.


I played/completed Directive 8020 and it was okay/fine, but it got me interested in going back and playing the Dark Pictures Anthology as I'd never played through those games. I started Man of Medan and it's shocking how good it looks and how well it runs compared to Directive 8020.

Same dev, same type of game, Man of Medan had a significantly smaller budget and sold for a lower MSRP than 8020, but someone this 7 year old game looks 90% as good as Directive 8020 and runs waayyyyyy better.

On my 9800X3D / 9070 XT rig:

• Native 4K Man of Medan w/RT - 85-110 FPS
• Native 4K Directive 8020 w/RT - 25-40 FPS

So Man of Medan was ~3-4x more performant than Directive 8020 while looking 90% as good.

Animations are definitely better in Directive 8020, and the density of overall detail is improved. Nothing even remotely justifies the performance delta though. And frankly the depth of field in Directive 8020 looks worse and the shadows do too. DoF doesn't trace in-focus objects as precisely which makes them look more like cutouts. Shadows can look inconsistent and flickery in Directive 8020 but are pretty much perfect in Man of Medan.


A few of my comparison shots:

Directive 8020

Man of Medan

Directive 8020

Man of Medan

 

[SlimySnake]

Gotta say, Directive 8020 is one of the first games to make me really think that maybe the UE5 haters aren't completely off-base.


I played/completed Directive 8020 and it was okay/fine, but it got me interested in going back and playing the Dark Pictures Anthology as I'd never played through those games. I started Man of Medan and it's shocking how good it looks and how well it runs compared to Directive 8020.

Same dev, same type of game, Man of Medan had a significantly smaller budget and sold for a lower MSRP than 8020, but someone this 7 year old game looks 90% as good as Directive 8020 and runs waayyyyyy better.

On my 9800X3D / 9070 XT rig:

• Native 4K Man of Medan w/RT - 85-110 FPS
• Native 4K Directive 8020 w/RT - 25-40 FPS

So Man of Medan was ~3-4x more performant than Directive 8020 while looking 90% as good.

Animations are definitely better in Directive 8020, and the density of overall detail is improved. Nothing even remotely justifies the performance delta though. And frankly the depth of field in Directive 8020 looks worse and the shadows do too. DoF doesn't trace in-focus objects as precisely which makes them look more like cutouts. Shadows can look inconsistent and flickery in Directive 8020 but are pretty much perfect in Man of Medan.


A few of my comparison shots:

Directive 8020

Man of Medan

Directive 8020

Man of Medan

Its not just rt, its path tracing. And like with most path traced games, its likely implemented by nvidia so take it up with them.

From the benchmarks I've seen, hradware lumen in this game is over 60 fps at native 4k which isn't too bad considering man of medan did not have rtgi, only reflections and AO.

I do agree that the game doesn't feel like a big leap over their previous games but that's a developer issue.

 

[Msamy]

I read a theory on twitter that the main reason behind announcing UE6 just 4 years after the launch of UE5 is that they are aware of all the UE5 hate/backlash and are hoping for a clean slate.

UE5.6 has already fixed most of the original issues like single threaded CPU utilization, shader compilations, traversal stutter, and lack of hardware lumen and nanite foliage support. UE5.7 drastically improved nanite foliage performance by completely reworking it and UE5.8 now runs Megalights at 60 fps on PS5 and is finally production ready. They even added a probe based Lumen system that is twice as fast as software lumen just for lower end hardware.

My guess is that they are like lets just start over and ship all these features we introduced in later UE5 versions that no one really bothered to use.

UE6 will be masterpiece

 

[dgrdsv]

Gotta say, Directive 8020 is one of the first games to make me really think that maybe the UE5 haters aren't completely off-base.


I played/completed Directive 8020 and it was okay/fine, but it got me interested in going back and playing the Dark Pictures Anthology as I'd never played through those games. I started Man of Medan and it's shocking how good it looks and how well it runs compared to Directive 8020.

Same dev, same type of game, Man of Medan had a significantly smaller budget and sold for a lower MSRP than 8020, but someone this 7 year old game looks 90% as good as Directive 8020 and runs waayyyyyy better.

On my 9800X3D / 9070 XT rig:

• Native 4K Man of Medan w/RT - 85-110 FPS
• Native 4K Directive 8020 w/RT - 25-40 FPS

So Man of Medan was ~3-4x more performant than Directive 8020 while looking 90% as good.

Animations are definitely better in Directive 8020, and the density of overall detail is improved. Nothing even remotely justifies the performance delta though. And frankly the depth of field in Directive 8020 looks worse and the shadows do too. DoF doesn't trace in-focus objects as precisely which makes them look more like cutouts. Shadows can look inconsistent and flickery in Directive 8020 but are pretty much perfect in Man of Medan.


A few of my comparison shots:

Directive 8020

Man of Medan

Directive 8020

Man of Medan

Same thing, precompute vs realtime, saves a lot of dev time, less storage space, gives generally better results visually, solves issues with dynamic lighting, etc.
The problem is that to make full use of all the advantages UE5 provides over UE4 a game must do things with lighting which weren't really possible in UE4 without some serious artistic hacking.
And most games don't do that, they just do whatever they did in UE4 while simply switching to the new UE5 rendering systems. In such cases using UE4 could be advantageous for performance reasons.
But one should also remember how badly UE4 ran when you've tried adding RT to it - Jedi Survivor and all. So if you want to do a direct comparison this would be a better way to do it. Or compare UE4 vs UE5 with Lumen off, also works.

 

[archangelmorph]

It's bizarre that we got an UE6 announcement from a game using UE6 and not from the creators of UE6 itself. I'm curious how they will go about revealing it, with UE5 they released and paired it with the PS5, and I assumed they were going to do the same when the PS6 comes out but clearly not.

I'm hoping UE6 has spent a decent amount of time in the oven, and not a half baked unoptimised mess like UE5 was upon its release.

EDIT:

It's very weird

 

[GymWolf]

yep. the physics are there if the devs want to use them. they just dont.

Tbf that is just basic havok not destruction.

We have to wait kingmakers for that.

 

[dgrdsv]

I read a theory on twitter that the main reason behind announcing UE6 just 4 years after the launch of UE5 is that they are aware of all the UE5 hate/backlash and are hoping for a clean slate.

UE5.6 has already fixed most of the original issues like single threaded CPU utilization, shader compilations, traversal stutter, and lack of hardware lumen and nanite foliage support. UE5.7 drastically improved nanite foliage performance by completely reworking it and UE5.8 now runs Megalights at 60 fps on PS5 and is finally production ready. They even added a probe based Lumen system that is twice as fast as software lumen just for lower end hardware.

My guess is that they are like lets just start over and ship all these features we introduced in later UE5 versions that no one really bothered to use.

So most recent UE5 changes seem to be going into the opposite directions. Megalights, Nanite foliage are performance heavy and require HWRT. Cheaper s/w Lumen is aimed at smartphones basically.
Maybe they want to do a clean break of features and introduce UE6 as the high end only option, with HWRT requirement? While UE5 will remain for the lower end h/w.
It does seem weird to introduce UE6 in a MP title running at 200+ fps though.